Admin Guide

Guidewire PolicyCenter™
System Administration Guide

Release 9.0.6
©2018 Guidewire Software, Inc.
For information about Guidewire trademarks, visit http://guidewire.com/legal-notices.
Guidewire Proprietary & Confidential — DO NOT DISTRIBUTE
Product Name: Guidewire PolicyCenter

Product Release: 9.0.6
Document Name: System Administration Guide
Document Revision: 27-November-2018
Guidewire PolicyCenter 9.0.6 System Administration Guide
Contents
About PolicyCenter documentation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16

Conventions in this document . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17
Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17
Part 1
Application Administration
1 Managing Users . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .21
PolicyCenter Default System Users . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .21
Default Owner . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .21
PolicyChange Daemon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .22
Renewal Daemon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .22
Super User. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .22
System User. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .23
Change the Unrestricted User . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .24
Minimum and Maximum Password Length. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .24
2 Application Logging. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .25

Overview of Application Logging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .25
The Logging Properties File. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .25
The Logging Directory Path. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .26
Changes to the Logging Level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .26
Logging and the env Environment Property . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .26
The Log Files Directory and the View Logs Screen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .26
Adding Loggers to the Logging Properties File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27
Logging Level Reference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27
Understanding Logging Categories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .28
Logging Category Inheritance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .28
Viewing a List of Logging Categories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .28
Enabling Logging Categories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .29
Logging Category Reference. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30
Logger Category API . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .31
Formatting a Log Message . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .32
Conversion Character Reference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .32
Format Modifier Reference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .34
Configure Logging in a Multiple Instance Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .34
Dynamic Changes to Logging Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .35
Setting the Logging Configuration Dynamically. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .35
Reloading the Logging Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .36
Changing a Logging Level Temporarily . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .36
Configuring Archive Logging Operations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .37
Print Details of the Archive Graph . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .37
Generate a Separate Archive Log . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .38
PolicyCenter Logging Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .38
Batch Processes that Generate Archive Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .38
3
Part 2
Server Administration
3 PolicyCenter Server Configuration. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .41
Important PolicyCenter Server Configuration Files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .41
Understanding the PolicyCenter Server Environment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .41
Understanding the Configuration Registry Element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .42
The Configuration Registry Element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .42
Example Syntax for Registry Server Element. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .44
Example Syntax for the Registry System Property Element . . . . . . . . . . . . . . . . . . . . . . . . . . . . .45
Assigning Server Roles to PolicyCenter Cluster Servers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .45
Defining a New Server Role . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .46
JVM Options and Server Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .46
JVM Options for gwb Build Commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .47
JVM Options Specific to the runServer Build Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .48
Setting JVM Options in PolicyCenter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .48
Configuration Parameters by Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .49
Default Configuration Values and Environment Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . .49
4 PolicyCenter Server Administration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .51

Starting Guidewire PolicyCenter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .51
Start PolicyCenter on QuickStart (Jetty) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .51
Start the Application Server from Guidewire Studio for PolicyCenter . . . . . . . . . . . . . . . . . . . . . .52
Stopping GuidewirePolicyCenter. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .52
Stop Guidewire PolicyCenter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .52
Server Startup Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .53
Server Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .53
Important Server Mode Caveats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .53
Server Modes as a Safety Feature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .54
Server Test Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .54
Setting the Server Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .54
Determining the Server Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .55
Server Run Levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .55
Server Modes and Server Run Levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .56
Setting the Quick Start Server Run Level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .56
Set the QuickStart Run Level at Server Start . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .57
Set the Server Run Level Through System Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .57
Set the Server Run Level Through Web Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .57
Determining the Server Run Level. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .57
Use System Tools to Determine the Server Run Level. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .58
Use Web Services to Determine the Server Run Level. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .58
Server Maintenance Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .58
Place the Server in Maintenance Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .59
Monitoring Server Performance in PolicyCenter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .59
Monitoring Server Status with WebSphere. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .59
Monitoring and Managing Event Messages. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .59
About Message Failure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .60
Messages Screen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .60
Message Queues Screen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .61
Message Handling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .62
Access the Messaging Editor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .62
Purge completed messages. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .63
Session Timeout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .63
User Session Replication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .64
Server Caching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .64
4
Cache Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .64

Caching and Stickiness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .65
Concurrent Data Change Prevention . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .65
Caching and Clustering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .65
Cache Behavior and Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .65
Server Cache Tuning Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .67
Special Caches for Rarely Changing Objects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .69
Server Memory Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .69
Memory Usage Logging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .69
Enabling Garbage Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .70
Understanding Possible Memory Leak . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .71
Heap Dump Generation and Analysis Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .73
JVM Profiling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .73
Large Object Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .74
5 Geocoding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .75
Understanding Geocoding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .75
Configuring Geocoding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .76
About Guidewire Geocoding Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .77
Enable the Geocode Plugin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .77
Schedule Geocode Batch Processing Runs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .78
Configure the Number of Worker Instances for Geocode Batch Processing . . . . . . . . . . . . . . . . . .78
6 Administering Batch Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .81

Overview of Batch Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .81
Work Queues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .82
Batch Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .85
Working with batch processing types in Studio. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .85
Working with Work Queue Writers and Batch Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .86
Run a Writer from PolicyCenter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .86
Run a Batch Process from PolicyCenter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .87
Run a Writer or Batch Process from the Command Prompt . . . . . . . . . . . . . . . . . . . . . . . . . . . . .87
Terminate a Writer or Batch Process from the Command Prompt . . . . . . . . . . . . . . . . . . . . . . . . .87
Check Status of a Writer or Batch Process from the Command Prompt . . . . . . . . . . . . . . . . . . . . .88
The Work Queue Scheduler . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .89
Understanding a Work Queue Schedule Specification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .89
Determine If It Is Possible to Schedule a Batch Process. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .91
View a Batch Process Schedule in PolicyCenter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .91
Scheduling Batch Processes for Specific Environments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .92
Disabling the PolicyCenter Scheduler . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .92
Configuring Work Queues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .92
The Work Queue Configuration File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .92
General Work Queue Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .93
Worker Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .94
Worker Task Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .94
Work Queues and Server Roles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .95
Performing Custom Actions After Batch Processing Completion. . . . . . . . . . . . . . . . . . . . . . . . . . .96
Schedule the Process Completion Monitor Batch Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .96
Implement the IBatchCompletedNotification Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .97
Register a Custom Batch Notification Plugin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .97
Troubleshooting Work Queues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .98
Work Queues and Batch Processes, a Reference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .98
Account Holder Count Work Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .98
Account Withdraw Evaluation Work Queue. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .99
Activity Escalation Work Queue. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
Apply Pending Account Data Updates Batch Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
5
Archive Policy Terms Work Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

Archive Reference Tracking Sync work queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
Audit Task Work Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
Bound Policy Exception Work Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
Cleanup ETLPurgeRoot Work Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
Clean Up Account Contact Role Table Work Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
Clear Policy Renewal Check Dates Batch Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
Closed Policy Exception Work Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
Create UWRules for UWIssue Types Work Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
Data Distribution Batch Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
Database Consistency Check Work Queue. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
Database Statistics Work Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
Deferred Upgrade Tasks Batch Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
Destroy Contact For Personal Data work queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
Extract Rating Worksheets Work Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
Fix JobGroup on Moved Policies Work Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
Form Text Data Delete Batch Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
Geocode Writer Work Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
Group Exception Work Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
Handle Unresolved Contingency Work Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
Impact Testing Export Work Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
Impact Testing Test Case Preparation Work Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
Impact Testing Test Case Run Work Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
Job Expiration Work Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
Notify External System For Personal Data work queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
Open Policy Exception Work Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
Overdue Premium Report Work Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
Phone Number Normalizer Work Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
Policy Hold Job Evaluation Work Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
Policy Locations Risk Assessment Work Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
Policy Renewal Start Work Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
Populate Search Columns Batch Processing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
Premium Ceding Work Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
Process Completion Monitor Batch Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
Process History Purge Batch Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
Product Model Pattern Activation Batch Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
Purge Batch Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
Purge Cluster Members Batch Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
Purge Failed Work Items Batch Process. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
Purge Message History Batch Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
Purge Old Contact Destruction Request work queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
Purge Old Transaction IDs Batch Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
Purge Orphaned Policy Period Work Queue. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
Purge Profiler Data Batch Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
Purge Quote Clones Work Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
Purge Rate Book Export Result Work Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
Purge Rating Worksheets Work Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
Purge Risk Assessment Temporary Store Work Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
Purge Temporary Policy Periods Work Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
Purge Workflow Batch Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
Purge Workflow Logs Batch Process. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
Rate Book Export Work Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
Recalculate Contingency Action Start Date Work Queue. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
Reset Purge Status and Check Dates Work Queue. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
Retrieve Policy Terms Work Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
6
Retire Activities Batch Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124

Retrieve Policy Terms Work Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
Solr Data Import Batch Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
Team Screens Batch Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
User Exception Work Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
Work Item Set Purge Batch Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
Work Queue Instrumentation Purge Batch Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
Workflow Work Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
Unused and Internal Batch Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
Part 3
Server Clustering Administration
7 Understanding PolicyCenter Server Clustering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
Cluster Terminology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
Guidewire PolicyCenter Cluster Installations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
Cluster Communication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
Configuring Cluster Communication. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
Cluster Plugin Parameter Reference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
Cluster Plugin System Properties Reference. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
Cache eviction messages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
Logging cluster plugin parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
Server Roles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
batch Server Role . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
Messaging Server Role . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
scheduler Server Role . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
startable Server Role . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
workqueue Server Role . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
ui Server Role . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
Example PolicyCenter Cluster Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
Cluster Member Startup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
Cluster Member Shutdown . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
ui-Role Cluster Member Shutdown. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
Non-ui Role Cluster Member Shutdown . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
8 Working with a PolicyCenter Cluster . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147

Planning a PolicyCenter Cluster . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
Create a Guidewire PolicyCenter Cluster. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
Managing a PolicyCenter Cluster . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
Enabling Guidewire Clustering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
Adding a Server to a PolicyCenter Cluster . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
Performing a Cluster Configuration Upgrade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
Add a Server to a PolicyCenter Cluster . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
Restarting the PolicyCenter Cluster Servers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
Monitoring Cluster Health . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
Cluster Member Health . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
Cluster Member Monitoring in PolicyCenter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
Monitor Cluster Members Using System Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
Using the PolicyCenter ping Utility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
9 Working with Component Leases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157

Component Lease Management. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
Simple Lease Management Lifecycle for a Batch Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
Simple Lease Management Lifecycle for a Message Destination . . . . . . . . . . . . . . . . . . . . . . . . 158
Component Lease Failover . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
7
Automatic Failover of a Component Lease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158

The Background Task Failover Plugin. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
Batch Process Prioritization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
Messaging and Startable Service Load Balancing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
10 Deploying Configuration Changes in a Clustered Environment . . . . . . . . . . . . . . . . . . . . . . . . 165

Rolling Upgrade Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
Risks Associated with Rolling Upgrade of Cluster Server Members . . . . . . . . . . . . . . . . . . . . . . 165
Assumptions Around a Rolling Upgrade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
Differences between a Rolling Upgrade and a Full Upgrade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
Configuration Compatibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
Making changes to Gosu code in a rolling upgrade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
Making Changes to Typelists in a Rolling Upgrade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
Updating Product Model Patterns in a Rolling Upgrade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
Coverage, Condition and Exclusion Patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
Offering Patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
Question Set and Question Patterns. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
Modifier Patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
Lookup, Grandfathering, and Modifier MinMax Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
Verification of Configuration Compatibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
Performing a Rolling Upgrade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
Prepare for a Rolling Upgrade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
Perform a Rolling Upgrade in a Test Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
Perform a Rolling Upgrade in a Production Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
Unexpected Upgrades . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
Backing Out a Rolling Upgrade. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
Configuration Upgrade of a Production Stand-alone PolicyCenter Server . . . . . . . . . . . . . . . . . . . . 177
Part 4
Security Administration
11 Managing Secure Communications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
PolicyCenter and the Transport Layer Security Protocol. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
PolicyCenter and Secure Communications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
The PolicyCenter Connection Address . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
Restricting access to a PolicyCenter screen by server mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
Multifactor Authentication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
Add multifactor authentication field to PolicyCenter Login screen . . . . . . . . . . . . . . . . . . . . . . . 183
12 Securing Access to PolicyCenter Objects. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185

Understanding the Object Access Infrastructure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
Understanding the PolicyCenter Permission Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
Permission Class Generation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
Beyond Roles and Permissions to Access Control. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
Access Control Configuration Files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
The Security Configuration File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
Static Handler Elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
Wrap Handler Elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
13 Securing Access to Notes and Documents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193

Note Permission Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
Note Permission Elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
Document Permission Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
Document Permission Elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
8
14 Securing Access to Accounts, Jobs, and Policy Periods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197

Account Producer Code Handler Elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
Job Producer Code Handler Elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
PolicyPeriod Producer Code Handler Elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
Part 5
Database Administration
15 Database Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
Accessing the Database Configuration File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
The Database Configuration File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
The Database autoupgrade Attribute . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208
The databasestatistics Database Configuration Element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
The dbcp-connection-pool Database Configuration Element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
The reset-tool-params Database Configuration Element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
The jndi-connection-pool Database Configuration Element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
Configuring JNDI Connection Initialization for Oracle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214
The oracle-settings Database Configuration Element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216
The sqlserver-settings Database Configuration Element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
The upgrade Database Configuration Element. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
The mssql-db-ddl Database Configuration Element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
The ora-db-ddl Database Configuration Element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226
The versiontriggers Database Configuration Element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
16 Database Maintenance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237

About the Upgrade and Versions Screen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
Database Best Practices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
Understanding Data Model Updates. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238
Run a Schema Verification Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239
Guidewire Database Direct Update Policy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239
PolicyCenter Database Back Up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240
Database Consistency Checks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240
Recommendations for Running Database Consistency Checks . . . . . . . . . . . . . . . . . . . . . . . . . . 241
Consistency Checks and Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
............................................. . . . . . . . . . . . . . . . . . . . . . . . . . 241
Running Consistency Checks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
Configuring the Number of Threads for Consistency Checks . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
Resize Database Columns. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244
Purging Unwanted Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
About Purging Activity-related Workflow Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
Purging Batch Process History Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
Purging Cluster Member Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
Purging Failed Work Items . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 246
Purging Job and Policy Period Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 246
Purging Message History Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
Purging Old Transaction ID Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
Purging Orphaned Policy Period Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248
Purging Profiler Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248
Purging Quote Clones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249
Purging Rate Book Export Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249
Purging Rating Worksheets Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250
Purging Temporary Risk Assessment Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250
Purging Temporary Policy Period Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251
Purging Workflow Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251
Purging Workflow Log Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252
9
Purging Work Item Set Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252

Purging Work Queue Instrumentation Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
17 Database Statistics Generation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255

Understanding Database Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255
Database Statistics Generation for Oracle Databases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255
Database Statistics Generation for SQL Server Databases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256
Updates to Database Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256
Understanding Database Statistics Batch Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257
Automatic Generation of Database Statistics During Upgrade . . . . . . . . . . . . . . . . . . . . . . . . . . 257
Managing Database Statistics using System Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257
Perform a Full Database Statistics Update . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258
Update Statistics on Tables that Exceed a Change Threshold . . . . . . . . . . . . . . . . . . . . . . . . . . . 258
Check the Database Statistics Updating Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258
Cancel the Database Statistics Updating Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259
Generate Statistics SQL for All Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259
Generate Statistics SQL for Tables that Exceed a Threshold . . . . . . . . . . . . . . . . . . . . . . . . . . . 259
Configuring Database Statistics Generation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 260
Configuring the Number of Threads for Statistics Generation . . . . . . . . . . . . . . . . . . . . . . . . . . 260
Statistics and the Database Configuration File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 260
The Database Statistics Element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
The Table Statistics Database Element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264
The Histogram Statistics Database Element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265
18 Importing and Exporting Administrative Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267

Ways to Import Administrative Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267
About the import Directory. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268
Administration Data Import at Initial Server Startup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268
Security Zone Data Import at Initial Server Startup. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268
Business Rules Import at Initial Server Startup. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
Controlling the Import of Data During Server Startup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
About Adding Admin Data after Initial Server Startup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
Add Additional Roles and Privileges after Initial Server Startup. . . . . . . . . . . . . . . . . . . . . . . . . 270
Character Set Encoding for File Import . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270
Maintain Data Integrity During Administrative Data Import . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270
Administrative Data and the PolicyCenter Data Model. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270
Public ID Prefix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271
Support for Unique Public IDs in a Development Environment . . . . . . . . . . . . . . . . . . . . . . . . . 271
Constructing a CSV File for Import . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
Constructing a Heading Line in CSV-formatted Files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
Constructing Data Lines in CSV-formatted Files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
Construct an XML File for Import . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275
Constructing the XML for the Administrative Data Import File . . . . . . . . . . . . . . . . . . . . . . . . . 275
Using Tools to Import Administrative Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
Import Administrative Data Using Import Tools. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
Importing and Exporting Administrative Data from PolicyCenter . . . . . . . . . . . . . . . . . . . . . . . . . 278
About Importing PolicyCenter Administrative Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278
About Exporting PolicyCenter Administrative Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279
Understanding Roles and Permissions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280
Role Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280
Importing Security Zone Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280
Understanding the Security Zones File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
View the XSD for Security Zones. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
Import Security Zones. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282
About Importing Zone Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282
Importing a Zone Data File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
10
Importing Custom Zone Data Files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284
19 Free-text Batch Load Command. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285

When to Run the Free-text Batch Load Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285
Prerequisites for Running the Free-text Batch Load Command . . . . . . . . . . . . . . . . . . . . . . . . . . . 286
Run the Free-text Batch Load Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286
Recovering from Solr Batch Load Errors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287
Clean-up Tasks after Running the Free-text Batch Load Command . . . . . . . . . . . . . . . . . . . . . . . . 287
Free-text Batch Load Command and Native SQL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287
20 Production Data Fix Tool . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289

Overview of the Production Data Fix Tool . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289
Typical Use of the Production Data Fix Tool. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 290
Writing Gosu Data Change Code . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291
Registering Data Change Code. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291
Run Gosu Data Change Code. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293
Logging Data Change Operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293
Data Change Command Tool Reference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 294
Data Change Web Service Reference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295
Part 6
Business Rules Administration
21 Administering Business Rules. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299
Business Rules in Guidewire PolicyCenter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299
Business Rule Roles and Permissions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299
Business Rule Configuration Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 300
Business Rule Production Server Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 300
Business Rule Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301
Business Rule Versioning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301
Rules for Deleting a Business Rule Version . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 302
Business Rule Deployment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 302
Business Rule State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 303
Business Rule Lifecycle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304
Business rule logging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306
Invalid Business Rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306
22 Business Rules Import and Export . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309

Overview of Business Rule Export and Import . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309
About Rule Version Conflicts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310
Business Rule Import Failure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310
Source and Target Data Models and Rule Export and Import . . . . . . . . . . . . . . . . . . . . . . . . . . . 310
Export Business Rules from Guidewire PolicyCenter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310
Import Business Rules into Guidewire PolicyCenter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311
The Import/Export Status Screen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311
About the Business Rules Export File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312
The Review Import/Complete Import Screens. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312
The Rule Import Disposition Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313
The Rule Import Manage Synchronization Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313
Business Rule Validation Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 314
The Compare Rules Screen. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 314
Resolve Rule Import Conflicts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315
Automatic Import of Business Rules at Server Startup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315
Import PolicyCenter Business Rules at Server Startup. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 316
Correct Duplicate BizRuleDeploymentId Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 316
11
Business Rule Import after Application Upgrade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317
Part 7
Administration Tools
23 Server Tools. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321
Accessing the Server Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321
Batch Process Info . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321
Processes Table Columns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 322
Chart and History Tabs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323
Download a Batch Process History Report. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323
Work Queue Info . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 324
Work Queue Table Columns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 324
Item Statistics Tabs and Columns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325
Work Queue Reports. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327
The Work Queue Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327
Download a Work Queue Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327
The Work Queue Raw Data Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 328
Download the Work Queue Raw Data Report. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 328
The Work Queue History Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 328
Download the Work Queue History Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 328
About Work Queue Efficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329
Set Log Level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329
View Logs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 330
Info Pages. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 330
Configuration. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 330
Archive Info . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 331
Domain Graph Info. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 332
Consistency Checks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 333
Database Table Info . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337
Database Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339
Database Storage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 340
Data Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342
Database Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343
Oracle Statspack. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 346
Oracle AWR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347
Oracle AWR Unused Indexes Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 349
Oracle Outlines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 349
SQL Server DMV Snapshot. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 350
Microsoft JDBC Driver Logging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 351
Load History . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 351
Load Integrity Checks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353
Load Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 354
Runtime Environment Info . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 354
Safe Persisting Order . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 354
Loaded Gosu Classes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355
Serialization Info . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355
Management Beans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356
Startable Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356
Cluster Members and Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356
Cluster Members – This Application Server Instance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357
Cluster Members – Application Server Instances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 358
Cluster Members – Components Tab. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359
Cluster Members – History Tab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 360
12
The Cluster Components Screen. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 360

Download a Server Component Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 362
Schedule a Planned Cluster Member Shutdown . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 362
Upgrade and Versions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 362
The Upgrade Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 364
Understanding Guidewire Software Versioning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365
Cache Info . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 367
The Cache Summary View . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 367
The Historical Performance View . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 368
The Cache Details View . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 368
Understanding the Cache Data Report. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 369
Guidewire Profiler . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 369
Guidewire Profiler Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 370
Profiler Entry Points . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 370
Web Session Profiling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 371
PolicyCenter Application Profiling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 372
Profiler Trace Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 372
Guidewire Profiler Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 373
Save a Profiler Analysis Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 374
Ways to View a Guidewire Profiler Analysis Reports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 374
Review Profiler Upgrade Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375
Profiler Tags . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 376
Profiler Frames . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 376
Profiler Stacks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 376
Using Custom Profile Tags with Guidewire Profiler . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 376
Understanding Properties and Counters on a Frame . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 377
Understanding the Guidewire Profiler API. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 378
Profiler-related Configuration Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 378
Product Model Info . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 378
24 Internal Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 381

Accessing the Internal Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 381
Reload . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 381
Testing System Clock. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 382
PC Sample Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 382
Free-text Search . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 382
Archiving Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 383
25 Command Prompt Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 385

Administration Tools Overview. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 385
User Credentials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 385
Accessing Administration Tool Help. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 386
Administrative Tool Command Syntax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 386
PolicyCenter Command Prompt Tools Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 386
Data Change Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 387
Data Change Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 387
Import Tools Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 388
Import Tools Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 389
Maintenance Tools Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 390
Maintenance Tools Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 390
Messaging Tools Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 391
Messaging Tools Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 391
System Tools Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 393
System Tools Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 393
Table Import Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 399
Table Import Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 399
13
Template Tools Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 401

Template Tools Options. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 402
Workflow Tools Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 402
Workflow Tools Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 403
Zone Import Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 403
Zone Import Options. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 403
14
15
About PolicyCenter documentation
The following table lists the documents in PolicyCenter documentation:
Document Purpose
InsuranceSuite Guide If you are new to Guidewire InsuranceSuite applications, read the InsuranceSuite Guide for informa‐
tion on the architecture of Guidewire InsuranceSuite and application integrations. The intended read‐
ers are everyone who works with Guidewire applications.
Application Guide If you are new to PolicyCenter or want to understand a feature, read the Application Guide. This guide
describes features from a business perspective and provides links to other books as needed. The in‐
tended readers are everyone who works with PolicyCenter.
Database Upgrade Guide Describes the overall PolicyCenter upgrade process, and describes how to upgrade your PolicyCenter
database from a previous major version. The intended readers are system administrators and imple‐
mentation engineers who must merge base application changes into existing PolicyCenter application
extensions and integrations.
Configuration Upgrade Guide Describes the overall PolicyCenter upgrade process, and describes how to upgrade your PolicyCenter
configuration from a previous major version. The intended readers are system administrators and im‐
plementation engineers who must merge base application changes into existing PolicyCenter applica‐
tion extensions and integrations. The Configuration Upgrade Guide is published with the Upgrade
Tools and is available from the Guidewire Community.
New and Changed Guide Describes new features and changes from prior PolicyCenter versions. Intended readers are business
users and system administrators who want an overview of new features and changes to features. Con‐
sult the “Release Notes Archive” part of this document for changes in prior maintenance releases.
Installation Guide Describes how to install PolicyCenter. The intended readers are everyone who installs the application
for development or for production.
System Administration Guide Describes how to manage a PolicyCenter system. The intended readers are system administrators re‐
sponsible for managing security, backups, logging, importing user data, or application monitoring.
Configuration Guide The primary reference for configuring initial implementation, data model extensions, and user inter‐
face (PCF) files for PolicyCenter. The intended readers are all IT staff and configuration engineers.
PCF Reference Guide Describes PolicyCenter PCF widgets and attributes. The intended readers are configuration engineers.
Data Dictionary Describes the PolicyCenter data model, including configuration extensions. The dictionary can be gen‐
erated at any time to reflect the current PolicyCenter configuration. The intended readers are configu‐
ration engineers.
Security Dictionary Describes all security permissions, roles, and the relationships among them. The dictionary can be
generated at any time to reflect the current PolicyCenter configuration. The intended readers are con‐
figuration engineers.
Globalization Guide Describes how to configure PolicyCenter for a global environment. Covers globalization topics such as
global regions, languages, date and number formats, names, currencies, addresses, and phone num‐
bers. The intended readers are configuration engineers who localize PolicyCenter.
Rules Guide Describes business rule methodology and the rule sets in Guidewire Studio for PolicyCenter. The in‐
tended readers are business analysts who define business processes, as well as programmers who
write business rules in Gosu.
Contact Management Guide Describes how to configure Guidewire InsuranceSuite applications to integrate with ContactManager
and how to manage client and vendor contacts in a single system of record. The intended readers are
PolicyCenter implementation engineers and ContactManager administrators.
16 About PolicyCenter documentation

Document Purpose
Best Practices Guide A reference of recommended design patterns for data model extensions, user interface, business
rules, and Gosu programming. The intended readers are configuration engineers.
Integration Guide Describes the integration architecture, concepts, and procedures for integrating PolicyCenter with ex‐
ternal systems and extending application behavior with custom programming code. The intended
readers are system architects and the integration programmers who write web services code or plu‐
gin code in Gosu or Java.
Java API Reference Javadoc‐style reference of PolicyCenter Java plugin interfaces, entity fields, and other utility classes.
The intended readers are system architects and integration programmers.
Gosu Reference Guide Describes the Gosu programming language. The intended readers are anyone who uses the Gosu lan‐
guage, including for rules and PCF configuration.
Gosu API Reference Javadoc‐style reference of PolicyCenter Gosu classes and properties. The reference can be generated
at any time to reflect the current PolicyCenter configuration. The intended readers are configuration
engineers, system architects, and integration programmers.
Glossary Defines industry terminology and technical terms in Guidewire documentation. The intended readers
are everyone who works with Guidewire applications.
Product Model Guide Describes the PolicyCenter product model. The intended readers are business analysts and implemen‐
tation engineers who use PolicyCenter or Product Designer. To customize the product model, see the
Product Designer Guide.
Product Designer Guide Describes how to use Product Designer to configure lines of business. The intended readers are busi‐
ness analysts and implementation engineers who customize the product model and design new lines
of business.
Conventions in this document

Text style Meaning Examples
italic Indicates a term that is being defined, A destination sends messages to an external system.
added emphasis, and book titles. In Navigate to the PolicyCenter installation directory by running the
monospace text, italics indicate a variable to following command:
be replaced.
cd installDir
bold Highlights important sections of code in for (i=0, i<someArray.length(), i++) {

examples. newArray[i] = someArray[i].getName()
}
narrow bold The name of a user interface element, such Click Submit.
as a button name, a menu item name, or a
tab name.
monospace Code examples, computer output, class and The getName method of the IDoStuff API returns the name of the
method names, URLs, parameter names, object.
string literals, and other objects that might
appear in programming code.
monospace italic Variable placeholder text within code Run the startServer server_name command.
examples, command examples, file paths, Navigate to http://server_name/index.html.
and URLs.
Support
For assistance, visit the Guidewire Community.
About PolicyCenter documentation 17
Guidewire customers
https://community.guidewire.com
Guidewire partners
https://partner.guidewire.com
18 About PolicyCenter documentation

part 1
Application Administration
chapter 1
Managing Users
This topic discusses the default system users that Guidewire provides in the base PolicyCenter configuration.
PolicyCenter Default System Users

In the base configuration, PolicyCenter provides the following default system users:
• Default Owner
• PolicyChange Daemon
• Renewal Daemon
• Super User
• System User
Note: PolicyCenter default system users are separate and distinct from the users in the sample data
that Guidewire provides for testing and development purposes.
Guidewire provides each of these default users for a specific purpose and to fulfill one or more specific roles. In
most cases, it is not possible to delete a default user as internal code uses that user for a specific purpose. It is
possible to edit various aspects of user information, for example, setting a new password for that user.
Default Owner
Default system user defaultowner has the following characteristics.
User Default Owner

User name defaultowner
Group Default Root Group
Can delete No
User defaultowner has first name Default and last name Owner. In the base configuration, PolicyCenter does not
assign roles to this user.
If PolicyCenter cannot assign certain business objects to a specific user, PolicyCenter assigns that object to user
defaultowner. PolicyCenter performs this assignment internally.
IMPORTANT Do not make direct assignments to user defaultowner.
Guidewire recommends, as a business practice, that someone in the organization periodically search for outstanding
work assigned to user defaultowner. If the search finds one of these assignments, the searcher must reassign these
Managing Users 21
items to a proper owner. Guidewire also recommends that the rule administrator investigate why PolicyCenter did
not assign an item of that type and correct any errors in the rules.
PolicyChange Daemon
Default system user policychange_daemon has the following characteristics.
User PolicyChange Daemon

User name policychange_daemon
Can delete Yes
User policychange_daemon has first name PolicyChange and last name Daemon. In the base configuration,
PolicyCenter assigns the Superuser role to the policychange_daemon user. The Superuser role has all permissions.
Thus, user policychange_daemon has unrestricted access to the entire PolicyCenter application.
PolicyCenter defines user policychange_daemon as the default user for automatic policy changes started through
the PolicyChangeAPI. The Super User role assigned to this user provides overriding authority to auto-approve all
underwriting issues that occur in automated policy changes. Guidewire recommends that you review the roles and
permissions associated with this user in a production environment.
Deleting User policychange_daemon

Although it is possible to delete user policychange_daemon, the code in PolicyChangeAPI.gs explicitly names
this user. If you delete this user, you must update the code in PolicyChangeAPI with the name of a user with the
equivalent permissions.
Renewal Daemon
Default system user renewal_daemon has the following characteristics.
User Renewal Daemon

User name renewal_daemon
Can delete Yes
User renewal_daemon has first name Renewal and last name Daemon. In the base configuration, PolicyCenter
assigns the Superuser role to the renewal_daemon user. The Superuser role has all permissions. Thus, user
renewal_daemon has unrestricted access to the entire PolicyCenter application.
PolicyCenter defines user renewal_daemon as the default user for automated renewal policy transactions started
through PolicyRenewalStart batch processing. The Super User role assigned to this user provides overriding
authority to auto-approve all underwriting issues that occur in automated policy renewals. Guidewire recommends
that you review the roles and permissions associated with this user in a production environment.
Deleting User renewal_daemon

Although it is possible to delete user renewal_daemon, the code in PolicyRenewalPlugin.gs (called by
PolicyRenewalStart) explicitly names this user. If you delete this user, you must update the code in
PolicyRenewalPlugin with the name of a user with the equivalent permissions.
Super User
Default system user su has the following characteristics.
22 chapter 1: Managing Users
User Super User

User name su
Can delete No
User su has first name Super and last name User. Guidewire configures user su as an unrestricted user in
configuration file config.xml. PolicyCenter does not evaluate permissions for unrestricted users. As a consequence
of PolicyCenter bypassing permission checking, the unrestricted user has access to the entire PolicyCenter
application. Thus, an unrestricted user behaves as if it has all permissions.
In the base configuration, PolicyCenter assigns the Superuser role to the su user.
IMPORTANT Guidewire strongly recommends that you change the Super User password from its
default value.
User su and the Command Line Tools

To run the PolicyCenter command prompt tools, you must supply a user name and password. If you do not supply
the -user parameter, the command defaults to user su and you must supply that user’s password.
See also
• “Change the Unrestricted User” on page 24
• “Command Prompt Tools” on page 385
• Installation Guide
System User
Default system user sys has the following characteristics.
User System User

User name sys
Can delete No
User sys has first name System and last name User. In the base configuration, PolicyCenter does not assign roles to
this user.
PolicyCenter requires user sys to exist. PolicyCenter uses this user to perform automated work such as running
batch processing, messaging polling, and server startup. Each time PolicyCenter needs to do such work, it creates a
session with the sys user. This is why there might appear to be many sessions with the sys user. Session in this
sense is not a web session. Rather, it represents the authentication of a user.
WARNING Do not rename or delete the sys user. Deleting or renaming this user disables
PolicyCenter.
Temporary System Users

PolicyCenter creates system users in addition to the standard users who log into PolicyCenter.
PolicyCenter gives a designation of Temporary system user to an unauthenticated user session. PolicyCenter
creates such sessions for login. By design, PolicyCenter does not associate a user with the login screen. The
system_tools -sessioninfo command does not filter out this user. The Server Tools Management Beans screen
does filter out this user.
Managing Users 23
Change the Unrestricted User

About this task
By default, PolicyCenter uses su as the user with unrestricted access to the entire PolicyCenter application. It is
possible to change the unrestricted user.
Procedure
1. In the Studio Project window, expand configuration→config:
2. Open file config.xml for editing.
3. Set configuration parameter UnrestrictedUserName to the user name of the new unrestricted user:
<param name="UnrestrictedUserName" value="userName"/>
Minimum and Maximum Password Length

The MinPasswordLength and MaxPasswordLength parameters in config.xml control the minimum and maximum
number of characters for passwords. For example, if you want all users in your system to have a password length of
at least six characters and a maximum of sixteen, set the following in config.xml:
<param name="MinPasswordLength" value="6"/>

<param name="MaxPasswordLength" value="16"/>
24 chapter 1: Managing Users

chapter 2
Application Logging
PolicyCenter creates automatic logs of many actions by users and operations by the server.
Overview of Application Logging

Guidewire uses the slf4j API, in conjunction with Apache log4j libraries and internal Guidewire libraries, to
provide logging in the PolicyCenter application. A logging properties file controls the behavior of the logging
activity in the application.
The logging properties file uses Apache log4j formatting to define the following logging components:
Logger Logical file name. It is possible to configure each logger independently to log information at a certain level.
Appender Output point (destination) for a logger. This can be, for example, the application console or a specific logging file.
Layout Log entry formatting instructions. Each logger category can have its own layout format.
These logging component types work together to log messages according to the message type and severity level.
These components also define the format and the output destination for the various logging categories.
For more information on slf4j, see the following web site:
http://slf4j.org/index.html
For more information on Apache log4j loggers, appenders, and layouts, see the following web site:
http://logging.apache.org/log4j/1.2/manual.html
The Logging Properties File

Properties file logging.properties specifies a number of system logging options for the PolicyCenter server. You
access this file from the following location in Guidewire Studio:
configuration→config →logging
The logging.properties file uses the format specified by Apache log4j. The entries in logging.properties
control what to log and to which file to write the log. In the base configuration, PolicyCenter sets the basic logging
configuration to the following:
log4j.rootCategory=INFO, Console, DailyFileLog
Application Logging 25
This setting:
• Instructs PolicyCenter to send system-wide informational messages to two output points: the PolicyCenter
console (Console) and a log file (with name DailyFileLog in the default configuration).
• Sets the default logging level to INFO.
Within file logging.properties, entries such as log4j.appender.* indicate the parameters of each output point.
These entries identify properties such as location or output format options. As PolicyCenter starts, it attempts to
write a log file in the location specified by log4j.appender.DailyFileLog.File. By default, PolicyCenter writes
to the following log file:
tmp/gwlogs/PolicyCenter/logs/pclog.log
PolicyCenter creates the log file automatically. However, if the directory specified by DailyFileLog.File does not
exist, PolicyCenter writes log information to the console only.
For more information about how to create and manage log4j entries in logging.properties, see the following
Apache web site:
http://logging.apache.org/log4j/1.2/manual.html
The Logging Directory Path

In file logging.properties, express all file locations using an absolute path. Regardless of operating system, you
must use forward slashes and not backslashes. The directory path that you specify must exist. PolicyCenter creates
the log file itself automatically.
Note: If you specify a non-existent logging directory path, PolicyCenter writes logging information to
the application console only.
Changes to the Logging Level

If you edit the logging.properties file and do not restart the server, the new logging level only take effects for
new database connections. However, you can make an immediate logging change in the PolicyCenter server without
having to first redeploy PolicyCenter through the use of one of the following:
• Command prompt administration tool system_tools
• Web service SystemToolsAPI
Logging and the env Environment Property

If the env environment property is non-existent (null), PolicyCenter reads the logging configuration from the
default logging.properties file.
However, if the env environment property is non-null, PolicyCenter tries to obtain the logging configuration from an
env-logging.properties file in the following PolicyCenter directory:
modules/configuration/config/logging
For example, if you have an environment called test, PolicyCenter looks for logging properties in test-
logging.properties. If this file does not exist, then PolicyCenter reads the logging configuration from the default
logging.properties file.
The Log Files Directory and the View Logs Screen

PolicyCenter includes a log viewer on the Server Tools View Logs screen, accessible to administrators. The
guidewire.logDirectory property in logging.properties specifies the location of log files for the PolicyCenter
log viewer.
In the base configuration, Guidewire sets this property to the following value:
guidewire.logDirectory = /tmp/gwlogs/ClaimCenter/logs/
Set the log file locations for the individual log4j.appender.category.File entries to the same directory as that
used for guidewire.logDirectory. This ensures that these log files are visible as well from the View Logs screen.
26 chapter 2: Application Logging
Adding Loggers to the Logging Properties File

File logging.properties provides a number of the examples of how to set up a logger for a specific logging
category. However, it is very likely that you will want to add additional logging categories to this file. To add
additional logging categories to file logging.properties, do one of the following:
• Uncomment an existing example logger
• Add a new logging category
Make an Example Logger in the Logging Properties File Active
About this task
Guidewire comments out many of the example loggers in file logging.properties. This action makes the logger
inactive. To an example logger active, uncomment the logger entry by removing the hash mark (#) at the beginning
of the line. For example, in the base configuration, the Messaging logging category is inactive in Guidewire
PolicyCenter.
##### Messaging #####

#
# Defines the log file for messaging.
#
#log4j.category.Messaging=DEBUG, MessagingLog
#log4j.category.Messaging.Events=DEBUG, MessagingLog
log4j.additivity.Messaging=false
log4j.appender.MessagingLog=org.apache.log4j.DailyRollingFileAppender
log4j.appender.MessagingLog.encoding=UTF-8
log4j.appender.MessagingLog.File=/tmp/gwlogs/BillingCenter/logs/messaging.log
log4j.appender.MessagingLog.DatePattern = .yyyy-MM-dd
log4j.appender.MessagingLog.layout=org.apache.log4j.PatternLayout
log4j.appender.MessagingLog.layout.ConversionPattern=%-10.10X{server}
%-8.24X{userID} %d{ISO8601} %p %m%n
To make the Messaging logging category active, remove the hash mark from the following line of code.
log4j.category.Messaging=DEBUG, MessagingLog
Add a New Logging Category to the Logging Properties File
About this task
Guidewire supports a number of logging categories for which there are no existing examples in the PolicyCenter
base configuration. To add a new logging category to file logging.properties, do the following:
Procedure
1. Determine whether Guidewire supports the logging category.

2. Copy an existing example logger in the logging properties file and paste it at the end of the file.
3. Modify the copy of an existing logger to create your new logger.
Logging Level Reference

The logging level determines how much information PolicyCenter records in the log. The following list describes
the different levels of information that PolicyCenter records in the log, in the increasing order of the severity of the
information.
Level Description
TRACE Messages about processes that are about to start or that completed. These types of messages provide flow‐of‐control
logging. Trace logging has no or minimal impact on system performance. Typical messages might include:
• Calling plugin.
• Returned from plugin call.
DEBUG Messages that test a provable and specific theory intended to reveal some system malfunction. These messages need
not be details but include information that would be understandable by an administrator. For example, dumping the
contents of an XML tag or short document is acceptable. However, exporting a large XML document with no line
breaks is usually not appropriate. Typical messages might include:
• Length of Array XYZ = 2345.
• Now processing record with public ID ABC:123456.
INFO Messages that convey a sense of correct system operation. Typical messages might include:
• Component XYZ started.
• User X logged on to PolicyCenter.
WARN Messages that indicate a potential problem. Examples include:
• An assignment rules did not end in an assignment.
• Special setting XYZ was not found, so PolicyCenter used the default value.
• A plugin call took over 90 seconds.
ERROR Messages that indicate a definite problem. Typical messages might include:
• A remote system refused a connection to a plugin call.
• PolicyCenter can not complete operation XYZ even with a default.
Understanding Logging Categories

File logging.properties contains examples of the component categories available for logging in Guidewire
PolicyCenter. It is possible for both internal PolicyCenter code or custom integration code to define additional
logging categories that do not show in the logging properties file. It is also possible for third-party components, such
as Apache, to provide their own logging categories. Each Guidewire application has its own unique categories as
well.
Logging Categories for Plugins
In some cases, PolicyCenter does not use a particular plugin interface. In those cases, the logging category exists in
file logging.properties, but, PolicyCenter does not use the category for logging.
Logging Category Inheritance

The RootLogger category in logging.properties is the parent of all other logging categories. Setting the logging
level on RootLogger causes all categories to inherit the same logging level, provided those categories do not set a
different logging level. Set the logging level property of the different logging entries to set how much information
you want sent to each type of log.
Logging levels inherit from parent categories unless the child category defines a different level. For example, setting
Messaging to DEBUG also sets Messaging.Email and Messaging.Events to DEBUG, unless those categories
individually define another logging level.
In the base configuration, Guidewire sets the value of rootCategory to INFO:
log4j.rootCategory=INFO
Viewing a List of Logging Categories

You can use any of the following means to view a list of the currently enabled logging categories.
View More information

Set Log Level screen “View Logging Categories from the Set Log Level Screen” on page 29
system_tools command “View Logging Categories Using a System Tools Command” on page 29
SystemToolsAPI web service “View Logging Categories Using Web Services” on page 29
View Logging Categories from the Set Log Level Screen

Procedure
1. Log into PolicyCenter as a user with administrative privileges.
2. Navigate to the Server Tools Set Log Level screen.
3. Expand the Logger drop-down.
In the drop-down list, you can view the PolicyCenter standard logging categories, logging categories for
internal Guidewire code, and logging categories for third-party software.
View Logging Categories Using a System Tools Command

About this task
The system_tools command lists logging categories defined by PolicyCenter only. It does not list third-party
logging categories.
Procedure
1. Ensure that the PolicyCenter server is running.
2. Navigate to the following directory in the PolicyCenter installation directory:
admin/bin
3. Enter the following command:
system_tools -user user -password password -loggercats
The PolicyCenter administration command prompt tools all require that you enter the password of an
administrative user for the tool to work. The use of a user name is optional.
View Logging Categories Using Web Services

About this task
The SystemToolsAPI method lists logging categories defined by PolicyCenter only. It does not list third-party
logging categories.
Procedure
2. Call the following method on the SystemToolsAPI web service:
SystemToolsAPI.getLoggingCategories
Enabling Logging Categories

The logging.properties file contains additional logging categories for system components such as plugins or
integration code. In the base configuration, Guidewire comments out these entries by default. To enable a logging
category, uncomment the appropriate log4j.category.* entry.
For example, to log the processing of Gosu code by the PolicyCenter rule engine, uncomment the first line in the
following example code.
# log4j.category.RuleEngine=INFO, RuleEngineLog
log4j.additivity.RuleExecution=false
log4j.appender.RuleExecutionLog=org.apache.log4j.DailyRollingFileAppender
log4j.appender.RuleExecutionLog.encoding=UTF-8
log4j.appender.RuleExecutionLog.File=/tmp/gwlogs/PolicyCenter/logs/ruleexecution.log
In this example, notice the following:

• PolicyCenter ignores all logging entries related to a specific category if the log4j.category.* entry for that
category has a hash mark (#) at the beginning of the line. To make a specific logging category active, uncomment
the log4j.category.* entry for that category.
• PolicyCenter writes the log information to a different output point, as defined by the following line in
logging.properties:
log4j.appender.RuleEngineLog.File=/tmp/gwlogs/PolicyCenter/logs/ruleengine.log
Just as with the daily log, the directory location for any specialized log file must exist. The most important settings
to change are the location and name of the log file and the logging threshold for that logging category.
Related information
Apache Logging Services
Logging Category Reference

The following list describes many of the major Guidewire logging categories that are available in PolicyCenter.
Logging category Description

Api.* Base logging category for calls for all SOAP APIs.
Application.* Base logging category for internal Guidewire application logging.
Application.Addressbook Logging for Guidewire platform code that interacts with ContactManager. Guidewire
ContactManager does not use this category.
Assignment.* Base logging category for the assignment subsystem.
Availability Logging of the determination of availability of an item in PolicyCenter. PolicyCenter bases the
availability criteria for each item on the item type.
For example, suppose that one criterion is the effective date of an item. If the effective date is
not prior to or equal to today’s date, the item is not available.
BillingIntegration Logging of integration between PolicyCenter and BillingCenter.
BizRules Base logging category for business rule activity and events.
Configuration.* Base logging category for configuration problems in such areas as in security configuration, PCF
configuration, locale configuration and so forth.
Datagen Logging of internal Guidewire code used for testing.
Geodata.* Base logging category for the Geodata daemon.
Globalization.* Base logging category for the globalization subsystem.
Import Logging for import of XML data into PolicyCenter.
Integration.* Base logging category for general integration issues.
LDAP Logging of issues related to the LDAP subsystem.
Messaging.* Base logging category for the messaging system. PolicyCenter writes logging information in this
category to a separate messaging.log file.
OSGi.* Base logging category for calls to OSGi plugins.
PXLOGGER Logging for the internal Guidewire test platform.

Logging category Description

PerfAction.* Base logging category for issues related to performance.
PerfAnalyzer Logging of issues related to the performance analyzer.
Plugin.* Base logging category for all calls into any plugin. For child categories of Plugin, use the plugin
name (PluginName), for example:
Plugin.PluginName
PolicyCenter writes logging information in this category to a separate plugins.log file.
Profiler Base logging category for the Guidewire Profiler. See “Guidewire Profiler” on page 369 for more
information on the Guidewire Profiler.
RuleEngine Do not use. Use RuleExecution instead.
RuleExecution.* Base logging category for PolicyCenter rule execution. Only rule execution actions generate log‐
ging events in this category. This category does not contain any information from the rules en‐
gine itself.
PolicyCenter writes logging information in this category to a separate ruleexecution.log file.
RuleExecutionUI Logging of rule execution activity in the PolicyCenter interface.
Rules Do not use. Use the RuleExecution logging category instead.
Security Internal Guidewire base logging category for security logging.
Server.* Internal Guidewire base logging category for server and platform logging.
Studio Logging of Guidewire Studio activity. This category applies to non‐Gosu‐related activities in
Guidewire Studio only. However, if you execute Gosu code within the Studio Gosu Scratchpad,
Studio executes the Gosu code on the server. Thus, it is possible to trigger Rule Engine logging on
the server as well.
PolicyCenter writes logging information in this category to a separate studio.log file.
Test.* Internal Guidewire base logging category for the test subsystem.
User Logging of each user’s log in and log out of PolicyCenter.
UserInterface.* Internal Guidewire base logging category for user interface logging.
Workqueue.* Base logging category for work queue functionality. For more information on work queues, see
“Administering Batch Processing” on page 81.
Logger Category API

Guidewire uses the following API to trigger logging action by category:
gw.api.system.PCLoggerCategory
For example, to use this API in Gosu code to perform assignment logging, do something similar to the following:
uses gw.api.system.PCLoggerCategory
...
var logger = PCLoggerCategory.ASSIGNMENT
...
logger.debug("Print out this message.")
Notice the following:

• The code defines the logger category, in this case, Assignment.
• The code defines the logging level, in this case, debug.
• The code defines the statement to print in the log file or to the console, as configured.
The category that you define in this code must exist as one of the base configuration logging categories.
Base Configuration Logging Categories

A logging category defines how the Apache log4j logging utility handles different types of logging requests. In the
base configuration, Guidewire provides a number of readily usable logging categories,
There are several ways to view the base configuration logging categories:
• From the Set Log Level Server Tools screen.
• By running the system_tools command from a command prompt and adding the -loggercats option.
Logging Levels
You can use the logger category API to generate logging messages for any valid PolicyCenter logging level. The
following are all valid log levels:
• Trace
• Info
• Debug
• Warn
• Error
The following code is an example of the use of a debug logging statement in a Gosu rule.
logger.debug( "##### This is the Global Pre-renewal Assignment rule " +
actions.getRule().DisplayName )
The log level that you set here overrides the default logging level set for this category in logging.properties.
Formatting a Log Message

Each logger category in file logging.properties associates a log4j layout with the logging category appender.
The log4j layout definition controls the formatting of the logging message. The appender controls the output point
or destination.
The following example illustrates the layout definition for the application console log.
log4j.appender.Console.layout=org.apache.log4j.PatternLayout
log4j.appender.Console.layout.ConversionPattern=%-10.10X{server} %-8.24X{userID} %d{ISO8601} %p %m%n
Notice that:
• Apache utility class PatternLayout, a standard part of the Apache log4j distribution, provides the means of
handling string patterns.
• Conversion patterns use control characters, similar to the C language printf function, to specify the output
format for the message.
You can modify the log4j.appender.log.layout.ConversionPattern value to change the information included
in log messages for a log type. For example, to list the logging category for console logs, add %c to the
log4j.appender.Console.layout.ConversionPattern value. You can then filter logs by category.
Conversion Character Reference

Use the conversion characters listed in the following table to specify the formatting of logging category conversion
patterns in file logging.properties.
Character Description
%% Writes the percent sign to output.
%c Name of the logging category. See “Understanding Logging Categories” on page 28 for categories provided with
PolicyCenter.

%C Name of the Java class. Because the PolicyCenter logging API is a wrapper around log4j, %C returns the class
name of the logger. If you want class names in your log messages, include them specifically in the message rath‐
er than by using %C in the conversion pattern.
%d Date and time. Acceptable formats include:
• %d{ISO8601}
• %d{DATE}
• %d{ABSOLUTE}
• %d{HH:mm:ss,SSS}
• %d{dd MMM yyyy HH:mm:ss,SSS}
• ...
PolicyCenter uses %d{ISO8601} by default.
%F Name of the Java source file. Because the PolicyCenter logging API is a wrapper around log4j, %F returns a file
name for the PolicyCenter logging API. If you want file names in your log messages, include them specifically in
the message rather than by using %F in the conversion pattern.
%l Abbreviated format for %F%L%C%M. This outputs the Java source file name, line number, class name and method
name. Because the PolicyCenter logging API is a wrapper around log4j, the information returned is for the
PolicyCenter logging API. If you want information such as class and method names in your log messages, include
them specifically in the message rather than by using %l in the conversion pattern.
%L Line number in Java source. Because the PolicyCenter logging API is a wrapper around log4j, %L returns a line
number from the PolicyCenter logging API. If you want line numbers in your log messages, include them specifi‐
cally in the message rather than by using %L in the conversion pattern.
%m The log message.
%M Name of the Java method. Because the PolicyCenter logging API is a wrapper around log4j, %M returns info
string. If you want method names in your log messages, include them specifically in the message rather than by
using %M in the conversion pattern.
%n New line character of the operating system. This is preferable to entering \n or \r\n as it works across platforms.
%p Priority of the message. Typically, either FATAL, ERROR, WARN, INFO or DEBUG. You can also create custom priorities
in your own code.
%r Number of milliseconds since the program started running.
%t Name of the current thread.
%throwable Include a throwable logged with the message. Available format is:
• %throwable – Display the whole stack trace.
• %throwable{n} – Limit display of stack trace to n lines.
• %throwable{none} – Equivalent of %throwable{0}. No stack trace.
• %throwable{short} – Equivalent of %throwable{1}. Only first line of stack trace.
%X The nested diagnostic context. You can use this to include server and user information in logging messages.
Specify a key in the following format to retrieve that information from the nested diagnostic context: %X{key}.
The following keys are available:
• server
• user
• userID
• userName
For example, to include the server name, add %X{server}. For example, to include the server name, add
%X{server}.
There are three options for logging user information in logging patterns:
user – prints the numeric opaque ID for the user
userID – a unique user ID string, such as "aapplegate"
userName – a real name, such as "Andy Applegate"
For any of these, specify the minimum and the maximum size of the field. For example: %-16.16X{userName}.
If the actual value is shorter than the minimum field size, the user identifier gets padded with spaces on the
right. If the actual value is longer than the maximum size of the field, the user identifier gets truncated from the
left.
The user key lists a sequence number assigned to the user by the server and is not very informative. To include
user login ID information, instead use the userID key.
Related information
Log4j Pattern Layout
Format Modifier Reference

File logging.properties defines how Guidewire PolicyCenter handles application logging. In the logging
properties file, you can specify the format of information in log messages by using a conversion pattern for the
characters.
In using the conversion characters to define an output format, you can also add a format specification between the
percent sign and the letter in the conversion pattern. The following table describes conversion patterns available with
Apache log4j libraries.
Pattern Description
%N Specifies a minimum width of N for the output. N is an integer. If the output is less than the minimum width, the
logger pads the output with spaces. Text is right‐justified.
For example, to specify a minimum width of 30 characters for the logging category, add %30c to the conversion pat‐
tern.
%-N Left‐justifies the output within the minimum width of N characters. N is an integer.
For example, to have the logging category left justified within a minimum width of 30 characters, add %-30c to the
conversion pattern.
The default output is right‐justified.
%.N Specifies a maximum width of N for the output. N is an integer.
For example, to have the logging category output have a maximum width of 30 characters, add %.30c to the conver‐
sion pattern. The logger truncates output from the beginning if it exceeds the maximum width.
%M.N Pads with spaces to the left if output is shorter than M characters. If output is longer than N characters, then the
logger truncates from the beginning.
%-M.N Pads with spaces to the right if output is shorter than M characters. If output is longer than N characters, then the
logger truncates from the beginning.
Configure Logging in a Multiple Instance Environment

About this task
Within file logging.properties, it is possible to use variables to specify log file names and locations. This is
particularly useful if there are multiple PolicyCenter instances on the same physical server. This enables you to use a
common logging properties file and generate log files for each separate PolicyCenter instance.
Procedure
1. In file config.xml, add an entry to the <registry> element for each cluster server that you want to log.
For example:
<registry roles="…" >

<server env="test" serverid="testserver1" roles="…" />
<server env="test" serverid="testserver2" roles="…" />
</registry>

2. In file logging.properties, do the following:

a. Set the logging directory.
For example:
guidewire.logDirectory = /tmp/gwlogs/PolicyCenter/logs/
Set this variable to an absolute path to a directory that already exists. You must use forward slashes as the
path separator.
b. Set a value for log4j.appender.DailyFileLog.File that uses the guidewire.logDirectory and that
uses a -D JVM property to set the server ID.
For example, enter something similar to the following:
log4j.appender.DailyFileLog.File=${guidewire.logDirectory}/${gw.serverid.noroles}-
pclog.log
The use of the serverid.noroles property suppresses the names of the roles associated with each server.
Otherwise, PolicyCenter lists the server roles associated after the server ID in the log name.
3. Start each server using the system properties that set the environment and server ID values.
For example, if using development Jetty test servers, use the following commands:
gwb runServer -Denv=test -Dserverid=testserver1
gwb runServer -Denv=test -Dserverid=testserver2
As each server starts, the server writes a log file to the common log file directory specified by the value of
guidewire.logDirectory. Each log file includes the value of serverid in the log file name.
In this example, after starting the servers, the /tmp/gwlogs/PolicyCenter/logs directory contains two log
files:
• testserver1-pclog.log
• testserver2-pclog.log
Next steps
If you edit file config.xml, you must rebuild and redeploy PolicyCenter for the changes to take effect. If you update
the logging configuration file, you must also reload this file before your changes take effect.
Dynamic Changes to Logging Configuration

It is possible, under certain circumstances, to make dynamic changes to the PolicyCenter logging configuration
without having to redeploy PolicyCenter. Changing the database logging level dynamically does not make any
change to existing database connections. The new logging level only take effects for new database connections.
For example, if the database log level is set to debug, PolicyCenter logs all SQL statements. However, if you set the
debug logging level dynamically, PolicyCenter only logs SQL statements for new connections created within the
connection pool. For an existing connection, dynamically changing the logging level to debug has no affect.
Logging levels that you change dynamically remain in effect only while the server is running. If you restart the
server, PolicyCenter resets the logging behavior to what file logging.properties specifies.
Note: To set the logging level consistently for all database connections, you must set the log level in
logging.properties and restart the server.
Setting the Logging Configuration Dynamically

To reset the logging configuration dynamically, do one of the following:
• Update file logging.properties and reload the file.
• Reset the logging level for a given logging category temporarily.
You must use one of these methods to propagate logging configuration changes dynamically. Otherwise, you must
rebuild your PolicyCenter WAR or EAR and deploy it to the application server for the configuration changes to take
effect.
Reloading the Logging Configuration

It is possible to make an immediate change in the configuration of PolicyCenter logging without having to first
redeploy the PolicyCenter application server. You can do this in several different ways:
• Use a system_tools command option
• Use a SystemToolsAPI web service method
Either approach reloads the current logging configuration from the logging.properties file.
Reload the Logging Configuration Using a System Tools Command

Procedure
1. In the Studio Project window, update file logging.properties with your logging configuration changes.
2. Ensure that the PolicyCenter application server is running.
3. Open a command prompt in the following location in the PolicyCenter installation directory:
admin/bin
4. Enter the following at the command prompt:
system_tools -user user -password password -reloadloggingconfig
Reload the Logging Configuration Using Web Services

Procedure
1. In the PolicyCenter Studio Project window, update file logging.properties with your logging configuration
changes.
SystemToolsAPI.reloadLoggingConfig
Changing a Logging Level Temporarily

You can temporarily change the logging level for a logger by using one of the following options:
• By using the system_tools command prompt utility
• By using the SystemToolsAPI web service
• By setting the log level on the Server Tools Set Log Level info screen
Set the Log Level Using a System Tools Command

Procedure
2. Open a command prompt in the following location:
admin/bin
3. Enter the following at the command prompt:
system_tools -updatelogginglevel logger level

Result
The change in the logging level persists while the PolicyCenter application server is running only.
Set the Log Level Using a Web Service

Procedure
SystemToolsAPI.updatelogginglevel(logger,level)
Result
The change in the logging level persists while the PolicyCenter application server is running only.
Set the Log Level from PolicyCenter

It is possible to set the log level temporarily from the PolicyCenter Set Log Level screen.
Procedure
1. Log into Guidewire PolicyCenter using an administrative account.
2. Navigate to the Server Tools Set Log Level screen.
3. Select a logging level from the drop-down list.
4. Enter the new logging level for that category.
Result
The change in the logging level persists only while the PolicyCenter server is running.
Configuring Archive Logging Operations

To configure archive logging in Guidewire PolicyCenter, do the following:
• Set configuration parameter ArchiveEnabled in file config.xml to true.
• Uncomment the archiving-related logging code in file logging.properties by removing the hash mark (#) at
the beginning of the code line.
Print Details of the Archive Graph

About this task
As the PolicyCenter server starts, the server builds the archive domain graph. To print out the details of how
PolicyCenter determines the graph, perform the following steps.
Procedure
1. In the PolicyCenter Studio Project window, open file logging.properties for editing.
2. Add code similar to the following example to the file.
log4j.category.Server.Archiving.Graph=DEBUG,Console
Generate a Separate Archive Log

About this task
Guidewire recommends as a general practice that you create a separate log file for archive information.
Procedure
1. In the PolicyCenter Studio Project window, open file logging.properties for editing.
2. Add code similar to the following example to the file.
log4j.category.Server.Archiving=INFO, ArchiveLog
log4j.appender.ArchiveLog.File=/tmp/gwlogs/PolicyCenter/logs/archivelog.log
PolicyCenter Logging Example

To configure an archive log, add code similar to the following example in file logging.properties.
##### Archiving Loggers #####
# Root archiving logger

log4j.category.Server.Archiving=DEBUG
# PolicyPeriod level detail logger

log4j.category.Server.ArchivePeriodDetail=DEBUG
# Successfully archived policy logger

log4j.category.Server.Archiving.Success=INFO
# Policy domain graph logger

log4j.category.Server.Archiving.Graph=DEBUG, ArchiveLog, Console
# Archiving upgrade process logger

log4j.category.Server.Archiving.DocumentUpgrade=INFO
log4j.appender.ArchiveLog=org.apache.log4j.RollingFileAppender
log4j.appender.ArchiveLog.encoding=UTF-8
log4j.appender.ArchiveLog.File=/tmp/gwlogs/PolicyCenter/ArchivedPolicyTerms.log
log4j.appender.ArchiveLog.MaxFileSize=1024KB
log4j.appender.ArchiveLog.MaxBackupIndex=10
log4j.appender.ArchiveLog.layout=org.apache.log4j.PatternLayout
log4j.appender.ArchiveLog.layout.ConversionPattern=
%-15.15X{server} %-8.24X{userID} %d{ISO8601} %p %m%n
Batch Processes that Generate Archive Data

In Guidewire PolicyCenter, the following batch processes generate archive-related data:
• Archive Policy Term batch processing
• Retrieve Policy Terms batch processing

part 2
Server Administration
chapter 3
PolicyCenter Server Configuration
This topic discusses ways to set up your PolicyCenter server environment.
Important PolicyCenter Server Configuration Files

PolicyCenter uses the following files to manage application and database functionality.
File Description
config.xml File config.xml contains global system parameters that you use to control the behavior of Guidewire
PolicyCenter. These configuration parameters govern large‐scale system options, such as authentication, serv‐
er clustering, and the business calendar. You access file config.xml in Guidewire Studio under configura-
tion→config.
database- File database-config.xml stores database connection information and Data Definition Language (DDL) op‐
config.xml tions. You access file database-config.xml in Guidewire Studio under configuration→config.
See also
For more information on file config.xml and basic application configuration, see the following:
• “Configuration” on page 330
• Configuration Guide
For more information on file database-config.xml and basic database configuration options, see the following:
• “Database Configuration” on page 205
• “Database Maintenance” on page 237
Understanding the PolicyCenter Server Environment

During startup, PolicyCenter calculates key environment properties. These property values describe the environment
in which the server runs. After you set environment properties, you can access these properties through the
following methods:
• PolicyCenter commands
• Gosu code
• Java code
PolicyCenter Server Configuration 41
It is possible to specify environment property values either through JVM (Java Virtual Machine) options or through
the <registry> element:
• You define environment properties in the <registry> element in file config.xml.
• You set JVM options through command line options as you to start a PolicyCenter server.
Depending on how many PolicyCenter servers your environment requires, you might find it necessary to adjust the
environment properties significantly. You can use environment properties together with configuration file
config.xml to specify and control one or multiple PolicyCenter server environments.
See also
• “Important PolicyCenter Server Configuration Files” on page 41
• “Understanding the Configuration Registry Element” on page 42
• “JVM Options and Server Properties” on page 46
• “Configuration Parameters by Environment” on page 49
Understanding the Configuration Registry Element

It is possible to set the PolicyCenter server environment using the <registry> element in config.xml. Through the
<registry> element, you can define the following:
• The set of server roles that are valid for this cluster.
• The set of server roles that are valid for each individual server instance in the cluster.
• The environment in which a specific server operates, for example, a development environment or a production
environment.
Also through the <registry> element, you can redefine certain system properties that you specify at server startup.
See also
• “The Configuration Registry Element” on page 42
• “Example Syntax for Registry Server Element” on page 44
• “Assigning Server Roles to PolicyCenter Cluster Servers” on page 45
• “Defining a New Server Role” on page 46
The Configuration Registry Element

The <registry> element in file config.xml has the following syntax.
<registry roles="role1, role2, …" >

<server env="environment" roles="role1, role2, …" serverid="serverID" />
<systemproperty name="name" value="value" default="default" />
</registry>
File config.xml contains exactly one required <registry> element. The <registry> element can contain zero to
many <server> and <systemproperty> elements.
The attributes on the various elements have the following meanings.
Element Attribute Required Description

registry roles Yes List of valid server roles.
server env No The name of the environment in which the server operates. The default is null. The
value of env is immutable while the server is running.
42 chapter 3: PolicyCenter Server Configuration


serverid No Unique server name of a server instance in the cluster. If you do not specify this value
explicitly, PolicyCenter sets it to the hostname of the PolicyCenter server. Log entries
display only the first 10 characters of the serverid value.
roles No Specific role or roles assigned to the server specified by serverid.
systemproperty name Yes Specifies the name of the Java -D system property that you want to redefine. This val‐
ue can be one of the following:
• env
• serverid
If you run multiple server instances on the same host machine, define a serverid
value for each server instance with a separate systemproperty entry.
value Yes New name of the system property.
default Yes Default property value to use if you do not specify a value at the command prompt.
The following code shows an example <registry> element.
<registry roles="batch, scheduler, workqueue, messaging, startable, ui, custom1">

<server env="prod" serverid="p0001" roles="batch, scheduler, workqueue, messaging" />
<server env="prod" serverid="p0002" roles="batch, workqueue, messaging, startable" />
<server env="prod" serverid="p0003" roles="ui, custom1" />
<server env="test" serverid="t0001" roles="batch, scheduler, workqueue, messaging, startable,
ui, custom1" />
<systemproperty name="env" value="my.env" default="production"/>
<systemproperty name="serverid" value="my.id" default="myDefault"/>
</registry>
Notice that:
• The set of valid server roles includes a custom role (custom1).
• The <server> elements define three separate server instances in the production environment, each of which has
specific server roles.
• There is a single server instance in the test environment that has all server roles.
• There are two system property redefinitions, one for env and one for serverid.
Property serverid.noroles
In standard usage, the value of property serverid includes also the list of server roles defined for that server. If you
want to use this property in code, without the list of server roles, use serverid.noroles instead. For example,
instead of using gw.pc.serverid, use gw.pc.serverid.noroles to suppress the list of server roles associated with
this server ID.
Environment and Logging

The env property has special significance for logging behavior. If the env value is non-null, PolicyCenter tries to
obtain the logging configuration from a config/logging/env-logging.properties file. If this file does not exist
or if env is null, then the logging configuration is taken from the default logging.properties file.
Creating Special Server Roles

It is possible to define new server roles by adding the role to the list of roles specified by the roles attribute on the
<registry> element. For an example of how to define and use a specialized server role, see “Work Queues and
Server Roles” on page 95.
Accessing System Properties in Code

Gosu class gw.api.system.server.ServerUtil contains methods for working with system properties associated
with servers. See the Integration Guide for information on how to use this library.
See also
• “Server Roles” on page 139
Example Syntax for Registry Server Element

The following example illustrates how to use the <server> subelement of the <registry> element in config.xml
to set various server properties.
<registry roles="…">
<server env="production" serverid="prodserver" roles="batch, messaging, workqueue" />
</registry>
Notice that the defined <server> element:

• Associates the prodserver server with a production environment.
• Assigns the prodserver server the batch, messaging, and workqueue server roles.
PolicyCenter shows the serverid and role values for each server in a PolicyCenter cluster on the Server Tools
Cluster Members and Components screens.
It is also possible to set server system properties using a -D JVM option at server start up from a command prompt.
For example, you can use JVM options to set the environment variable (env) and the server ID (serverid) at
runtime using the following syntax. The option syntax varies by server type.
QuickStart (Jetty) -Denv=…

-Dserverid=…
Tomcat -Dgw.pc.env=…
-Dgw.pc.serverid=…
If you are starting the Jetty development server in Guidewire Studio™ for PolicyCenter, use the syntax for Tomcat
in the Run - Server configuration dialog, for example:
-Dgw.pc.serverid=testServer
For more information, see “Start the Application Server from Guidewire Studio for PolicyCenter” on page 52.
How PolicyCenter uses a ‐D JVM option

The values of env and serverid are immutable while the server is running.
PolicyCenter determines the value of a -D option in the following manner, using -Dserverid (on Jetty) as an
example:
• If you specify a -Dserverid=prodserver JVM option at the command prompt at server startup, PolicyCenter
sets the value of serverid for that server to prodserver.
• If you do not specify a -Dserverid JVM option at server start, PolicyCenter checks the server registry for a
serverid value defined by a server entry. If found, PolicyCenter uses that value. In the example, the serverid
value is prodserver.
• If you do not specify the JVM option, and no serverid value defined by a server entry exists, PolicyCenter sets
serverid to the host name of the computer. Under some extreme security settings, this value is not available, in
which case PolicyCenter sets the serverid to localhost.
Note: Log entries display only the first 10 characters of the serverid value.
See also
• “Cluster Members and Components” on page 356
Example Syntax for the Registry System Property Element

The following example illustrates how to set the server environment using the <systemproperty> element of the
<registry> element in config.xml.
<registry roles="…" >

<systemproperty name="env" value="my.env" default="production"/>
</registry>
Notice that the defined <systemproperty> element:

• Redefines the name of the env property to be my.env. Thus, in places in which you formerly used env, you now
use my.env. For example, -Denv becomes -Dmy.env.
• Sets a default value (production) for the system property.
To use this system property, you specify its value as a -D JVM option at server startup. The option syntax varies by
server type.
QuickStart (Jetty) -Dmy.env=…

Tomcat -Dgw.pc.my.env=…
The value of my.env is immutable while the server is running.

PolicyCenter determines the value of the redefined env system property in the following manner, using -Dmy.env
(on Jetty) as an example:
• If you specify the -Dmy.env=test JVM option at the command prompt at server startup, PolicyCenter sets the
value of env to test.
• If you do not specify a -Dmy.env option at server start, PolicyCenter sets env to the value of default that you
specified in the systemproperty, in this example, production.
• If you do not set the value of my.env either through a default registry property or with a JVM option,
PolicyCenter sets the value of the env property to null.
PolicyCenter ignores any attempt to set the environment property through the command prompt if you have
previously defined that property using a <systemproperty> element. For example, suppose that you set the
following <registry> element in config.xml:
<systemproperty name="env" value="my.env" default="standalone" />
Then, at server startup, you specify a -Denv="test" JVM option. PolicyCenter ignores any -Denv option that you
specify on the command prompt and sets the env value to standalone.
See also
Assigning Server Roles to PolicyCenter Cluster Servers

You assign specific server roles to individual servers using the roles attribute on the <server> element (on
<registry>) in config.xml. For example, suppose that the PolicyCenter cluster contains the following servers.
Server name Assigned roles

prodserver1 ui
prodserver2 ui
prodserver3 batch, workqueue, scheduler, messaging, startable

Server name Assigned roles

prodserver4 batch, workqueue, scheduler, messaging, startable
testserver1 batch, workqueue, scheduler, messaging, startable, ui
The <registry> element in config.xml defines these servers and server roles as follows:
<registry roles="batch, workqueue, scheduler, messaging, startable, ui" >

<server env="prod" roles="ui" serverid="prod1" />
<server env="prod" roles="ui" serverid="prod2" />
<server env="prod" roles="batch, workqueue, scheduler, startable" serverid="prod3" />
<server env="prod" roles="batch, workqueue, scheduler, startable" serverid="prod4" />
<server env="test" roles="batch, workqueue, scheduler, messaging, startable, ui" serverid="test1" />
</registry>
Notice the following:

• The roles attributes on <registry> defines the valid server roles for the PolicyCenter cluster servers.
• The env attribute on <server> defines a server environment for each individual server.
• The roles attribute on <server> defines the server roles assigned to each individual server.
• The serverid attribute on <server> associates a specific server with one or more specific server roles.
Server Mode and Server Roles

PolicyCenter servers in development mode default to having all existing, defined server roles. PolicyCenter servers
in test or production mode default to having no assigned server roles. Thus, you must manually add any desired role
to a test or production PolicyCenter server using one of the following methods:
• Use the <registry> element to assign roles to a production server.
• Use the -Dserverid JVM option at server startup to assign a specific role to the server.
See “JVM Options Specific to the runServer Build Command” on page 48 for information about the -Dserverid
JVM option.
Defining a New Server Role

You define server roles using the roles attribute on the <registry> element in file config.xml. In the base
configuration, Guidewire defines the following server roles:
<registry roles="batch, workqueue, scheduler, messaging, startable, ui" />
To add a specialized server role, say, one to use in managing activities, you need merely to add the new server role
to the list of roles:
<registry roles="batch, workqueue, activity, scheduler, messaging, startable, ui" />
See also
• “Defining a New Work Queue Role” on page 95
JVM Options and Server Properties

It is possible to set certain PolicyCenter server properties by using -D JVM options on the PolicyCenter gwb build
commands. Some of these JVM options work with all the gwb build commands. Other JVM options work only with
the gwb runServer build command. In all cases, the exact manner in which you define and use the JVM option
depends on the type of server involved.
Once set, a server property is immutable while the server is running.
See also
• “The Configuration Registry Element” on page 42

• “Example Syntax for Registry Server Element” on page 44
• “Example Syntax for the Registry System Property Element” on page 45
JVM Options for gwb Build Commands

The following JVM options work with many of the PolicyCenter gwb build commands.
JVM option syntax Description

-Denv=aaaa Starts the PolicyCenter server using the specified environment variable (env).
-Dgw.passthrough.systemProperty=dddd Starts the PolicyCenter server using the provided system property value. The op‐
tion parameters have the following meanings:
• systemProperty – The name of a system property defined in the <registry>
element in file config.xml.
• dddd – The value of the system property.
JVM env string value always lower case
If you use a JVM option to set the server environment, PolicyCenter converts the provided string to lower case
automatically. For example, the conversion to lower case occurs if you use the following JVM options:
• Jetty - -Denv
• Tomcat - -Dgw.pc.env
Thus, if you enter -Dgw.pc.env=TEST, PolicyCenter converts the string TEST to the string test. Be aware that the
automatic conversion of the env string to lower case can cause a referenced env value to not match the defined
environment name.
Reference table of core gwb commands
It is possible to use the -Denv and the -Dgw.passthrough JVM options with some, but not all, of the gwb build
commands. The following table indicates whether the listed JVM command options work with the core gwb
commands (tasks).
Core task Can use JVM option Cannot use JVM option
clean •
cleanIdea •
codegen •
compile •
dropDb •
genDataDictionary •
idea •
runServer •
stopServer •
studio •
JVM Option Examples

For example, to pass -DmySystemProperty=someValue to build command dropDB, use the following command
option.
gwb dropDb -Dgw.passthrough.mySystemProperty=someValue
Then, to make the dropDB command specific to a test environment, use the following command (for a Jetty server).
gwb dropDb -Denv=test -Dgw.passthrough.mySystemProperty=someValue
See also
• “Setting JVM Options in PolicyCenter” on page 48
JVM Options Specific to the runServer Build Command

The following JVM options work with the PolicyCenter gwb runServer command only.
JVM option syntax Description

-Dgw.port=nnnn Starts the PolicyCenter server on the specified port (nnnn). The server URL reflects this port num‐
ber, for example:
localhost:nnnn/pc/PolicyCenter.do
-Dgw.server.mode=xxxx Starts the PolicyCenter server in the specified server mode. Valid values are:
• dev
• prod
The default is dev.
-Dserverid=aaaa Sets the server ID, and possibly, one or more server roles for the PolicyCenter server:
-Dserverid=aaaa#bbbb • Server ID – Without the hash mark (#), aaaa represents a server ID only.
• Server role – With the hash mark, #bbbb assigns the bbbb server role to the server with aaaa
server ID. Use a comma‐separated list, with no spaces, to list multiple server roles.
The exact JVM syntax to use depends on the server type, for example:
• Quickstart (Jetty) – Use -Dserverid=aaaa
• Tomcat – Use -Dgw.pc.serverid=aaaa
See also
• “Setting JVM Options in PolicyCenter” on page 48
Setting JVM Options in PolicyCenter

It is possible to set server system properties using the -D JVM option syntax. How you set a command option
depends on the server type:
JBoss Pass the options as arguments to the JBoss run script.

QuickStart (Jet‐ Set the options at server start using the following syntax:
ty) gwb runServer -Denv=…
gwb runServer -Dserverid=…
Tomcat Set the options using the CATALINA_OPTS environment variable. Use the following syntax:
-Dgw.pc.env=…
-Dgw.pc.serverid=…

WebLogic Edit the startManagedWebLogic file.

WebSphere Open the Administrative Console and add the option to Generic JVM arguments. If you have multiple servers, set
the option for each server.
See also
• “JVM Options for gwb Build Commands” on page 47
• “JVM Options Specific to the runServer Build Command” on page 48
Configuration Parameters by Environment

Typically, you need to support more than one server environment. Guidewire recommends you maintain at least
development, test, and deployment environments. So that you do not have to change configuration parameters each
time you switch between environments, PolicyCenter provides the ability to set a configuration parameters for a
specific environment.
Environment-specific parameters can reference environment properties to indicate in which environment they are
valid. You specify the environment for a parameter in config.xml by adding one or both of an env or server
attribute. For example:
<param name="BusinessDayStart" value="9:00 AM" server="dev1" />

<param name="BusinessDayStart" value="7:00 AM" env"=test" />
<param name="BusinessDayStart" value="8:00 AM"/>
Then, you can have PolicyCenter use the environment-specific parameter by specifying the environment in JVM
options at server startup. Continuing the example, to have BusinessDayStart resolve to 7:00 a.m., specify the test
environment in your JVM options:
gwb runServer -Denv=test
However, if PolicyCenter resolves serverid to dev1, PolicyCenter sets BusinessDayStart to the 9:00 a.m. value.
If you define environment-specific parameters, PolicyCenter applies the setting if either the env or server resolves
to a known value. For example, suppose that you specify the BusinessDayStart parameter as follows:
<param name="BusinessDayStart" value="9:00 AM" env="test" server="prodserver" />

<param name="BusinessDayStart" value="8:00 AM" />
PolicyCenter sets BusinessDayStart to 9:00 a.m. if either env resolves to test or serverid resolves to
prodserver. Thus:
• If PolicyCenter resolves env to test and serverid to chicago, the BusinessDayStart is 9:00 a.m.
• Similarly, if PolicyCenter resolves env to production and serverid to prodserver, the BusinessDayStart is
also 9:00 a.m.
• If env does not resolve to test and server does not resolve to prodserver, PolicyCenter uses the default
BusinessDayStart of 8:00 a.m.
For a list of configuration parameters, including information about which parameters can be set by environment, see
the Configuration Guide.
See also
Default Configuration Values and Environment Properties

At startup, PolicyCenter requires many parameters. Consider this carefully if you specify parameters by
environment. In some cases, you might want to specify a configuration parameter without any env or server
attribute to assure the parameter always resolves to some value. In the following example, the last line always
resolves to a known value if the other two parameter lines do not resolve:
<param name="ClusteringEnabled" value="false" server="chicago" />

<param name="ClusteringEnabled" value="true" env="test" />
<param name="ClusteringEnabled" value="true"/>
The last line setting in this example acts as the default value for the parameter. Of course, you might want the server
to start only if a certain environment is available. In this case, a default is inappropriate.
See also
• “Configuration Parameters by Environment” on page 49

chapter 4
PolicyCenter Server Administration
This topic discusses the PolicyCenter server, run levels, modes, monitoring servers, and server caching.
Starting Guidewire PolicyCenter

How you start PolicyCenter depends on the application server type.
Application Server More information...

QuickStart (Jetty) from a command prompt “Start PolicyCenter on QuickStart (Jetty)” on page 51
QuickStart (Jetty) from PolicyCenter Studio “Start the Application Server from Guidewire Studio for PolicyCenter” on page
52
JBoss Installation Guide
Tomcat Installation Guide
WebLogic Installation Guide
WebShpere Installation Guide
Start PolicyCenter on QuickStart (Jetty)

About this task
Guidewire recommends the following sequence in starting PolicyCenter on the QuickStart server.
Procedure
1. Open a command prompt and navigate to the root of PolicyCenter application directory.
2. Run the following command to compile the needed application resources and move them to the correct
location in the application server:
gwb compile
3. Run the following command to start the application server.
gwb runServer -x compile
PolicyCenter Server Administration 51

There is a dependency between the runServer command and the compile command. Guidewire recommends
that you always use the -x compile option with the runServer command to remove that dependency after
you initially run the compile command. Otherwise, PolicyCenter must first verify what resources, if any, need
to be recompiled, then perform an incremental recompile of those resources before starting the server.
See also
Start the Application Server from Guidewire Studio for PolicyCenter

It is possible to start the PolicyCenter development server (Jetty) from Guidewire Studio™ PolicyCenter.
Procedure
1. Navigate to the following location in Guidewire Studio™ for PolicyCenter, using the menu bar at the top of
the screen:
Run→Run...
Studio opens a Run drop-down with a list of server-related options.
2. Select Edit Configurations.. from the list.
Studio opens the Run - Server dialog.
3. Select Server in the left-hand navigation pane and verify the configuration options set for the server.
It is possible to modify the base configuration server settings by adding additional VM options. If you do so,
use the following format (using server ID an example):
-Dgw.pc.serverid=testServer
4. Click Run.
Studio opens a pane at the bottom of the screen to display the server log. This area also has controls (icons) for
stopping and starting the server.
Stopping GuidewirePolicyCenter
Before you stop Guidewire PolicyCenter, you must stop all work queues. Distributed workers run on daemon
threads. As the JVM (Java Virtual Machine) exits, it destroys these threads. This can cause issues if the JVM
destroys a thread while that thread is processing a work item. For example, suppose that a work queue calls a plugin
that makes a blocking call to an external system or otherwise take a long time to return. In that case, if you do not
shut down the work queue threads correctly, it is possible to end up with inconsistent data.
Stop Guidewire PolicyCenter

Procedure
1. Login into Guidewire PolicyCenter as a user with administrative privileges.
2. Within PolicyCenter, press ALT+SHIFT+T to display the Server Tools screens.
3. Navigate to Batch Process Info screen:
a. For any process that has a Last Run Status that indicates that the process is running, click the Stop button in
the Action column.
b. For any process that has a Next Scheduled Run time that is before the time that you intend to stop
PolicyCenter, click Stop in the Schedule column.
All processes must have a Status of Completed before you stop the PolicyCenter server.
52 chapter 4: PolicyCenter Server Administration
4. Stop Guidewire PolicyCenter:

• To stop PolicyCenter in a production environment, stop the server on which it is running.
• To stop PolicyCenter in a development environment, run the following command from the PolicyCenter
installation directory:
gwb stopServer
Server Startup Tests

The PolicyCenter server performs a series of tests during start up. For some of these tests, a test failure prevents the
server from starting. For other tests, a test failure is simply a warnings and PolicyCenter permits the server to start.
In many cases, these checks warn about potential problems with the archive and domain graphs, but which might not
be an issue depending on business logic.
See also
• “Domain Graph Info” on page 332

• “Consistency Checks” on page 333
• Product Model Guide
Server Modes
Server mode determines what functionality is available at various server run levels. All PolicyCenter server types,
except for QuickStart, can run in any of the following server modes:
• Development
• Test
• Production
PolicyCenter starts in production mode on all supported servers by default, except for the QuickStart server.
PolicyCenter on the QuickStart server always runs in development mode. You cannot run PolicyCenter on the
QuickStart server in production or test mode.
See also
• “Server Test Mode” on page 54

• “Server Run Levels” on page 55
• “Setting the Server Mode” on page 54
• “Set the QuickStart Run Level at Server Start” on page 57
Important Server Mode Caveats

The following caveats are important in working with the Guidewire development and production databases:
• It is not permissible to start a server in development mode using a production mode database.
• It is not permissible to start a server in production mode using a development mode database. Starting the server
in production mode expressly does not upgrade the development mode database to production mode.
Server Modes as a Safety Feature

Guidewire provides the server modes as a safety precaution so that it is not possible to use development tools on a
production server. Some system functions are useful for development, but are not appropriate, or, are dangerous if
used in a production environment. Thus:
• In development and test modes, it is possible to access both the Server Tools and Internal Tools screens.
• In production mode, it is possible to access the Server Tools screens only.
For more information on these tools, see “Server Tools” on page 321 and “Internal Tools” on page 381.
Setting or resetting the system time is also not safe to do in a production environment. The use of the
ITestingClock plugin is critical for testing time-sensitive processes. You can also use this plugin to modify the
current time in a running server for demonstrations. However, it is possible for the use of this plugin in a production
environment to cause disastrous results. Therefore, you can only use this plugin if the server is in development or
test mode.
It is not possible to deploy certain types of product model changes to a server set to production mode. Guidewire
provides these locks to protect data integrity. Any time that is important to upgrade (or preserve) the database, place
the server into production or test mode to minimize the risk of data corruption.
See also
Server Test Mode

Other than the exceptions listed in the following table, server test mode is identical to server production mode.
Functionality Test mode Production mode

System clock You can adjust the testing system clock by using the Not available
setCurrentTime method on the ITestingClock plugin.
See also
• “Testing System Clock” on page 382
• Integration Guide
Server Tools screens Available Available
Internal Tools screens Available Not available
Console and log file PolicyCenter prints a message to the console during startup indi‐ PolicyCenter prints a message to the
cating that the server is running in test mode. console during startup indicating
that the server is running in produc‐
tion mode.
PolicyCenter title bar The title bar for a browser connected to PolicyCenter indicates Not applicable
that PolicyCenter is in test mode.
Setting the Server Mode

You can change the server mode while using any server type except QuickStart. The QuickStart server always runs
PolicyCenter in development mode.
To set the server mode at server start, use the following system parameter:
-DserverMode={dev|prod|test}
To change the mode of a running server, restart the server and set the -DserverMode parameter to dev, test or prod.
PolicyCenter ignores this parameter on the QuickStart server.
Determining the Server Mode

You can determine the server mode through one of the following methods:
• By reading the console log as you start PolicyCenter
• By checking the title bar of the browser in which PolicyCenter is running
Note: Whenever the server starts in development mode, PolicyCenter logs a warning.
Server Run Levels

Guidewire PolicyCenter supports the concept of a server run level. A run level provides the ability to start the
PolicyCenter server with a specific level of functionality. For example, during a full application and database
upgrade, Guidewire recommends that you start PolicyCenter with limited functionality, in MAINTENANCE mode.
Guidewire specifies the server run levels in several different ways:
• A numeric run level, ranging from 0 to 5, used with the development QuickStart server.
• A server run level name that corresponds to the numeric run levels
• A system run level name used with the administrative system tools to set the server to one of the server run levels
The following table lists the correspondence between the various ways of describing the server run levels.
QuickStart Server run level System run level

0 NONE —
1 GUIDEWIRE_STARTUP —
2 SHUTDOWN —
3 NODAEMONS maintenance
4 DAEMONS daemons
5 MULTIUSER multiuser
The following list describes each type of server run level in more detail.
Type Description
QuickStart run Set at QuickStart server start using the following command, with n being a specific run level number:
level gwb runServer --run-level n
See “Set the QuickStart Run Level at Server Start” on page 57.
Server run level Shown in the server log. The server starts at level 0 and proceeds to move through each server run level in
the sequence until arriving at the requested run level.
System run level Set through command prompt system_tools options, for example:
system_tools -maintenance
See “System Tools Command” on page 393 for details.
Server run levels are independent of the server mode. The combination of mode and run level determines the
availability of functionality, such as the user interface and web services.
See also
• “Server Modes” on page 53
• “Server Modes and Server Run Levels” on page 56
• “Place the Server in Maintenance Mode” on page 59
Server Modes and Server Run Levels

The following table shows which functionality is available for the possible combinations of server modes and server
run levels. For a description of the numeric values used with the development QuickStart server, see “Server Run
Levels” on page 55.
Quick‐ Server run level Production mode Development mode

Start
0 NONE • Nothing available. • Nothing available.
1 GUIDEWIRE_STARTUP • User interface not available. • User interface not available.
• Web services not available. • Web services not available.
• Database not available. • Database not available.
2 SHUTDOWN • User interface not available. • User interface not available.
• Web services not available. • Web services not available.
• Database not available. • Database not available.
3 NODAEMONS • User interface not available. • User interface available. All logins allowed.
• Web services available. • Web services available.
• Staging table loading available. • Staging table loading available.
• Product Model checked for illegal • Batch processes available.
changes.
• Batch processes available.
4 DAEMONS • User interface not available. • User interface available. All logins allowed.
• Product Model checked for illegal • Work queues (including workflow) available.
changes. • Workflow Stat Manager available.
• Work queues (including workflow) • Scheduler available.
available. • Daemons started by PolicyCenter, such as those
• Workflow Stat Manager available. used to dispatch messages.
• Scheduler available.
• Daemons started by PolicyCenter,
such as those used to dispatch
messages.
5 MULTIUSER • User interface available. All logins • User interface available. All logins allowed.
allowed. • Server Tools available to all users if
• Server Tools available for users with EnableInternalDebugTools is set to true in
admin permission only. config.xml
• Internal Tools not available. • Internal Tools available.
• Product Model checked for illegal
changes.
Setting the Quick Start Server Run Level

It is possible to set the run level of a server to one of the following system run levels:
• daemons
• maintenance
• multiuser
If you run PolicyCenter in a clustered environment, you cannot place all the computers in a particular run level with
a single command. Instead, you must run the command individually on each cluster member.

See also
• “Server Modes and Server Run Levels” on page 56
• “System Tools Command” on page 393
Set the QuickStart Run Level at Server Start

Procedure
1. Open a command prompt and navigate to the PolicyCenter installation directory.
2. Run the following command:
gwb runServer -run-level n
The value of n is the numeric value of the run level as defined in “Server Run Levels” on page 55.
Set the Server Run Level Through System Tools

Procedure
2. Open a command prompt and navigate to the following location in your PolicyCenter installation directory:
admin/bin
3. Enter one of the following commands:
system_tools -user user –password password –daemons

system_tools -user user –password password –maintenance
system_tools -user user –password password –multiuser
You must supply the username (user) and password (password) for a user with administrative privileges on
the PolicyCenter server. The run level is a value as defined in “Server Run Levels” on page 55.
Set the Server Run Level Through Web Services

Procedure
SystemToolsAPI.setRunLevel
Next steps
If you run PolicyCenter in a clustered environment, you cannot place all the computers in a particular run level with
a single method call. Instead, you must call the method individually on each cluster member.
Determining the Server Run Level

You can determine the server run level through any of the following methods.
Type More information

System tools “Use System Tools to Determine the Server Run Level” on page 58
Web services “Use Web Services to Determine the Server Run Level” on page 58
Server ping “Using the PolicyCenter ping Utility” on page 153
The returned message indicates the server run level. The possible responses are:
• MULTIUSER
• DAEMONS
• MAINTENANCE
• STARTING
See also
Use System Tools to Determine the Server Run Level

Procedure
2. Open a command prompt and navigate to the following location in the PolicyCenter installation directory:
admin/bin
3. Enter the following command:
system_tools -user user -password password –ping
the PolicyCenter server.
Use Web Services to Determine the Server Run Level

Procedure
SystemToolsAPI.getRunLevel
See also
Server Maintenance Mode

Periodically, you need to perform maintenance on PolicyCenter, for example, importing new security roles. To
prevent users from logging into PolicyCenter during these activities, place PolicyCenter into the maintenance run
level.
The maintenance run level effectively disables the PolicyCenter web interface if the server is in production mode.
PolicyCenter stops allowing new user connections and halts existing user sessions to production mode instances
while running at the maintenance run level.
PolicyCenter still allows connections made through APIs or command prompt tools for any daemons with a
minimum run level equal or lower than NODAEMONS. Restricting the run level permits integration processes to
proceed without interference from non-administrator users.
See also
• “Set the QuickStart Run Level at Server Start” on page 57
• “Set the Server Run Level Through System Tools” on page 57
• “Set the Server Run Level Through Web Services” on page 57
Place the Server in Maintenance Mode

Procedure
1. Open a command prompt and navigate to the following PolicyCenter installation directory:
admin/bin
system_tools -password password -user user -maintenance
the PolicyCenter server.
Monitoring Server Performance in PolicyCenter

Use standard operating system tools to monitor memory usage, CPU usage, and disk space to verify that the servers
run smoothly. In particular, monitor disk space for log files, so PolicyCenter does not run out of disk space for logs.
Archive and truncate system logs periodically to prevent the PolicyCenter logs from growing too large.
If the server crashes with the following JVM error, increase the maximum heap size (-Xmx setting) of the JVM.
Internal Error (53484152454432554E54494D450E43505001A8)
See also
• “Monitoring Server Status with WebSphere” on page 59
• “Monitoring Cluster Health” on page 151
• Documentation specific to the application server
Monitoring Server Status with WebSphere

You can monitor the status of the PolicyCenter server from the WebSphere console. To view the server’s status,
select the PolicyCenter cluster member from the main console. WebSphere displays the Show Status option if the
server is running. WebSphere also generates and displays system logs. Also, you can start an Export for Backup from
the WebSphere console. Guidewire recommends that you back up the server before performing any major system
maintenance.
Monitoring and Managing Event Messages

PolicyCenter generates a large number of events. In a PolicyCenter installation, it is often helpful, or even necessary,
for PolicyCenter to notify other applications of these events. PolicyCenter integration developers create message
destination objects that provide the means for passing information between PolicyCenter and a particular
destination. Rule writers can write Gosu rules to generate messages in response to events of interest. PolicyCenter
queues these messages and dispatches them to receiving systems by using the destination objects.
Guidewire recommends that you monitor message traffic to verify that the integration is running smoothly.
See also
• For more information about messages, including how to create message destination objects, see the Integration
Guide.
About Message Failure

If PolicyCenter receives an unrecoverable or unexpected exception from a send attempt, or reaches the retry limit, it
does not send messages to that destination until you clear the error. If PolicyCenter receives a processing error that is
not retryable, PolicyCenter also suspends the destination and waits for you to clear the error.
To clear a message error, do one of the following:
• Manually retry to send the message
• Skip the message
You can retry or skip messages:
• Through the Message Queues screen available to application administrators
• Through messaging_tools command prompt options
If PolicyCenter becomes completely out of synchronization with an external system, such that skipping or retrying a
message is insufficient to re-synchronize the two systems, re-synchronize the entire destination. A re-
synchronization causes PolicyCenter to drop all pending and failed messages and resend all the messages associated
with a particular item.
See also
• “Message Queues Screen” on page 61
• “Messaging Tools Command” on page 391
Messages Screen
Use the PolicyCenter Messages screen to review information about messages. An administrator can access the
Messages screen in PolicyCenter by using the following navigation path:
Administration→Monitoring→Messages
You can search and filter message by any of the following:
• Destination
• Transaction Type
• Transaction Number
• Product
• Policy Number
• Account Number
• Message Status – For each status, the corresponding status is:
◦ Failed messages – MessageStatus.ERROR_STATES
◦ Messages needing retry – MessageStatus.RETRYABLE_STATES
◦ Unfinished messages – Messages that do not fall into ERROR_STATES or RETYRABLE_STATES
The messages in Results have the following information:
• Transaction Number
• Transaction Type
• Policy Number
• Product
• Primary Named Insured

• Message Status
• Error Description
See also
Message Queues Screen

PolicyCenter lists each message destination in the Message Queues screen. To open this screen, click the Administration
tab and then navigate in the left Sidebar to Monitoring→Message Queues. The screen contains cumulative information
about message destinations.
From this screen, you can select a message destination and perform the following actions:
Suspend
Click to suspend the operation of the selected message destination.
Resume
Click to resume the operation of the selected message destination if that destination is in a state of suspension.
Restart
Click to restart the operation of a selected message destination.
You can also restart messaging by clicking Restart Messaging Engine.
If a message destination is running correctly, you do not see any accumulation of information in the columns on this
screen. If there is a problem and messages begin to accumulate, you can drill down into a message destination by
clicking the destination name. This action opens a Destination screen. From the Destination screen, you can see
additional detail about any issues with a destination. This information can assist you in diagnosing the error. In
particular, you can use the Error Message column to see the possible cause of a particular issue.
Destination Screen
The Destination screen lists all failed or in-process messages for an account for all destinations. You can search for a
particular account and then open the account’s detail view. From this screen, you can select one or more objects and
instruct PolicyCenter to skip, retry, or re-synchronize any message that is still in process or in a failed state.
The Detailed Statistics table column headers have the following meanings.
Column header Meaning

Safe Ordering Object Name PolicyCenter groups messages for each messaging destination based on their associated primary
object. (PolicyCenter processes messages associated with objects other than the primary object as
non‐safe‐ordered messages.)
Send Time Time the PolicyCenter sent the message.
Failed A message can fail for several reasons, for example:
• The message destination did not process the message successfully due to a processing error.
• The message destination returns a negative acknowledge (NAK) indicating that the message
delivery failed.
• The message was part of a series of messages, one or more of which failed.
Retryable Error Waiting to attempt a retry. PolicyCenter attempted to send the message but the destination threw
an exception. If the exception was retryable, PolicyCenter automatically attempts to retry the send
before turning the message into a failure. PolicyCenter attempts to send an event message several
times. Typically, you can configure the number of retries and the interval between them for an
integration. Review documentation for the specific destination to find out how to configure it.
In Flight PolicyCenter is waiting for an acknowledgment.

Column header Meaning

Unsent The message has not been sent, for example:
• The message is waiting on a prior message.
• The destination is not processing messages because the destination is in a state of suspension.
• The destination is falling behind in processing messages.
Error Message Error message returned if a message fails.
It is possible to filter the messages that show in the table by selecting a filtering characteristic from the filtering
drop-down list.
Message Handling
A PolicyCenter server reads integration messages from a queue and dispatches them to their destinations. However,
there is no guarantee that messages in the queue are ready for dispatching in the same order in which PolicyCenter
places the messages in the queue.
For example, suppose that a messaging server starts writing message 1 to the queue, and then starts writing message
2 to the same queue. It is possible that the server completes and commits message 2 while still writing message 1.
This does not, in itself, present an issue. However, if the server attempts to read messages off the queue at this
moment, then it skips the uncommitted message1 and reads message 2. You are most likely to encounter this
situation in a clustered PolicyCenter environment.
To address this situation, PolicyCenter provides the IncrementalReaderSafetyMarginMillis parameter in file
config.xml. This parameter determines how long after detecting a skipped message that PolicyCenter attempts to
read messages again. This waiting period gives PolicyCenter a chance to commit the skipped message. If it is not
possible to commit the message before the expiration of the waiting period:
• PolicyCenter assumes the message is lost and that it is not possible to commit the message.
• PolicyCenter skips the message permanently, thereafter.
For example, in the previous scenario, PolicyCenter waits 10 seconds (the default parameter value) before
attempting to read messages again, beginning with the skipped message 1. If message 1 has still not been committed
at that time, PolicyCenter skips it permanently.
Set the IncrementalReaderSafetyMarginMillis parameter sufficiently long so that the server can commit the
messages, but without prematurely marking messages as permanently skipped. As the server does not read any other
messages during this waiting period, do not set IncrementalReaderSafetyMarginMillis so long as to delay the
delivery of messages.
You can also set the following configuration parameters in config.xml to configure the messages reading
environment:
• IncrementalReaderPollIntervalMillis
• IncrementalReaderChunkSize
Access the Messaging Editor

About this task
You create and configure message environments and destinations in file messaging-config.xml.
Procedure
1. In the PolicyCenter Studio Project window, navigate to configuration→config→Messaging.
2. Double-click file messaging-config.xml to open the file in the Studio Messaging editor.
See also
Purge completed messages

About this task
If PolicyCenter sends an event message, the expected response from the destination is a positive acknowledgment
(ACK) indicating that the destination received and processed the message. PolicyCenter retains completed messages
until you purge them. Because the number of messages in PolicyCenter can become very large, Guidewire
recommends that you purge completed messages on a regular basis.
Procedure
1. Ensure that the Guidewire PolicyCenter server is running.

2. Open a command prompt in the following location in the PolicyCenter installation directory:
admin/bin
3. Enter the following messaging_tools command, replacing the variables as needed:
messaging_tools -user user -password password -purge MM/DD/YY
For example, the following command purges all completed messages received prior to 11/30/18.
messaging_tools -user su -password gw -purge 11/30/18
You must supply the username (user) and password (password) for a user with administrative privileges.
Session Timeout
PolicyCenter creates a session for each browser connection. PolicyCenter uses the server’s session management
capability to manage the session. Each individual session receives a security token that the PolicyCenter server
preserves across multiple requests. The server validates each token against an internal store of valid tokens.
You configure the timeout value for a session by setting the SessionTimeoutSecs parameter in config.xml. This
value sets the session expiration timeout globally for all PolicyCenter browser sessions.
Typically, the server determines the session timeout value according to the following hierarchy.
Level Description
Server The session timeout to use for all applications on the server if the timeout value is not set at a higher level.
Enterprise ap‐ The session timeout specified at the enterprise application level. You can specify this value at the EAR file
plication level. You can set the enterprise application session timeout value to override the server session timeout val‐
ue.
Web applica‐ The session timeout specified at the web application level. You can specify this value at the WAR file level.You
tion can set the web application session timeout value to override the enterprise application and server session
timeout values.
Application The session timeout specified in the application web.xml file. PolicyCenter does not specify a session time‐
level out in web.xml.
Application An application can override any other session timeout value. PolicyCenter uses the session timeout value
code specified by the SessionTimeoutSecs parameter in config.xml.
See also
User Session Replication

Do not attempt to replicate sessions across PolicyCenter cluster members. PolicyCenter does not implement or
support user session replication for a number of reasons, including the following:
• PolicyCenter sessions are not serializable. Therefore, you cannot replicate a PolicyCenter session, either with or
without persistence to the database.
• PolicyCenter sessions hold the user state in memory and contain large amounts of information. Guidewire
estimates that this information amounts to 1 MB of data on average for a 32-bit server and close to 2 MB for a
64-bit server. Session replication would create significant cross-member communication that is detrimental to
performance.
• PolicyCenter commits changes to the database on almost all transactions. Notable exceptions are some wizards
for which PolicyCenter commits data changes only after the user completes all necessary entries.
• PolicyCenter scales horizontally almost linearly. The implementation of a session replication solution would very
likely impede that linear scalability.
Instead, Guidewire recommends that you implement a PolicyCenter cluster consisting of multiple PolicyCenter
application instances with failover and a load balancing solution. The load balancer must implement session affinity,
meaning that it must route connections from the same user session to the same PolicyCenter server.
See also
• “Planning a PolicyCenter Cluster” on page 147
Cache Management
Objects do not remain forever present or valid in the PolicyCenter database cache. Guidewire provides several
caching mechanisms to verify that cache entries are still relevant. They are:
• Stale timeout
• Eviction timeout
• Cluster member object tracking
Stale Timeout
A stale timeout mechanism ensures that the server instance does not use old object entries excessively. An object is
stale if it has not been refreshed from the database within a configurable amount of time. Upon accessing a cache
entry, the server instance calculates the duration since the object was last read from the database. If that duration
exceeds the stale time, the server instance refreshes the cache entry from the database.
To avoid increased complexity, PolicyCenter prefers this mechanism over evicting objects upon stale timeout. You
can set a default stale time by adjusting the GlobalCacheStaleTimeMinutes parameter in config.xml.
Eviction Timeout
A PolicyCenter evict timeout mechanism removes old objects from the cache. For example, an object has an evict
time of 15 minutes and a stale time of 30 minutes. If the server uses the object a single time every 14 minutes,
PolicyCenter never evicts the cache entry, but the entry does eventually become stale.
You can set the default evict time by adjusting the GlobalCacheReapingTimeMinutes parameter in config.xml. In
the base configuration Guidewire sets the value of GlobalCacheReapingTimeMinutes to 15 minutes. The minimum
value for this parameter is 1 minute. The effective maximum value for parameter
GlobalCacheReapingTimeMinutes is the lesser of its set value or the value of GlobalCacheStaleTimeMinutes
parameter.
Cluster Member Object Tracking

Upon receipt of an inter-cluster message indicating that a cluster member changed an object value, the other cluster
members mark the corresponding entry in the cache as obsolete. The object value then becomes available for reuse.
See also
• “Guidewire PolicyCenter Cluster Installations” on page 132
• “Cache Info” on page 367
Caching and Stickiness

PolicyCenter is fully leveraging the caching mechanism if a user returns to the same server instance across different
HTTP requests. In a clustered environment, the load balancer must direct requests to the same PolicyCenter server
upon consecutive interactions. This mechanism, referred to as stickiness, enables a true horizontal scalability
solution. For more information on load balancing options for PolicyCenter, consult Guidewire Services.
Each server manages its own cache. It is possible for the same object to live in the cache of two or more servers at
the same time. Some object sets, such as users, likely live in the global cache of all servers in a cluster.
Concurrent Data Change Prevention

It is possible for different users, either on the same PolicyCenter server or on a different server, to attempt to change
data objects concurrently. To prevent data corruption, Guidewire implements a data object versioning mechanism.
During the update to an object value to the database, PolicyCenter compares a counter associated with the object to
the counter in the database. A counter value mismatch indicates that two different user changed the object
concurrently. In such case, PolicyCenter rejects the change and the cache mechanism throws a concurrent data
change exception. PolicyCenter presents the user who initiated the concurrent change with the error and reloads the
latest data. The user can then reapply the changes.
Furthermore, PolicyCenter commits changes in an atomic bundle, ensuring transactional integrity. Therefore,
PolicyCenter enforces protection against concurrent data changes across the whole transaction. This mechanism is a
standard design pattern called optimistic locking.
Concurrent data change exceptions occur only if two users modify the same object. A proper organization of the
user community avoids this. Nevertheless, if two users modify the same object, any automatic resolution carries a
significant risk of causing unwanted modifications. The optimistic locking mechanism causes very few concurrent
data change exceptions and users can easily resolve those exceptions.
Other design patterns exist for concurrent data changes. The pessimistic locking pattern prevents all other users from
modifying an object while one user or batch process is making a change. In many cases, pessimistic locking
becomes completely dysfunctional. For example, if a user or batch process cannot complete a change, the locking
mechanism blocks any other user of batch process from making a change. Pessimistic locking systems generally
become impractical. Therefore, PolicyCenter uses the optimistic locking mechanism.
Caching and Clustering

For information on how Guidewire clusters handle caching, see “Guidewire PolicyCenter Cluster Installations” on
page 132.
Cache Behavior and Performance

Guidewire generally distinguishes two types of caches:
• Server cache – A cache that is purely local to the PolicyCenter server
• Database cache – A cache used by a database to hold data retrieved from storage
Proper database caching behavior is critical to performance.
The server cache is purely local to each individual PolicyCenter server. Therefore, one server instance might contain
information on a specific object while another server might not contain that information.
For example, if a PolicyCenter user works on a policy, PolicyCenter loads corresponding objects on the server to
which the user connects. If another user must approve the action of the first user, the approver user might interact
with another PolicyCenter server. In that case, the second server likely does not have the corresponding information
in cache. Therefore, the approver might experience slower performance as the server must populate the cache.
Cache content is lost every time that you stop a PolicyCenter server. Therefore, after you start a PolicyCenter server,
expect lower performance during a ramp-up phase.
Batch processes leverage the cache mechanism. Batch jobs can work on many objects. Therefore, batch jobs can use
the cache extensively. This can have the adverse effect of prematurely evicting objects from the cache, thereby
forcing additional cache loads. If you run many intensive batch processes, Guidewire recommends that you have a
dedicated server with the batch role with no online traffic directed to it.
See also
• “Server Cache Tuning Parameters” on page 67
Cache Thrashing
Cache thrashing is a phenomenon whereby evictions remove cache entries prematurely and force additional database
reloads that are detrimental to performance. There are several cases that can lead to cache thrashing:
• A single data set can be too large to reside in the global cache. This forces the server to load the same data from
the database and subsequently evict the data, potentially thousands of times, while loading a single web page.
This results in serious performance issues.
• Some concurrent actions result in thrashing. For example, a user logs onto a server that has the batch server role.
A batch job, which can load many objects into the cache, can remove objects from the cache. This forces the
server to reload the cache as the user again needs those objects. For this reason, Guidewire recommends that you
have separate servers for handling batch processing and user interface transactions.
If an individual cache reports hundreds or thousands of evictions and a low cache hit rate, then that cache is
experiencing cache thrashing. If you notice cache thrashing on a server that is not processing batch jobs, re-size the
cache. Otherwise, dedicate the server to batch jobs.
See also
• “Detect Cache Thrashing” on page 66.
Detect Cache Thrashing

Procedure
1. Log into Guidewire PolicyCenter as a user with administrative privileges.
2. Press ALT+SHIFT+T to open the Server Tools screens.
3. Navigate to the Cache Info screen.
4. Use the information in the Cache Info screen to analyze the number of evictions.
5. Click Clear Global Cache to clear the cache.
6. Reproduce the operation that you suspect caused the cache thrashing.
7. Reanalyze the information on the Cache Info screen.
Next steps
After you have taken the proper action, repeat the analysis to verify that the change yielded the results you expected
to occur.
See also
• “Cache Thrashing” on page 66.
Cache Impact on Memory Utilization

The maximum size of the cache is dependent on cache parameters. The server cache grows in size to reach a
maximum specified by cache sizing parameters. Java does not provide a good means to estimate the memory usage
of objects. Therefore, it is difficult to reliably estimate the maximum size of a cache. If the cache size exceeds the
maximum heap size, the application eventually runs out of memory.
Larger caches increase memory starvation issues. Larger caches expand the memory footprint of the application.
Performance decreases as garbage collection becomes more frequent and analyzes more objects.
Set the cache as large as needed, but no larger. Monitor garbage collection to extrapolate memory usage patterns and
garbage collection statistics.
See also
• “Server Cache Tuning Parameters” on page 67.
• “Server Memory Management” on page 69
Server Cache Tuning Parameters

File config.xml contains cache parameters for PolicyCenter. Access this file from Guidewire Studio at
configuration→config.
Parameter Description
ExchangeRatesRefreshIntervalSecs The number of seconds between refreshes of the exchange rates cache.
PolicyCenter uses this specialized cache for exchange rates only.
GlobalCacheActiveTimeMinutes Time, in minutes, that PolicyCenter considers cached objects active. You can think
of this as a period in which PolicyCenter is heavily using an item, for example, how
long a user stays on a screen. The cache mechanism gives higher priority to pre‐
serving these higher‐use objects.
Set GlobalCacheActiveTimeMinutes to a value less than
GlobalCacheReapingTimeMinutes.
GlobalCacheDetailedStats Boolean value that specifies whether to collect detailed statistics for the global
cache. Detailed statistics are data that PolicyCenter collects to explain why the
caching mechanism evicted items from the cache. PolicyCenter collects basic sta‐
tistics, such as the miss ratio, regardless of the value of
GlobalCacheDetailedStats. Disabling collection of detailed cache statistics can
sometimes improve performance.
Guidewire sets the value of GlobalCacheDetailedStats to false by default. Set
the parameter to true to help tune your cache.
If the GlobalCacheDetailedStats parameter is set to false, the Cache Info screen
does not include the Evict Information and Type of Cache Misses graphs.
At runtime, use the Management Beans screen to enable the collection of detailed
statistics for the global cache.
GlobalCacheReapingTimeMinutes Time, in minutes, since the last use of a cached object before PolicyCenter consid‐
ers the object eligible for reaping. This can be thought of as the period during
which PolicyCenter is most likely to reuse an object.
An evict timeout mechanism removes old objects from the cache. Once per mi‐
nute, a thread evicts cache entries that have not been used for a period equal to
or greater than GlobalCacheReapingTimeMinutes. This mechanism differs from
the stale timeout mechanism. The stale timeout mechanism refreshes from the
database those cache entries that have exceeded the stale time. This process oc‐
curs as the server accesses a cached object. The evict timeout mechanism deletes
any cache entries that are older than the default evict time. An object can become
stale but not evicted if it is continually in use. For example, a bean has an evict
time of 15 minutes and a stale time of 30 minutes. If the server uses the object
once every 14 minutes, PolicyCenter never evicts the cache entry, but the entry
does eventually become stale.
GlobalCacheReapingTimeMinutes is initially set to 15 minutes. The minimum val‐
ue for this parameter is 1 minute. Since the eviction thread only runs once per mi‐
nute, a smaller value would not make sense. The maximum value for this parame‐
ter is 15 minutes.
GroupCacheRefreshIntervalSecs The number of seconds between refreshes of the groups cache. PolicyCenter uses
this specialized cache for group‐related data only.

GlobalCacheSizeMegabytes Maximum amount of heap space used to store cached entities, expressed as a
number of megabytes. This parameter supersedes the value of
GlobalCacheSizePercent.
At runtime, you can use the Cache Info or Management Beans screen to modify this
value.
GlobalCacheSizePercent Maximum amount of heap space used to store cached entities, expressed as a per‐
centage of the maximum heap size.
GlobalCacheStaleTimeMinutes Time, in minutes, after which PolicyCenter considers an object in the cache stale if
it has not been refreshed from the database.
A stale timeout mechanism ensures that the server does not use excessively old
object entries. An object is stale if it has not been refreshed from the database
within a configurable amount of time. Upon accessing a cache entry, the server
calculates the duration since the object was last read from the database. If that
duration exceeds the stale time, the server refreshes the cache entry from the da‐
tabase. To avoid increased complexity, PolicyCenter prefers this mechanism over
evicting objects upon stale timeout.
At runtime, you can use the Cache Info or Management Beans screen to modify this
value.
GlobalCacheStatsWindowMinutes This parameter denotes a period of time, in minutes, that PolicyCenter uses for
two purposes:
• The period of time to preserve the reason that PolicyCenter evicted an object,
after the event occurred. If a cache miss occurs, PolicyCenter reports the
reason on the Cache Info screen.
• The period of time to display statistics on the chart on the Cache Info screen.
ScriptParametersRefreshIntervalSecs The number of seconds between refreshes of the script parameters cache.
PolicyCenter uses this specialized cache for script parameters only.
ZoneCacheRefreshIntervalSecs The number of seconds between refreshes of the zones cache. PolicyCenter uses
this specialized cache for zones only.
See also
• “Special Caches for Rarely Changing Objects” on page 69

• “Cache Info” on page 367
• “Management Beans” on page 356
Understanding Cache Metrics

Cache hit ratio metric information intrinsically depends on the workflow that is using the object. Some workflows
involve reading an object only one time while others involve reading the object many times. The cache hit varies
depending on these workflows.
Also, it is also for the data be appear skewed to the low side if a server started recently, or if the server has not had
much user load. For example, if you recently started the server, and users have only visited a few screens, the hit rate
is very low because PolicyCenter encountered only a few cache hits. As users visit more pages, the hit rate
increases.
Thus, there are no good default cache hit ratios. Experimentation combined with performance measurements
constitutes the only approach to identifying appropriate cache sizes.
See also
• For information on how to view cache performance, see “Cache Info” on page 367.
Special Caches for Rarely Changing Objects

PolicyCenter includes specific individual caches for the following rarely changing objects:
• Exchange rates
• Groups
• Script parameters
• Zones
These specialized caches periodically refresh the entire set of the rarely changing object. This prevents the server
from querying the database each time PolicyCenter accesses one of these objects, thereby improving performance.
For each of these special caches, you can set the refresh interval through a configuration parameter in file
config.xml.
The following table lists the objects with specialized caches, along with the configuration parameter that controls the
refresh rate of each specialized cache.
Cache objects Cache refresh rate set by…

ExchangeRate ExchangeRatesCacheRefreshIntervalSecs
Group GroupCacheRefreshIntervalSecs
ScriptParameter ScriptParametersRefreshIntervalSecs
Zone ZoneCacheRefreshIntervalSecs
See also
• “Server Cache Tuning Parameters” on page 67.
Server Memory Management

Java provides platform-side memory management that significantly simplifies coding. The JVM (Java Virtual
Machine) periodically identifies unused objects and reclaims associated memory. In general, computer science calls
this term garbage collection. Garbage collection can have a significant impact on server performance.
This topic describes Java platform garbage collection analysis.
Memory Usage Logging

If configured in file logging.properties, Guidewire logs contain memory usage messages that provides
information on the use of memory by the JVM. To enable memory usage logging, you must set parameter
MemoryUsageMonitorIntervalMins in config.xml to a value other than the default of 0.
A memory usage message looks similar to the following:
serverName 2016-04-09 16:44:14,423 INFO Memory usage: 80.250 MB used, 173.811 MB free, 254.062 MB total,
2048.000 MB max
The following list describes the different types of information that you see in a memory logging message.
Memory usage Meaning

used Amount of memory allocated to objects. This includes memory for the following memory types:
• Memory used by active objects still in use
• Memory used by stale objects (on which the JVM eventually performs garbage collection)
free Amount of un‐allocated memory
total Amount of memory that the JVM process reserves from the operating system.

Memory usage Meaning

max Amount of maximum total memory that PolicyCenter allows the JVM to use.
It is common for a server to use up the maximum amount of memory fairly quickly, so that used and total are at or
near the max value. This indicates normal operation. If the server needs more memory, it triggers garbage collection
to free up the memory used by stale objects.
Memory Usage Messages and Garbage Collection

It is not possible to configure the content of the logging messages to provide enough detail to indicate or warn of
memory issues. The only way to obtain a more accurate picture of memory usage is through running the garbage
collector.
Guidewire does not support running the garbage collector merely for the sake of more detailed logging. However, to
obtain more detailed information on PolicyCenter memory usage in the current configuration, enable garbage
collector logging.
You only need to worry about memory issues if the server throws an OutOfMemoryError exception. If that happens,
Guidewire recommends that you configure the garbage collector to print out detailed memory information.
See also
• “Enabling Garbage Collection” on page 70
• “Understanding Possible Memory Leak” on page 71
Enabling Garbage Collection

The garbage collector can provide additional information on collection statistics. Careful analysis helps understand
garbage collector behavior.
Verify that the performance analysis tool you choose supports the version of the JVM that you use for PolicyCenter.
To determine the supported JVM versions, visit the Guidewire Community and search for knowledge article 1005,
Supported Software Components.
See also
• “Enabling Verbose Garbage Collection for IBM JVM” on page 70
• “Enabling Verbose Garbage Collection for Oracle Java Hotspot VM” on page 71
Enabling Verbose Garbage Collection for IBM JVM

To enable verbose garbage collection for IBM Java Virtual Machines, add the -verbose:gc flag to the JVM options.
Other options exist for the same functionality.
IBM estimates that the overhead associated with verbose garbage collection is minimal and estimated to be 2% of
the garbage collection time. In other words, if the JVM spends 5% of its time garbage collecting without verbose
garbage collection, it would spend 5.1% of the time garbage collecting with verbose garbage collection.
The output provided is in XML format. This output is generally rich enough for a thorough analysis. In general,
there is no need for additional levels of logging.
Used with WebSphere, the IBM JVM outputs garbage collection logs into a file called native_stderr.log.
IBM provides the IBM Support Assistant. You can install multiple plugins within this tool. Several plugins are
available for the IBM JVM and WebSphere. These tools provide deep analysis of JVM behavior, spot issues, and
recommend how to tune the JVM.
IBM Support Assistant

IBM provides the IBM Support Assistant Workbench. It is possible install multiple plugin tools within the
workbench. The “IBM Monitoring and Diagnostic Tools for Java – Garbage Collection and Memory Visualizer” is
the tool to use to analyze garbage collection logs.
The tool provides some tuning recommendations. The recommendations work more for the IBM JDK than the
HotSpot JDK. Additionally, the tool provides graphs with hints on JVM behavior that help identify issues such as
memory shortages or excessive pauses.
Refer to the following web site for more information about the IBM Support Assistant:
http://www.ibm.com/software/support/isa/
Enabling Verbose Garbage Collection for Oracle Java Hotspot VM

To enable verbose garbage collection on an Oracle Java Hotspot JVM, add the -verbose:gc flag to the Java
HotSpot VM options. Several levels of logging exist, providing more or less output.
The garbage collection time logs can time stamp the various entries with the exact date. Guidewire recommends the
following options:
-XX:+PrintGCDetails -XX:+PrintGCDateStamps -XX:+PrintHeapAtGC -XX:+PrintGCApplicationConcurrentTime

-XX:+PrintGCApplicationStoppedTime
These options provide the following:

• Nature of the garbage collection (minor or full)
• Amount of memory reclaimed
• Time elapsed since JVM start or date corresponding to the event, depending on available options
• Before and after state of the different memory pools (nursery, tenured and permanent)
• Amount of time the application runs between collection pauses
• Duration of the collection pause
The level of information can be overwhelming, though it is necessary in some cases.
Add the -Xloggc:file option to redirect output to the specified file.
HPjmeter and GCViewer

HPjmeter and GCViewer are tools that enable you to visually analyze the HotSpot JDK garbage collection logs.
Both tools generate:
• Key metrics about the period (number of minor/major collections, percent of time spent paused, and so forth)
• Visual representation of the different garbage collections
These tools might require different verbose garbage collection options. Otherwise, HPjmeter or GCViewer might not
be able to analyze the corresponding output.
For information on these utility tools, refer to the respective company web sites.
Refer to the following web sites for more information:
• HPjmeter – http://www.javaperformancetuning.com/tools/hpjmeter/index.shtml
• GCViewer – http://www.tagtraum.com/gcviewer.html
Understanding Possible Memory Leak

Guidewire applications are memory-intensive. Guidewire applications generally require larger heaps than most other
Java applications.
Garbage collection logs might show that memory usage grows significantly over time, resulting in a lack of
available memory. Computer science commonly describes this condition as a memory leak. To diagnose the
problem, it is necessary to collect a dump of all used objects, called a heap dump, to identify all objects in the heap.
Developers familiar with PolicyCenter can then analyze the heap dump. Such analysis helps identify excessive
memory usage, identify its root cause and possibly find a change that will avoid such issues.
The investigation of memory leaks differs slightly depending on the JVM platform.
Common Approaches for Heap Dump Generation

In general, you can use the following options to generate a heap dump:
1. Set specific flags to force the following behaviors:
• Heap dump generation if the heap is full and an out-of-memory condition occurs
• Heap dump generation if an application CTRL-BREAK or SIGQUIT occurs
Combine these options with options instructing the JVM to generate the heap dump at a specific directory
location.
2. Use tools to connect to a running JVM. Such tools provide the option to trigger a heap dump.
Generating Heap Dumps with IBM JDK

In general, the capability for generating heaps dumps is satisfactory in all of the IBM JVM release levels. For
information on generating heap dumps for IBM JDK 1.6, refer to the following document:
IBM Developer Kit and Runtime Environment, Java Technology Edition, Version 6
Related information
IBM Developer Kit and Runtim Environment - Diagnostics Guide
Generating Heap Dumps with Oracle HotSpot JDK

Use the following flags for heap dump generation:
• HeapDumpOnCtrlBreak
• HeapDumpOnOutOfMemoryError
For more information about analyzing heap dumps, refer to the following Oracle document:
Troubleshooting Guide for Java SE 6 with HotSpotVM
Related information
Troubleshooting Guide for Java SE 6 with HotSpot VM
Additional Heap Dump Recommendations

While generating heap dumps, pay attention to the following facts:
• Heap dump generation frequently fails because the single file it generate is very large and the configuration of
the supporting environment prevents regular accounts from creating such large files. Therefore, it is usually the
case that you must provide some configuration to allow the account running the PolicyCenter instance to create
such large files.
• The generation of a heap dump during out-of-memory conditions is sometimes challenging. As a JVM is
reaching maximum memory utilization, it generally experiences severely degraded performance. As the pace of
the leak decreases gradually, the occurrence of the out-of-memory condition might take an inordinate amount of
time. This length of time might be incompatible with the need to restore performance for users or processes.
• Windows only: Windows does not support signals. Therefore, generating a heap by starting the JVM with a heap
dump on CTRL-BREAK, depends on the capacity to send a CTRL-BREAK. You cannot send a CTRL-BREAK to a JVM
started as a background process. Therefore, for the time of the investigation, start the JVM from a command
prompt rather than as a background process.
• The JVM generally provides optional flags that prevent it from listening to signals. Disable these flags while
trying to generate a heap dump through signals.
• Heap dump analysis is very memory intensive. Assume that the tool used to analyze the heap dump might need a
heap two to three times larger than the amount of objects captured in the heap dump. Host the heap dump
analyzer on a server with a 64-bit JVM and a significant amount of memory. If such a configuration is not
available, you might want to reduce the heap size so that the memory leak reaches an identifiable threshold
sooner. This method allows the generation of smaller, easier to analyze heap dumps.
• Heap dump analysis tools generally point to the CacheImpl class as the largest memory consumer. This class
corresponds to the Guidewire cache. It is normal that the cache consumes a few hundred megabytes. In this case,
the memory issue is likely not caused by cache growth. If the cache consumes significantly more memory, you
might need to be downsize the cache. See .
Heap Dump Generation and Analysis Tools

Several tools are available for heap dump generation and analysis. Both IBM and Oracle provide some tools to assist
with these tasks on their respective JVMs. Other vendors also provide some tools that aim to assist with these tasks
on the most common JVM platforms.
IBM‐only Tools for Heap Dump Generation

IBM provides tools that can assist with heap dump analysis. You can find information on these tools on the IBM
Support Assistant web site.
IBM also provides the IBM DTJF feature for Eclipse Memory Analyzer for use in analyzing IBM JVM heap dumps.
You can tune the tool to use a larger heap, which is frequently necessary to analyze very large heap dumps.
Related information
IBM Support Assistant
IBM DTJF feature for Eclipse Memory Analyzer
Oracle‐only Tools for Heap Dump Generation

Oracle Java Monitoring and Management Console, or JConsole, is a management tool that connects to a running
Java HotSpot VM. You can trigger a heap dump by using jConsole. Refer to the following Oracle document for
more information:
Using JConsole to Monitor Applications
Oracle bundles the Java Heap Analysis Tool (jhat) with Java HotSpot VM 1.6. Therefore, if you want to analyze a
heap with jhat, you can install Java HotSpot VM 1.6 and use the jhat release provided. Refer to the following
Oracle document for more information:
Java Heap Analysis Tool
Oracle jVisualVM is another multipurpose tool that you can use to analyze heap dumps. Refer to the following
Oracle document for more information:
Java VisualVM
Related information
Using JConsole to Monitor Applications
Java Heap Analysis Tool
Java VisualVM
Generic Tools for Heap Dump Generation

YourKit is a commercial product that provides many functions. You can use YourKit to connect to the JVM, analyze
the JVM, and trigger heap dumps. It also provides some very interesting heap dump analysis tools.
JProbe is another commercial product providing many capabilities, including heap dump analysis.
Guidewire development mainly uses YourKit with good success. Guidewire Support uses YourKit and several other
products like jVisualVM, the IBM DTJF feature for Eclipse Memory Analyzer, and JProbe.
JVM Profiling
Java profilers are available for two main purposes:
Memory profil‐ Identifies memory usage, and, more specifically, memory leaks due to referenced but unused objects.
ing
CPU profiling Helps identify programmatic hot spots and bottlenecks. This analysis might help remove the corresponding
bottlenecks, thereby increasing performance.
Guidewire has used two profiling tools internally and found each to be of good quality. Both tools provide both
memory and CPU profiling:
• Guidewire recommends YourKit for memory profiling.
• Guidewire recommends JProfiler for CPU profiling.
To profile PolicyCenter, load the profiler agent into the PolicyCenter JVM either as it is starting PolicyCenter or by
attaching the profiler agent to a running JVM. Refer to the profiler documentation for instructions.
Large Object Analysis

Large Java objects cause an extra strain on the JVM for various reasons. If garbage collection analysis shows that
the JVM is allocating very large objects, investigate this further and understand the source of the objects.

chapter 5
Geocoding
Geocoding is the process of assigning a latitude and longitude to an address. Guidewire supports geocoding in
ClaimCenter, PolicyCenter, and ContactManager.
Understanding Geocoding
The geocoding process assigns latitudes and longitudes to addresses. Software then uses geocoded addresses to
present users with geographic information, such as the distance between two addresses. All primary addresses in
PolicyCenter, PolicyCenter, and ContactManager are candidates for geocoding.
The Geocode Plugin Interface

To implement geocoding, PolicyCenter provides a default GeocodePlugin plugin interface. Implementations of the
plugin connect with specific external geocoding services, which provide geocode coordinates for specific addresses.
Typically, plugin implementations also use an external mapping service to calculate and return proximity
information, driving instructions, and maps. You enable the plugin implementation in Guidewire Studio and specify
the parameters for the implementation you choose.
PolicyCenter provides a fully functioning and supported GeocodePlugin implementation, the BingMapsPlugin
Gosu class. This plugin implementation connects to the Microsoft Bing Maps Geocode Service. If you intend to use
the BingMapsPlugin implementation, your organization must have a valid account with Microsoft.
Microsoft Bing Maps

Before you can use the default Bing Maps plugin implementation, your organization must have its own account,
login, and application key with Microsoft Bing Maps. To obtain these items, access the Bing Maps Dev Center.
You must set up a Bing Maps account and obtain an application key. After you create an application key, the
application name is arbitrary and there is no need for an application URL.
Geocoding 75
Working with Geocode Batch Processing

Guidewire implements geocode batch processing as a work queue. The Geocode writer in PolicyCenter and the
ABGeocode writer in ContactManager use BatchGeocodeBatchGeocode and other criteria to filter which addresses
to pass to the plugin for geocoding.
• ContactManager verifies that the BatchGeocode property is true and the GeocodeStatus property is none.
• PolicyCenter verifies the following:
◦ The address is an AccountLocation or subtype.
◦ The address is a primary address of a UserContact.
◦ The BatchGeocode property is true and the GeocodeStatus property is none.
The geocoding process submits qualifying addresses to your implementation of the GeocodePlugin plugin. After the
GeocodePlugin plugin adds geocode coordinates to an address, the geocoding process updates the address in the
database.
Geocode Status Typelist

The GeocodeStatus typelist defines the set of status codes returned from the default GeocodePlugin plugin
implementation. As the typelist is final, you cannot edit it. You access the GeocodeStatus typelist in Guidewire
Studio in the following location:
configuration→config→Metadata→Typelist
See also
Related references
“Geocode Writer Work Queue” on page 108
Related information
Bing Maps Dev Center
Configuring Geocoding
Configuring geocoding involves the following tasks:
• Enabling your implementation of the GeocodePlugin plugin
• Setting geocoding feature parameters
• Scheduling the Geocode work queue
• Configuring the number of Geocode batch processing workers
Enabling the Geocode Plugin

By default, the base configuration GeocodePlugin plugin implementation uses the Bing Maps implementation.
Guidewire disables this plugin in the base configuration. You must enable the geocode plugin before you can use the
geocoding functionality. If you intend to use geocoding in multiple Guidewire applications, you must make these
changes in each application separately.
Scheduling the Geocode Plugin

By default, Guidewire does not schedule geocode batch processing in the base configuration. To schedule geocode
batch processing, you need to uncomment the relevant section in file scheduler-config.xml. Guidewire
recommends that you schedule the geocoding process to run during periods of minimal activity on the PolicyCenter
servers.
76 chapter 5: Geocoding
IMPORTANT Schedule geocode batch processing for PolicyCenter and ContactManager with
sufficient processing windows between runs to assure sufficient time for runs to fully process the work
items in the work queues. If you find duplicate work items in the work queues for the same address
ID, extend the interval between runs.
Configuring the Number of Geocoding Batch Processing Workers

At the time you first start PolicyCenter, it is possible for your database to have many addresses to geocode,
especially if you imported many new addresses into your production database. A large number of new addresses can
cause the GeocodePlugin plugin implementation to take a long time to process these new addresses. Guidewire
recommends that you configure the geocoding process with a sufficient number of worker instances before you start
your production servers.
The default configuration specifies one worker instance. Worker instances pass addresses from the work queue to
the GeocodePlugin plugin implementation. Consider increasing the number of worker instances to improve
throughput. To further improve throughput, assign worker instances to run on multiple servers.
See also
• “Understanding Geocoding” on page 75
About Guidewire Geocoding Parameters

Configure geocoding features in the PolicyCenter user interface with the following parameters in config.xml.
UseGeocodingInPrimaryApp If true, PolicyCenter enables searching for nearby locations in the Rein‐
surance Management user interface. ContactManager does not respond
to this parameter.
The default is false.
UseMetricDistancesByDefault If true, PolicyCenter uses kilometers and metric distances instead of miles
and United States distances for location searches.
Set this parameter identically in Guidewire applications that use geocod‐
ing.
ProximitySearchOrdinalMaxDistance The maximum distance to use while performing an ordinal (nearest n
items) proximity search for locations. This distance is in miles, unless
UseMetricDistancesByDefault is true.
The default is 300.
ProximityRadiusSearchDefaultMaxResultCount The maximum number of results to return if performing a radius (within n
miles or kilometers) proximity search. This parameter has no effect on or‐
dinal (nearest n items) proximity searches.
The default is 1000.
Enable the Geocode Plugin

Procedure
1. In the PolicyCenter Project window, expand configuration→config→Plugins→registry.
2. Double-click GeocodePlugin to open it.
3. If the Disabled check box is checked, un-select the check box to enable the plugin.
4. Ensure that the Class field specifies the Bing Maps implementation class:
gw.plugin.geocode.impl.BingMapsPluginRest
5. Under Parameters, specify the following:
Geocoding 77
applicationKey The application key that you obtained from Bing Maps.
geocodeDirec- The locale for geocoded addresses and routing instructions returned from Bing Maps. For example, use
tionsCulture the locale code ja-JP for addresses and instructions for Japan. The plugin uses en-US if you do not
specify a value. For a current list of codes that Bing Maps supports, refer to the following web site
http://msdn.microsoft.com/en-us/library/cc981048.aspx
imageryCulture The language for map imagery. For example, use the language code ja for maps labeled in Japanese.
The plugin uses en if you do not specify a value. For a current list of codes that Bing Maps supports,
refer to the following web site
http://msdn.microsoft.com/en-us/library/cc981048.aspx
mapUrlHeight Height of maps, in pixels. The plugin uses 500 if you do not specify a value.
mapUrlWidth Width of maps, in pixels. The plugin uses 500 if you do not specify a value.
6. Save your changes.
See also
Schedule Geocode Batch Processing Runs

About this task
By default, the schedule runs geocode batch processing at 1:30 AM daily. If you regularly add many new contacts,
especially in ContactManager, tune the schedule to match your expected daily load of new addresses.
Procedure
1. In Guidewire Studio, navigate to configuration→config→scheduler and open scheduler-config.xml for
editing.
2. Uncomment the following section in scheduler-config.xml file.
<ProcessSchedule process="Geocode">
<CronSchedule hours="1" minutes="30"/>
</ProcessSchedule>
See also
• “The Work Queue Scheduler” on page 89
• “Understanding a Work Queue Schedule Specification” on page 89
Configure the Number of Worker Instances for Geocode Batch Processing

Procedure
1. In Guidewire Studio, navigate to configuration→config→workqueue and open work-queue.xml for editing.
2. Modify the number of worker instances in the following section in file work-queue.xml to meet your business
needs.
<work-queue workQueueClass="com.guidewire.pc.domain.geodata.geocode.PCGeocodeWorkQueue"
progressinterval="600000">
<worker instances="1"/>
</work-queue>
3. Update database statistics by running the Database Statistics batch process.

See also
• “Worker Configuration” on page 94
• “Database Statistics Work Queue” on page 105
Geocoding 79
chapter 6
Administering Batch Processing
PolicyCenter provides tools for configuring and managing various forms of batch processing. You can schedule
batch processing to run regularly or on demand.
See also
• “Batch Process Info” on page 321
• “Work Queue Info” on page 324
Overview of Batch Processing

PolicyCenter supports two modes of batch processing:
• Work queue
• Batch process
Work queue
A work queue operates on a batch of items in parallel. PolicyCenter distributes work queues across all servers in a
PolicyCenter cluster that have the appropriate role. In the base configuration, Guidewire assigns this functionality to
the workqueue server role.
A work queue comprises the following components:
• A processing thread, known as a writer, that selects a group (batch) of business records to process. For each
business record (a policy record, for example), the writer creates an associated work item.
• A queue of selected work items.
• One or more tasks, known as workers, that process the individual work items to completion. Each worker is a
short-lived task that exists in a thread pool. Each work queue on a cluster member shares the same thread pool.
By default, each work queue starts a single worker on each server with the appropriate role, unless configured
otherwise.
Work queues are suitable for high volume batch processing that requires the parallel processing of items to achieve
an acceptable throughput rate.
Batch process
A batch process operates on a batch of items sequentially. Batch processes are suitable for low volume batch
processing that achieves an acceptable throughput rate as the batch process processes items in sequence. For
example, writers for work queues operate as batch processes because they can select items for a batch and write
them to their work queues relatively quickly.
Administering Batch Processing 81
See also
• “Work Queues” on page 82
• “Batch Processes” on page 85
Work Queues
A work queue comprises the following components.
Writer A writer thread selects units of work for batch processing and writes their identities to a work queue.
Work A work queue is a database table that the writer loads with a batch of work items and from which workers check
queue out work items for processing.
Worker One or more worker tasks that check out work items from the work queue and process them to completion. By
default, each work queue starts a single worker on each cluster member with the workqueue role, unless config‐
ured otherwise.
Starting the writer initiates a run of batch processing on a work queue. The batch is complete if the workers exhaust
the queue of all work items, except those that they failed to process successfully.
Work Queue Architecture

The following diagram illustrates the components of a work queue and how they function.
82 chapter 6: Administering Batch Processing

YES
Wake Table empty?

Locate items
Sleep NO
Notify Workers Load table

Writer Server with ‘batch’ Role
Server A Server C
Server B
C C
A B
Wait for notification Check out item

Process item
NO
YES
Workers
Table empty?
Work Queue Writers

Whenever the writer thread awakes or starts on demand, it checks the work queue table for any work items that
remain from a prior batch. The specific configuration of each work queue determines how the work queue creates
and processes work items.
The following workflow examples provides a simple description of how PolicyCenter work queues operate.
Work Items Remaining

If work items remain from a previous batch, the following sequence of events occur:
1. The writer thread notifies the workers that they have work to process.
2. PolicyCenter terminates the writer thread.
No Work Items Remaining

If no work items remain from a previous batch, the following sequence of events occurs:
1. The writer thread selects items for a new batch.

2. The writer writes the identifiers of the selected items to the work queue table.
3. The writer notifies the workers that they have work to process.
4. PolicyCenter terminates the writer thread.
Work Queue Workers

Typically, each work queues shares the standard work item table (StandardWorkQueue) for its work items. However,
a worker task operates only on work items that its associated writer inserts into the table. For example, suppose that
PolicyCenter distributes the Activity Escalation work queue across a Guidewire cluster, with six workers operating
on three different servers. Those workers work only on activity escalation items in the standard work item table.
Typically, you configure work queues for multiple workers. Thus, typically, some number of workers operate
throughout the day on work items in the standard work queue table.
At specified intervals (defined through the maxpollinterval attribute on worker in work-queue.xml), a worker
awakens and checks the work item table for work items from its associated writer. If a worker finds work items
available for processing, the worker checks out its quota from the work queue. For each item, the worker sets the
following attributes.
Status Set to checkedout.

The value of the Status attribute can be any one of the following:
• available
• checkedout
• failed
LastUpdateTime Set to the time at which the worker checks out the work item.
CheckedOutBy Set to the worker.
After it checks out a quota of work items, the worker task processes them sequentially. Whenever a worker
completes a work item successfully, it deletes the item from the table and begins to process the next item. The
standard work item table (StandardWorkQueue) is retireable, so successfully completed work items remain in the
table for historical reference.
Note: In rare cases, it is possible for PolicyCenter to notify a worker of work, but, the worker finds no
work is available after it awakens. For example, for small batch runs, a worker can check out all items
in the batch with its first check out quota of items. This action can occur between the time the writer
notifies the workers and other workers awaken. If a worker awakens and finds no work items, the
worker goes back to sleep.
Work Queue Scheduling and Processing Intervals

A writer for a work queue starts at the interval specified in scheduler-config.xml. Typically, you schedule the
writer to start several times during the day or once at night. Access schedule configuration file schedule-
config.xml in Guidewire Studio at the following location:
configuration→config →scheduler
Worker tasks awaken much more frequently than their writers start. One writer awakens at least as frequently as a
specified maximum polling interval, if not more frequently.
You do not schedule worker tasks. Instead, they awaken in response to notification from the writer or upon
expiration of the polling interval. After a worker awakens, if there are work items to process, it processes up to the
number of batchsize items, as defined in work-queue.xml. If there are more items than the batch size to process,
the worker awakens another worker. This worker repeats the process of checking out work items and waking up
another worker if necessary, until maximum allowed number of workers are active.
You configure the number of workers, polling interval, and batch size in work-queue.xml. Access work queue
configuration file work-queue.xml in Guidewire Studio at the following location:
configuration→config→workqueue
You can view and manage work queues from the Server Tools Work Queue Info screen in PolicyCenter, if you have
the appropriate administrative privilege.
See also
• “Performing Custom Actions After Batch Processing Completion” on page 96
Batch Processes
PolicyCenter distributes batch processes across all PolicyCenter cluster members that have the batch server role.
Each server with the batch role also has a batch process lease manager that acquires and manages the batch process
leases on that server. In this context, a lease represents a single run of a single batch process.
Available servers with the batch role compete for available batch processing leases. After a server acquires a lease,
that server runs the batch process to completion.
How aggressively the cluster servers compete with each other depends on how many batch processes each one is
individually already running. Those servers running fewer or no batch processes are more likely to acquire a new
batch process lease than other servers already busy running processes. It is possible to configure this behavior.
For scheduled batch processes, a scheduler component, running on a cluster member with the scheduler role,
decides to start a batch process according to the published schedule. The scheduler first creates a new lease for the
batch process in the database. All cluster members with the batch server role then compete to acquire that lease.
The cluster member that wins the competition starts the batch process.
The Batch Process Info Screen

Generally, a batch process runs to completion and then reports its result back to a log or to the administrative user
interface. You can view and manage batch processing from the Server Tools Batch Process Info screen in
PolicyCenter, if you have the appropriate administrative privilege. From this screen, you can view the batch process
schedule, if any, and start or stop a scheduled batch process.
Batch Process Caching

Batch processes that run concurrently on the same server share a common cache. The cache demands of each
process end up flushing the cache more frequently, so fewer cache hits occur for each process. That increases the
amount of physical reads from the relational database, thus degrading performance. Concurrent data change
exceptions can occur also if multiple batch processes (on the same or different servers) attempt to update the same
cached entity instances. This requires one or the other batch process to retry an item, leading to further performance
degradation.
See also
• “Working with Work Queue Writers and Batch Processes” on page 86
• “Component Lease Management” on page 157
• “Batch Process Prioritization” on page 160
Working with batch processing types in Studio

To access information about the PolicyCenter batch processing types, open the following files in Studio:
• BatchProcessType.ttx – Provides the name and codes for each batch processing type.
• work-queue.xml – Provides the name of the backing class and number of work instances for a processing type.
• scheduler-config.xml – Provides the schedule for any batch process type that is Schedulable.
The number of worker tasks is important if zero (0). Setting the number of worker tasks to 0 disables the work queue
as there are no workers to perform the work. To enable the work queue, set the number of workers to greater than
zero.
Working with Work Queue Writers and Batch Processes

You can run many batch processes, including writers for work queues, from PolicyCenter or from the command
prompt.
How to Run a Writer from the Work Queue Info Screen
It is possible to run a writer for a work queue from the Server Tools Work Queue Info screen. PolicyCenter enables the
Run Writer button on this screen for all work queue types that belong to the BatchProcessTypeUsage category
UIRunnable.
To access the Work Queue Info screen, you must have the internaltools permission. The Admin role has this
permission by default. Alternatively, if the EnableInternalDebugTools parameter is set to true in config.xml and
the server is running in development mode, all users can access the Work Queue Info screen.
How to Run a Batch Process from the Batch Process Info Screen
It is possible to run batch processes from the Server Tools Batch Process Info screen in PolicyCenter. PolicyCenter
enables the Run button on this screen for all batch process types that belong to the BatchProcessTypeUsage
category UIRunnable.
To access the Batch Process Info screen, you must have the internaltools permission. The Admin role has this
permission by default. Alternatively, if the EnableInternalDebugTools parameter is set to true in config.xml and
the server is running in development mode, all users can access to the Batch Process Info screen.
How to Terminate a Writer or Batch Process from the Command Prompt
It is possible to terminate some in-progress processes, including writers for work queues, from a command prompt.
To determine if it is possible to terminate an in-progress batch process, consult the reference topic for that particular
batch processing type. The information for each individual batch process includes whether it is a single phase or
multiphase process. You can only stop those processes listed as being multiphase.
Note: It is not possible to terminate a single phase batch process. Single phase processes run in a
single transaction. Thus, there is no convenient place to terminate the process.
See also
• “Work Queues and Batch Processes, a Reference” on page 98

• “Maintenance Tools Command” on page 390
Run a Writer from PolicyCenter

Start, or stop, a work queue from the Work Queue Info screen.
Before you begin
You must have the internaltools user permission to access the Work Queue Info screen.
About this task

Note: You can run writers for work queues from either the Work Queue info screen or the Batch Process
Info screen.
Procedure
1. Log in to PolicyCenter.
2. Press ALT+SHIFT+T to display the Server Tools tab.
3. Navigate to Work Queue Info.
4. Click Run Writer in the Actions column of the work queue that you want to run.
Run a Batch Process from PolicyCenter

Start, or stop, a batch process from the Batch Process Info screen.
Before you begin

You must have the internaltools user permission to access the Batch Process Info screen.
Procedure
1. Log in to PolicyCenter.
2. Press ALT+SHIFT+T to display the Server Tools tab.
3. Navigate to Batch Process Info.
4. Click Run in the Action column of the batch process that you want to run.
Run a Writer or Batch Process from the Command Prompt

Start a batch processing type using a maintenance_tools command option.
About this task

It is possible to run many batch processes, including writers for work queues, from a command prompt.
Procedure
1. Start the PolicyCenter server if it is not already running.
2. Open a command prompt.
3. Navigate to the following location in the PolicyCenter installation directory:
admin/bin
maintenance_tools -password password -startprocess process
For the process value, specify a valid process code. For the process code for each batch processing type,
including writers for work queues, consult the reference topic for the individual batch processing type.
See also
Terminate a Writer or Batch Process from the Command Prompt

It is possible to stop a process type using a maintenance_tools command option.
About this task

It is not possible to use the following procedure to terminate the operation of the table_import command.
Procedure
admin/bin
maintenance_tools -password password -terminateprocess process
See also
Check Status of a Writer or Batch Process from the Command Prompt

Check the status of a batch process using a maintenance_tools command option.
About this task

It is possible to check the status of processes, including writer processes for work queues, from a command prompt.
Procedure
admin/bin
maintenance_tools -password password -processstatus process
Result
For work queues, executing this command returns the status of the writer process. The command does not check
whether any work items remain in the work queue. Thus, it is possible for the process status to report as being
complete after the writer finishes adding items to the work queue. However, it is possible that there are work items
that need processing that remain in the work queue.
See also
The Work Queue Scheduler

The PolicyCenter scheduler launches many batch processes, including writer processes for work queues, according
to a schedule defined in scheduler-config.xml. Access this file in Guidewire Studio at the following location:
configuration→config→scheduler
The scheduler-config.xml file contains entries in the following format:
<ProcessSchedule process="process_code" env="environment">

<CronSchedule schedule_attributes/>
</ProcessSchedule>
The process attribute sets the process to run. The env attribute is an optional attribute that specifies the environment
in which the schedule definition for the process applies. The schedule_attributes value is a valid schedule
specification.
If needed, you can list multiple ProcessSchedule entries for the same process. The process then runs according to
each specified schedule. If you schedule a process to run while the same process is already running, then
PolicyCenter skips the overlapping process. If the scheduler-config.xml file does not list a process, then the
process does not run.
Generally, schedule the amount of time between batch process runs in hours as opposed to minutes. This is because
some batch processes require a lot of server resources. Schedule such processes to wake infrequently or at times that
the server is less busy, such as late at night or very early in the morning.
You may want to schedule some PolicyCenter batch processes to run periodically throughout the business day. For
example, the default configuration of PolicyCenter schedules the ActivityEsc batch process to run every 30
minutes. Exclude running such batch processes periodically during your nightly batch processing window. Instead,
wait until the end of the batch window to run them. For example, schedule the ActivityEsc batch process to run
every 30 minutes except during your nightly batch window. Alternatively, run such batch processes at prescribed
places in your chain of nightly batch process.
The PolicyCenter scheduler uses the PolicyCenter server time for reference.
See also
Understanding a Work Queue Schedule Specification

The <CronSchedule> element in file scheduler-config.xml describes when PolicyCenter is to run the process.
Use this element to define a schedule_attributes value to specify the exact timing, such as once every hour or
every night at a certain time. The schedule_attributes value is a combination of one or more of the following
attributes:
Attribute Standard Values Default Example

seconds 0-59 0 seconds="0"
minutes 0-59 0 minutes="15"
hours 0-23 * hours="12"
dayofmonth 1-31 * dayofmonth="1"

Attribute Standard Values Default Example

month 1-12 or JAN-DEC * month="2"
dayofweek 1-7 or SUN-SAT ? dayofweek="1"
IMPORTANT If you do not provide a value for a defined schedule attribute, the scheduler uses its
default value in determining the work queue schedule. For example, if you do not specify a value for
the hours attribute, the scheduler assumes a value of * and PolicyCenter runs the work queue process
every hour. Thus, Guidewire recommends that you provide a value for each scheduler attribute. If you
do not provide a value for a specific attribute, carefully review that attribute's default value and
determine if the default value meets your business needs.
Useful Attribute Characters

Along with the standard values listed, there are some special characters that give you more flexible options.
Charac‐ Meaning
ter
* Indicates all values. For example, minutes="*" means run the process every minute.
? Indicates no specific value. Used only for dayofmonth and dayofweek attributes. See the examples for clarification.
- Specifies ranges. For example, hour="6-8" specifies the hours 6, 7 and 8.
, Separates additional values. For example, dayofweek="MON,WED,FRI" means every Monday, Wednesday, and Fri‐
day.
/ Specifies increments. For example, minutes="0/15" means start at minute 0 and run every 15 minutes.
L Specifies the last day. Used only for dayofmonth and dayofweek attributes. See the examples for clarification.
W Specifies the nearest weekday, use only with dayofmonth. For example, if you specify 1W for dayofmonth, and that
day is a Saturday, the trigger then fires on Monday the 3rd. You can combine this with L to schedule a process for
the last weekday of the month by specifying dayofmonth="LW".
# Specifies the nth day of the week within a month. For example, a dayofweek value of 4#2 means the second Wed‐
nesday of the month (day 4 = Wednesday and #2 = the second Wednesday in the month).
These represent only some of the values that you can use for setting schedule.
Scheduler Examples
The following table lists a few examples of how to work with the <CronSchedule> element.
Example Description
<CronSchedule hours="10" /> Run every day at 10 a.m.
<CronSchedule hours="0" /> Run every night at midnight.
<CronSchedule minutes="15,45" /> Run at 15 and 45 minutes after every hour.
<CronSchedule minutes="0/5" /> Run every five minutes.
<CronSchedule hours="0" Run at midnight on the first day of the month.
dayofmonth="1" />
<CronSchedule hours="12" Run at noon every weekday (without regard to the day of the month).
dayofweek="MON-FRI" dayofmonth="?" />
<CronSchedule hours="22" Run at 10 p.m. on the last day of every month.

dayofmonth="L" />

Example Description
<CronSchedule hours="22" Run at 10 p.m. on the second‐to‐last day of every month.
dayofmonth="L-2" />
<CronSchedule minutes="3" Run 3 minutes after every other hour, 8:03 a.m. to 6:03 p.m., Monday through
hours="8-18/2" dayofweek="1-5" Friday.
dayofmonth="?"/>
<CronSchedule minutes="*/15" Run every 15 minutes after the hour, 12:15 a.m. to 8:45 a.m. and 6:15 p.m. to
hours="0-8,18-23"/> 11:45 p.m.
<CronSchedule hours="0" Run at midnight on the last Friday of the month.
dayofmonth="6L" />
<CronSchedule hours="4" Run at 4 a.m. on the second Wednesday of the month.

dayofmonth="4#2" />
The Quartz Enterprise Job Scheduler

These characters represent only some of the values that you can use for setting a schedule. Guidewire bases the
PolicyCenter scheduler on the open source Quartz Enterprise Job Scheduler. The scheduler therefore uses the same
specification for schedule attributes that Quartz uses. To determine the exact Quartz version, check the filename of
the Quartz JAR file in PolicyCenter/admin/lib.
Related information
Quartz Documentation
Determine If It Is Possible to Schedule a Batch Process

Use the Batch Process Info screen to determine if it is possible to schedule a batch process.
About this task

It is possible to schedule many batch processes, including work queue writers. It is not possible, however, to
schedule all batch processes.
Procedure
1. Log into PolicyCenter as an administrative user.
2. Press ALT+SHIFT+T to access Server Tools.
3. Navigate to the Batch Process Info screen.
4. Select Schedulable from the processes drop-down filter.
PolicyCenter displays only those batch processes, including work queue writers, that it is possible to schedule
in file scheduler-config.xml.
View a Batch Process Schedule in PolicyCenter

Review information in the Batch Process Info screen.
Procedure
1. Log in to PolicyCenter as an administrative user.
2. Press ALT + SHIFT + T to access Server Tools.
3. Navigate to the Batch Process Info screen.
4. Click the Next Scheduled Run column header to sort processes by schedule.
If there is no current schedule for a process, the Next Scheduled Run field is blank.
Scheduling Batch Processes for Specific Environments

You can define a schedule for each batch process for different environments. To specify an environment for the
process schedule, include the env attribute on the <ProcessSchedule> element in file scheduler-config.xml.
<ProcessSchedule process="process_code" env="environment">

</ProcessSchedule>
In this way, you can have different results for batch processing based on environment.
Disabling the PolicyCenter Scheduler

It is possible to disable the internal scheduler by setting the SchedulerEnabled configuration parameter to false in
file config.xml. If you do so, then you need to implement your own mechanism for running PolicyCenter batch
processes. For example, you can use your own scheduling application to trigger batch processing execution along
with one of the following:
• The startBatchProcess method on the MaintenanceToolsAPI web service
• The maintenance_tools -startprocess process command option
Configuring Work Queues

In working with work queues, there are multiple general areas that you can configure.
Configuration area Configuration file More information

Work queue scheduling scheduler-config.xml “The Work Queue Scheduler” on page 89
Workers and work queues work-queue.xml “The Work Queue Configuration File” on page 92
Work queue configuration parameters config-xml Configuration Guide
The Work Queue Configuration File

You may want to modify the configuration of Guidewire-provided work queues to improve performance. You
configure attributes of a work queue and its workers in file work-queue.xml. For custom work queues, you must
modify work-queue.xml to enable your work queue to operate.
File work-queue.xml contains one top-level element, which is <work-queues>. This element has one required
attribute, defaultServer. In the base configuration, Guidewire sets the value of this attribute to workqueue.
<work-queues xmlns="http://guidewire.com/work-queue" defaultServer="#workqueue">
Note: The hash mark in front of workqueue (#workqueue) indicates that the value that follows the
hash mark is a server role and not a server ID.
Attribute defaultServer requires a value. There is no default. The PolicyCenter server refuses to start if you do not
provide a value for this attribute. The server also refuses to start if you set defaultServer to a role that does not
exist in <registry> element in config.xml.
Work Queue Definitions

Within the top-level <work-queues> element in work-queue.xml, use subelement <work-queue> to define
individual work queues.
<work-queue workQueueClass="string" progressinterval="decimal">

<worker instances="integer" throttleinterval="decimal" env="string" server="string"/>
</work-queue>

The <work-queue> subelement has attributes to configure a named work queue in general. The <worker>
subelement has attributes to configure worker tasks on specific servers. You can declare as many workers as you
want for a work queue by specifying on which servers the workers run.
Access work-queue.xml in Guidewire Studio at the following location:
configuration→config →workqueue
See also
• “General Work Queue Configuration” on page 93
General Work Queue Configuration

The <work-queue> element in work-queue.xml contains attributes for configuring the general characteristics of a
work queue.
Attribute Description
Required attributesdon't
progressinterval The progressinterval value is the amount of time, in milliseconds, that PolicyCenter
allots for a worker to process the number of batchsize work items. If the time a worker
has held a batch of items exceeds the value of progressinterval, then PolicyCenter con‐
siders the work items to be orphans. PolicyCenter reassigns orphaned work items to a
new worker instance. The progressinterval value must be greater than the time to
process the slowest work item, or that work item never completes.
Guidewire recommends that you set the progressinterval value greater than the proc‐
essing time for an entire batchsize of work items:
• If a worker takes more time than the time specified by progressinterval to
processes its assigned work items, PolicyCenter reverts the remaining work items to
available from checkedout.
• If many worker batches take longer than the time specified by progressinterval, the
repeated checking out and reverting to available of work items can have a negative
impact on performance.
workQueueClass (Required) The workQueueClass value must be one of the following:
• A Guidewire‐provided work queue class listed in the base configuration version of
work-queue.xml
• A custom work queue class derived from Gosu class WorkQueueBase
You cannot configure Guidewire‐provided batch processes or custom batch processes de‐
rived from the Gosu class BatchProcessBase.
Optional attributes
blockWorkersWhenWriterActive If the work queue workers start execution before the work queue writer completes writ‐
ing work items to the work queue, it can possibly cause performance issues under certain
circumstances.
If set to true, PolicyCenter blocks the work queue workers from acquiring new work
items until the writer completes writing work items to the queue. After the writer com‐
pletes writing any new work items, the workers automatically start acquiring work items
again.
The default is false. Only enable this attribute for the work queues for which you require
this capability. Guidewire recommends that you consider setting this attribute to true if
the work queue writer can run for extensive periods of time due to the work load gener‐
ated.
logRetryableCDCEsAtDebugLevel If the value of logRetryableCDCEsAtDebugLevel is set to true for a work queue,
PolicyCenter logs any retryable Concurrent Data Change Exception (CDCE) at the DEBUG
level. The log message includes a prepended string to indicate that the error is non‐fatal.
PolicyCenter logs any CDCE that pushes the retry count over the value of retryLimit, or
the value of workItemRetryLimit if retryLimit is not set, at the ERROR level.

retryInterval How long in milliseconds to wait before retrying a work item that threw an exception. The
default value is 0, meaning PolicyCenter retries processing the item immediately.
retryLimit The number of times PolicyCenter retries a work item that threw an exception or that be‐
came an orphan for this work queue.
If you do not specify a retryLimit value for a work queue, PolicyCenter uses the value of
the WorkItemRetryLimit configuration parameter in config.xml as the default value.
IMPORTANT: Guidewire generally recommends that you increase, never decrease, the
number of retries for a work queue.
Worker Configuration
The use of the <worker> element in work-queue.xml is optional. However, in actual practice, it is necessary for
there to be at least one <worker> element for each <work-queue> element for the work queue to operate properly.
The <worker> element contains an instances attribute that has a default value of 1. Without a <worker> element to
provide this default, the processing logic does not allocate any workers for the work queue.
All of the following attributes are optional.
instances The number of workers to create. By default, PolicyCenter sets the values of this attribute to 1.
If a worker wakes up and detects work items, it checks out those work items from the work queue. If
there are more work items than the value specified by the batchsize attribute, the worker starts another
worker. Each new worker checks out up to the maximum batchsize number of work items. If there are
more work items remaining, the new worker starts another worker. The creation of workers continues
until the number of workers reaches the maximum limit of workers as specified by the instances attrib‐
ute.
maxpollinterval How often a worker wakes up automatically and queries for work items, even if the worker receives no
notification. You might need to increase the value of maxpollinterval to prevent excessive numbers of
queries for work items. The default value of maxpollinterval is 60,000 milliseconds.
throttleinterval The delay between processing work items in milliseconds. The value controls how long the process
sleeps. A value of 0 (zero) means worker tasks process work items as rapidly as possible. To reduce the
CPU load, set the value of throttleinterval to a positive non‐zero value.
batchsize How many work items the worker attempts to check out while searching for more work items. Larger
batch sizes are more efficient, but might not result in good load distribution. The default value for
batchsize is 10.
env The environment in which this particular worker is active.

server The serverid of the server on which this particular worker is active.
See also
• For information about the definition of the env and the serverid values in the cluster registry in config.xml,
see “Understanding the Configuration Registry Element” on page 42.
Worker Task Management

An executor manages the worker tasks on each cluster server with the appropriate role. In the base configuration,
Guidewire assigns this functionality to the workqueue server role. Each server with the worqueue role creates one
executor for each work queue on that server.
Each work queue executor periodically creates a worker. (The executor can also create a worker upon receiving a
notification from the writer.) This worker checks the work queue for items to process. If necessary, the initial worker
creates an additional worker if there is more work to process than it can handle. This new worker can also create a
worker if there is still more work to process. After there is no more work to process, all active workers stop.
It is possible to control the maximum number of workers, for all work queues on a server, by setting the value of
configuration parameter WorkQueueThreadPoolMaxSize in config.xml. It is possible to set this value individually
on each PolicyCenter server in the cluster.
See also
• “General Work Queue Configuration” on page 93
• “Work Queues and Server Roles” on page 95
Work Queues and Server Roles

In the base configuration, Guidewire assigns work queue functionality to servers with the workqueue role. File
work-queue.xml associates the workequeue server role with the application work queue functionality.
<work-queues xmlns="http://guidewire.com/work-queue" defaultServer="#workqueue" />
Note: The hash mark in front of workqueue (#workqueue) indicates that the value that follows the
hash mark is a server role and not a server ID.
The workqueue role is merely the default role, however. You are free to create and assign new work queue
management roles. You can also use server roles to enable or disable certain work queues on a specific PolicyCenter
server.
Defining a New Work Queue Role

You define server roles using the roles attribute on the <registry> element in file config.xml. In the base
configuration, Guidewire defines the following server roles:
<registry roles="batch, workqueue, scheduler, messaging, startable, ui" />
To add a specialized work queue role, say, one to use in managing activity work queues, you need merely to add the
new server role to the list of roles:
<registry roles="batch, workqueue, activityworkqueue, scheduler, messaging, startable, ui" />
See also
• “Defining a New Server Role” on page 46
Assigning a Work Queue to Specific PolicyCenter Cluster Servers

Suppose that you want to assign the management of the activity work queues in the PolicyCenter cluster to a subset
of the cluster servers with the activityworkqueue role. By default, PolicyCenter distributes all other work queues
to those servers with the default workqueue role.
For example, in file work-queue.xml, you define the following:
<?xml version="1.0"?>
<work-queues xmlns="http://guidewire.com/work-queue" defaultServer="#workqueue">
<work-queue workQueueClass="com.guidewire.pl.domain.escalation.ActivityEscalationWorkQueue" … >
<worker server="#activityworkqueue"/>
</work-queue>
<work-queue workQueueClass="com.guidewire.pl.domain.geodata.geocode.GeocodeWorkQueue" … >
<worker/>
</work-queue>
</work-queues>
In this example:
1. If a server has the workqueue role only, then that server:

a. Starts an executor for the GeocodeWorkQueue work queue.
b. Does not start an executor for the ActivityEscalationWorkQueue work queue.
2. If a server has the activityworkqueue role only, then that server:
a. Starts an executor for ActivityEscalationWorkQueue work queue.
b. Does not start an executor for the GeocodeWorkQueue work queue.
3. If a server has both the activityworkqueue and workqueue roles, then that server starts executors for both
work queues.
4. If a server has neither the activityworkqueue nor the workqueue role, then the server does not start an
executor for either of these work queues.
See also
Performing Custom Actions After Batch Processing Completion

You can use Process Completion Monitor batch processing to launch custom actions after a work queue or batch
process completes a batch of items. For example, you might want to start the writer of a follow-on work queue
during nightly batch processing.
Process Completion Monitor batch processing runs at schedulable intervals and examines the ProcessHistory table
for all completed work queues and batch processes.
For each completed work queue that it finds, Process Completion Monitor:
• Determines if all the work items in that work queue have either completed or failed.
• Calls the IBatchCompletedNotification plugin implementation on a process if the process is complete and has
no remaining available or checked-out work items.
• Sets ProcessHistory.NOTIFICATIONSENT to true to invoke the IBatchCompletedNotification plugin
implementation a single time only for any given process.
The IBatchCompletedNotification interface has a completed method that you can override to perform specific
actions if a work queue or batch process finishes a batch of work. The parameters of the completed method are the
ProcessHistory entity and the number of failed items. PolicyCenter considers work queue processing as complete
if no work items remain on the queue, other than work items that failed. PolicyCenter considers a batch process as
complete if the process stopped and its process history is available.
See also
• “Schedule the Process Completion Monitor Batch Process” on page 96
• “Implement the IBatchCompletedNotification Interface” on page 97
• “Register a Custom Batch Notification Plugin” on page 97
• “Process Completion Monitor Batch Process” on page 115
Schedule the Process Completion Monitor Batch Process

Process Completion Monitor batch processing runs at schedulable intervals and examines the ProcessHistory table
for all completed work queues and batch processes.
About this task

In the base configuration, Guidewire does not schedule the Process Completion Monitor batch process. To enable
this batch process, you need to add the batch process to file scheduler-config.xml.
Procedure
1. In the PolicyCenterStudio Project window, expand configuration→config→scheduler.
a. Open scheduler-config.xml.
b. Add the following <ProcessSchedule> element:
<ProcessSchedule process="ProcessCompletionMonitor">
<CronSchedule minutes="*/5"/>
</ProcessSchedule>
See also
Implement the IBatchCompletedNotification Interface

The Process Completion Monitor calls the IBatchCompletedNotification plugin implementation on a process if
the process is complete and has no remaining available or checked-out work items.
Procedure
1. In the PolicyCenterStudio Project window, expand configuration→gsrc.
2. Do one of the following:
• If a package for your plugin implementation classes already exists within gsrc, navigate to that package, then
skip to step 6.
• If a package for your plugin implementation classes does not exist, continue to step 4.
3. Right-click gsrc, then click New→Package.
4. Enter a package name, such as workqueue.
5. Right-click your implementation class package and click New→Gosu Class.
6. Enter the name IBatchCompletedNotification for the gosu class.
7. Click OK.
8. Define your Gosu class, using the following framework:
package myCompany.plugin.workqueue
uses gw.plugin.workqueue.IBatchCompletedNotification
class IBatchCompletedNotification implements IBatchCompletedNotification {
construct() { }
override function completed(batch : ProcessHistory, numFailed : int) {
//do something
return
}
}
9. Save your work.
Register a Custom Batch Notification Plugin

After you create your IBatchCompletedNotification plugin implementation, you need to register the plugin with
PolicyCenter.
Procedure
1. In the PolicyCenterStudio Project window, expand configuration→config→Plugins.
2. Right-click registry and click New→Plugin.
3. In the Plugin dialog, enter IBatchCompletedNotification for the name of your plugin.
4. In the Plugin dialog, click …
5. In the Select Plugin Class dialog, type IBatchCompletedNotification and select the IBatchCompletedNotification
interface.
6. In the Plugin dialog, click OK.
Studio creates a GWP file under Plugins→registry with the name that you entered.
7. Click the Add Plugin icon and select Add Gosu Plugin.
8. For Gosu Class, enter your class, including the package.
See also
Troubleshooting Work Queues

It is possible for a work queue worker to encounter a problem that causes it to fail, before the worker completes all
the items it checked out from the queue. For example, it is possible for a server to die, killing its workers in the
middle of processing. This action can result in orphan work items. A work item becomes an orphan if a worker has
an item checked out but does not complete processing the item within an allotted amount of time. The
progressinterval attribute on the <worker> element in work-queue.xml defines this time span.
Workers treat orphans just as they do available items. The next worker that encounters the orphan item in the table
adopts it for processing and resets the LastUpdateTime, CheckedOutBy, and Status fields on the orphan work item.
If a work queue is experiencing a large number of orphans, review log files to locate timeouts during processing. For
example, a timeout can occur while a worker waits for an external server to return a value. If the log contains these
type of timeouts, increase the progressinterval value for the work queue in work-queue.xml to give workers
more processing time.
Sometimes, a problem inherent in the item itself causes the processing of the item to fail. For example, a batch
process throws an exception. In such cases, the worker stops processing the item and goes on to the next. The item
becomes an orphan and the next worker attempts to process it. In this way, a work queue attempts to process each
item multiple times up to a limit configured for the work queue. If a work item exceeds the limit of processing
attempts, PolicyCenter changes the status of the work item to failed. Workers ignore items with a status of failed
and no longer attempt to process them.
See also
• “Purge Failed Work Items Batch Process” on page 117
Work Queues and Batch Processes, a Reference

In the base configuration, PolicyCenter provides a number of work queues and batch processes.
See also
• “Working with Work Queue Writers and Batch Processes” on page 86
• “Terminate a Writer or Batch Process from the Command Prompt” on page 87
• “Unused and Internal Batch Processes” on page 127
Account Holder Count Work Queue

Code AccountHolderCount

Categories UIRunnable, Schedulable
Implementation Work queue

Schedule Not scheduled
Class AccountHolderCountWorkQueue.java
On Contact objects, the AccountHolderCount values sometimes are incorrect. Account Holder Count batch
processing finds Contact objects with incorrect AccountHolderCount values, and updates the value. If the value of
AccountHolderCount is set incorrectly, it can affect search performance.
Based on the number of contacts that need updating, decide whether to run Account Holder Count manually or
automatically as a scheduled batch process.
• If the number of contacts to update is relatively small and does not change often, consider running it manually.
Periodically check the data distribution in case something later causes Contact.AccountHolderCount fields to
be incorrect.
• If the number of contacts to update is large, or if contacts regularly have incorrect AccountHolderCount values,
consider scheduling Account Holder Contact to run on a regular basis.
Determine the Number of Incorrect AccountHolderCount Contacts
About this task
In the base configuration, Guidewire does not schedule Account Holder Count batch processing. To decide whether
to schedule the batch process or run it manually, do the following:
Procedure
1. Log into Policy using an administrative account.

2. Press ALT+SHIFT+T to open the Server Tools screen.
3. Navigate to Info Screens→Data Distribution.
4. In the Collect distributions for all tables field, select Specify tables.
5. Add the pc_contact table.
6. Click Submit Data Distribution Batch Job.
The pc_contact data distribution table reports the Number of Contacts with incorrect AccountHolderCount
field.
See also
• “Data Distribution” on page 342
Account Withdraw Evaluation Work Queue

Code AccountWithdraw

Schedule Once a day, at 2 a.m.
Class AccountWithdrawWorkQueue.gs

The Account Withdraw Evaluation process marks the account status as withdrawn (Withdrawn) if:
• There are no policies associated with the account.
• The Account.CreateTime or Account.OriginationDate is older than a configurable number of months in the
past. The AccountsWithdrawnAfterMonths parameter in config.xml specifies the number of months. In the
base configuration, this parameter is set to 37 months.
• There are no open activities associated with the account.
Activity Escalation Work Queue

Code ActivityEsc

Schedule Every 30 minutes
Class ActivityEscalationWorkQueue.java
The Activity Escalation work queue finds activities that meet certain escalation criteria and marks the activity for
escalation. The process logic looks for activities that meet each of the following criteria:
• The activity has an escalation date.
• The escalation date is prior to today’s date.
• PolicyCenter has not previously escalated the activity.
If the Activity Escalation work queue finds an activity that meets all the criteria, it marks the activity as escalated
and calls the activity escalation rules to determine any actions.
If you set your escalation deadline in days, then there is no reason to run activity escalation more than daily.
However, if your escalation deadline is shorter, then run this process more frequently to take action on overdue
activities in a timely manner. By default, PolicyCenter runs Activity Escalation batch processing every 30 minutes.
As indicated, you can change this schedule as needed.
See also
• Application Guide
• Rules Guide
Apply Pending Account Data Updates Batch Process

Code ApplyPendingAccountDataUpdates
Implementation Batch process

Stoppable No (single phase process)
Schedule Once a day, at 12:10 a.m.
Class ApplyPendingAccountDataUpdates.gs
The Apply Pending Account Data batch process synchronizes changes made during future-dated jobs with the
backing account information. For examples, suppose that a customer updates her personal auto policy to show that
her marital status changes on some future date. This process applies that change to the corresponding account on the
edit-effective date of the policy change job.
See also
Archive Policy Terms Work Queue

Code ArchivePolicyTerm
Categories Schedulable, UIRunnable

Class ArchivePolicyTermWorkQueue.java
The Archive Policy Term work queue archives policy terms. The process calls the IPCArchivingPlugin
implementation to determine whether a policy is eligible for archiving. For a policy period to be eligible for
archiving, the server time must have reached the PolicyTerm.NextArchiveCheckDate date.
It is possible for the Archive Policy Term work queue to make large changes to the database tables. After running
the Archive Policy Term process, Guidewire recommends that you update database statistics. Updating database
statistics enables the optimizer to pick better queries based on more current data.
Scheduling the Archive Policy Terms Work Queue

By default, Guidewire disables the scheduling of the Archive Policy Terms Work Queue in the base configuration by
commenting out the scheduling information for this process in file scheduler-config.xml. To schedule this work
queue, remove the comment marks from around the schedule information.
See also
• “Understanding Database Statistics” on page 255
• “Archive Info” on page 331
Archive Reference Tracking Sync work queue

Code ArchiveReferenceTrackingSync

Class ArchiveReferenceTrackingSyncWorkQueue.java
The Archive Reference Tracking Synch work queue finds all references from any archived documents to any object
instances in the entity graph. You run this process a single time only to build the initial table of archived objects.
To schedule this work queue, you must add the appropriate information to file scheduler-config.xml.
See also
• For a full description of the ArchiveReferenceTrackingSync work queue, see the Configuration Guide.
Audit Task Work Queue

Code AuditTask

Class AuditTaskMonitorWorkQueue.java
The Audit Task work queue starts scheduled audits.
See also
Bound Policy Exception Work Queue

Code BoundPolicyException

Class BoundPolicyExceptionWorkQueue.java
The Bound Policy Exception work queue runs policy exception rules on every bound PolicyPeriod that has not had
exception rules run on it for more than a defined number of days.
Parameter BoundPolicyThresholdDays in config.xml defines the minimum number of days that must pass before
PolicyCenter runs the Exception rules again on a bound PolicyPeriod. To be eligible for processing, the
PolicyPeriod must be either in force or expired within the last year.
Note: You must uncomment the entry for this work queue in file scheduler-config.xml to schedule
this work queue.
See also
• Rules Guide
Cleanup ETLPurgeRoot Work Queue

Code CleanupETLPurgeRoot

Schedule Every Monday, at 4:00 a.m.
Class PurgeOldETLPurgedRootWorkQueue.java
The Cleanup ETLPurgeRoot work queue deletes old ETLPurgeRoot entities that have a purge date that is older than a
calculated date, which is the current date minus a set number of days.
Parameter KeepPurgedRootsForDays in config.xml sets the minimum number of days that must pass after the
purge event has occurred for the request to be eligible for delete by the Cleanup ETLPurgeRoot process. In the base
configuration, Guidewire sets the value of this configuration parameter to 180 days.
Clean Up Account Contact Role Table Work Queue

Code CleanupAccountContactRole

Schedule Saturday, at 4 a.m.
Class CleanupAccountContactRoleWorkQueue.gs
The Clean Up Account Contact Role Table work queue deletes retired AccountContactRole entity instances and
any associated AccountContactRoleReplacement entity instances. In addition, the process marks
AccountContactRole entity instances that are no longer in use as Retired.
Clear Policy Renewal Check Dates Batch Process

Code PolicyRenewalClearCheckDate
Categories APIRunnable, MaintenanceOnly

Schedule Not schedulable
Class PolicyRenewalClearCheckDate.gs
The Clear Policy Renewal Check Dates batch process clears (null out) the existing check date for all policies
through a single direct database update statement. Because this is a direct update statement, this batch process is
only available in maintenance mode. In other words, the server must be a run level of NO_DAEMONS or lower.
Only run this process if a configuration change to automated renewal contains a possible risk that the automated
renewal process picks up some policies unacceptably late for renewal. Ultimately, the implementation of the
PolicyRenewalPlugin plugin controls the lead time of any individual policy needs for renewal. By default, this
process depends on the code of that plugin and the NotificationConfig system table accessed through the
NotificationConfigPlugin plugin implementation. As you can overwrite the functionality of either plugin, which
types of changes might significantly change the renewal start date can vary.
This process is not available from the PolicyCenter interface. It is also not possible to schedule this batch process.
Instead, you must start this process using the maintenance_tools -startprocess command option with
PolicyRenewalClearCheckDate or its equivalent web service command.
Run the Clear Policy Renewal Check Dates Process

Run Clear Policy Renewal Check Dates batch processing to clear (null out) the existing dates for all policies through
a single, direct, database update.
Before you begin

Only run this process if one or more of the following is true:
• If a configuration change to automated renewal contains a possible risk that the automated renewal process picks
up some policies unacceptably late for renewal.
• After you bring the server back up after applying a configuration update.
Procedure
1. Place the server in maintenance mode (NO_DAEMONS or lower).
2. Run the following maintenance_tools command a single time only.
maintanance_tools -password password -startprocess PolicyRenewalClearCheckDate
3. After batch processing completes, finish bringing the server all the way up to production mode.
See also
Closed Policy Exception Work Queue

Code ClosedPolicyException

Class ClosedPolicyExceptionWorkQueue.java
The Closed Policy Exception work queue runs policy exception rules on every closed PolicyPeriod that has not
had exception rules run on it for a defined number of days.
Parameter ClosedPolicyThresholdDays in config.xml defines the minimum number of days that must pass
before PolicyCenter runs the Exception rules again on a closed PolicyPeriod.
See also
• Rules Guide
Create UWRules for UWIssue Types Work Queue

Code CreateUnmappedUWRules
Categories UIRunnable

Class CreateUnmappedUWRulesWorkQueue
After an upgrade, the Create UWRules for UWIssue Types work queue runs at server start. For each UWIssueType
object that does not have a corresponding UWRule object, this process does the following:
• It creates a new UWRule object.
• It links the new UWRule object with its corresponding UWIssueType object.
• It marks the associated business rule as being externally managed.
Data Distribution Batch Process

Code DataDistribution

Categories APIRunnable

Class PCDataDistributionBatchProcess.java
The Data Distribution batch process generates data on the distribution of various items in the PolicyCenter database.
It is not possible to schedule this process. You must run this process from the Data Distribution screen of the Server
Tools Info Pages or by using the maintenance_tools command line utility.
As this type of batch process can be very resource intensive, it has the possibility of adversely affecting the
performance of the application. Before you run this process in a production environment, Guidewire recommends
that you run the process first against a test environment that contains a full copy of production data.
See also
Database Consistency Check Work Queue

Code DBConsistencyCheck
Categories Schedulable

Non‐exclusive Yes
Class DBConsistencyCheckWorkQueue.java
The Database Consistency Check work queue runs consistency checks on the PolicyCenter database.
To schedule the consistency checks process, use the following system_tools command, adding the optional
information on which checks to run against which tables, if you choose:
system_tools -user user -password password -checkdbconsistency ...
See also
• “Work Queues” on page 82 for a discussion of how PolicyCenter handles work queues.
• “Database Consistency Checks” on page 240 for an overview of database consistency checks
• “Consistency Checks” on page 333 for details of the Consistency Checks screen in PolicyCenter
• “Command Prompt Tools” on page 385 for an explanation of command prompt options
Database Statistics Work Queue

Code DBStats


Class DBStatisticsWorkItemWorkQueue.java
The Database Statistics work queue generates database statistics about how the PolicyCenter application and data
model interact with the physical database. For example, database statistics store row counts in a table, how a table
distributes the data, and much more. A database management system uses statistics to determine query plans to
optimize performance.
IMPORTANT Do not run or schedule this process if you set <databasestatistics> attribute
useoraclestatspreferences to true in file database-config.xml.
Development Mode
In development mode, it is possible to run Database Statistics batch processing in any of the following ways:
• From a command prompt, using the -updatestatistics option of the system_tools command
• From the Execution History tab of the Server Tools Database Statistics screen
• As a scheduled batch process
Production Mode
In production mode, it is possible to run Database Statistics batch processing in the following ways only:
• From a command prompt, using the -updatestatistics option of the system_tools command.
Guidewire recommendation
Guidewire specifically recommends that you collect full statistics in the following circumstances:
• If there are significant changes to data such as after a major upgrade.
• If using the zone_import command.
• If you are trying to troubleshoot performance problems.
In all other cases, Guidewire recommends that you collect INCREMENTAL database table statistics only.
See also
• “Managing Database Statistics using System Tools” on page 257
• “Database Statistics” on page 343
Deferred Upgrade Tasks Batch Process

Code DeferredUpgradeTasks

Class UpgradeDeferredRecreateIndexesBatchProcess.java
The Deferred Upgrade Tasks batch process creates the nonessential performance indexes and indexes on archived
entities.
PolicyCenter runs the Deferred Upgrade Tasks batch process after an upgrade automatically if you set the following
attribute on <upgrade> in database-config.xml to true:
defer-create-nonessential-indexes
If DeferredUpgradeTasks batch processing fails, manually run the batch process again during non-peak hours. To
manually run the Deferred Upgrade Tasks batch processing, use the following command:
maintenance_tools -server url -password password -startprocess DeferredUpgradeTasks
Note: To run this command, you must have permission to create indexes until after the
DeferredUpgradeTasks batch process completes.
To check the status of the DeferredUpgradeTasks batch processing, review the upgrade logs and the PolicyCenter
Server Tools Upgrade and Versions screen.
Production Mode
Do not go into full production while the Deferred Upgrade Tasks process is still running. The lack of so many
performance-related indexes can likely make PolicyCenter unusable.
Until the Deferred Upgrade Tasks batch process has run to completion, PolicyCenter reports errors during schema
validation while starting. These include errors for column-based indexes existing in the data model but not in the
physical database and mismatches between the data model and system tables.
Do not use the PolicyCenter archiving feature until the Deferred Upgrade Tasks batch processing completes
successfully.
See also
• “Upgrade and Versions” on page 362
• “Maintenance Tools Options” on page 390
• Upgrade Guide
Destroy Contact For Personal Data work queue

Code DestroyContactForPersonalData

Class PersonalDataContactDestructionWorkQueue.gs
The Destroy Contact For Personal Data work queue finds all PersonalDataContactDestructionRequest objects
that have a status set to New or ReRun (category ReadyToAttemptDestruction). How far the destruction process
went for the found contacts is determined by the ContactDestructionStatus returned from the Destroyer, the class
that implements the PersonalDataDestroyer interface.
The process sets the contact destruction status to the returned status. If that status is Completed, Partial, or
NotDestroyed (category DestructionStatusFinished), the process also populates the date of completion
information.
The process throws an exception if the return status is New or if you try to change the status from a typecode in the
DestructionStatusFinished category.
See also
Extract Rating Worksheets Work Queue

Code ExtractWorkSheets

Class WorksheetExtractWorkQueue.java
The Extract Rating Worksheets work queue extracts rating worksheet data from worksheet container
(WorksheetContainer) objects to files in a specified directory on the server with the batch role. The process also
marks WorksheetContainer objects for purging.
This process calls the implementation of the WorksheetExtractPlugin plugin, which sets the target directory.
See also
• “Purge Rating Worksheets Work Queue” on page 121
Fix JobGroup on Moved Policies Work Queue

Code FixJobGroupOnMovedPolicies

Class FixJobGroupOnMovedPoliciesWorkQueue.gs
The Fix JobGroup on Moved Policies work queue assigns the correct job group category to Job objects associated
with policies moved between accounts.
Form Text Data Delete Batch Process

Code FormTextDataDelete

Class FormTextDataDeleteBatchProcess.java
The Form Text Data Delete batch process deletes orphaned, purged, or archived FormTextData objects. A
FormTextData object stores the text data that comprises the XML data for a Form.
Geocode Writer Work Queue

Code Geocode


Class GeocodeWorkQueue.java
The Geocode Writer work queue runs periodically to update geocoding information on user contact (UserContact)
primary addresses and account locations. The UserContact entity represents a PolicyCenter user.
Scheduling the Geocode batch process
To schedule the Geocode Writer work queue, uncomment the entry for it in scheduler-config.xml.
See also
• “Understanding Geocoding” on page 75

• “Configuring Geocoding” on page 76
Group Exception Work Queue

Code GroupException

Class GroupExceptionWorkQueue
The Group Exception work queue executes any defined group exception business rules on all groups in the system.
See also
• Rules Guide
Handle Unresolved Contingency Work Queue

Code HandleUnresolvedContingency
Class HandleUnresolvedContingencyWorkQueue.gs

Schedule Once a day at 2:00 a.m.
The Handle Unresolved Contingency work queue initiates action on pending contingencies with an action start date
of either today or a past date on which action has not yet started.
Adding Contingency Actions
You can also add contingency actions by modifying Gosu class HandleUnresolvedContingencyWorkQueue. Add a
type code to the ContingencyAction type list. In ContingencyEnhancement.gsx, add an action start date for this
action in the updateActionStartDate method. Add a property getter for the new job using
isPolicyChangeAction as a model.
See also
Impact Testing Export Work Queue

Code ImpactTestingExport

Class ImpactTestingExportWorkQueue.java
The Impact Testing Export work queue exports test periods to an Excel file if you click Create Excel Export File on the
PolicyCenter Impact Results screen.
It is only possible to run Impact Testing Export batch processing from the Server Tools Batch Process Info screen.
See also
Impact Testing Test Case Preparation Work Queue

Code ImpactTestingTestPrep

Class ImpactTestingTestPrepWorkQueue.java
The Impact Testing Test Case Preparation work queue generates baseline policy periods on the selected policies if
you click Create Baselines from the PolicyCenter Create Baseline screen.
It is only possible to run Impact Testing Test Case Preparation batch processing from the Server Tools Batch Process
Info screen.
See also
Impact Testing Test Case Run Work Queue

Code ImpactTestingTestRun

Class ImpactTestingTestRunWorkQueue.java
The Impact Testing Test Case Run work queue generates test policy periods rated using the selected rate books if
you click Quote Test Periods from the PolicyCenter Testing Periods screen.
It is only possible to run this process from the Server Tools Batch Process Info screen.
See also
Job Expiration Work Queue

Code JobExpire

Class JobExpirationWorkQueue.java
The Job Expire batch work queue expires a job if the job has had no action taken on it for a configurable period of
time. Job Expire batch processing changes jobs from New, Draft, or Quote status to Expired. In the default
configuration, the process expires submissions in these statuses that are at least seven days past the effective date of
the policy.
To configure the expiration threshold as the number of days past the effective date or the creation date, modify the
corresponding configuration parameter in config.xml:
• JobExpirationEffDateThreshold
• JobExpirationCreateDateThreshold
Enable Expiration for a Specific Job Type
About this task
Job Expire batch processing examines all jobs meeting the date criteria. but only expires those jobs for which
job.canExpireJob returns true.
Procedure
1. In the PolicyCenter Studio Project window, expand configuration → config:

a. Open config.xml.
b. Search for JobExpireCheck<JobType>, for example, JobExpireCheckAudit.
c. Change the value from false to true.
2. (Audit only) In the Project window, expand configuration→gwrc→gw→job:
a. Open AuditProcess.gs.
b. Modify the canExpireJob method to return true instead of false.
Often, business requirements do not permit the expiration of Audit jobs. To expire the audit job type, you
need override the AuditProcess.canExpireJob method.
Result
Forcing the expiration of all jobs for which expiration is possible improves the performance of queries related to
jobs.
See also
Notify External System For Personal Data work queue

Code NotifyExternalSystemForPersonalData

Class PersonalDataDestructionNotifyExternalSystemsWorkQueue.gs
The Notify External System For Personal Data work queue finds all PersonalDataDestructionRequest objects
that have a status typecode in the DestructionStatusFinished category and RequestersNotified set to false.
The work queue processes found requests by sending a notification to all associated requesters, and marking
RequestersNotified as true. If the notification fails, the process throws an exception and RequestersNotified
remains false.
Note: The class that implements this work queue is
PersonalDataDestructionNotifyExternalSystemsWorkQueue. In your implementation, you must
verify that the notification was successful before marking the RequestersNotified property as true.
A method on the PersonalDataDestruction plugin, notifyExternalSystemsRequestProcessed, is called by the
PersonalDataDestructionNotifyExternalSystemsWorkQueue to notify external systems when a personal data
destruction request completes. The process passes the original RequestID to the method, which does nothing by
default. You are expected to implement this method to notify systems of interest. The process receives the
RequestID through the PersonalDataDestructionAPI web service when the destruction request is originally
created.
Note: In the base configuration, the class that implements the PersonalDataDestruction plugin is
PCPersonalDataDestructionPlugin.
See also
Open Policy Exception Work Queue

Code OpenPolicyException

Class OpenPolicyExceptionWorkQueue.java
The Open Policy Exception work queue runs policy exception rules on open, unlocked PolicyPeriod objects that
have not had exception rules run on them for a definable number of days.
Parameter OpenPolicyThresholdDays in config.xml defines the minimum number of days that must pass before
PolicyCenter runs the Exception rules again on an open and unlocked PolicyPeriod.
By default, Guidewire comments out the entry for this work queue in scheduler-config.xml. To schedule this
work queue, remove the comment characters from around the entry for the work queue.
See also
• Rules Guide
Overdue Premium Report Work Queue

Code OverduePremiumReport

Class OverduePremiumReportWorkQueue.java
The Overdue Premium Report work queue runs premium report escalation rules on overdue premium reports.
Phone Number Normalizer Work Queue

Code PhoneNumberNormalizer

Schedule Once only, after an upgrade to 8.0.0+
Class CompactPhoneNormalizerWorkQueue.java
IMPORTANT The Run Phone Number Normalizer work queue once only, after upgrading from earlier
application versions to 8.0.0+. Disable the Phone Number Normalizer work queue in a production
environment.
The Phone Number Normalizer work queue calls the registered plugin that implements the
IPhoneNumberNormalizer interface. Use the Phone Number work queue to normalize phone numbers after
upgrading from earlier versions of PolicyCenter to 8.0.0+.
Guidewire recommends that you use a substantial number of workers with the Phone Number Normalizer work
queue. Using a small number of workers to normalize the phone numbers in a large database can take a very long
time. The optimal number of workers to use varies according to the available hardware and the volume of the data
involved. It is also possible to allocate workers to several different PolicyCenter servers rather then simply
increasing the number of workers on a single server.
Disable this work queue after the process completes normalizing all old phone numbers by setting the number of
workers in work-queue.xml to 0. You never need to run the Phone Number Normalizer work queue more than once,
after an upgrade to 8.0.0+.
See also
• “Configuring Work Queues” on page 92
• “Upgrading phone numbers” in the Upgrade Guide
• “Phone number normalizer plugin” in the Integration Guide
Policy Hold Job Evaluation Work Queue

Code PolicyHoldJobEval


Class PolicyHoldJobEvalWorkQueue.java
The Policy Hold Job Evaluation work queue evaluates each job against the policy holds blocking it. The process
finds policy hold jobs that have the following characteristics:
• Jobs that are open
• Jobs that have policy periods with an active blocking hold
• Jobs that have not been evaluated since the time the policy hold changed
The PolicyHoldJobEvalPlugin plugin implementation determines the actions to take on each job found.
Policy Locations Risk Assessment Work Queue

Code PolicyLocationsRiskAssessment

Class PolicyLocationsRiskAssessmentWorkQueue.gs
The Policy Locations Risk Assessment work queue starts risk assessments for all policy locations associated with a
policy period. This process sends an activity to the user requesting risk assessment that indicates whether the risk
assessment succeeded or failed.
In the base configuration, PolicyCenter calls this work queue if you select Update Risk Evaluations on the PolicyCenter
Risk Analysis screen in a commercial property policy transaction.
This work queue retrieves risk assessment results from the Guidewire Spotlight risk assessment service for each
location on a given PolicyPeriod and creates LocationRiskAssessment objects. Because PolicyCenter sets an
effective date on each LocationRiskAssessment entity that it creates, it is only possible to run this process on an
unlocked PolicyPeriod object.
If the process fails, it does the following:
• It rolls back the entire transaction and does not persist any of the risk assessment results.
• It creates an activity and sends the activity to the list of configured user roles on the current job. You can
configure the list of roles in the SpotlightNotificationActivityCreator class.
• It sends a notification activity to the requesting user, if the work item includes a reference to that user.
If the process succeeds, it does the following:
• It commits the transaction and persists the risk assessment results on the PolicyPeriod.
• It creates an activity and sends the activity to the list of configured user roles on the current job.
• It sends a notification activity to the requesting user, if the work item includes a reference to that user.
See also
Policy Renewal Start Work Queue

Code PolicyRenewalStart

Class PolicyRenewalStartWorkQueue.java

The Policy Renewal Start work queue starts renewal processing for policies set to expire. The process determines
the date on which to start the renewal process by subtracting the value of configuration parameter
RenewalProcessLeadTime in config.xml from the policy expiration date.
Populate Search Columns Batch Processing

Code PopulateSearchColumns

Class PopulateSearchColumnBatchProcess.java
The Populate Search Columns batch process populates denormalized searchColumn columns from their designated
sourceColumn columns.
It is only possible to run this process from the maintenance_tools command.
See also
Premium Ceding Work Queue

Code PremiumCeding

Class RICedingWorkQueue
The Premium Ceding work queue performs calculations related to premium ceding.
See also
Integration Guide
Process Completion Monitor Batch Process

Code ProcessCompletionMonitor

Class ProcessCompletionMonitor.java
The Process Completion Monitor batch process runs at configurable intervals and examines the ProcessHistory
table for all completed work queues and batch processes.
For each completed work queue that it finds, Process Completion Monitor:
• Determines if all the work items in that work queue have either completed or failed.
• Calls the IBatchCompletedNotification plugin implementation on a process if the process is complete and has
no remaining available or checked-out work items.
• Sets ProcessHistory.NOTIFICATIONSENT to true to prevent the process from invoking the
IBatchCompletedNotification plugin implementation more than a single time for any given process.
The IBatchCompletedNotification interface has a completed method that you can override to perform specific
actions after a work queue or batch process completes a batch of work.
By default, Guidewire does not add an entry for this batch process to scheduler-config.xml in the base
configuration. To schedule this batch process, add an entry for it to scheduler-config.xml.
See also
• “Performing Custom Actions After Batch Processing Completion” on page 96.
Process History Purge Batch Process

Code ProcessHistoryPurge

Schedule On the third day of the month, at 1:30 a.m.
Class ProcessHistoryPurge.gs
The Process History Purge batch process purges batch process history data from the ProcessHistory table. It is
important to periodically delete process history data. A large number of history records in the database can slow
performance during use of the Server Tools Batch Process Info or Work Queue Info screens.
This process uses Gosu class ProcessHistoryPurge to read the value of the BatchProcessHistoryPurgeDaysOld
parameter in config.xml. The process then uses this value to determine the history data to purge.
Note: PolicyCenter also uses configuration parameter BatchProcessHistoryPurgeDaysOld to
determine how many days to retain process history records, which the separate “Work Item Set Purge
Batch Process” on page 126 process removes.
Product Model Pattern Activation Batch Process

Code ProductModelPatternActivation
Implementation Batch Process

Class ProductModelPatternActivationProcess.java
The Product Model Pattern Activation batch process makes active newly defined product model patterns added
during a rolling upgrade to a PolicyCenter server cluster. In a rolling upgrade, you bring down a single server at a
time and perform a configuration upgrade on that server. After bringing that server back online, you do the same
with the next server in the cluster and so on.
To provide for data integrity during this process, Guidewire disables any newly added product model patterns until
all servers in the cluster have the same application configuration. Run the Product Model Pattern Activation batch
process after you complete the rolling upgrade and after you flip the Rolling Upgrade Complete flag in the Server Tools
Upgrade and Versions screen.
It is only possible to run this batch process using web services or by using the following maintenance_tools
command option:
maintenance_tools -password password -startprocess productmodelpatternactivation
See also
• “Updating Product Model Patterns in a Rolling Upgrade” on page 169
• “Performing a Rolling Upgrade” on page 172
Purge Batch Process

Code Purge

Class PurgeWorkQueue.java
The Purge batch process purges jobs and prunes policy periods that meet the purge and prune criteria. This process
deletes jobs and other entities from the database. By default, Guidewire disables the Purge batch process in the base
configuration.
See also
• To enable Purge batch processing, the Configuration Guide.
• For information on the criteria that the batch process uses to determine eligibility for deletion from the database,
see the Configuration Guide.
Purge Cluster Members Batch Process

Code PurgeClusterMembers

Schedule On the first day of each month, at 2:00 a.m.
Class PurgeClusterMembers.gs
The Purge Cluster Members batch process purges ClusterMemberData entities. This process uses Gosu class
PurgeClusterMembers to read the value of the ClusterMemberPurgeDaysOld parameter in config.xml. The
process then uses this value to purge the ClusterMemberData entities that have a LastUpdate value prior to the
current date minus the value of the ClusterMemberPurgeDaysOld.
Purge Failed Work Items Batch Process

Code PurgeFailedWorkItems


Schedule On the first day of the month, at 1:00 a.m.
Class PurgeFailedWorkItems.gs
The Purge Failed Work Items batch process purges failed work items from all PolicyCenter work queues. This
process uses Gosu class PurgeFailedWorkItems to determine which work items to delete.
During a scheduled execution of the batch process, the batch process deletes failed work items that are older than the
last run date of the batch process. It then set the last run date to the current date. Thus, if the scheduled execution of
this batch process is monthly, the process deletes work items that are older than a month only.
If you run this batch process manually and there are work items that are newer than the last run date, the batch
process does not delete them. If you then run the batch process a second time on the same day, the process deletes
work items that are older than the current date. This is the expected behavior.
Purge Message History Batch Process

Code PurgeMessageHistory

Schedule On the 20th of the month, at 1:00 a.m.
Class PurgeMessageHistory.gs
The Purge Message History batch process purges old messages from the message history table. The
KeepCompletedMessagesForDays parameter in config.xml specifies how many days a message can remain in the
message history table before Purge Message History batch processing removes the message.
Purge Old Contact Destruction Request work queue

Code RemoveOldContactDestructionRequest

Class RemoveOldContactDestructionRequestWorkQueue.gs
The Purge Old Contact Destruction Request work queue finds all PersonalDataDestructionRequest,
PersonalDataContactDestructionRequest, and PersonalDataDestructionRequester objects that have the
following values:
• The RequestersNotified property is set to true.
• The PersonalDataContactDestructionRequest.DestructionDate value plus the value of the
ContactDestructionRequestAgeForPurgingResults configuration parameter is less than or equal to today’s
date
PolicyCenter removes each found request that has AllRequestsFulfilled equal to true.
By default, Guidewire disables the entry for this work queue in scheduler-config.xml. To schedule this work
queue, remove the comment marks from around the entry in scheduler-config.xml.
See also
Purge Old Transaction IDs Batch Process

Code PurgeTransactionIDs

Class PurgeTransactionIds.gs
The Purge Old Transaction IDs batch process deletes SOAP header transaction IDs generated by systems external to
Guidewire PolicyCenter. This process uses Gosu class PurgeTransactionIds to read the value of the
TransactionIdPurgeDaysOld parameter in config.xml. The process then purges transaction IDs that have a
creation date prior to the current date minus the value of the TransactionIdPurgeDaysOld parameter.
Guidewire does not schedule this batch process in the base configuration as the table that stores the transaction IDs
takes very little space in the database. Unless there is a constant buildup of these transaction IDs, there is no real
need to continually purge this data. In fact, if you do purge this data, it is then not possible to determine if a new
transaction is a duplicate of a transaction sent by the external system at an earlier date. There are other alternatives to
purging this data. For example, you can partition the table by date.
See also
• For information on SOAP headers, see the Integration Guide.

• For information on how to check for duplicate transaction IDs, see the Integration Guide.
Purge Orphaned Policy Period Work Queue

Code PurgeOrphanedPolicyPeriod

Class PurgeOrphanedPolicyPeriodWorkQueue.java
The Purge Orphaned Policy Period work queue finds orphaned policy periods (policy periods not associated with a
specific job) and deletes them. The process deletes policy periods and other entities from the PolicyCenter database.
See “Purge Orphaned Policy Period Work Queue” on page 119 for information on how PolicyCenter selects a policy
period for deletion.
By default, Guidewire disables the Purge Orphaned Policy Period work queue in the base configuration. To enable
this work queue, see the Configuration Guide.
See also

Purge Profiler Data Batch Process

Code PurgeProfilerData

Class ProfilerDataPurgeBatchProcess.java
The Purge Profiler Data batch process purges profiler data at regularly specified intervals. This process uses the
read-only ProfilerDataPurgeBatchProcess class to read the value of the ProfilerDataPurgeDaysOld parameter
in config.xml. The process then uses the value of this parameter to determine how many days to retain profiler data
before the Purge Profiler Data process removes it.
Purge Quote Clones Work Queue

Code PurgeQuoteClones

Class PurgeQuoteClonesWorkQueue.java
Quote cloning creates copies of policy period quotes, known as quote clones. The Purge Quote Clones work queue
deletes quote clones that are marked as processed from the PolicyCenter database.
By default, Guidewire disables the entry for this work queue in scheduler-config.xml in the base configuration.
To schedule this work queue, remove the comment marks from around its entry in scheduler-config.xml.
See also
Purge Rate Book Export Result Work Queue

Code PurgeRateBookExportResult

Schedule Every day at 3 a.m.
Class RateBookExportResultPurgeWorkQueue.gs
The Purge Rate Book Export Result work queue removes spreadsheets and XML files resulting from exporting rate
books to spreadsheet or to XML. The process removes the files from the PolicyCenter database. This processing
does not remove files downloaded to the user’s computer.
This process removes RateBookExportResult Excel and XML files that are more than 60 days old. Specify the
number of days in the RateBookExportResultAgeForPurging parameter in config.xml.
It is possible to disable this process by setting the PurgeRateBookExportResultEnabled parameter in config.xml
to false.
See also
Purge Rating Worksheets Work Queue

Code PurgeWorksheets

Class WorksheetPurgeWorkQueue
The Purge Rating Worksheets work queue removes worksheet container (WorksheetContainer) objects that meet
certain criteria. Configuration parameter RatingWorksheetContainerAgeForPurging in config.xml specifies the
minimum number days after the closure of a job before the WorksheetContainer associated with its policy is
eligible for purging.
By default, Guidewire disables the entry for the Purge Rating Worksheets work queue in scheduler-config.xml in
the base configuration. To schedule this work queue, remove the comment marks from around the entry for it in
scheduler-config.xml.
See also
• “Extract Rating Worksheets Work Queue” on page 108
Purge Risk Assessment Temporary Store Work Queue

Code PurgeRiskAssessmentTempStore

Schedule Every Monday at 1:00 a.m.
Class PurgeRiskAssessmentTempStoreWorkQueue.gs
The Purge Risk Assessment Temporary Store work queue deletes temporary objects created for risk assessment. The
process deletes objects if they are older than the number of days specified by configuration parameter
PurgeRiskAssessmentTempStoreDays in config.xml. In the base configuration, Guidewire sets the value of this
parameter to 30 days.
The Purge Risk Assessment Temporary Store work queue purges temporary objects if the course of standard
PolicyCenter operations does not remove them.
See also
Purge Temporary Policy Periods Work Queue

Code PurgeTemporaryPolicyPeriods


Class PurgeTemporaryPolicyPeriodsWorkQueue.java
Sometimes policy period (PolicyPeriod) objects in a temporary (Temporary) status exist in the database. The
Purge Temporary Policy Periods work queue removes these temporary policy periods from the database.
Configuration parameter PurgeTemporaryPolicyPeriodsAfterDays sets the minimum number of days that must
pass before it is possible to purge a temporary policy period. In the base configuration, Guidewire set the value of
this parameter to 14 day.
To enable batch processing, set configuration parameter PurgeTemporaryPolicyPeriodsEnabled to true. If set to
false, the work queue write does not remove any temporary policy periods.
In the base configuration, Guidewire does not provide an entry for this work queue in scheduler-config.xml. To
schedule this work queue, add an entry for it to scheduler-config.xml.
Purge Workflow Batch Process

Code PurgeWorkflows

Schedule On the first day of the month, at 1:30 a.m. and on the 10th day of the month, at 1:00 a.m.
Class PurgeWorkflows.gs
The Purge Workflow batch process purges completed workflows after resetting any referenced workflows. This
process uses Gosu class PurgeWorkflow to read the value of configuration parameter WorkflowPurgeDaysOld days
in config.xml. The process then uses this value to determine the number of days to retain workflow data before
purging it.
Purge Workflow Logs Batch Process

Code PurgeWorkflowLogs

Schedule On the first day of the month, at 2:30 a.m.
Class PurgeWorkflowLogs.gs
The Purge Workflow Logs batch process purges completed workflows logs. This process uses Gosu class
PurgeWorkflowLogs to read the value of configuration parameter WorkflowLogPurgeDaysOld in config.xml. The
process then uses this value to determine the number of days to retain workflow logs before purging them.
Rate Book Export Work Queue

Code RateBookExport


Schedule Every day, at 3:00 a.m.
Class RateBookExportWorkQueue.gs
The Rate Book Export work queue exports a rate book and its included rate table and rate routines to a file in either
Microsoft Excel or XML format. PolicyCenter initiates this process if the user selects Export to Spreadsheet in the
PolicyCenter Rate Book screen.
See also
Recalculate Contingency Action Start Date Work Queue

Code RecalculateContingencyActionStartDate

Class RecalculateContingencyActionStartDateWorkQueue.gs
The Recalculate Contingency Action Start Date work queue recalculates the action start date on all unresolved
contingencies. It is only necessary to run this work queue if you modify the ContingencyEnhancement.gsx code or
want to update the action start date of existing contingencies.
See also
Reset Purge Status and Check Dates Work Queue

Code ResetPurgeStatusAndCheckDates

Class ResetPurgeStatusAndCheckDatesWorkQueue.java
The Reset Purge Status and Check Dates work queue resets the purge status and purge or prune dates on the jobs.
Run this process if you have a need to reset the purge status and purge date.
For each job, this batch process:
• Sets the Job.PurgeStatus property to Unknown
• Sets the Job.NextPurgeCheckDate to null
For example, if a job has a PurgeStatus of NoActionRequired or Pruned, the batch process resets the
PurgeStatus to Unknown.
It is only possible to run this work queue using the following maintenance_tools command:
maintenance_tools -password password -startprocess resetpurgestatusandcheckdates
Retrieve Policy Terms Work Queue

Code RestorePolicyTerm
Categories UIRunnable Schedulable

Class RestorePolicyTermWorkQueue.java
The Retrieve Policy Terms work queue processes requests to restore archived policy terms.
To schedule this process, uncomment the existingentry for the work queue in file scheduler-config.xml.
Retire Activities Batch Processing

Code ActivityRetire

Schedule Once a day, at 30 minutes past midnight
Class RetireActivityWorkQueue.java
The Retire Activities work queue retires Activity entity records that are either:
• Canceled
• Dismissed
Retrieve Policy Terms Work Queue

Code RestorePolicyTerm

Class RestorePolicyTermWorkQueue
The Retrieve Policy Terms work queue retrieves archived policy terms marked for retrieval. It is possible for a user
to request the retrieval of an archived policy term. Retrieving a policy term generates an activity for the user who
requested the retrieval of the policy term.
To schedule this work queue, remove the comment marks from around the entry for it in scheduler-config.xml.
See also
Solr Data Import Batch Process

Code SolrDataImport


Class SolrDataImportBatchProcess.gs
The Solr Data Import batch process tests the operation of the free-text batch load command, especially its embedded
SQL query. Only run Solr Data Import batch processing on development-mode servers.
IMPORTANT Do not run this process in production to load and re-index the Guidewire Solr Extension.
Instead, run the free-text batch load command (batchload) on the host on which the Guidewire Solr
Extension resides.
See also
• “Free-text Batch Load Command” on page 285
Team Screens Batch Process

Code TeamScreens

Schedule Every hour from 6:00 a.m. to 9:00 p.m., at three minutes past the hour
Class TeamScreensProcess.java
The Team Screens batch process collects statistics for the PolicyCenter Team tab screens.
See also
User Exception Work Queue

Code UserException

Class UserExceptionWorkQueue.java
The User Exception work queue runs the user exception rule set on all users in the system.
See also
• Rules Guide
Work Item Set Purge Batch Process

Code WorkItemSetPurge

Schedule On the second day of the month, at 1:30 a.m.
Class WorkItemSetPurge.gs
The Work Item Set Purge batch process purges work item sets from the database. This process uses Gosu class
WorkItemSetPurge to read the value of the BatchProcessHistoryPurgeDaysOld parameter in config.xml. The
process then uses this value to determine the number of days to retain work item sets before purging them.
Note: The BatchProcessHistoryPurgeDaysOld parameter also configures how many days to retain
process history records, which the separate “Process History Purge Batch Process” on page 116
process removes.
Work Queue Instrumentation Purge Batch Process

Code WorkQueueInstrumentationPurge

Schedule On the second day of the month, at 2:30 a.m.
Class WorkQueueInstrumentationPurge.gs
The Work Queue Instrumentation Purge batch process purges instrumentation data for work queues. This process
uses Gosu class WorkQueueInstrumentationPurge to read the value of the
InstrumentedWorkerInfoPurgeDaysOld parameter in config.xml. The process then uses this value to determine
how long to retain work queue instrumentation data.
See also
Workflow Work Queue

Code Workflow

Schedule Every 10 minutes, starting on the hour (0, 10, 20, 30, 40, 50)
Class WorkflowDistributedWorkQueue.java
The Workflow work queue runs PolicyCenter workflow tasks. Workflow cannot advance any faster in the
background than the work queue schedule. To increase the speed of PolicyCenter workflow, change the schedule of
this work queue.
See also
Unused and Internal Batch Processes

PolicyCenter uses some batch process types either internally, or, in a few cases, not at all. You cannot run these
processes from PolicyCenter, maintenance_tools, or a web services API.
Unused processes
The Batch Process Type typelist (BatchProcessType.tti) includes a few Guidewire platform processes that
PolicyCenter has officially retired. Guidewire indicates this status by setting the retired flag on the process
typecode to true and placing a line through the typecode the typelist. You can ignore these processes.
In addition, there are other processes listed in the Batch Process Type typelist that Guidewire no longer uses but
which you can still see in the typelist, depending on the application version. Examples of these types of processes
include the following:
• Archive
• Bulk Purge
• Contact Auto Synch
• Stat Report
Internal processes
PolicyCenter uses some Guidewire batch processes internally. For example, PolicyCenter runs some processes to
generate database performance reports only. You cannot run internal processes separately. Do not attempt to use the
following internal processes:
• Back Out Rolling Upgrade
• Business Rule Export/Import
• Business Rule Validator
• Create Perf Only Indexes
• Find Usages Of Upgraded Typecodes
• Microsoft DMV Report
• Oracle AWR Report
• Staging Table Import


part 3
Server Clustering Administration

chapter 7
Understanding PolicyCenter Server

Clustering
To improve performance and reliability, you can install multiple PolicyCenter servers in a configuration known as a
cluster. A PolicyCenter cluster distributes client connections among multiple PolicyCenter servers, reducing the load
on any one server. If one server fails, the other servers seamlessly handle its traffic. This topic describes how a
PolicyCenter cluster functions.
See also
Cluster Terminology
Guidewire uses the following terminology in discussions involving Guidewire PolicyCenter clusters.
Term Meaning
Host The physical machine on which one or more Guidewire applications run.
Application instance An individual PolicyCenter deployment. It is possible to run multiple application instances on the same
host.
PolicyCenter server The server associated with each application instance. For production environments, Guidewire supports
JBoss, Tomcat, WebLogic, and WebSphere servers. If running multiple servers on the same host, each
server must map to a different physical port.
Server role A categorization of each application instance in the cluster by its function, as defined by its role. Exam‐
ples of server roles are ui (manages user interface requests), batch (manages batch processing), and
messaging (manages messaging and message destinations).
You define server roles using the cluster <registry> element in config.xml. See “Server Roles” on page
139 for more information. See also “Understanding the Configuration Registry Element” on page 42.
Cluster A grouping of two or more Guidewire application instances that have a common configuration and func‐
tion as an integrated unit. Typically, each server in the cluster has one or more roles or function. Most
often, if there are multiple servers assigned the ui role, the cluster contains a third‐party load balancer
as well.
The individual application instances in the cluster must all connect to a common database.
Cluster member A single application instance within a Guidewire cluster.
Understanding PolicyCenter Server Clustering 131

Guidewire PolicyCenter Cluster Installations

The typical PolicyCenter cluster consists of two or more PolicyCenter servers, each of which has one or more server
roles, and a load balancer. In general, a PolicyCenter cluster contains the following types of servers:
• One or more user interface servers (online servers) that process web requests, perform business transactions, and
render web pages.
• One or more non-user interface servers (offline servers) that manage batch processing, work queues, scheduling,
message destinations, and startable services (plugins).
• A load balancer that manages user interface requests if there are multiple user interface servers.
Technically, non-user interface servers can also process web requests from business users, meaning that they can
also act as user interface servers. However, Guidewire recommends that you avoid redirecting business user web
requests to non-user interface servers to simplify the management of the cluster.
Cluster Membership
As a PolicyCenter server joins the cluster, it updates a membership table in the PolicyCenter database. All cluster
servers periodically poll this table to determine cluster membership.
Cluster Availability
To ensure a high degree of availability, Guidewire recommends that the cluster configuration include two or more
servers with each specific server role. You also need to provide ample capacity for running role-constrained items
such as message destinations or batch processes.
Cluster Monitoring
Guidewire provides cluster monitoring screens that are available to those with privileges to view the Server Tools
screens:
• The Server Tools Cluster Members screens provide information on each server in the cluster.
• The Server Tools Cluster Components screen provides information on the components running on a given server.
Also, there are system_tools command options that provide information on cluster members and components in
the PolicyCenter cluster.
Cluster Cache Usage

Cluster inter-communication ensures that if an object changes in one cluster member cache, that member sends a
cache invalidation message to the other members of the cluster. This message instructs the other cluster members to
tag the cache entry for the object as obsolete and evict it from the cache. The next time a PolicyCenter server needs
the object, it reloads the object value directly from the database. This mechanism is different from full cache
synchronization, in which a server broadcasts the new value of the object to other cluster members.
It is possible to lose message packets. Network failures or other issues also can disrupt communication between
cluster members. Such cases can result in cache eviction messages not being propagated to all cluster members. As a
result, one or more cluster members can contain stale cache entries.
Guidewire applications implement a data versioning mechanism to prevent data corruption. One or more version
mismatches indicates that one or more objects have changed since PolicyCenter last accessed the entities. This
mismatch results in PolicyCenter issuing a Concurrent Data Change Exception (CDCE). The user or batch job can
then re-issue a change based on the latest values entered.
See also
• “PolicyCenter Server Configuration” on page 41
• “Cache Management” on page 64
• “Concurrent Data Change Prevention” on page 65
132 chapter 7: Understanding PolicyCenter Server Clustering

• “Messaging and Startable Service Load Balancing” on page 162
• “System Tools Options” on page 393
Cluster Communication
In the base PolicyCenter configuration, PolicyCenter clusters use the following types of transport mechanisms for
sending messages between cluster members:
• Reliable broadcast without replies
• Unreliable fast broadcast without replies
• Reliable unicast with reply
Guidewire provides a default plugin implementation to support each of these transport types. However, it is possible
to implement your own unicast/multicast transport by implementing the corresponding plugin. Guidewire disables
fast broadcast messaging in the base configuration.
Unicast Communications
PolicyCenter clusters use PPP protocol over TCP for direct server-to-server communication. For example, it is
possible for a PolicyCenter Server Tools screen function to create a message request that directly targets a specific
server. In this case, server A, on which the message request originate, sends a unicast message to server B, who
receives and processes the request. PolicyCenter server lease management also leverages unicast communication to
speed up certain actions, such as lease transfers.
Multicast Communications
PolicyCenter clusters leverage the database for distributing broadcast messages.
Cluster Plugin Implementations

In the base configuration, Guidewire provides the following plugin interfaces to support cluster communication:
ClusterBroadcastTransportFactory Provides a single factory method for creating a cluster transport for reliable
broadcast of messages, with no replies. PolicyCenter stores broadcast messages
in the database and then periodically loads any new broadcast messages onto
each node in the cluster. This type of cluster transport guarantees the delivery
order and the reliable delivery of the broadcast message.
PolicyCenter uses this mechanism for default message broadcast if you do not
enable the ClusterFastBroadcastTransportFactory plugin implementation.
ClusterFastBroadcastTransportFactory Provides a single factory method for creating a cluster transport for fast broad‐
cast of messages, with no replies. This type of transport:
• Uses UDP multicast protocol
• Does not guarantee the delivery order or even the actual delivery of the
broadcast message
PolicyCenter typically uses this type of cluster transport for broadcasting cache
eviction notifications to cluster members.

The use of this transport type is optional. In the base configuration, Guidewire
disables the ClusterFastBroadcastTransportFactory plugin implementation
due to its use of the UDP protocol. If you do not enable the plugin implementa‐
tion, PolicyCenter uses the ClusterBroadcastTransportFactory cluster trans‐
port for broadcast messages instead.
Note: This type of cluster communication uses the UDP multicast protocol.
Most (but, not all) cloud environments disable multicast communication with
no option to reenable it. Because of this fact, Guidewire does not recommend
the use of a ClusterFastBroadcastTransportFactory implementation in a
cloud environment. Only use this plugin type in an environment that supports
multicast communication.
ClusterUnicastTransportFactory Provides a single factory method for creating a cluster transport for point‐to‐
point unicast messages between specific servers in the cluster. The default plugin
implementation uses TCP for the transport protocol.
Guidewire provides internal Java classes for these cluster-related plugin implementations. It is not possible to
modify these Java classes. To see the plugin definitions, open PolicyCenter Studio and navigate to the following
location in the Studio Project window:
configuration→config→Plugins→registry
Configuring Cluster Communication

The cluster-related plugin implementations that Guidewire provides in the base PolicyCenter configuration are
sufficient for most purposes. However, if you need more fine grained control over cluster communications, it is
possible to use one of the following methods to provide that control:
• Plugin parameters
• System property overrides
For example, if you need precise control over binding ports, then use one of these methods to configure the ports
directly.
Plugin Parameters
The cluster plugin implementations that Guidewire provides in the base configuration all support plugin parameters
that you can use to reconfigure the plugin. All plugin parameters are optional. Guidewire provides default values for
each of the plugin parameters. See “Cluster Plugin Parameter Reference” on page 134 for more information.
To define a plugin parameter, you manually add that parameter to the plugin definition in the PolicyCenter plugin
editor. For example, suppose that you want to directly control the number of threads in the thread pool that handle
inbound requests in the ClusterUnicastTransportFactory plugin. In this case, you manually add the poolSize
parameter and value to the plugin definition for ClusterUnicastTransportFactory, using the Studio plugin editor.
System Property Overrides

PolicyCenter provides the ability to reconfigure your plugin configuration using system properties set from a
command prompt as you start the application server. One advantage to using system property overrides to set
transport values is that you do not have to modify the configuration inside a WAR/EAR file to do so. This makes it
easier to use the same WAR/EAR file in different environments.
The exact syntax to use in setting system parameters at application server start is dependent on the application server
type. See “Setting JVM Options in PolicyCenter” on page 48 for more information. See “Cluster Plugin System
Properties Reference” on page 137 for a list of the system parameters that you can use with the clustering plugins.
Cluster Plugin Parameter Reference

In the base configuration, PolicyCenter provides the following plugin implementations that manage message
transport in a PolicyCenter cluster. Each of these plugin implementations provide a number of configuration
parameters that you can set to precisely control such transport variables as server address and bind port number.
Plugin implementation For more information

ClusterBroadcastTransportFactory “Configuration parameters for ClusterBroadcastTransportFactory” on page 135
ClusterFastBroadcastTransportFactory “Configuration Parameters for ClusterFastBroadcastTransportFactory” on page
136
ClusterUnicastTransportFactory “Configuration Parameters for ClusterUnicastTransportFactory” on page 137
Defining a plugin parameter

To define a plugin parameter, you manually add that parameter to the plugin definition in the PolicyCenter plugin
editor. For example, suppose that you want to directly control the number of threads in the thread pool that handle
inbound requests in the ClusterUnicastTransportFactory plugin. In this case, you manually add the poolSize
parameter and value to the plugin definition for ClusterUnicastTransportFactory, using the Studio plugin editor.
Configuration parameters for ClusterBroadcastTransportFactory

Use the configuration parameters in the following table to provide precise control over the plugin parameter values
available in the implementation of ClusterBroadcastTransportFactory.
batchesDeleteInterval Average time (in milliseconds) between the execution of a SQL statement that deletes old
message batches from the database. Each server node in the cluster executes this SQL state‐
ment. Therefore, if the cluster installation contains many nodes, Guidewire recommends that
you increase this value.
The default is 60000 milliseconds (1 minute).
batchKeepPeriod Maximum amount of time for PolicyCenter to retain a batch in the database before deleting it.
The default is 600000 (10 minutes).
batchReadInterval Maximum time interval (in milliseconds) between reading and receiving new batches. The de‐
fault is 3000 milliseconds (3 seconds).
batchWriteAttempts Maximum number of attempts to write to a batch queue. If the number of consecutive errors
exceeds this threshold, the transport switches to ERROR mode in which each new messages
pops the oldest message out of the in‐memory queue. The purpose of this parameter value is
to avoid out‐of‐memory issues. The default is 30.
batchWriteInterval Maximum time interval to wait (in milliseconds) before PolicyCenter writes, or sends, the cur‐
rent batch of messages. The default is 2000 milliseconds (2 seconds).
maxOutboundBufferSize Maximum size of outbound buffer (in megabytes). The purpose of this parameter value is to
prevent out‐of‐memory issues if a transport is having problems writing or sending messages.
The default is 25 megabytes.
preferredBatchDataSize Maximum size of the batch (in bytes). If the size of the current batch (the sum of all of the
message batch sizes) reaches this threshold, PolicyCenter writes, or sends, the current batch of
messages immediately.
This value must be less than or equal to the largest possible integer value supported by your
hardware.
preferrredBatchMessageCnt Maximum number of pending messages allowed in the batch queue. If the number of messag‐
es in the batch queue reaches this threshold, PolicyCenter writes, or sends, the current batch
of messages immediately.
The value must be less than or equal to the largest possible integer value supported by your
hardware.
receiverPoolSize Number of threads in the thread pool that handle inbound messages. The default is 4.
To set a plugin parameter, you must manually add that parameter to the plugin definition in the PolicyCenter plugin
editor. To access the plugin editor, navigate to the following location in PolicyCenter Studio and double-click the
plugin name:
See also
Configuration Parameters for ClusterFastBroadcastTransportFactory

Note: Guidewire disables the implementation of the plugin in the base PolicyCenter configuration.
You must enable the plugin implementation before PolicyCenter recognizes any of these parameters.
Use the configuration parameters in the following table to provide precise control over the plugin parameters in
ClusterFastBroadcastTransportFactory.
bindAddr Inet address to which PolicyCenter is to bind. This parameter can be useful if there are
multiple‐NIC hosts. The default fallback for this parameter is the first non‐loopback interface
found on the host as defined by the NetworkInterface Java API.
bindPort Port number to which PolicyCenter is to bind. This parameter can be useful for server hosts
behind a firewall.
maxMessageSize Maximum allowable size of message. PolicyCenter calculates the default value of this parame‐
ter using the following algorithm:
(Maximum IP datagram size) ‐ (UDP header size) ‐ (IP header size)
The maximum IP datagram size is 65,535. The UDP header size is 8. The IP header size is one of
the following values:
• IPv4 = 20
• IPv6 = 40
Thus, if using IPv6, the default value for this parameter is 65,535 ‐ 8 ‐ 40, which is 65,487.
messageKeepPeriod Time (in milliseconds) to keep messages in memory in order to skip retransmitted messages
and to combine divided messages. The default is one of the following:
• 2 * (maximum retransmit interval)
• 10 seconds, if not using retransmit
messageSalt Integer value that PolicyCenter uses in calculating the sending message checksum. This value
must be the same on all servers in the PolicyCenter cluster. The default is 12345.
multicastAddress Multicast Inet address. The default is 228.8.8.8.
multicastPort Multicast port. The default is 38180.
nodeStatisticsKeepPeriod Time (in milliseconds) to keep node statistics in memory after last activity. The default is
3,600,000 (1 hour).
oldMessagesDeleteInterval Average time between the removal old messages from the memory (in milliseconds). The de‐
fault is 1,000.
receiverPoolSize Number of threads in the thread pool that handle inbound messages. The default is 4.
receiverQueueCapacity Thread pool queue capacity. The default is 100.
retransmitIntervals Comma‐separated list of retransmit intervals (in milliseconds). The default is 10000.
sendHeartbeatInterval Time (in milliseconds) between sending heartbeat messages. The default is 30,000 (30 sec‐
onds).
ttl Time‐to‐live (TTL) for multicast datagram packets. The default is 8.
plugin name:
See also
Configuration Parameters for ClusterUnicastTransportFactory

Use the configuration parameters in the following table to provide precise control over the plugin parameters in
ClusterUnicastTransportFactory.
bindAddr Inet address to which PolicyCenter is to bind. This parameter can be useful if there are multiple‐NIC
hosts. The default fallback for this parameter is the first non‐loopback interface found on the host as
defined by the NetworkInterface Java API.
bindPort Port number to which PolicyCenter is to bind. This parameter can be useful for server hosts behind a
firewall. The default fallback for this parameter is an ephemeral port, a free port above 1024 that is
within a range supplied by the host operating system.
poolQueueCapacity Thread pool queue capacity. The default is 50.
poolSize Number of threads in the thread pool that handle inbound requests. The default is 4.
plugin name:
See also
Cluster Plugin System Properties Reference

You can use system properties, set at server startup from a command prompt, to modify certain parameters that
affect message transport in the PolicyCenter cluster. Using these system properties, you can precisely control such
transport variables as server address and bind port number to override the plugin parameter settings in the following
plugin implementations:
• ClusterFastBroadcastTransportFactory
• ClusterUnicastTransportFactory
The following list describes the plugin parameters that you can set through system properties at application start.
Plugin type Plugin parameter System property

ClusterFastBroadcastTransportFactory • bindAddr • gw.cluster.fudp.bind_addr
ClusterUnicastTransportFactory.bindAddr • bindAddr • gw.cluster.nbtcp.bind_addr
• bindPort • gw.cluster.nbtcp.bind_port
The exact syntax to use in setting system parameters at application server start is dependent on the application server
type. See “Setting JVM Options in PolicyCenter” on page 48 for more information.
Cache eviction messages

Guidewire does not guarantee the delivery of cache eviction messages in a PolicyCenter cluster. However, with that
said, the following is true:
Message plugin type Notes

ClusterBroadcastTransportFactory It is unlikely for there to be a missed eviction message with this type of message
transport as the message plugin reads the messages in order from the database.
It is possible in extreme circumstances such as losing database connectivity for a
time period longer than the configured cleaning interval for the broadcast table.
However, it is not likely.
ClusterFastBroadcastTransportFactory If enabled, PolicyCenter typically uses this type of cluster transport for
broadcasting cache eviction notices to cluster members. As this type of message
transport uses UDP packets, it is possible to drop a UDP packet during
transmission. Is is also possible for PolicyCenter to miss a packet if a garbage
collection operation takes longer than the wait time for retransmitting a packet.
If dropped packets become an issue, try setting the retransmitIntervals
parameter on the plugin to a smaller value than its 10 second default value.
Logging cluster plugin parameters

By default, PolicyCenter logs cluster-related information at the INFO level in the server log. The cluster information
that PolicyCenter provides includes information on where the bind parameter values come, for example, from the
bind property default or from a parameter definition in the plugin editor.
Plugin Logging header

ClusterFastBroadcastFactory Server.Cluster.FastUdpMulticast
ClusterUnicastTransportFactory Server.Cluster.PointToPoint
Ephemeral port values

Unless you provide a value for the bindPort value on the ClusterUnicastTransportFactory plugin, PolicyCenter
randomly selects a free port above 1024 that is within a range supplied by the host operating system. You can view
information on an ephemeral port in the server log. Look for information that looks similar to the following:
Server.Cluster.PointToPoint Listener(host, port=#####): Listening
Cluster logging examples

In the first logging example, the cluster installation enables both the ClusterFastBroadcastFactory and
ClusterUnicastTransportFactory plugins. The installation does not, however, provide separate values for the
bind parameters for these plugins. Notice, therefore, that the following logging example provides information for
both plugins and indicates that bind values are the default values (first address of the first network
interface).
19:50:30,131 INFO Server.Cluster Starting cluster channel...

19:51:45,885 INFO Server.Cluster.FastUdpMulticast Bind address is /nn.nn.n.nnn
(first address of the first network interface)
19:51:45,915 INFO Server.Cluster.PointToPoint com.guidewire.pl.cluster.internal.ptp.
PointToPointUnicastTransport@xxxxxxxx: Starting...
19:51:48,047 INFO Server.Cluster.PointToPoint Bind address is /nn.nn.n.nnn
(first address of the first network interface)
19:51:48,887 INFO Server.Cluster.PointToPoint Listener(serverid=prod1,port=53872): Listening...
19:51:48,954 INFO Server.Cluster.FastUdpMulticast Started on /nn.nn.n.nnn
19:51:48,959 INFO Server.Cluster Members joined the cluster: prod1 (xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx)
19:51:49,290 INFO Server.Cluster Inserted new record for prod1 node into cluster table
19:51:49,303 INFO Server.Cluster Cluster channel started.
In the second logging example, the cluster installation again enables both the ClusterFastBroadcastFactory and
ClusterUnicastTransportFactory plugins. In this case, however, the installation provide a value for the
ClusterUnicastTransportFactory.bindPort parameter, which is 53870, defined in the plugin editor for this
plugin.
11:51:23,059 INFO Starting cluster channel...

11:51:23,250 INFO Bind address is /nn.n.n.n (first address of the first network interface)
11:51:23,273 INFO com.guidewire.pl.cluster.internal.ptp.PointToPointUnicastTransport@xxxxx: Starting...
11:51:23,375 INFO Bind address is /nn.n.n.n (first address of the first network interface)
11:51:23,375 INFO Bind address is 53870 (specified via plugin configuration)
11:51:23,377 INFO Listener(serverid=prod1,port=53870): Listening...
11:51:23,465 INFO Started on /nn.n.n.n
11:51:23,470 INFO Members joined the cluster: prod1 (xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx)
11:51:23,869 INFO Inserted new record for prod1 node into cluster table
11:51:23,877 INFO Cluster channel started.
Server Roles
In general, Guidewire application cluster contains servers (cluster members) that manage the following types of
functionality.
Function Description
Online processing Server interactively manages requests from users logged into Guidewire PolicyCenter.
Background processing Server manages batch process execution, work queue processing, message destination processing,
lease management, and other similar items.
Guidewire categorizes each individual server instance in the cluster by its function, as defined by its role. In the base
configuration, PolicyCenter defines server roles to handle the following functionality. In a typical installation, only
those servers that support external requests such as user input use the ui server role.
Server func‐ Server role Servers with these roles manage...

tion
Online proc‐ “ui Server Role” on page 142 • Interactions (user requests) between the PolicyCenter user interface
essing and PolicyCenter itself.
• Web service calls; there is no distinct server role for web services.
However, any cluster member that receives a web service request can
process that request by default.
• Work queue processing, if there are distributed workers. Guidewire
recommends that a server with the ui role handle these types of
requests only during periods of low‐load and during off‐peak hours.
Background “batch Server Role” on page 140 • Batch scheduling.
processing • Batch process execution.
“Messaging Server Role” on page Message destination processing.
140
“scheduler Server Role” on page Schedule handling.
141
“startable Server Role” on page Startable service management.
142
“workqueue Server Role” on page Work queue management, if there are distributed workers.
142
It is possible for multiple servers in the PolicyCenter clusters to have the same server role. Servers that have the
same role type typically have similar resource allocations and configuration. Conversely, servers with different
server role types typically have different workloads and allocate their resources differently.
Server Roles and Lease Managers

Guidewire associates a specific lease manager type with each server role. Thus, a message destination lease manager
on a server with the messaging role manages message destinations for that server.
Each Guidewire server contains all types of lease managers. However, a lease manager only becomes active on a
server with a server role that matches its lease type.
Server Role Validation

PolicyCenter performs the following server role validation at server start up:
• Validation of roles used in the application configuration files – If PolicyCenter detects the use of an unknown
role in any of the configuration files, it throws an exception and refuses to start.
• Validation of roles assigned a specific server – If PolicyCenter detects the assignment of an unknown role to a
server, it prints a warning to the server log and ignores the unknown role.
See also
• “Component Lease Management” on page 157
batch Server Role

Guidewire PolicyCenter distributes batch processing across all server instances in the cluster that have the batch
server role. At least one server in the PolicyCenter cluster must have the batch server role. It is possible to request
the start of a batch process from any server in the cluster. However, only a batch server is capable of performing the
work involved in the batch process. If the start request originates from a server that does not have the batch role,
that server communicates the request to the other members of cluster. A server with the batch role accepts the
request and performs the work.
Exclusive and non‐exclusive batch processing

Guidewire categorizes batch processes into exclusive and non-exclusive batch processes:
• For exclusive batch processes, Guidewire guarantees that the batch process runs on exactly one cluster member
with the batch role at a time.
• For non-exclusive batch processes, Guidewire guarantees that a single submission of a batch process runs exactly
once. However, it is possible to submit non-exclusive batch processes for execution multiple times. In this case, it
is possible for the batch process to run once for each submission, possibly concurrently, on multiple servers that
have the correct server role.
Cluster members with the batch role use a batch-specific lease manager. Guidewire provides a configurable load
balancing strategy for those servers with the batch role.
See also
Messaging Server Role

Message destinations and messaging server roles
In Guidewire PolicyCenter, it is possible to associate a specific server role with each message destination. In the
base configuration, PolicyCenter does not specify a server role for any of the predefined message destinations. For
all message destinations without a defined server role, PolicyCenter automatically sets the associated server role to
the base configuration role of messaging.
The following table describes the relationship between server roles and message destinations.
Server Messaging destination

Assigned messaging role Can acquire the lease of any message destination
Assigned role or host Can only acquire the lease to a message destination with an assigned role that matches that of
name the server. For example, suppose that there is a message destination with an assigned role of A.
Only a server that has role A can acquire the destination lease. A server with only role B cannot
acquire the lease.
It is possible to associate a specific host name with a specific message destination as well. Then,
only that server host can acquire the lease to the message destination.
No assigned messaging Cannot acquire the lease of any message destination.
role
You associate a specific server role or host name with a message destination in the Guidewire Studio™ message-
config.xml file. If you do not set a server role for a message destination, the Messaging editor shows a default role
of messaging at the top of the screen.
You set a specific role on a PolicyCenter server in one of the following ways:
• Through an option on the gwb runServer command used to start the server from a command prompt.
• Through the registry metadata definitions in file config.xml.
Message destinations and leases

Guidewire PolicyCenter distributes message processing across all server instances that have the messaging server
role (which is any role assigned to a message destination). Cluster members with these roles use a messaging-
specific lease manager. In this case, a lease corresponds with a message destination, with each destination having a
server declaration.
PolicyCenter creates message destination leases during messaging initialization. As each PolicyCenter starts, if it
has a messaging role, it looks for a message destination to start. This message destination is any available message
destination whose assigned role matches a role assigned to the server. PolicyCenter repeats this process until there
are no more qualified destinations to assign.
Each messaging lease manager accepts a listener from the messaging system. Upon the activation of a lease
operation by the lease manager such as deleting or expiring a messaging lease (and similar operations), the listener
notifies the messaging system. PolicyCenter then stops messaging destinations that are deleted or acquired by other
servers.
A shutdown command for a messaging destination initiates a request to expire the lease on the server for that
destination. The destination listener accepts that request and passes the request to the lease manager for that
destination.
PolicyCenter periodically reviews all destination leases to determine if the leases are still valid and to look for new
leases to acquire. If a lease expires or the lease manager creates a new lease, PolicyCenter again searches for new
messaging destinations to assign.
Guidewire provides a configurable load balancing strategy for those servers with the messaging role.
See also
• “Messaging Tools Command” on page 391
scheduler Server Role

Guidewire PolicyCenter typically utilizes only a small number of cluster members with the scheduler server role.
These server instances run multiple synchronized instances of the scheduler in parallel.
IMPORTANT Each server with the scheduler role must also have configuration parameter
SchedulerEnabled set to true in config.xml.
Servers with the scheduler role
Guidewire recommends that even small clusters have at least two servers with the scheduler role. This mitigates
the possibility of a single server with the scheduler role becoming a single point of failure. However, Guidewire
does not recommend more than four servers with this role in a cluster as large number of servers competing for
database resources can possibly increase database contention.
See also

• “Startable Services” on page 356
startable Server Role

Guidewire PolicyCenter implements certain plugins (services) as cluster singletons. These plugins implement the
IStartablePlugin interface and do not carry the distributed annotation. PolicyCenter runs a single instance of
these plugins only, on a server with the startable server role. Cluster member with this role use a plugin-specific
lease manager.
Guidewire provides a configurable load balancing strategy for those servers with the startable role.
Note: Guidewire defines an additional type of cluster singleton plugins, known as inbound
integrations, in file inbound-integration-config.xml.
See also

• “Startable Services” on page 356
workqueue Server Role

Guidewire PolicyCenter distributes work queues across all server instances that have the workqueue server role. By
default, each work queue starts a single worker on each server with the appropriate role, unless configured
otherwise.
See also
• “Work Queues” on page 82

• “Work Queues and Server Roles” on page 95
ui Server Role
Guidewire PolicyCenter uses the ui server role as a placeholder role only. Guidewire PolicyCenter servers typically
operate in conjunction with a non-Guidewire load balancer that manages the user interface transactions.
PolicyCenter distributes web requests to the various cluster members according to the rules specified for the load
balancer. Any cluster server that receives a web request processes that request, regardless of role assignment.
See also
Example PolicyCenter Cluster Configuration

You assign a role to an individual server using the <registry> element in file config.xml. See “Understanding the
Configuration Registry Element” on page 42 for details.
The following sample code illustrates a PolicyCenter cluster definition with five cluster members.
<registry roles="batch, scheduler, workqueue, messaging, startable, ui>

<server env="prod" serverid="pcnode1"roles="batch, scheduler, messaging, startable"/>
<server env="prod" serverid="pcnode2"roles="batch, scheduler, messaging, startable"/>
<server env="prod" serverid="pcnode3"roles="workqueue, ui" />
<server env="prod" serverid="pcnode4"roles="workqueue, ui" />
<server env="prod" serverid="pcnode5"roles="ui" />
</registry>
Cluster Member Startup

As each member of the cluster starts, a number of events occur:
• During startup, each PolicyCenter server runs a checksum on its own configuration. The server then compares its
checksum with the configuration checksum stored in the database. If the startup server determines that the two
configurations are not compatible, it does not start.
• As each PolicyCenter server joins the cluster, it updates a membership table in the PolicyCenter database. All
cluster members periodically poll this table to determine cluster membership.
• During a rolling upgrade of the individual server members in the cluster, the startup server verifies that it is
running either the source (original) or target (new) configuration. If the server detects any other configuration, it
does not start.
See also
• “Restarting the PolicyCenter Cluster Servers” on page 149
• “Configuration Compatibility” on page 167
Cluster Member Shutdown

In general, a PolicyCenter cluster contains the following types of servers:
• One or more user interface servers that process web requests, perform business transactions, and render web
pages.
• One or more non-user interface servers that manage batch processing, work queues, scheduling, message
destinations, and startable services (plugins).
The shutdown procedure for a Guidewire PolicyCenter cluster member is dependent on the role of server involved.
See also
• “ui-Role Cluster Member Shutdown” on page 144
• “Non-ui Role Cluster Member Shutdown” on page 144
ui‐Role Cluster Member Shutdown

Guidewire PolicyCenter provides the means to schedule a planned PolicyCenter server shutdown. You schedule a
server shutdown through the Server Tools Cluster Components screen.
After you schedule a planned shutdown on a specific server:
• PolicyCenter shows an on-screen banner for that server instance that warns of the impending shutdown.
• The banner shows a count-down timer to the planned shutdown and recommends that any logged-in user log out
of PolicyCenter.
See “Schedule a Planned Cluster Member Shutdown” on page 362 for more information.
Non‐ui Role Cluster Member Shutdown

As with PolicyCenter user interface servers, you schedule a shutdown of a server cluster member with a non-user
interface role through the Server Tools Cluster Components screen. However, there are additional requirements for
shutting down a non-user interface server.
Stop any running background tasks in the following order:
Background Shutdown considerations

task
1. Scheduler The scheduler typically runs on several servers, as a distributed component. Thus, stopping the scheduler on
one server does not affect the cluster.
2. Batch proc‐ It is necessary for all running batch processes to complete before starting the server shutdown. In many cas‐
essing es, it is better to let a batch process complete, without attempting to terminate the process. However, as
some batch processes can take significant amount of time to complete, potentially hours, you often need to
make an individual determination for each batch process.
See “About Planned Server Shutdowns” on page 357 for more information.
3. Work queue It is necessary to stop all work queue workers on the server that you intend to shut down. As PolicyCenter
distributes work queues across cluster members, stopping workers on one server does not typically affect
the cluster. However, it is possible for the overall performance of work items processing to drop.
Stop work queues after stopping batch processing to stop generating business transactions that can poten‐
tially generated messages.
4. Messaging It is necessary to stop messages destinations on the server that you intend to shut down. Other servers in
destinations the cluster must take over the stopped message destination.
5. Startable It is necessary to stop startable plugins (services) on the server that you intend to shut down. Other servers
plugins in the cluster must take over the stopped services.
Most background tasks, except batch processes, stop quickly as their units of work are small. The actual task
managers, for example the Batch Process Manager of the Message Destination Manager, do not instantly stop in a
server shutdown. Instead, each lease manager moves to a passive mode in which it does not start new background
tasks and moves to stop or complete any currently running tasks.
After all components stop their background tasks, you can shut down the batch server safely.
Server Shutdown Monitoring

As it may take some time to stop all background tasks, you can use the following API method to track the
background task shutdown progress:
ISystemTools.getClusterState()
You can also use the following system_tools command options to gather information about a server and the state
of the server components:
system_tools -components
system_tools -nodes

See “System Tools Options” on page 393 for a discussion of these command options.


chapter 8
Working with a PolicyCenter Cluster
This topic discusses ways to implement, manage, and monitor a Guidewire PolicyCenter cluster.
Planning a PolicyCenter Cluster

Plan the cluster so that if any one server fails, the other servers in the cluster can handle its traffic without being
overwhelmed. PolicyCenter servers in the cluster can run on separate computers, or you can run multiple servers on
the same computer. Guidewire recommends you maintain at least three PolicyCenter servers in the cluster, whether
on the same or different physical computers. With multiple servers running on the same computer, in the event of a
failure, then all servers are unusable. Of course, the exact configuration depends on specific usage needs.
To establish a cluster, you must also install your own load balancing solution. The load balancer acts as the bridge
between client connections and the private network. Clients send a connection request to the load balancer and it
routes the request to a PolicyCenter server. The load balancer must implement session affinity, meaning that it must
route connections from the same user session to the same PolicyCenter server. If the load balancer directed a user to
a different server, the session is reset. This can result in loss of unsaved data.
See also
Create a Guidewire PolicyCenter Cluster

To create a PolicyCenter cluster, first define the individual servers in configuration, then deploy your configuration
to each server in the cluster.
Procedure
1. In your source configuration, for use on all PolicyCenter servers in the cluster, open config.xml for editing.
2. In config.xml, set the following clustering-related configuration parameters appropriately:
ClusteringEnabled
ClusterMemberPurgeDaysOld
ConfigVerificationEnabled
PDFMergeHandlerLicenseKey
3. In config.xml, using the <registry> element, define the following:
a. The set of valid server roles for use in this cluster.
Working with a PolicyCenter Cluster 147
b. The server roles for each individual PolicyCenter server on the cluster.
Note: Guidewire does not require that you use the server <registry> element in config.xml to
define the individual server instances in the cluster. You can also set these values through JVM
parameters at server start up.
4. In config.xml, set the value for KeyGeneratorRangeSize.
5. Create a deployment WAR or EAR file for Guidewire PolicyCenter.
6. Install a PolicyCenter cluster server in the same way that you install a standalone PolicyCenter server.
IMPORTANT If you install multiple PolicyCenter servers on the same host machine, each
PolicyCenter server must run in its own JVM instance.
7. Start each member of the cluster in turn.

8. In PolicyCenter, navigate to the Server Tools Cluster Members screen and verify the information.
See also

Managing a PolicyCenter Cluster

This topic discusses the ongoing management of a clustered environment.
Enabling Guidewire Clustering

To enable clustering in a Guidewire installation, set the ClusteringEnabled parameter in config.xml to true on
all cluster members, for example:
<param name="ClusteringEnabled" value="true"/>
To disable clustering and remove a server from a cluster, set this parameter to false on that server. After the server
is no longer in a cluster, it behaves as any other standalone PolicyCenter server.
See also
Adding a Server to a PolicyCenter Cluster

PolicyCenter does not require that you use the cluster registry in file config.xml to define cluster members. Its use
is optional in that regard. Instead:
• You can specify the server ID and server roles from the command prompt at server startup using JVM command
options.
• Or, you can specify a subset of the cluster server members in the cluster registry and set other server properties
from the command prompt.
For example, you can define the set of servers assigned the batch server role in the cluster registry and specify the
remaining server IDs through JVM options at server start.
In adding a server to the PolicyCenter cluster, keep the following scenarios in mind.
148 chapter 8: Working with a PolicyCenter Cluster
Server configuration Notes

Not specified in config.xml There is no issue. You can add a server to the cluster without any change of configuration or
stopping and restarting servers.
Specified in config.xml It is possible to perform a rolling upgrade of each server instance in the cluster to upgrade the
cluster registry of each individual server instance in turn.
See also
• “Add a Server to a PolicyCenter Cluster” on page 149
Performing a Cluster Configuration Upgrade

Guidewire provides a way to upgrade individual servers in a PolicyCenter cluster individually. This type of upgrade,
called a rolling upgrade, is in contrast to a full database and application upgrade. Guidewire permits only a few
specific changes to files, file types, and installation folders during a rolling upgrade of the individual servers in a
PolicyCenter cluster.
See also
Add a Server to a PolicyCenter Cluster

Procedure
1. Deploy the current cluster configuration WAR or EAR file to the server that you intend to add to the
PolicyCenter cluster.
2. Verify that this configuration has parameter ClusteringEnabled set to true in the configuration’s
config.xml file.
3. If using the config.xml server registry, you must set the serverid attribute for this server to a unique value
within the cluster.
If you use a non-unique serverid value, the server with the duplicate ID does not start and does not join the
cluster. See “Understanding the PolicyCenter Server Environment” on page 41.
4. Start the new server.
Result
After you start the new server, it connects to the cluster and compares its configuration with the cluster configuration
stored in the PolicyCenter database. It performs a checksum of the config.xml file and checks the config
subdirectories. If the configurations differ, the server fails startup and PolicyCenter writes failure messages to the
log file.
Restarting the PolicyCenter Cluster Servers

There are several different scenarios that require you to bring down one or more members of the PolicyCenter
cluster and then restart the servers.
Server Restart after Server Maintenance

In performing maintenance on the servers in the PolicyCenter cluster, do the following for each PolicyCenter server
in the Guidewire cluster, as required:
• Set the server run level to maintenance.
• Perform the needed work.
• Restart the server.
See also
• “Place the Server in Maintenance Mode” on page 59
Server Restart after Rolling Upgrade

In an application configuration upgrade (known as rolling upgrade), you bring down a single server, in the cluster,
upgrade its application configuration, and bring that server back up. You then repeat this process with each member
of the PolicyCenter cluster in turn, until all cluster members use the upgraded configuration.
IMPORTANT Start the configuration upgrade on a single cluster server and let it fully initialize before
starting the upgrade process on the other cluster members.
Before starting a rolling upgrade, click Start Rolling Upgrade in the Server Tools Upgrade and Versions screen. You can
do this on any server in the PolicyCenter cluster. This action indicates that a rolling upgrade of the individual cluster
members is in progress.
After completing the upgrade of all cluster servers, click Rolling Upgrade Complete in the Server Tools Upgrade and
Versions screen. This action indicates that all servers in the cluster now use the upgrade WAR/EAR file and that the
rolling upgrade process is complete.
See also
• “Unexpected Upgrades” on page 175
Server Restart after Full PolicyCenter Upgrade

In a full upgrade, you bring down all members of the PolicyCenter cluster completely. Before starting a full upgrade,
click Start Full Upgrade in the Server Tools Upgrade and Versions screen. You can do this on any server in the
PolicyCenter cluster. This action sets a flag in the PolicyCenter database to indicate that a full upgrade is in progress.
As long as the upgrade-in-progress flag exists, it is not possible to start a PolicyCenter cluster member that uses the
old (non-upgrade) WAR/EAR file.
In a full database upgrade, you must start a single cluster server and allow it to complete the full upgrade cycle
before starting the upgrade on the remaining servers. After all the servers in the PolicyCenter cluster start with the
upgrade WAR/EAR file, PolicyCenter automatically deletes the upgrade-in-progress flag from the PolicyCenter
database.
Use the following guidelines as you bring up the individual cluster members:
• After the first cluster server completes the upgrade cycle, it is possible to bring up all other servers in the
PolicyCenter cluster in parallel. However, if starting a large numbers of servers causes resource contention, insert
a short interval of time between each server start, for example, 10 seconds.
• As a general rule, start servers that manage back-end processes first. For example, start servers with the batch
and messaging roles before starting servers with the ui role.
See also
• Upgrade Guide
Server Restart after Database Restore

It is sometimes the case that you want to restore the PolicyCenter cluster database from one of the following:
• A non-cluster server
• A server in a different cluster environment but with the same code base
Attempting to start the PolicyCenter server after the database restore can generates database errors sufficient to
cause the server to not start. To prevent these errors, it is necessary to start a single cluster node and let it cycle
through a full database upgrade before starting the other members of the PolicyCenter cluster.
To set the full upgrade flag on a server as you start the server, use the following JVM system property, with
YYYYMMDD being today's date:
-Dgw.pc.full.upgrade.intended.date=YYYYMMDD
For example:
gwb runServer -Dgw.pc.full.upgrade.intended.date=20180726
After the first server completes the database upgrade operation, start the other servers in the PolicyCenter cluster.
See also
Monitoring Cluster Health

You can monitor the health of a Guidewire application cluster in multiple ways.
Cluster Member Health

Typically, hardware or software load balancers check the health of the various cluster members and stop directing
traffic to a cluster member that stops responding. This check is very summary and simply verifies that the
corresponding network port responds. Therefore, it is possible that a load balancer redirects traffic to a cluster
member that is not capable of processing that traffic appropriately.
Some examples are that PolicyCenter is:
• Not fully started yet
• At the MAINTENANCE run level
• Experiencing significant issues, such as an out-of-memory condition
Some other infrastructure constraints exist. For example, some environments cannot use source IP stickiness to load-
balance conversational SOAP calls. In this situation, if using Guidewire ContactManager, you can instantiate one
ContactManager instance on each server machine hosting one or more PolicyCenter instances. You then configure
each PolicyCenter instance on that machine to direct contact requests to that local ContactManager instance. In this
scenario, if the local ContactManager instance is not functioning properly, it is advisable to stop directing traffic to
any of the PolicyCenter instances on that server.
Guidewire applications include a simple HTTP ping utility that enables you to check the application status with a
web browser. For instance, to check the status of an instance of PolicyCenter running on port 8080 of the local
computer, you would enter the following URL into a web browser:
http://localhost:8080/pc/ping
There are three possible responses by the web browser:

• If the application is running at the default MULTIUSER run level, the browser displays the number 2.
• If, however, the application is running in any mode other than MULTIUSER, the browser displays a specific ASCII
character, depending on the circumstances.
• If the PolicyCenter server is not running, the browser displays an HTTP failure message, depending on the
configuration of the server.
See the Integration Guide for a list of the specific ASCII characters that the ping utility can return. Invoking this
utility programmatically provides more granular information on the server’s status.
It is possible to configure the load balancers access this URL on a regular basis to determine the health of each
member of the cluster. You can then use these results to create redirection logic.
For example, suppose that you have an environment in which PolicyCenter directs traffic to a local ContactManager
instance. You then configure the load balancer to only redirect traffic to a PolicyCenter instance if both that instance
and the ContactManager instance on that server are accessible for user requests.
Cluster Member Monitoring in PolicyCenter

The Server Tools Cluster screens, accessible to system administrators, provide information on your PolicyCenter
cluster installation. For PolicyCenter to make this screen visible, the value of cluster configuration parameter
ClusteringEnabled in config.xml must be true.
The Cluster screens provide information on the server cluster installation that includes the following items.
Information about the local server Its server ID, the server roles assigned to this PolicyCenter server, and similar
member information
Information about individual members Their server IDs, number of active user sessions on each server instance, the
recognized by the cluster server roles assigned to each application instance, date and time of each server
start, and similar information
Information on the components running The component state, date and time of each component start, and similar
on each cluster member information
Information on any component lease The date and time of the deadline in which to complete the failover process
failover in progress
History of each member in the cluster The start and stop times for each cluster member, server roles, run level, and
similar information
From this screen, on any cluster member, you can start or cancel a planned shutdown for any recognized server
instance in the cluster.
See also
• See “Cluster Members and Components” on page 356 for more information on the Server Tools Cluster screens.
• See “Schedule a Planned Cluster Member Shutdown” on page 362 for information on starting or canceling a
planned cluster member shutdown.
Monitor Cluster Members Using System Tools

About this task
Guidewire provides several system_tools command options that provide information about individual members of
the PolicyCenter cluster.
-components Provides information about the components that exist on each PolicyCenter server in the cluster.
-nodes Provides information about each PolicyCenter server in the cluster.
Procedure
2. Open a command prompt and navigate to the PolicyCenter installation directory:
admin/bin
3. Enter the following commands:

system_tools -user user -password password -components

system_tools -user user -password password -nodes
See also
Using the PolicyCenter ping Utility

Guidewire provides an unauthenticated web page that you can ping to access information about a PolicyCenter
server. To access this web page, use the following URL:
http://server:port/pc/ping?v=2
The ping utility returns the following types of information, depending on various factors, including whether the
server is a production server and whether the server start was successful.
Information type Result description

Planned shutdown status Shutdown status (for example, planned, ready) if the server is involved in a planned shutdown.
Server run level code ASCII decimal code, for example, 45, which represents the ‘(‘ character.
Server run level name Server run level, for example, MULTIUSER.
Server run level ordinal QuickStart (Jetty) run level, for example, 5, which corresponds to the MULTIUSER run level.
Server ID Name of the server host.
Server up time Time, in seconds, for which the server has been operational.
Server start up exception Exception that caused the server to fail start up, if applicable. By default, the ping utility does not
display this information on a production server.
Thread stack trace Stack trace of the thread performing the transition from one server run level to another. By default,
the ping utility does not display this information on a production server.
Using the ping Utility with a Production Server

Due to security concerns, the PolicyCenter ping utility does not return certain information if used with a production
server. In production mode, the ping utility does the following:
• It replaces the text of any exception that occurs at server start up with <not null>.
• It removes entirely the stack trace of the thread that performs the work of transitioning from one server run level
to another.
To view this information for a production mode server, start the production server with the following JVM command
option:
-Dgw.ping.servlet.show.stack=true
ping Utility Examples

The following examples illustrate the kind of information that the server generates for the different server scenarios:
• Successful server start
• Failed server start
• Server transition between run levels
• Planned server shutdown
Successful server start

The following code is an example of the ping utility return values if the PolicyCenter server starts successfully.
{
"runLevelCode": 50,
"runLevelName": "MULTIUSER",
"runLevelOrdinal": 5,
"serverId": "PolicyCenterServer1",
"uptimeSeconds": 45
}
Failed server start

The following code is an example of the ping utility return values if the PolicyCenter development server fails to
start.
{
"runLevelCode": 40,
"runLevelName": "NODAEMONS",
"startupException": "java.lang.RuntimeException: Test Startup Exception\n\tat
com.guidewire.pl.system.server.PingServerServletTest.
testInitTabStateJSONObjectShowsStartupException(PingServerServletTest.java:52)\n
\tat... ",
"uptimeSeconds": 40
}
The following code is an example of the ping utility return values if the PolicyCenter production server fails to start.
By default, PolicyCenter does not show the actual exception text and instead replaces the text with <not null>.
{
"runLevelCode": 40,
"runLevelName": "NODAEMONS",
"serverId": "PolicyCenterServerPROD1",
"startupException": "<not null>",
"uptimeSeconds": 40
}
Server transition between run levels

The following code is an example of the ping utility return values on a development server while it is transitioning
from one run level to another. In this example, the server is transitioning from MULTIUSER run level to the DAEMONS
run level.
{
"attemptingTransition": {
"fromRunLevelName": "MULTIUSER",
"fromRunLevelOrdinal": 5,
"threadStackTrace": "Thread-142:TIMED_WAITING\n\tat sun.misc.Unsafe.park(Native Method)
\n\tat
java.util.concurrent.locks.LockSupport.parkNanos(LockSupport.java:215)\n\tat...”,
"toRunLevelName": "NODAEMONS",
"toRunLevelOrdinal": 3
},
"runLevelCode": 45,
"runLevelName": "DAEMONS",
"uptimeSeconds": 6814
}
Note: Use system_tools command options to transition a PolicyCenter server from one run level to
another.
Planned server shutdown

The following code is an example of the ping utility return values.
{
"runLevelCode": 50,
"serverId": "testsrv1",
}
The following code is an example of the ping utility return values after starting a planned server shutdown from the
(Server Tools) Info Pages→Cluster Members page.
{
"plannedShutdownStatus": "activated",
"runLevelCode": 50,
"serverId":
"testsrv1",
}
The following code is an example of the ping utility return values after the server shutdown completes.
{
"plannedShutdownStatus": "ready",
"runLevelCode": 50,
"serverId": "testsrv1",
}
See also
• “Using the ping Utility with a Production Server” on page 153


chapter 9
Working with Component Leases
This topic discusses server component lease managers, lease management, and component lease load balancing.
Component Lease Management

A lease represents the right to perform some job for some period of time. Within a PolicyCenter application cluster,
a lease specifically represents one of the following:
• A single run of a batch process
• An instance of a message destination
• An instance of a startable plugin, if it is a single instance plugin
Each server instance in the PolicyCenter cluster has a lease manager for each lease type. However, some
functionality require a server with a specific server role. For example, only a server with the messaging role can
acquire and manage a lease for a message destination.
Each type of lease manager can perform the following actions:
• Create, acquire, renew, or retire a lease
• Release a lease or request the release of a lease from another lease manager
• Transfer a lease to, or request the transfer of a lease from, another lease manager
Lease Management Contention

PolicyCenter lease management provides for minimal contention if two lease managers attempt to acquire the same
lease at approximately the same time. Guidewire guarantees that only one server can acquire a specific lease.
Guidewire also guarantees that, for an exclusive batch process, only a single such lease can exist at any one time.
Simple Lease Management Lifecycle for a Batch Process

The following is a general description of the lease management lifecycle for a batch process:
• PolicyCenter translates the request to execute a batch process into the creation of an available lease and
broadcasts the lease availability to all servers in the cluster.
• A lease manager on a server with the batch role discovers the available lease and acquires it. No other server in
the cluster can acquire the lease. The lease is exclusive.
• The server with the active lease executes the batch process until processing is complete, with the server
extending the lease as needed.
• At the completion of the batch process, the server lease manager releases the lease.
Working with Component Leases 157
Simple Lease Management Lifecycle for a Message Destination

The following is a general description of the lease management lifecycle for a message destination:
• At message initialization, PolicyCenter creates each enabled destination as an available lease.
• A lease manager on a server with the messaging role discover the available leases and attempts to acquire a
lease.
• As a server succeeds in acquiring a lease, it initializes the corresponding destination and places the destination
into the proper state (started, stopped, or suspended).
• A destination runs as intended, based on its state.
• A server periodically extends its destination lease until the messaging daemon stops, such as during server
shutdown or if you set the server run level to a level below DAEMONS.
• The server releases its destination lease.
Component Lease Failover

In the course of standard cluster operation, meaning no network issues, no lost cluster members or similar problems,
PolicyCenter periodically and automatically renews all component leases. The expiration of a lease is an exceptional
condition that requires attention. If a component lease expires, PolicyCenter considers the owner of that lease to
have failed. To detect this situation, the cluster lease managers periodically search for expired leases and initiate
failover of the expired lease to another member of the cluster.
After a lease manager detects an expired lease, it does the following:
• It terminates the lease.
• It updates the lease history.
• It deletes the lease or recreates the lease for acquisition by another cluster member, depending on the type of the
lease.
Automatic and Manual Failover
Guidewire PolicyCenter supports both automatic and manual failover of a component lease from one cluster
member to another member of the cluster. However, there are situations in which an automatic component failover is
not desirable, for example:
• There is a need for further additional configuration of an external third-party product before it is possible to start
a destination or plugin on a different computer.
• There is a need for further diagnostic testing to determine the exact cause of failure before initiating the failover
process.
See also
• See “Automatic Failover of a Component Lease” on page 158 for a description of the automatic failover process
that occurs after a lease expires.
• See “The Background Task Failover Plugin” on page 160 for a description of the default plugin implementation
that Guidewire provides for handling component lease failover.
Automatic Failover of a Component Lease

The following state diagram illustrates the different states in the automatic failover of a component lease.
PolicyCenter defines these states in the FailoverState typelist. The FailoverState typelist does not contain the
Completed state. The Completed state is implicit after the failover process completes and the lease manager deletes
the original lease.
158 chapter 9: Working with Component Leases
Postponed
Not Started In Progress Completed
Failed
Initially, each component lease starts in the Not Started failover state. If a lease expires, the first lease manager that
discovers the expired lease does the following:
• It sets the lease to the In Progress failover state. After set to this state, the component associated with the lease
cannot run anywhere until there is a resolution of the issue that caused the lease to expire.
• It sets the Retry Failover field in the Server Tools Cluster Components screen to the following value.
CurrentTime + BackgroundTaskFailoverPlugin.FailoverTimeout
If more than one lease manager discovers the expired lease at the same time, only the first lease manager continues
the failover handling. The other lease managers detect that their SQL updates do not change anything and do not
continue the failover process for that lease.
The lease manager that started the failover calls the handleComponentNameFailover method on the
BackgroundTaskFailoverPlugin plugin to determine what to do next with the lease. The method returns one of the
following actions to handle the component lease failover.
Possible Description
actions
Com‐ The BackgroundTaskFailoverPlugin plugin logic confirms the lease failure and instructs the lease manager to
plete the complete the failover. In this case, the lease manager completes the failover process, either by deleting or expiring
failover the lease.
Postpone It is possible that the BackgroundTaskFailoverPlugin plugin logic cannot reliably confirm the lease failure. In this
the fail‐ case, it can postpone the failover process by returning an associated action to take and the time duration to wait
over before taking that action. The lease manager updates the Retry Failover field in the Server Tools Cluster Components
screen with the following value:
Current Time + FailoverHandlingResult.Duration
After the updated retry failover time expires, the lease manager considers the lease expired and starts the process
of lease failover again.
Dismiss It is possible that the BackgroundTaskFailoverPlugin plugin logic decides the specified background task did not
the fail‐ fail, or, that this particular task requires some manual action. In this case, the BackgroundTaskFailoverPlugin
over plugin logic dismisses or fails the automatic failover of the lease. The lease with its FailoverState set to Failed
remains in the database until there is some kind of manual intervention. The failover process does not attempt to
retry the automatic failover.
Use ex‐ The BackgroundTaskFailoverPlugin plugin logic returns a failover handled acton. This action instructs the lease
ternal manager to do nothing with the lease. An external tool either deletes or renews the lease.
tool Calling an external tool to complete the failover can happen in any of the following ways:
• Programmatically calling the SystemToolsAPI.nodeFailed method.
• Programmatically calling the SystemToolsAPI.completeFailedFailover method.
• Clicking the Complete Failover button on the Server Tools Cluster Components screen.
If the cluster member that started the failover does not complete the failover in the specified retry failover time,
another cluster member detect this condition. The second cluster member then restarts the failover.

If at any point the original lease manager for the lease takes action to renew the lease, it does the following:
• It sets the FailOver state for the lease to Not Started.
• It resets the Retry Failover value to null.
At this point, the renewal of the lease resets the automatic failover process and negates any previous failover action
undertaken for the renewed lease.
The Background Task Failover Plugin

In the base configuration, Guidewire provides a BackgroundTaskFailoverPlugin plugin implementation that you
configure to manage component lease failover. The default implementation class for this plugin is Gosu class
DefaultBackgroundTaskFailoverPlugin. This default implementation makes the implicit assumption that there is
no manual cleanup work required for any batch process, message destination, or startable plugin after a component
lease failover.
At the failover of a component lease, the failover logic postpones the start of the failover process by several minutes,
the value of static variable INITIAL_POSTPONE_TIMEOUT. Failover logic implements the timeout to handle the case
in which the database was unavailable for a relatively long period of time, and then comes back online. The failover
postponement provides some time for the cluster members to recover and renew their leases.
After the failover postponement completes:
• If the cluster member that owns the lease does not return to the cluster, the automatic failover process continues.
• If the cluster member that owns the lease does return to the cluster, the automatic failover process fails.
Active External Monitoring

Some clusters use external monitoring and management software to watch the JVM processes of cluster members.
Guidewire provides Gosu class ActiveExternalMonitoringBackgroundTaskFailoverPlugin as a template that
you can to use in such installations. You must implement your own logic for the
notifyExternalMonitoringAboutExpiredlease method in this class.
Passive External Monitoring

Guidewire provides Gosu class PassiveExternalMonitoringBackgroundTaskFailoverPlugin as a template to
use in requesting a cluster member status report from external monitoring and management software. In this case,
the external monitoring and management software is not actively watching the JVM processes of cluster members.
Instead, it provides cluster member data on request.
Any monitoring and management software that you use for this purpose needs to be able to do the following:
• Check if a JVM process on a specified cluster member is alive and return the process uptime in seconds if the
process is running.
• Terminate a specified JVM process.
• Start a new JVM process.
The plugin implementation manages the failover.
Batch Process Prioritization

In the base configuration, Guidewire provides an implementation of a lightweight strategy to prioritize the running
of batch processes within the PolicyCenter cluster. File batch-process-config.xml provides a way to control the
execution of multiple batch processes.
This file has the following syntax:
<batch-process-config>
<settings defaultServer="server_ID" env="environment" startupDelay="delay_list" startup-Timeout=n

pollInterval=n/>
<param name="parameter_name" value="parameter_value" env="environment"/>
<batch-process typecode="type_code" env="environment" server="server_ID"/>
</batch-process-config>

The following table describes the attributes on the <settings> element.
Element Attribute Re‐ Description

quired
settings defaultServer Yes Prefaced by the # (hashtag) symbol, this value specifies a server role. For example,
defaultServer="#batch" means that every cluster member with the batch role is ca‐
pable of running batch processes.
Without the # symbol, this value specifies a server ID. For example,
defaultServer="node1" means that the cluster member defined in the <registry>
element in config.xml as serverid=node1 is capable of running batch processes.
pollInterval Yes Time, in seconds, between polling the database for new available leases.
PolicyCenter also broadcasts information on new batch process leases. See “Simple
Lease Management Lifecycle for a Batch Process” on page 157 for more information.
startupDelay Yes Number of seconds the batch process manager has to wait before starting the next
process. The delay is dependent on the number of already running batch processes on
the current server. In the base configuration, Guidewire sets this value to the following:
"0, 5, 15, 90, 180"
These values have the following meaning:
0 – Start the first process immediately
5 – Start the second process after 5 seconds
15 – Start the third process after 15 seconds
90 – Start the fourth process after 90 seconds
180 – Start the fifth and all subsequent processes after 180 seconds
startupTimeout Yes Number of seconds allotted for a batch process to acquire a lease. If the time expires
without the batch process acquiring the lease, PolicyCenter generates an error mes‐
sage similar to the following text (600 is the default value):
ERROR Server.BatchProcess No one started batch process after 600 seconds
env No Use to assign values for distinct installation environments. If used, this <settings> ele‐
ment applies only to the listed environments. To apply a specific <settings> element
to multiple environments, input multiple comma‐separated values for the env attrib‐
ute. A multi‐valued env attribute eliminates the need for a separate <settings> ele‐
ment for each environment.
The following table describes the attributes on the optional <param> element.

param name Yes Name of the parameter. The only valid value is RetryIfInitialConditionsFail.
value Yes Boolean value of RetryIfInitialConditionsFail.
env No Use to assign a different value for RetryIfInitialConditionsFail to distinct installation
environments.
The following table describes the attributes on the optional <batch-process> element.

quired
batch- typeCode Yes Typecode value of a batch processes defined in the BatchProcessType typelist.
process
env No Use to assign values for distinct installation environments. If used, this <batch-process>
element applies only to the listed environments. To apply a specific <batch-process> ele‐
ment to multiple environments, you can input multiple comma‐separated values for the env
attribute. A multi‐valued env attribute eliminates the need for a separate <batch-process>
element for each environment.


quired
server No Specifies a server on which to run the batch process specified by the typeCode attribute.
The RetryIfInitialConditionsFail parameter
The optional RetryIfInitialConditionsFail parameter governs the behavior of a batch process depending on the
return value of batch process class method checkInitialConditions. The name of the parameter is case-
insensitive.
This parameter takes the following values.
true If the batch process checkInitialConditions method returns false, the batch process does not terminate
immediately. Instead, it continues to execute until its component lease expires. This is the batch process default
behavior if the attribute is missing from batch-process-config.xml.
false If the checkInitialConditions method returns false, the batch process terminates immediately.
The <param> element supports the env attribute, meaning that it is possible to set this parameter for different
application environments, for example:
<paramname="RetryIfInitialConditionsFail" value="false"/>
<paramname="RetryIfInitialConditionsFail" value="false" env="dev"/>
<paramname="RetryIfInitialConditionsFail" value="true" env="prod"/>
Notice the following in this code example:

• The example code contains one parameter definition that has no defined env value set for the
RetryIfInitialConditionsFail parameter.
• The example code does not contain a defined RetryIfInitialConditionsFail value for the staging
environment
If a particular environment does not set the RetryIfInitialConditionsFail parameter, PolicyCenter does the
following:
• It sets the value of the missing parameter for that environment to that of the parameter definition that does not
have an env attribute set, if that parameter definition exists.
• If there is no parameter definition that does not have the env attribute set, PolicyCenter sets the value of the
missing parameter to the RetryIfInitialConditionsFail default value, which is true.
Messaging and Startable Service Load Balancing

In the base configuration, Guidewire provides several default strategies for managing the load balancing of
messaging destinations and startable services (plugins) within the PolicyCenter cluster.
Gosu class DefaultBackgroundTaskLoadBalancingPlugin provides the implementation details for plugin
BackgroundTaskLoadBalancingPlugin. Using this class, you can implement the following strategies for load
balancing.
Strategy Description
Work stealing Periodically transfers a component lease from an over‐utilized server to a server that is under‐utilized.
Work acquisition Provides under‐utilized servers with a chance to acquire a lease on an available component.
To enable or disable the use of each strategy, Guidewire provides the following plugin parameters.

messageDestinationLoadBalancingMode Manages the load balancing strategies for message destinations.
startablePluginLoadBalancingMode Manages the load balancing strategies for startable services.
Each of these plugin parameters can take one of the following values.
Value Description
disabled Disables both the work acquisition and work stealing strategies.
dynamic Enables both the work acquisition and work stealing strategies.
notransfer Enables the work acquisition strategy only.
It is possible to modify Gosu class DefaultBackgroundTaskLoadBalancingPlugin to meet your business needs.


chapter 10
Deploying Configuration Changes in a

Clustered Environment
Guidewire provides a way to deploy configuration changes to each individual server in a PolicyCenter cluster.
Guidewire calls this type of configuration deployment a rolling upgrade, in the sense that upgrade changes move
through the cluster, one server instance at a time. This type of configuration deployment is in contrast to a full
database and application upgrade. A full upgrade requires that you bring down all PolicyCenter servers in the cluster
to complete the upgrade. Typically, a full upgrade includes changes to the PolicyCenter database.
Rolling Upgrade Overview

To support the need to perform regular configuration updates without stopping the entire production installation,
Guidewire provides a way to update each application instance in a PolicyCenter cluster separately. Guidewire calls
this a rolling upgrade or configuration deployment. In a rolling upgrade, it is possible to stop an individual
PolicyCenter server and deploy configuration changes to it without stopping the other servers in the cluster. In this
way, PolicyCenter continues to function even through the configuration process.
For this to be possible, Guidewire restricts the types of configuration changes that are possible with a rolling
upgrade. The target (new) upgrade configuration must be compatible with the source (old) application configuration.
Guidewire provides a command line tool that you use to determine if the two application configurations are
compatible. If the two configurations are compatible, you can perform a rolling upgrade of each server instance in
the cluster. If not, you must perform a full application and database upgrade.
The requirements for a rolling upgrade are much simpler than for a full application upgrade. Thus, it is possible for a
PolicyCenter system administrator to perform a rolling upgrade of the cluster server members. Alternatively, a
member of your IT department can perform this type of upgrade.
IMPORTANT You cannot use a rolling upgrade to upgrade from a major or minor version of
Guidewire PolicyCenter to another major or minor version of PolicyCenter. In almost every case, a
rolling upgrade is not suitable for Guidewire application patches or maintenance releases. Only if an
application patch meets the compatibility criteria necessary for a rolling upgrade is a rolling upgrade
of that patch possible. A rolling upgrade is not a replacement or substitute for a full application and
database upgrade.
Risks Associated with Rolling Upgrade of Cluster Server Members

In performing a rolling upgrade, you trade safety for convenience. In particular, there are risks associated with
modifying any configuration related to distributed processing such as batch processes and work queues. Do not
Deploying Configuration Changes in a Clustered Environment 165
attempt to perform a rolling upgrade of the production PolicyCenter cluster if your changes affect those areas of the
application.
It is possible for individual PolicyCenter servers in the cluster to fail the configuration deployment process for a
variety of reasons. If so, you need to spend time testing and recovering from a failed attempt to return a server to a
safe state. In particular, you need to review any entities that possibly changed due to the deployment process.
Assumptions Around a Rolling Upgrade

As a rolling upgrade occurs in a live production environment, it is important to understand the assumptions
underlying a rolling upgrade.
After starting the rolling upgrade:
• Guidewire expects most PolicyCenter users to continue to interact with the cluster members running the source
build.
• Guidewire expects only a small set of users to interact with the cluster members running the target build.
• Guidewire expects a user to not switch between the source and target configurations as a matter of course. For
example, Guidewire does not expect a user to process the same policy on two different cluster members with
different configurations by switching back and forth between the configurations.
Guidewire expects you to execute and complete a rolling upgrade during a single scheduled maintenance window.
The maintenance window is a period of low activity that spans a duration of hours, not days.
See also
• “Verification of Configuration Compatibility” on page 172
Differences between a Rolling Upgrade and a Full Upgrade

Guidewire provides two different ways to modify the configuration of a PolicyCenter server in a cluster:
• Full upgrade
• Rolling upgrade
Each of these upgrade types has associated advantages and disadvantages. You use each upgrade type in different
circumstances.
Full Application and Database Upgrade

A full application upgrade has the following functionality:
• Supports either a single or multiple server configuration and upgrade.
• Supports making significant changes to both the PolicyCenter application and database.
• Requires that you stop the entire PolicyCenter production installation for a period of time.
• Requires a significant amount of testing before and after complex changes.
It is possible to start a full upgrade from any server instance in the PolicyCenter cluster, from the Server Tools
Rolling Cluster Member Upgrade

A rolling upgrade of the individual server members in the cluster has the following functionality:
• Supports configuration deployment to multiple server instances.
• Successively targets a single cluster member for configuration deployment while leaving the remaining cluster
members available for the processing of PolicyCenter operations.
• Supports a limited set of configuration changes.
• Requires that the new target configuration be compatible with the existing source configuration.
166 chapter 10: Deploying Configuration Changes in a Clustered Environment
• Supports a command-line utility that checks the compatibility of the source and target configurations before
deployment.
• Supports configuration deployment management and tracking from within PolicyCenter.
It is possible to start a rolling upgrade from any server instance in the PolicyCenter cluster, from the Server Tools
See also
Configuration Compatibility
Guidewire permits changes to the following files, file types, and installation folders in PolicyCenter Studio during
configuration deployment to the individual members of a cluster.
File or folder Safe actions

config.xml It is safe to change the value of a configuration parameter that Guidewire designates as possible to vary on
different servers in the same environment. Guidewire lists these parameters as Set for Environment: Yes in
the parameter description. To access a list of configuration parameters for review, see .
database- It is safe to change the value of attribute useoraclestatspreferences (from false to true or true to
config.xml false) during a rolling upgrade. However, if you reset the value of this attribute from true to false, you
must also perform manual steps to complete the process. See “Revert to DBStats Batch Processing for
Database Statistics” on page 263 for more information.
gsrc It is generally safe to add new Gosu classes during a rolling upgrade. However, it is unsafe to modify or re‐
move a Gosu class in a rolling upgrade. See “Making changes to Gosu code in a rolling upgrade” on page
168 for more information.
import It is safe for the import directory to differ across individual PolicyCenter instances, as long as the database
is not empty on server startup. This is because the import directory contains files that PolicyCenter imports
by default on server startup into an empty database.
Localizations It is safe to add, delete, or modify any of the display keys in the various
display_languageCode.properties files.
IMPORTANT It is not safe to make changes to either language.xml or localization_localeCode.xml as
these files must remain consistent across all cluster members.
plugin It is safe to modify a .gwp file in the root plugin directory, including pointing to a new implementation
class.
Guidewire does not permit the following with respect to plugins in a rolling upgrade:
• Modifications to non‐distributed, startable plugins
• Modifications to files in the plugins/Gosu or plugins/shared directories
Changes to plugin implementations can cause individual PolicyCenter instances to have different versions of
an object. Thus, it is possible that PolicyCenter instances running the source configuration to consider ob‐
jects create or updated on the target configuration to be invalid.
resources/ It is only safe to make changes to specific product model pattern types, as well as to very narrow changes to
productmodel those types. See “Updating Product Model Patterns in a Rolling Upgrade” on page 169 for more informa‐
tion.
rules It is safe to perform the following operations on PolicyCenter Gosu rules:
• Add a new rule
• Modify an existing rule, including enabling or disabling the rule
• Rename a rule, which is actually deleting a rule and adding the rule under a different name
Changes to validation rules can cause individual PolicyCenter instances to have different versions of an ob‐
ject. Thus, it is possible that PolicyCenter instances running the source configuration. to consider the ob‐
jects created or updated on the target configuration to be invalid.

File or folder Safe actions

servlet It is generally safe to modify existing Gosu servlet implementations in a rolling upgrade. However, it is not
permissible to modify config/servlet/servlets.xml in a rolling upgrade.
IMPORTANT Guidewire recommends that you undertake thorough testing after making a change to a Gosu
servlet implementation to verify that all product integrations continue to work as intended.
templates It is safe to make modifications to note, email, and document templates in the following directories as the
impact of a change to a template affects only that template:
• config/resources/doctemplates
• config/resources/emailtemplates
• config/resources/notetemplates
typelists It is generally safe to add typecodes to an existing typelist or to add an entirely new typelist. Also, it is safe
to edit the typelist description or change its category.
Note: If you add a typelist, the typelist shows on servers running the target (new) configuration only. On
servers running the source (old) configuration, the typelist shows as blank. See “Making Changes to Type‐
lists in a Rolling Upgrade” on page 168.
web It is safe to add or a delete a PCF or to modify an existing PCF.
webservices It is generally safe to make changes to web services. However, a change to an existing web service has the
potential to break integration with a third‐party product.
IMPORTANT Guidewire recommends that you undertake thorough testing after making changes to web
services configuration to verify that all product integrations continue to work as intended.
See also
Making changes to Gosu code in a rolling upgrade

It is generally safe to add new Gosu classes to Guidewire PolicyCenter during a rolling upgrade. However, as Gosu
is prevalent throughout PolicyCenter, modifications to Gosu code during a rolling upgrade can affect nearly every
part of the behavior of PolicyCenter. Thus, Guidewire explicitly recommends that you do not modify existing Gosu
classes during a rolling upgrade.
By definition, a rolling upgrade means that some nodes in a multi-node system pick up changes from the upgrade
before other nodes. Thus, changes in Gosu code can cause some nodes to behave differently than other nodes while
the rolling upgrade is in progress. For a full upgrade, this is not an issue, because you must shut down every node in
the cluster before you start the upgrade. However, for a rolling upgrade, the lack of a consistent behavior is a major
concern.
Do not modify, change, or delete Gosu code during a rolling upgrade of a PolicyCenter cluster.
Making Changes to Typelists in a Rolling Upgrade

Guidewire permits only certain changes to typelists during a rolling upgrade of application servers in a PolicyCenter
cluster.
Typelist Changes that Are Safe to Make in a Rolling Upgrade
In general, the following typelist configuration changes are safe to make in a rolling upgrade:
• Adding a new typelist
• Adding one or more typecodes to an existing typelist, either by extension or otherwise
• Modifying the name, description, priority, and sort order of an existing typecode
Typelist Changes that Are Not Safe to Make in a Rolling Upgrade

It general, the following typelist configuration changes are not safe to make in a rolling upgrade:
• Deleting an existing typecode
• Renaming an existing typecode, which is essentially a delete and add operation
• Adding a typecode to a typelist marked as final
Typelists that You Cannot Change in a Rolling Upgrade

There are specific typelists for which it is not safe to make a change of any kind in a rolling upgrade. These typelists
are:
• BatchProcessType
• GroupType
• LanguageType
• RoleType
• SystemPermissionType
• UserRole
• ValidationLevel
Guidewire maintains the list of blacklisted typelists in the following internal file:
• pl-rolling-upgrade-typelist-blacklist.txt
It is not possible to modify the list of blacklisted typelists in any way.
Updating Product Model Patterns in a Rolling Upgrade

Guidewire does not support all product model pattern changes during a rolling upgrade of the individual members of
a PolicyCenter server cluster. In general, Guidewire only supports the addition of new product model patterns for
types that support availability logic.
It is possible to make narrowly specific changes safely to the following types of product model patterns in a rolling
upgrade:
• “Coverage, Condition and Exclusion Patterns” on page 170
• “Offering Patterns” on page 170
• “Question Set and Question Patterns” on page 170
• “Modifier Patterns” on page 170
• “Lookup, Grandfathering, and Modifier MinMax Data” on page 171
Guidewire only supports changes to types of product model patterns and other product model information that the
above list explicitly describes.
Important
1. PolicyCenter provides the ability to reload product model availability data through the Server Tools Product
Model Info screen, by clicking Reload Availability. Guidewire disable this button during a rolling upgrade. The
button does not become active until a user clicks Rolling Upgrade Complete in the Server Tools Upgrade and
Versions screen.
2. Guidewire disables (makes inactive) all updates to product model patterns during a rolling upgrade in order to
preserve data integrity. After you complete the rolling upgrade, you must run Product Model Pattern
Activation batch processing to make active all product model patterns that you added in the configuration
update.
3. PolicyCenter upgrade logic automatically triggers the activation of product model patterns added during a
rolling upgrade if any node in the cluster restarts after the rolling upgrade completes. However, Guidewire
does not recommend that you uses this mechanism to activate product patterns. Run Product Model Pattern
Activation batch processing instead.
4. Guidewire specifically does not support the deletion of product model patterns in any type of upgrade.
See also
• “Product Model Pattern Activation Batch Process” on page 116.
Coverage, Condition and Exclusion Patterns

PolicyCenter stores coverage, condition and exclusion patterns in XML files, in Studio directories of the following
type under the resources directory:
productmodel/policylinepatterns/LOB/coveragepatterns/*.xml
PolicyCenter stores the coverage term and coverage term option patterns associated with each clause in a specific
XML file for its parent clause. In the file path:
• LOB refers to a specific line of business, for example, PersonalAutoLine.
• * represents the coverage or exclusion pattern, for example, PACollisionCov.xml or
ExcludeCustomEquipment.xml.
Guidewire supports only the addition of patterns for these types (coverages, conditions, exclusions, and their
associated coverage term and coverage term options) during a rolling upgrade. Do not attempt to modify the
attributes of any existing pattern of these types during a rolling upgrade. It is possible, however, to add new children
to existing patterns. For example, it is possible to add a new coverage term to an existing coverage, or, a new
coverage term option to an existing coverage term.
Offering Patterns
PolicyCenter stores offering pattern data in XML files, in Studio directories of the following type under the
resources directory:
productmodel/products/product/offerings/*.xml
In the file path:

• product refers to a specific product, for example, BusinessAuto.
• * represents the offering pattern, for example, BAStandardOffering.xml.
Guidewire supports only the addition of Offering patterns during a rolling upgrade. It is possible to add any type of
offering selection to an Offering pattern during the upgrade.
Question Set and Question Patterns

PolicyCenter stores question and question set pattern data in XML files, in the following Studio directory under the
resources directory:
productmodel/questionsets/*.xml
In the file path:

• * represents the question set pattern, for example, BABusinessAutoPreQualA.xml
Guidewire supports only the addition of question set patterns or questions during a rolling upgrade. Do not attempt
to modify or delete any existing question or question set pattern during a rolling upgrade. Guidewire specifically
does not support the addition, modification or deletion of question answers in a product model rolling upgrade.
Modifier Patterns
PolicyCenter stores modifier patterns in XML files, in Studio directories of the following types under the resources
directory:
productmodel/products/product/product.xml
productmodel/policylinepatterns/LOB/LOB.xml

In the file path:

• product refers to a specific product, for example, BusinessAuto in the products directory. For example, you
find the modifier patterns associated with the Business Auto product in the file of the same name,
BusinessAuto.xml, in that directory.
• LOB refers to a specific line of business, for example, BusinessAutoLine in the policylinepatterns directory.
Guidewire supports only the addition of modifier patterns during a rolling upgrade.
Lookup, Grandfathering, and Modifier MinMax Data

PolicyCenter stores grandfathering and jurisdiction-specific modifier MinMax data in XML files, in Studio
directories of the following types under the resources directory:
productmodel/policylinepatterns/LOB/coveragepatterns/*-grandfathering.xml
productmodel/policylinepatterns/LOB/jurisdictions/jurisdiction/coveragepatterns/*-grandfathering.xml
productmodel/policylinepatterns/product/jurisdictions/jurisdiction/*-modifierminmax
productmodel/policylinepatterns/LOB/jurisdictions/jurisdiction/*-modifierminmax
In the file path:

• product refers to a specific product, for example, BusinessOwners.
• LOB refers to a specific line of business, for example, BOPLine.
• jurisdiction refers to the two character state code for the specific jurisdiction to which the file applies, for
example, AK for Alaska.
• * represents the pattern for which the grandfathering and modifier MinMax file applies, for example,
BAComprehensive-grandfathering.xml or PersonalAutoLine-modifierminmax.xml.
Lookup files are all of the form *-lookups.xml, with * as the name of the pattern associated with the lookup, for
example, BAComprehensive-lookups.xml. PolicyCenter stores lookup files in multiple directories, in either of the
following:
• In the same directory as the associated pattern
• In the appropriate jurisdiction-specific directory if the lookup is specific to a particular jurisdiction
For example, PolicyCenter stores non-jurisdiction-specific lookup rows for the BAComprehensive coverage pattern
(or, any of its child coverage terms and coverage term options) in file BAComprehensive-lookups.xml.
PolicyCenter locates this file in the following Studio directory:
proudctmodel/policylinepatterns/LOB/BusinessAutoLine/coveragepatterns
Correspondingly, PolicyCenter stores Alaska-specific lookups for the BAComprehensive coverage in the following
directory in Studio:
productmodel/policylinepatterns/BusinessAutoLine/jurisdictions/AK/coveragepatterns
Guidewire supports modifications to the files in these directories in the following ways only:
• By adding new lookup, grandfathering, and modifier minmax files to the directory
• By adding new lookup and grandfathering elements to existing files
Note: It is not necessary to use product model rolling upgrade to add new lookups or grandfathering
availability data. It is possible also to update the PolicyCenter server with this information, as well as
update availability scripts, using the Server Tools Reload Availability Data functionality. However, do not
attempt to use this functionality during the rolling upgrade itself.
See also

Verification of Configuration Compatibility

Guidewire provides a command line build tool to use in determining if the source and target application
configurations are sufficiently compatible to support configuration deployment in a clustered environment. To use,
run the following command from a command prompt on any cluster member.
system_tools -verifyconfig filepath
This utility compares the following two server configurations:

• The new or target server configuration contained in a WAR/EAR file pointed to by the filepath parameter.
• The existing server configuration of the cluster member on which you run the system_tools command.
The tool provides an on-screen report that contains information about the feasibility of a configuration deployment
for the server members in the cluster. For example, the tool provides the following types of information:
Configurations are different Requires a full server upgrade. Guidewire does not permit a configuration deployment (rolling
upgrade) using the target configuration.
Configurations are identical No upgrade is necessary.
Configurations are compatible Guidewire permits a configuration deployment of these changes.
If a configuration deployment is not possible, the command lists the incompatible or missing files.
If a configuration deployment is in progress, there are two possible configurations active in the cluster. Each
individual server instance in the cluster is using either the source configuration or the target configuration.
The -verifyconfig command option checks for both configurations on the cluster member on which you run the
command and reports which of the configurations is active on this cluster member. If neither configuration is active,
the command reports that a configuration deployment is in progress and that it is not possible to verify the
configuration at this time.
See also
Performing a Rolling Upgrade

Performing a rolling upgrade (configuration deployment) on your production PolicyCenter cluster requires extensive
testing to minimize the risk of deploying new configuration changes in a production environment.
The following list describes the steps involved in performing a rolling upgrade.
Step Description For more information

Prepare Verify that your environment is set up correctly for a rolling upgrade. “Prepare for a Rolling Upgrade” on page
173
Test Test the PolicyCenter upgrade configuration in a non‐production envi‐ “Perform a Rolling Upgrade in a Test Envi‐
ronment before you implement the configuration in a production en‐ ronment” on page 173
vironment.
Upgrade Upgrade the servers in your production environment with the new “Perform a Rolling Upgrade in a Production
PolicyCenter configuration, one by one. Environment” on page 174
Rolling Upgrade Terminology
Guidewire uses the following term definitions in the discussion on rolling upgrade.
Instance An individual PolicyCenter server running in a VM (Virtual Machine) or JVM (Java Virtual Machine) or
stand‐alone PolicyCenter server.
Test instance A PolicyCenter instance with the same data model and EAR/WAR build file as that used on a production
instance. The test cluster does not need to have the same number of test instances as the production
cluster. However, there needs to be at least two instances in the test cluster to be able to test the rolling
upgrade process.
Production instance A member of the production cluster accessed and used by external PolicyCenter users.
Prepare for a Rolling Upgrade

Before you begin
Before you begin the rolling upgrade process, review “Performing a Rolling Upgrade” on page 172.
About this task

The following list describes the necessary actions that you need to undertake before performing the rolling upgrade.
Action How? More information

Verify that the database schema Use the database schema verification tool available “Database Table Info” on
matches the PolicyCenter data model from the Server Tools Database Table Info screen. page 337
as defined in the data model metada‐
ta files.
Verify that the source and target con‐ Use the system_tools -verifyconfig utility to veri‐ “Verification of Configuration
figurations are compatible. fy compatibility between the old and new configura‐ Compatibility” on page 172
tions.
Back up your current application con‐ Create a robust database backup and restore strategy Consult with your database
figuration. in case there are problems upgrading individual clus‐ administrator
ter members.
Create a custom version label for your This step is optional. However, a version label provides “Understanding Guidewire
configuration deployment. a way to identify this configuration deployment for fu‐ Software Versioning” on page
ture reference. 365
Next steps
After you complete these steps, continue to “Perform a Rolling Upgrade in a Test Environment” on page 173.
Perform a Rolling Upgrade in a Test Environment

Before you begin
Before you begin testing your rolling upgrade configuration, review “Prepare for a Rolling Upgrade” on page 173.
About this task

Guidewire recommends that you test your planned upgrade configuration in a test environment exhaustively before
upgrading a production installation.
Procedure
1. In file database-config.xml, verify that the <database> element autoupgrade attribute is set to manual (or
non-existent).
If the attribute is missing, the default value for this attribute is manual. The value cannot be full.
2. Create a new EAR/WAR PolicyCenter build that includes all the proposed configuration changes. The build
name must includes some identification such as a date or a version number.
3. Place the EAR/WAR build in a local directory on a test instance.

4. Run the following command option from the command prompt.
system_tools -verifyconfig filepath
This verification process tests if the target configuration is compatible with the source configuration on that
server. If the configuration verification tool indicates that the proposed changes are compatible with the
existing configuration, continue to the next step.
5. Back up the test database.
6. On any server instance in the cluster, navigate to the Server Tools Upgrade and Versions screen and click Start
Rolling Upgrade.
7. Bring down one of the servers in the test environment and deploy the new EAR/WAR file to this server
instance.
8. Bring the test instance back up.
9. Perform user acceptance testing on the test cluster, on both the old and new configurations. .
If testing indicates that there are no major issues in running the two configurations simultaneously, continue to
the next step.
10. Deploy the new EAR/WAR file to all test instances.
11. Navigate to the Server Tools Upgrade and Versions screen of any server instance in the test environment.
12. Click Rolling Upgrade Complete.
This action clears the upgrade flag indicating that a rolling upgrade is in progress. The rolling upgrade of the
new configuration changes is now complete.
13. Run Product Model Pattern Activation batch processing to activate the newly added product model patterns.
Running this batch process makes active all product model patterns that you added in the configuration
update. See “Product Model Pattern Activation Batch Process” on page 116 for more information.
IMPORTANT Restarting any node in the cluster after you complete the rolling upgrade (after
you click Rolling Upgrade Complete) automatically triggers the activation of the added product
model patterns. Guidewire does not recommend that you activate the product model patterns in
this manner.
14. Perform acceptance testing on all the test instances to verify that PolicyCenter works as intended.
Next steps
If testing indicates that there are no issues with the new configuration running on all test instances, continue to
“Perform a Rolling Upgrade in a Production Environment” on page 174.
Perform a Rolling Upgrade in a Production Environment

Before you begin
Before performing a rolling upgrade in a production environment, ensure that you have completed all steps in
“Perform a Rolling Upgrade in a Test Environment” on page 173.
About this task

Only upgrade a production PolicyCenter after testing the upgrade configuration exhaustively in a test environment.
Procedure
1. In file database-config.xml, verify that the <database> element autoupgrade attribute is set to manual (or
non-existent).
If the attribute is missing, the default value for this attribute is manual. The value cannot be full.
2. On any instance in the PolicyCenter production cluster, navigate to the Server Tools Upgrade and Versions
screen.
a. Click Start Rolling Upgrade.

b. Verify the checklist of upgrade prerequisites.
c. Click Start Rolling Upgrade.
This action sets a flag in the PolicyCenter database that indicates a rolling upgrade is in progress.
3. Navigate to the Server Tools Cluster Members screen on any server instance.
a. For the instance that you want to shutdown, click Start Planned Shutdown in the Actions column.
b. Set the appropriate shutdown parameters in the Schedule Planned Shutdown screen.
c. Click Schedule Shutdown.
This action schedules a shutdown of the specified instance. All users logged into this PolicyCenter
instance see an on-screen message indicating that a planned shutdown is in progress. After the scheduled
period of time elapses, there are no more user connections to this production instance.
4. Deploy the new PolicyCenter build to the production instance that you shut down.
5. Bring the instance with the configuration build back up.
6. Perform acceptance testing on the production instances to determine if there are any major issues with running
the two configurations in the same cluster.
If testing indicates that there are no major issues with the new configuration in the production cluster, continue
to the next step. If there are issues, repeat the previous steps until you have upgraded all the production
instances with the new configuration build.
7. Perform another round of acceptance testing.
If testing indicates that there are no major issues with the new configuration on the production instances,
continue to the next step.
8. Navigate to the Server Tools Upgrade and Versions screen of any instance in the PolicyCenter production cluster.
9. Click Rolling Upgrade Complete.
This action clears the upgrade flag indicating that a rolling upgrade is in progress. The rolling upgrade of the
new configuration changes is now complete.
10. Run Product Model Pattern Activation batch processing to activate the newly added product model patterns.
Running this batch process makes active all product model patterns that you added in the configuration
update. See “Product Model Pattern Activation Batch Process” on page 116 for more information.
IMPORTANT Restarting any node in the cluster after you complete the rolling upgrade (after
you click Rolling Upgrade Complete) automatically triggers the activation of the added product
model patterns. Guidewire does not recommend that you activate the product model patterns in
this manner.
11. Perform another round of acceptance testing to ensure that there are no issues with the new configuration.
Unexpected Upgrades
Any time that you deploy a new WAR/EAR file to a PolicyCenter server and restart the server, PolicyCenter
assumes that an upgrade is in progress. To prevent the unexpected upgrade of a server, Guidewire requires that you
set an upgrade flag in PolicyCenter before starting either a full or rolling upgrade.
Guidewire requires the use of this flag to mitigate the risk of accidentally triggering an unexpected upgrade. As a
consequence, however, it is possible to encounter situations in which the PolicyCenter server does not start. In that
case, you must undertake a recovery sequence to return the server to a state in which it can start.
Full Upgrade
For a full upgrade, Guidewire first requires that you click Start Full Upgrade in the Server Tools Upgrade and Versions
screen (on any cluster member). This action signals your intention to perform a full upgrade. PolicyCenter then sets
a database flag to indicate that a full upgrade is in progress. After you complete the upgrade, PolicyCenter deletes
the database flag. You must set the upgrade flag again before starting a new full upgrade.
It is possible to set the full upgrade in progress flag in the following ways as well.
System To set the upgrade flag through system tools, use the following command option:
tools system_tools -startfullupgrade
At least one cluster member must be running in order for you to use this option.
Web serv‐ To set the upgrade flag using web services, call the SystemToolsAPI web service method startFullUpgrade. At
ices least one cluster member must be running in order for you to use this option.
Java sys‐ To set the upgrade flag through a Java system property, use the following system parameter to set the expected
tem prop‐ date of the upgrade while starting one of the affected servers:
erty gwb runServer -Dgw.pc.full.upgrade.intended.date=date
The date parameter is the current date in yyyyMMdd format.
If you encounter a situation in which all cluster members refuse to start because the upgrade flag was not set, you
cannot set the upgrade flag through the server. Instead, you must set the upgrade flag using the Java system
parameter.
Rolling Upgrade
For a rolling upgrade, Guidewire first requires that you click Start Rolling Upgrade in the Upgrade and Versions screen
(on any cluster member). This action signals your intention to perform a rolling upgrade and sets a rolling upgrade
in progress database flag. If you do not set the upgrade flag, PolicyCenter refuses to start a rolling upgrade.
After you complete the upgrade of all servers in the cluster, you must click Rolling Upgrade Complete on the Upgrade
and Versions screen, which removes the upgrade flag. After you do so, it is not possible start a cluster member
running the source (old) configuration.
Guidewire permits a rolling upgrade of the individual members of a PolicyCenter cluster under certain conditions
only. In effect, the source (old) configuration and target (new) configuration must be compatible in very specific
ways.
Thus, during a rolling upgrade, if you mistakenly deploy an incompatible WAR/EAR file to a PolicyCenter server,
you can encounter a situation in which the server does not start. This is true whether you have set the rolling upgrade
in progress flag. In this case, remove the incompatible WAR/EAR file and deploy a compatible WAR/EAR file
before attempting to restart the server.
See also
Backing Out a Rolling Upgrade

Important Caveat
It is possible to back out configuration changes on a given server application instance, depending on the state of the
configuration deployment on that instance. Guidewire does not recommend that you back out an ongoing
deployment on a server instance except in extraordinary circumstances. If you do so, you need to test that the server
configuration returns to a safe state.
Initiating a Back Out
IMPORTANT Before you begin the process of backing out a rolling upgrade of a PolicyCenter, first
shut down any cluster node that is running the new, target, application configuration.
If a rolling upgrade of a PolicyCenter cluster is in progress, the PolicyCenter Server Tools Upgrade and Versions
screen shows an Initiate Backout button. Clicking this button opens the back-out wizard screen. This screen provides
on-screen instructions on how to proceed with the back out of the rolling upgrade.
As part of the back out process, PolicyCenter runs an internal batch process to verify there are no orphaned
typecodes. At least one cluster member must have the batch server role in order to execute this process.
Rolling Upgrade Failure

Under rare circumstances, it is possible for the failure of a rolling upgrade on a specific application server to result
in the inability to back out of the rolling upgrade on that server. To provide for this upgrade scenario, Guidewire
provides a Force Backout button on the Upgrade and Versions screen. Clicking this button brings up a dialog in which
you can select the server ID of the inactive server. The Force Backout button is only available if a rolling upgrade is in
progress.
Configuration Upgrade of a Production Stand‐alone PolicyCenter

Server
A stand-alone server has configuration parameter ClusteringEnabled set to false. The use of a rolling
(configuration) upgrade is not possible on a stand-alone server. However, some upgrade changes require a full
database upgrade, other upgrade changes do not.
A full upgrade on a production PolicyCenter stand-alone server causes the following behavior.
Upgrade type Example Behavior

Data model Changing the value of a semi‐permanent If you do not initiate a full upgrade from the Server Tools Upgrade
configuration parameter. PolicyCenter and Versions screen, the application server does not start.
writes these kinds of changes to the Conversely, if you do initiate a full upgrade from Server Tools, the
application database. upgrade completes and PolicyCenter updates the Upgrade and
Versions screen.
Configuration Changing the value of a general It does not matter whether you initiate a full upgrade from the
configuration parameter, one that is Server Tools Upgrade and Versions screen. In any case, the application
neither permanent nor semi‐permanent. server starts without any errors or warnings and PolicyCenter
PolicyCenter reads these kinds of changes updates the Upgrade and Versions screen.
from configuration files.


part 4
Security Administration
chapter 11
Managing Secure Communications
Guidewire products use a three-tier architecture:

• The browser tier presents the PolicyCenter interface to the user.
• The web/application tier processes business logic.
• The database tier stores data.
Encryption secures communication between computer systems. You can secure the communication between the
browser, web server and a PolicyCenter server to a level strong enough that it cannot be easily compromised.
IMPORTANT Computer security and encryption is a complex topic in which network architecture
plays a major role. Use this documentation as a starting point. Guidewire strongly recommends that
you also perform independent research and testing to develop a secure solution for your company
network and installed applications. Guidewire strongly recommends that you deploy PolicyCenter
over TLS (Transport Layer Security) for at least the login and change password pages. Ideally, deploy
PolicyCenter entirely under TLS to protect all sensitive transmitted data.
PolicyCenter and the Transport Layer Security Protocol

Guidewire applications use the TLS (Transport Layer Security) protocol while making WS-I and RPC web service
connections to HTTPS endpoints. TLS is the only communication protocol that Guidewire supports for this purpose.
There are multiple versions of the TLS protocol. All current releases of JDK 7 support TLS 1.0, TLS 1.1, and TLS
1.2.
The TLS default version is the underlying JDK default for the installed release of JDK 7. This is TLS 1.0 for the
public, free (non-paid support), releases. However, use of the TLS 1.0 default can cause the connection to a server to
fail if the server requires either TLS 1.1 or TLS 1.2.
Setting TLS version overrides
Guidewire provides several Java property overrides to set the default TLS version to use on outgoing secured
connections. You can use these property overrides with either the paid support or the free versions of JDK 7. Use
these property overrides to provide a comma-separated list of TLS protocol versions. PolicyCenter uses the first item
on the list as the preferred protocol. If that protocol is not available, PolicyCenter tries the subsequent protocols on
the list until the connection either succeeds or fails completely.
The following table lists the available property overrides.
Managing Secure Communications 181

Web service type Property Syntax

WS‐I gw.webservices.tls.protocols -D.gw.webservices.tls.protocols="a, b"
RPC gw.tls.protocols -Dgw.tls.protocols="a, b"
In the table, a and b refer to TLS versions, for example:
<java> ... -D.gw.webservices.tls.protocols="TLSv1.2, TLSv1.1"
Notice the following for this example:

• The property definition indicates that TLS1.2 is the preferred protocol. However, if TLS1.2 is not available,
PolicyCenter attempts to use TLS 1.1 instead.
• The property definition affects only client WS-I web service calls.
PolicyCenter and Secure Communications

A strong password policy is the first and best line of defense. Guidewire also recommends the following:
• Encrypt the communication between the Internet and the PolicyCenter server or cluster.
• Configure a separate server to act as an intermediary layer between the Internet and any PolicyCenter server or
servers. Typically, you locate this intermediary server in a DMZ that you establish through your network
architecture.
If you off-load encryption to a server, understand that non-native encryption processing is not as efficient. Native
applications generally use optimized encryption modules.
You can use a web server or proxy both to encrypt communications and to provide a layer between the Internet and a
PolicyCenter server. Computer network terminology generally calls a server working as an intermediary in this
manner a reverse proxy. There are multiple methods you can use to achieve an encrypted proxy solution.
See also
• “The PolicyCenter Connection Address” on page 182
The PolicyCenter Connection Address

In setting up a secure connection with Guidewire PolicyCenter, ensure that you use a connection address similar to
the following:
https://server:port/pc/PolicyCenter.do
You need to set the server name and port number to that on which the proxy server is running. Notice the use of the
https protocol instead of http, indicating that the server is connecting through a secure version of HTTP.
Set this address for every client that connects to the PolicyCenter application server, including web browsers,
PolicyCenter plugins, and applications that use PolicyCenter APIs.
In configuring this URL, also check PolicyCenter file config.xml for any URL specifications that you need to
specify.
Restricting access to a PolicyCenter screen by server mode

Guidewire uses PCF (Page Configuration Format) files to render PolicyCenter screens. Properties associated with
each PCF screen determine whether it is possible to visit, or edit, the screen.
It is possible to restrict access to a screen based on server mode. For example, an administrator can usually access an
Activity Patterns screen. This screen provides a list of all of the activity patterns available in PolicyCenter. Clicking
one of the listed activity patterns opens that pattern's detail screen.
182 chapter 11: Managing Secure Communications
Suppose, for some reason, that you want to restrict access to the Activity Patterns Detail screen on a production system.
Opening up ActivityPatternDetails.pcf in Studio shows the following properties:
• canEdit
• canVisit
To not permit access to the Activity Patterns Detail screen on a production system, enter the following value for the
canVisit property:
gw.api.system.server.ServerUtil.getEnv() != "PROD"
The canVisit property must evaluate to true for the Activity Patterns Detail screen to be accessible to a PolicyCenter
administrative user. If the server is in development or test mode, the expression evaluates to true and PolicyCenter
allows access to the screen.
Multifactor Authentication
Multifactor authentication, as the name implies, requires a user to provide two or more pieces of data (factors) to an
authentication mechanism as proof of identity. For example, a bank can require, in addition to name and password, a
code that the bank sends to the user's smart phone.
Guidewire provides the ability to pass multiple additional factors from the application login screen to PolicyCenter
for processing. To implement the simplest, most basic, form of multifactor authentication, which is the addition of a
multifactor entry field to the PolicyCenter login screen, you need to do the following:
• Make visible the multifactor entry field on the PolicyCenter login screen.
• Add additional Gosu classes to support the use of multifactor authentication in PolicyCenter.
Add multifactor authentication field to PolicyCenter Login screen

Guidewire provides a basic framework that you can use to implement multifactor authentication in PolicyCenter.
About this task

This procedure illustrates how to add one or more multifactor authentication fields to the PolicyCenter login screen.
The extra authentication fields perform no actual authentication. For multifactor authentication to work as intended,
you must perform additional configuration work.
IMPORTANT Any PolicyCenter application server that uses multifactor authentication must have the
messaging server role.
Procedure
1. In PolicyCenter Studio, open the display_xx.properties file for your locale and add something similar to
the following display keys to the file in appropriate place:
Web.Login.RSADongleNumber = RSA Dongle Number

Web.Login.RSADongleId = RSA Dongle ID
2. Open file Login.pcf in the following Studio directory:

configuration/config/web/pcf/util
a. Ensure that you select the entire PCF element so that the various tabs show in the Properties area of the
screen.
b. In the Variables tab, select the entry for loginForm.
c. Set the initial value (initialValue) for this variable to the following value (without the line breaks):
new gw.api.util.LoginForm(target, entryException)

.withFactor( "Web.Login.RSADongleNumber", "XXX")
.withFactor( "Web.Login.RSADongleId", "YYY")
Managing Secure Communications 183

XXX and YYY represent constants from the supporting multifactor authentication classes.
MFAAuthenticationSource.FACTOR_ATTRIBUTE_NAME, represents one such constant, for example.
3. Open file LoginDV.pcf in the following Studio directory:
configuration/config/web/login
a. Select the field with the label factor.Label. label.
b. In the Properties tab at the bottom of the screen, select property numCols and set its value to 20.
This action sets the length of all of the new entry fields on this screen to the same length as the existing
entry fields.
4. If using an application server other than the default Quickstart server, create and deploy a WAR or EAR file as
necessary.
5. Start the PolicyCenter server and let it compile the newly added code.
Result
Upon opening the PolicyCenter login screen, you see the additional RSA fields.
Next steps
To make the new multifactor authentication fields functionally useful, you must implement custom implementation
classes for the following plugins, among other configuration changes:
• AuthenticationServicePlugin
• AuthenticationSourceCreatorPlugin
See the Integration Guide for more information.
184 chapter 11: Managing Secure Communications

chapter 12
Securing Access to PolicyCenter

Objects
Guidewire designs its security infrastructure so that you can add custom permissions, automatically enforce
permissions, and easily map between users, permissions, and actions. This topic explains how to use the
PolicyCenter permission infrastructure to control access to key PolicyCenter objects.
Understanding the Object Access Infrastructure

You can assign each user one or more roles that contain permissions. These permissions control what the user can do
in Guidewire PolicyCenter. For example, a user in PolicyCenter with the correct role can create a note on a Policy.
The limitation of roles is that they do not distinguish among objects of the same type. In the previous example,
“note” means all notes and all note types. However, suppose that you want to restrict access to notes that contain
sensitive information. In this case, PolicyCenter provides access control features that you use to restrict access to
specific types of notes.
By implementing access control, you can subcategorize an object type and then restrict object access by these
subcategories. In the base configuration, you can apply access control to the following business objects:
• Account
• Activity
• Document
• Job – and all its subtypes
• Note
• PolicyPeriod
• User
It is possible to apply access control (permissions) to any PolicyCenter business object.
Understanding the PolicyCenter Permission Types

PolicyCenter contains the following types of permissions.
“System Permission Keys” on page 186 System permission keys apply to specific user interface elements or data model enti‐
ties.
“Application Permission Keys” on page Application permission keys represent a set of one or more system permissions.
186
Securing Access to PolicyCenter Objects 185

You can view a list of both system and application permission keys in the Guidewire Security Dictionary.
See also
System Permission Keys

Guidewire groups system permissions into the following categories:
• Screen-level permissions that apply to user interface elements
• Domain-level permissions that apply to data model entities
You can view a list of system permissions in the SystemPermissionType typelist. You can also extend this typelist
and add custom permissions.
Screen‐level Permissions
Screen-level permissions apply to user interface elements, for example, the permission to view the administrative
Server Tools screens. PolicyCenter defines many user interface permissions internally.
In general, screen-level permissions start with the word “view” followed by a reference to the user interface object
they protect. You can add custom screen-level permissions to Guidewire PolicyCenter by extending the
SystemPermissionType typelist.
PCF files define the point at which PolicyCenter calls user interface permissions. It is possible to change this point
by customizing the PCF file that calls it.
Domain‐level Permissions
Domain permissions apply to data model entities, such as permission to view Note objects. For example, as a user
attempts to access the summary for a sensitive note, PolicyCenter verifies that the user has the following
permissions:
• Permission to view the Policy screen
• Permission to access that particular note type
Most top-level objects in the PolicyCenter data model have associated domain-level permissions. PolicyCenter
defines all of an object’s domain-level permissions internally. It is not possible to add, remove, or edit domain
permissions. Similarly, PolicyCenter defines the points at which it checks these permissions in internal code and in
page configuration format (PCF) files. You cannot change the internal checks. You can, however, change the point at
which the PCF files calls these checks.
Application Permission Keys

Application permission keys represents a set of one or more system permissions. PolicyCenter defines application
permission keys internally as a method for improving system performance. For example, the Activity own
permission key represents the system permission for owning an activity. The Activity edit permission key
represents the system permission for the editing activities.
Guidewire defines all configurable application permissions in file security-config.xml. It is possible for you to
modifying this file and add new application permissions.
Guidewire defines many other access application permissions internally. It is not possible to change these
permissions.
See also
• “The Security Configuration File” on page 187
Permission Class Generation

PolicyCenter generates its permission classes at server startup by parsing file security-config.xml. The
PolicyCenter log shows this activity:
186 chapter 12: Securing Access to PolicyCenter Objects
...INFO SecurityManager finished parsing ...\modules\configuration\config\security\security-config.xml
PolicyCenter does not generate permissions automatically for the subtypes of an entity. You must explicitly add the
entity subtype to security-config.xml for PolicyCenter to generate permissions for that subtype.
Beyond Roles and Permissions to Access Control

You can group system permissions by adding them to roles and then assigning the role to a user. So, if a particular
role has a view policy document permission, any user with that role can view a document attached to a policy. And,
of course, the user must first have access to that policy.
In practice, however, you likely do not want all users to access all objects of the same type. For example, suppose
that an object has an associated document that contains information on a famous celebrity. You most likely want to
restrict access so that only certain people have access to the personal information contained in this document. You
use the PolicyCenter access control feature to make distinctions among objects of the same type and then secure
access to them.
While roles and permissions determine what actions a user can perform, access control determines the objects on
which the user can act. After you enable access control, a user requires both the correct role and the proper access.
To use access control, you apply a security attribute to an object and then determine which users have access to
objects with that attribute.
Access Control Configuration Files

You manage access control in PolicyCenter with the following files:
File Role the file plays in access control

security-config.xml Defines business object security handlers.
DocumentSecurityType.ttx Defines access control types related to documents.
MoneySecurityType.ttx Defines access control types related to money.

NoteSecurityType.ttx Defines access control types related to notes.
SystemPermissionType.ttx Contains customizable and custom system permissions.
You modify these files in Guidewire Studio.

• File config.xml is available under configuration→config.
• File security-config.xml is available are under configuration→config →security.
• The various typelist TTX files are available under configuration→config→Metadata→Typelist.
See also
• “Understanding the Object Access Infrastructure” on page 185
The Security Configuration File

PolicyCenter provides the means to specify user access to business objects through application configuration file
security-config.xml. This XML-formatted file contains a number of XML elements that you use to configure
user access to specific business objects.
You access security-config.xml in Guidewire Studio, in the following location:
configuration→config →security
Object Access and security

The following topics discuss the standard XML elements in security-config.xml that relate to PolicyCenter
object access and security:
• “Static Handler Elements” on page 188
• “Wrap Handler Elements” on page 189
Note and document objects

The following topics discuss the XML elements in security-config.xml that relate to PolicyCenter note and
document objects
• “Note Permission Overview” on page 193
• “Document Permission Overview” on page 194
Account, job and policy period objects

The following topics discuss the XML elements in security-config.xml that relate to PolicyCenter account, job,
and policy period objects:
• “Account Producer Code Handler Elements” on page 197
• “Job Producer Code Handler Elements” on page 199
• “PolicyPeriod Producer Code Handler Elements” on page 200
See also
• “Access Control Configuration Files” on page 187
Static Handler Elements

You use the <StaticHandler> element in security-config.xml to define security permissions on an entity. This
type of security permission is static and requires no object.
There is no limit to the number of <StaticHandler> elements that can exist in security-config.xml. Each
<StaticHandler> element can contain zero to many <SystemPermType> elements.
This element has the following syntax.
<StaticHandler entity="entity" permKey="perm" desc="..." noPermissionDisplayKey="key">

<SystemPermType code="code"/>
...
</StaticHandler>
You access this permission in code as perm.entity.perm. This syntax has the following meaning:
• entity – The business object or entity on which the permission acts.
• perm – The permission given for this entity.

StaticHandler entity Yes The entity type on which this security handler acts.
permKey Yes The application permission to grant.
desc No A human‐readable description of the permission.
noPermissionDisplayKey No A display key that provides the text to show if the user does not have
a required permission.
SystemPermType code Yes A code value defined in the SystemPermissionType typelist.

The following example shows a typical <StaticHandler> element.
<StaticHandler entity="User" permKey="ViewProfiler" noPermissionDisplayKey="No access to ViewProfiler.">

<SystemPermType code="internaltools"/>
<SystemPermType code="toolsprofilerview"/>
</StaticHandler>
Notice that:
• The security permissions work on a User entity.
• The application permission key is ViewProfiler.
• The handler lists a set of specific system permission types to which the handler grants the user access, if any of
the conditions are met.
To have the ViewProfiler application permission, the user must have an assigned role that contains one or more of
the listed system permissions.
Static Handlers Specify OR Conditions

Static security handlers define Boolean OR conditions. This means for the user to have a certain application
permission, the user must have an assigned role that contains at least one of the following:
• System permission A
• Or, system permission B
• Or, system permission C
• Or, …
Suppose that you have the following code that references the ViewProfiler static handler shown previously.
if (perm.User.ViewProfiler) ...
The sample code condition evaluates to true if the current user has an assigned role with either the internaltools
permission or the toolsprofilerview permission.
See also
Wrap Handler Elements

The <WrapHandler> element in security-config.xml defines complex security permissions on an entity. The
wrap handler “wraps around” the permission conditions of the associated handler. The associated handler type must
not be object-based, meaning that it must be one of the following:
• <StaticHandler>
• <WrapHandler>
It is not possible to create a security handler that takes an object using a <WrapHandler> element. A wrap security
handler always create a new static handler.
There is no limit to the number of <WrapHandler> elements that can exist in security-config.xml. Each
<WrapHandler> element can contain zero to many <SystemPermType> elements.
A <WrapHandler> element must come after the <Handler> element that defines the permission referenced by
wrapPermKey. The associated handler can be another <WrapHandler>. It is possible to cascade <WrapHandler>
elements.
<WrapHandler entity="entity" permKey="perm" wrapPermKey="wrapPerm" desc="..." noPermissionDisplayKey="key">

<SystemPermType code="code"/>
...
</WrapHandler>

You access this permission in code as perm.entity.perm. This syntax has the following meaning:
• entity – The business object or entity on which the permission acts.
• perm – The permission given for this entity.

WrapHandler entity Yes The entity type on which this security handler acts.
permKey Yes The application permission to grant.
wrapPermKey Yes The associated permission being wrapped. You must declare the per‐
mission referenced by the wrapPermKey earlier in security-
config.xml than this <WrapHandler> element.

noPermissionDisplayKey No A display key that provides the text to show if the user does not have
a required permission.
SystemPermType code Yes A code value defined in the SystemPermissionType typelist.
The following example illustrates a <StaticHandler> element with two cascading <WrapHandler> elements
following it.
// Static Handler - ViewProfiler permission

<StaticHandler entity="User" permKey="ViewProfiler" noPermissionDisplayKey="No access to ViewProfiler.">
<SystemPermType code="toolsProfilerview"/>
</StaticHandler>
//First Wrap Handler - EditProfiler permission

<WrapHandler entity="User" permKey="EditProfiler" wrapPermKey="ViewProfiler" noPermissionDisplayKey="No access to
EditProfiler.">
<SystemPermType code="toolsProfileredit"/>
</WrapHandler>
//Second Wrap Handler - EditWebserviceProfiler permission

<WrapHandler entity="User" permKey="EditWebserviceProfiler" wrapPermKey="EditProfiler" noPermissionDisplayKey="No
access to EditWebServiceProfiler.">
<SystemPermType code="toolsProfilerwebserviceedit"/>
</WrapHandler>
This sequence of handlers does the following:

1. The first wrap handler verifies that the user meets the security criteria defined in the handler specified by its
wrapPermKey attribute (ViewProfiler). If the user has an assigned role that contains any of the system
permissions specified by the ViewProfiler handler, the handler permits the user to have the EditProfiler
application permission. If the user does not have such a role, she receives an error message.
2. The second wrap handler checks to see that the user meets the security criteria defined in the handler specified
by its wrapPermKey attribute (EditProfiler). If the user has an assigned role that contains the
toolsProfilerwebserviceedit permission, the handler permits the user to have the
EditWebserviceProfiler application permission. If the user does not have such a role, she receives an error
message.
Wrap Handlers Specify AND Conditions

Wrap security handlers define Boolean AND conditions. Using the example shown previously, the sequence of
security handlers evaluates the following set of conditions:
(perm.System.internaltools OR perm.System.toolsProfilerview)
AND (perm.System.internaltools OR perm.System.toolsProfilerEdit)
AND (permission.System.toolsProfilerwebserivceedit)

For this compound condition to evaluate to true, all of the following conditions must be true:
• The user must have a role that contains either the interntools or toolsProfilerview system permission as
specified in ViewProfiler static handler.
• The user must have a role that contains either the interntools or toolsProfileredit system permission as
specified in the EditProfiler wrap handler.
• The user must have a role that contains the toolsProfilerwebservicesedit system permission as specified in
the EditWebServiceProfiler wrap handler.
Only if the user meets all sets of security criteria does the security handler permit the user to have the specified
application permission (EditwebserviceProfiler).
See also


chapter 13
Securing Access to Notes and

Documents
This topic explains how to use the PolicyCenter permission infrastructure to control access to document and note
objects.
See also

Note Permission Overview

In addition to using the standard system permissions for notes, you can control access to notes by configuring note
access permissions in security-config.xml. To use this feature, a note must have a note security type set. To see a
note of a particular type, a user must have both permission to view notes generally and the permission to access to
the note security type.
You set the security type for a note type by setting the note Security Type in PolicyCenter or through a Gosu class that
you write. In the base configuration, PolicyCenter provides the following note security types in the
NoteSecurityType typelist:
• Internal Only
• Sensitive
• Unrestricted
See also
• For information on the various security handler elements, see “The Security Configuration File” on page 187.
• For information on permissions, refer to the system permissions area of the PolicyCenter Security Dictionary.
• For information on typelists, refer to Guidewire Studio™ for PolicyCenter or the typelists area of the
PolicyCenter Data Dictionary.
Securing Access to Notes and Documents 193
Note Permission Elements

File security-config.xml must contain a <NoteAccessProfile> element for every note security type listed in the
NoteSecurityType typelist. If you add a new note security type to the typelist, then you must add a corresponding
<NoteAccessProfile> element to security-config.xml.
Thus, a <NotePermission> element in the security-config.xml file controls access to a note type.
The <NotePermission> element has the following syntax:
<NotePermissions>
<NoteAccessProfile securitylevel="level">
<NoteCreatePermission permission="perm"/>
<NoteDeletePermission permission="perm"/>
<NoteEditBodyPermission permission="perm"/>
<NoteEditPermission permission="perm"/>
<NoteViewPermission permission="perm"/>
</NoteAccessProfile>
</NotePermissions>

NoteAccessProfile securitylevel Yes A document security type defined in the NoteSecurityType typelist.
NoteCreatePermission permission Yes A system permission defined in the SystemPermissionType typelist.
NoteDeletePermission
NoteEditBodyPermission
NoteEditPermission
NoteViewPermission
The following code sample illustrates the security access levels for the Sensitive security access type.
<NotePermissions>
<NoteAccessProfile securitylevel="sensitive">
<NoteViewPermission permission="viewsensnote"/>
<NoteEditPermission permission="editsensnote"/>
<NoteDeletePermission permission="delsensnote"/>
</NoteAccessProfile>
</NotePermissions>
Note: PolicyCenter grants access permissions based on the roles assigned to a user only. It is not
possible to restrict Note access based on security zones or groups.
Document Permission Overview

In addition to using the standard system permissions for documents, you can control access to documents by
configuring document access permissions in security-config.xml. To use this feature, a document must have a
document security type set. To see a set of documents of a particular type, a user must have both permission to view
documents generally and access to the document security type.
You set a document type by using the document’s Security Type on the user interface or through a Gosu class that you
write. In the base configuration, PolicyCenter provides the following document security types in the
DocumentSecuritytype typelist:
• Internal Only
• Sensitive
• Unrestricted
194 chapter 13: Securing Access to Notes and Documents

See also
• For information on security handler elements, see “The Security Configuration File” on page 187.
• For information on permissions, refer to the system permissions area of the PolicyCenter Security Dictionary.
• For information on typelists, refer to Guidewire Studio™ for PolicyCenter or the typelists area of the
PolicyCenter Data Dictionary.
Document Permission Elements

File security-config.xml must contain a <DocumentAccessProfile> element for every document security type
listed in the DocumentSecurityType typelist. If you add a new document security type to the typelist, then you must
add a corresponding <DocumentAccessProfile> element to security-config.xml.
Thus, a <DocumentPermission> element in the security-config.xml file controls access to a document type.
This element has the following syntax:
<DocumentPermissions>
<DocumentAccessProfile securitylevel="level">
<DocumentCreatePermission permission="perm"/>
<DocumentDeletePermission permission="perm"/>
<DocumentEditPermission permission="perm"/>
<DocumentViewPermission permission="perm"/>
</DocumentAccessProfile>
</DocumentPermissions>

DocumentAccessProfile securitylevel Yes A document security type defined in the DocumentSecurityType
typelist.
DocuumentCreatePermission permission Yes A system permission defined in the SystemPermissionType type‐
DocumentDeletePermission list.
DocumentEditPermission
DocumentViewPermission
The following code sample illustrates the security access levels for the Unrestricted and Internal Only security
access types. Notice that unrestricted documents have no access controls set.
<DocumentPermissions>
<DocumentAccessProfile securitylevel="unrestricted"/>
<DocumentAccessProfile securitylevel="internalonly">
<DocumentViewPermission permission="viewintdoc"/>
<DocumentEditPermission permission="editintdoc"/>
<DocumentDeletePermission permission="delintdoc"/>
</DocumentAccessProfile>n="delsensdoc"/>
</DocumentAccessProfile>
</DocumentPermissions>
Note: PolicyCenter grants these permissions based on the user’s roles alone. You cannot restrict
document access based on security zones or groups.
Securing Access to Notes and Documents 195

196 chapter 13: Securing Access to Notes and Documents

chapter 14
Securing Access to Accounts, Jobs, and

Policy Periods
This topic explains how to use the PolicyCenter permission infrastructure to control access to PolicyCenter
accounts, jobs and policy periods.
See also

• “Note Permission Overview” on page 193
• “Document Permission Overview” on page 194
Account Producer Code Handler Elements

<AccountProducerCodeHandler> elements in security-config.xml determine account access by controlling the
account information and functionality to which a user has access. For example, an
<AccountProducerCodeHandler> element can define who has permission to create an account for a particular
produce code.
There is no limit to the number of <AccountProducerCodeHandler> elements that can exist in security-
config.xml. Each <AccountProducerCodeHandler> element can contain zero to many <SystemPermType> and
<ProducerStatus> elements.
<AccountProducerCodeHandler permKey="perm" desc="...">

<SystemPermType code="code" />
<ProducerStatus code="status" />
...
</AccountProducerCodeHandler>
Securing Access to Accounts, Jobs, and Policy Periods 197


quired
AccountProducerCodeHandler permkey Yes The application permission to grant. The permkey at‐
tribute becomes a property on the Account entity. To
access the entity property, use the following notation:
perm.Account.permkey

noPermissionDisplayKey No A display key that provides the text to show if the user
does not have a required permission.
SystemPermType code Yes A code value defined in the SystemPermissionType
typelist.
ProducerStatus code Yes The status of the producer associated with the policy.
This must be a value defined in the ProducerStatus
typelist.
PolicyCenter typically limits account visibility to users who have one of the producer codes associated with the
policies on the account. Producer status is an additional means of restricting access to account information. For
example, you can use a producer status of Suspended to limit the actions that user can take on that account or the
information that is visible on the account.
An <AccountProducerCodeHandler> element that does not contain a <ProducerStatus> element grants the
specified permission without restriction to producers with a status is Active. The following example illustrates this
concept.
<AccountProducerCodeHandler permKey="createForProducerCode"
desc="Permission to create account for a particular producer code">
<SystemPermType code="accountcreate"/>
As there is no <ProducerStatus> element in this example, the permission handler grants the ability to create a new
account to any producer with a status of Active.
The following example illustrates an <AccountProducerCodeHandler> element that contains multiple producer
status codes.
<AccountProducerCodeHandler permKey="view" desc="Permission to view an account">

<SystemPermType code="viewaccount"/>
<ProducerStatus code="Limited"/>
<ProducerStatus code="Suspended"/>
<ProducerStatus code="Terminating"/>
Notice that:
• The application permission key is view.
• The system permission type is viewaccount.
• The element explicitly grants permission to view an account to producers with a status of Limited, Suspended,
or Terminating. PolicyCenter gives any producer with a status of Active this permission automatically. Thus,
all producers can view accounts associated with their producer code except those producers with a status of
Terminated. As always, producers can only view accounts associated with their producer code.
See also

198 chapter 14: Securing Access to Accounts, Jobs, and Policy Periods
Job Producer Code Handler Elements

<JobProducerCodeHandler> elements in security-config.xml determine the access a producer has to the
different job types in PolicyCenter. For example, an <JobProducerCodeHandler> element can define who has
permission to create a submission job.
There is no limit to the number of <JobProducerCodeHandler> elements that can exist in security-config.xml.
Each <JobProducerCodeHandler> element can contain zero to many <SystemPermType> and <ProducerStatus>
elements.
<JobProducerCodeHandler jobType = "type" permKey="perm" desc="..." noPermissionDisplayKey="...">

...
</JobProducerCodeHandler>

quired
JobProducerCodeHandler jobType Yes The type of Job entity to which the permission applies, for
example, a Submission or Renewal job.
permkey Yes The application permission to grant. The permkey attribute
becomes a property on the specified job type (jobType). To
access the entity property, use the following notation:
perm.jobType.permkey

noPermissionDisplayKey No A display key that provides the text to show if the user does
not have a required permission.
SystemPermType code Yes A code value defined in the SystemPermissionType type‐
list.
ProducerStatus code Yes The status of the producer associated with the policy. This
must be a value defined in the ProducerStatus typelist.
The following example illustrates a JobProducerCodeHandler element.
<JobProducerCodeHandler jobType="Reinstatement" permKey="view"

desc="Permission to view a reinstatement">
<SystemPermType code="viewreinstate"/>
Notice that:
• The value of jobType is Reinstatement.
• The system permission type is viewreinstate.
• The element explicitly grants permission to view the details of a policy reinstatement job to producers with a
status of Limited. PolicyCenter gives any producer with a status of Active this permission automatically. Thus,
only producers with a status of Active or Limited can view the details of reinstatement submission job.
See also
PolicyPeriod Producer Code Handler Elements

<PolicyPeriodProducerCodeHandler> elements in security-config.xml determine access to policy information
and restrict the actions available on a policy. There is no limit to the number of
<PolicyPeriodProducerCodeHandler> elements that can exist in security-config.xml. Each
<PolicyPeriodProducerCodeHandler> element can contain zero to many <SystemPermType> and
<ProducerStatus> elements.
<PolicyPeriodProducerCodeHandler permKey="perm" desc="..."

isAllowedForPCOfRecord="true|false" noPermissionDisplayKey="...">
...
</PolicyPeriodProducerCodeHandler>

quired
PolicyPeriodProducerCodeHandler permkey Yes The application permission to grant. The permkey
attribute becomes a property on the
PolicyPeriod entity. To access the entity proper‐
ty, use the following notation:
perm.PolicyPeriod.permkey

isAllowedForPCOfRecord No A Boolean value which, if true, grants this per‐
mission to a produce whose producer code
matches that of the policy. The default is false.
In the base configuration, Guidewire sets this at‐
tribute value to true for the viewpolicyfile
permission only. Thus, only a producer with the
correct producer code can see the information in
the policy file.
noPermissionDisplayKey No A display key that provides the text to show if the
user does not have a required permission.
SystemPermType code Yes A code value defined in the
SystemPermissionType typelist.
ProducerStatus code Yes The status of the producer associated with the
policy. This must be a value defined in the
ProducerStatus typelist.
The following example illustrates an <PolicyPeriodProducerCodeHandler> element that contains multiple

producer status codes.
<PolicyPeriodProducerCodeHandler permKey="view" desc="Permission to view the policy file"

isAllowedForPCOfRecord="true">
<SystemPermType code="viewpolicyfile"/>
<ProducerStatus code="Suspended"/>
<ProducerStatus code="Terminating"/>
</PolicyPeriodProducerCodeHandler>
Notice that:
• The value of isAllowedForPCOfRecord is true. Thus, only a producer whose producer code matches the
producer code on the policy can see the policy file.
• The system permission type is viewpolicyfiles.
• The element explicitly grants permission to view a policy file to producers with a status of Limited.
PolicyCenter gives any producer with a status of Active this permission automatically. Thus, only producers
with a status of Active or Limited can view the details of policy file. As indicated by the
isAllowedForPCOfRecord element a producer can only view a policy file associated with its producer code.
See also

part 5
Database Administration
chapter 15
Database Configuration
This topic discusses database configuration file database-config.xml and how to use the file to configure
PolicyCenter database options.
See also
• “Database Best Practices” on page 237
• “Guidewire Database Direct Update Policy” on page 239
• “PolicyCenter Database Back Up” on page 240
• “Database Consistency Checks” on page 240
• “Resize Database Columns” on page 244
• “Purging Unwanted Data” on page 245
Accessing the Database Configuration File

You access file database-config.xml from the Guidewire Studio Project window, in the following location:
configuration→config
You can view, but not edit, many of the database configuration parameters from within PolicyCenter at Server
Tools→Info Pages→Database Parameters.
See also
• “The Database Configuration File” on page 205
• “Database Parameters” on page 339
The Database Configuration File

Guidewire provides the means to configure various aspects of the PolicyCenter database through configuration file
database-config.xml. This XML-formatted file contains a root <database> element with a number of
subelements that you use to implement database configuration options specific to your database type.
The <database> element has the following syntax. The following code sample shows required attributes in bold
font.
<database addforeignkeys="true|false" autoupgrade="full|manual" checker="true|false" dbtype="h2|oracle|sqlserver"

env="string"
Database Configuration 205

name="string" printcommands="true|false" versionchecksonly="true|false">

<databasestatistics databasedegree="integer" incrementalupdatethresholdpercent="integer"
numappserverthreads="integer" samplingpercentage="integer" useoraclestatspreferences="true|false">

<tablestatistics action="delete|keep|update" databasedegree="integer" name="string"
samplingpercentage="integer">

<histogramstatistics name="string" numbbuckets="integer"/>
</tablestatistics>
</databasestatistics>

<dbcp-connection-pool jdbc-url="string" max-idle="integer" max-total="integer" max-wait="integer"
min-evictable-idle-time="integer" num-tests-per-eviction-run="integer" password-file="string"
test-on-borrow="true|false" test-on-return="true|false" test-while-idle="true|false"
time-between-eviction-runs="integer" when-exhausted-action="block|fail|grow">
<reset-tools-params collation="string" oracle-tnsnames="string" system-password="string"
system-username="string"/>
</dbcp-connection-pool>

<jndi-connection-pool datasource-name="string"/>


<oracle-settings adaptive-optimization="OFF|REPORTING_ONLY" db-resource-mgr-cancel-sql="string"
query-rewrite="true|false" statistics-level-all="true|false"
stored-outline-category="string"/>


<sqlserver-settings jdbc-trace-file="string" jdbc-trace-level="string" unicodecolumns="true|false"/>


<upgrade>
<mssql-db-ddl>

<mssql-compression index-compression="NONE|PAGE|ROW" table-compression="NONE|PAGE|ROW/>
<mssql-filegroups admin="string" index="string" lob="string" op="string" staging="string"
typelist="string"/>

<mssql-table-ddl table-name="string">
<mssql-index-ddl filter-where="string"index-compression="NONE|PAGE|ROW"
index-filegroup="string" key-columns="string" partition-scheme="string"/>
<mssql-table-compression index-compression="NONE|PAGE|ROW" table-compression="NONE|PAGE|ROW"/>
<mssql-table-filegroups="string" index-filegroup="string" lob-filegroup
table-filegroup="string"/>
</mssql-table-ddl>
</mssql-db-ddl>
<ora-db-ddl>

<ora-compression index-compression="true|false" table-compression="ADVANCED|BASIC|NONE"/>
<ora-lobs caching="true|false" type="BASIC|SECURE|SECURE_COMPRESSED/>
<tablespaces admin="string" index="string" lob="string" op="string" staging="string"
typelist="string"/>

<ora-table-ddl table-name="string">
<ora-index-ddl index-compression="true|false" index-tablespace="string" key-columns="string"/>
<ora-table-compression index-compression="true|false" table-compression="ADVANCED|BASIC|NONE">
<ora-table-date-interval-partitioning datecolumn="string"
interval="DAILY|MONTHLY|QUARTERLY|WEEKLY|YEARLY">
<ora-table-hash-partitioning hash-columns="string" num-partitions="integer"/>
206 chapter 15: Database Configuration

<ora-table-tablespaces index-tablespace="string" lob-tablespace="string"

table-tablespace="string"/>
</ora-table-ddl>
</ora-db-ddl>
<versiontriggers dbmsperfinfothreshold="integer">

<versiontrigger extendedquerytracingenabled="true|false" name="string"
parallel-dml="true|false" parallel-query="true|false"
queryoptimizertracingenabled="true|false" recordcounters="true|false"
updatejoinorderedhint="true|false" updatejoinusemergehint="true|false"
updatejoinusenlhint="true|false"/>
</versiontriggers>
</upgrade>
</database>
File database-config.xml contains a single root-level <database> element that takes the following attributes.
name Required. String identifying the database for which PolicyCenter uses this connection specification.
dbtype Required. Database type, either h2 (for the QuickStart database), oracle, or sqlserver.
The following attributes are all optional.
addforeignkey Used only for development and testing. Do not use this attribute in production. The default is true.
autoupgrade Use to set how to upgrade the database. Valid values are:
• full – Takes precedence and initiates a full upgrade assuming all other necessary conditions are
met.
• manual – Requires that you set either the database upgrade type (in Server Tools Upgrade and
Versions screen) or the date system property.
checker Boolean. Whether PolicyCenter runs consistency checks before it starts:
• Development environments – For development environments with small data sets, you can enable
consistency checks to run each time the PolicyCenter server starts. Set the value of checker in the
database block to true to enable checks on server startup.
• Production environments – Running consistency checks upon server startup can take a long time,
impact performance severely, and possibly time out on very large datasets. Set the value of checker
in the database block to false to disable checks on server startup.
Valid values are:
• true – Guidewire recommends that you only set checker to true in development environments
with a small set of test data.
• false – Guidewire recommends that you set checker to false under most circumstances.
The default is true.
See the following for more information:
• “Configure Consistency Checks to Run at Server Start” on page 242
env Use of the env attribute to set a server environment enables you to provide different database configu‐
rations for different server environments. For example, you can set up different database configurations
for a production environment and a test environment.
See “Example Syntax for Registry Server Element” on page 44 for more information.
printcommands Boolean. Whether the server prints database upgrade messages to the console upon startup. Valid val‐
ues are:
• true ‐ By default, Guidewire sets the value of printcommands to true in the base configuration.
• false ‐ Do not set printcommands to false in a production environment.

versionchecksonly Boolean. Whether the PolicyCenter server runs only database version checks at startup, without per‐
forming any actual database upgrade steps:
• true ‐ PolicyCenter runs all version checks regardless of a failure in one of the checks.
• false ‐ PolicyCenter stops the upgrade if it detects an error.
The default is false. Changes to this attribute take effect only during an application upgrade.
The <database> element takes the following subelements. There is, at most, a single occurrence of each of these
subelements in the <database> element.
databasestatistics Specifies parameters that control the generation of database statistics. See “The databasestatistics
Database Configuration Element” on page 209 and “Database Statistics Generation” on page 255
for more information.
dbcp-connection-pool Specifies parameters for connection pool shared using dbcp. You must include this subelement if us‐
ing a dbcp data source. See “The dbcp‐connection‐pool Database Configuration Element” on page
210 and the Installation Guide for more information.
jndi-connection-pool Specifies parameters for a connection pool shared using JNDI. You must include this subelement if
using a jndi data source. See “The jndi‐connection‐pool Database Configuration Element” on page
213 and the Installation Guide for more information.
oracle-settings Specifies settings for Oracle databases. See “The oracle‐settings Database Configuration Element” on
page 216 and the Installation Guide for more information.
sqlserver-settings Specifies settings for SQL Server databases. See “The sqlserver‐settings Database Configuration Ele‐
ment” on page 217 and the Installation Guide for more information.
upgrade Specifies PolicyCenter behavior during a database upgrade. See “The upgrade Database Configura‐
tion Element” on page 217 for more information.
See also
The Database autoupgrade Attribute

You use the autoupgrade attribute on the <database> element in file database-config.xml to manage
PolicyCenter configuration and database upgrades. The autoupgrade attribute has the following syntax:
<database ... autoupgrade="full|manual" .../>
The attribute value has the following meaning. The default value is manual if the attribute is missing.
autoupgrade Description
attribute
full Whenever the application server starts, if it determines the need for a full upgrade, the presence of this
attribute set to full is sufficient permission to perform the upgrade. With this setting:
• If a rolling (configuration) upgrade is already in progress as the server starts, the server throws an
exception, to force the choice of an upgrade type.
• If a full upgrade is already in progress by other means as the server starts, there is no issue as this setting
is consistent with the a full upgrade.
manual This setting requires that you explicitly set the permission to upgrade through one of the following means:
• Setting the database upgrade type in the Server Tools Upgrade and Versions screen and initiating the
upgrade from that screen.
• Setting the following Java system property to the current date as the application server starts:
-Dgw.pc.full.upgrade.intended.date
See “Unexpected Upgrades” on page 175 for a discussion of the use of this Java system property.
You must set the value of autoupgrade to manual if performing a rolling (configuration) upgrade.

The behavior of this attribute depends on the following:

• Whether you are working with a single PolicyCenter server or with multiple PolicyCenter servers in a clustered
server environment.
• Whether you are attempting a full PolicyCenter database upgrade or using a rolling upgrade to implement
application configuration changes.
Single PolicyCenter Server Installations

The following tables describe the interactions between setting Start Full Upgrade on the Upgrade and Versions screen and
the value of the autoupgrade attribute.
Start Full Upgrade autoupgrade Result

Set Not set or full The server starts, the upgrade completes, and PolicyCenter updates the Upgrade and
Versions screen.
Not set Not set or manual The server does not start, the upgrade fails, and PolicyCenter logs an error message.
The following table describes the interactions between setting Start Full Upgrade on the Upgrade and Versions screen and
the value of the autoupgrade attribute during deployment of non-data model changes to a production mode server.
Start Full Upgrade autoupgrade Result

Set Not set or full The server starts, the upgrade completes, and PolicyCenter updates the Upgrade and
Versions screen.
Not set Not set or manual The PolicyCenter server starts.
Clustered PolicyCenter Server Installations

In clustered PolicyCenter server installations, the autoupgrade attribute has the following behavior.
autoupgrade Result
full If you set the value of autoupgrade attribute to full, any attempt to start a rolling upgrade fails. In addition,
you must always set the upgrade type (full) using the Server Tools Upgrade and Versions screen, or, by using the
system_tools -startfullupgrade command option, for example.
manual If you set the value of autoupgrade attribute to manual, you must always set the upgrade type (full or rolling) or
the upgrade fails. You can set the upgrade type in the Server Tools Upgrade and Versions screen or by setting a
JVM parameter at server startup. For a rolling upgrade, the new configuration (database-config.xml) must set
this value to manual. This new value overrides any value set in the old configuration.
Not set If you do not set the value of autoupgrade attribute, PolicyCenter assumes a default value of manual and
behaves accordingly.
The databasestatistics Database Configuration Element

The <database> element in file database-config.xml contains a single occurrence of subelement
<databasestatistics>. Use this element to control how PolicyCenter generates database statistics statements.
The <databasestatistics> element has the following syntax. The following code sample shows required
attributes in bold font.
<database>
<databasestatistics databasedegree="integer" incrementalupdatethresholdpercent="integer"

numappserverthreads="integer" samplingpercentage="integer" useoraclestatspreferences="true|false">

<tablestatistics action="delete|keep|update|force" databasedegree="integer" name="string"

samplingpercentage="integer">

<histogramstatistics name="string" numbbuckets="integer" />
</tablestatistics>
</database>
The following list describes the attributes that you can configure on the <databasestatistics> element. All of
these attributes are optional. See “The Database Statistics Element” on page 261 for more information on these
attributes.
databasedegree On Oracle, this attribute controls the degree of database parallelism that Oracle uses
in executing each individual statement. The default is 1. PolicyCenter uses the value
of this attribute for all statements.
SQL Server ignores the databasedegree attribute.
incrementalupdatethresholdpercent This attribute specifies the percentage of table data that must have changed since
the last statistics process for the incremental statistics generation batch process to
update statistics for the table.
numappserverthreads On both Oracle and SQL Server, the numappserverthreads attribute controls the
number of threads that PolicyCenter uses to update database statistics for staging ta‐
bles during import only.
samplingpercentage The behavior of this attribute depends on the database type. For Oracle, Guidewire
recommends that you always set this value to 0 to enable Oracle auto‐sampling.
useoraclestatspreferences On Oracle, this attribute sets the database statistics preferences to be able to use the
Oracle Autotask infrastructure instead of the DBStats batch process from
PolicyCenter. The default is false, which requires that you disable the Autotask and
schedule DBStats batch processing in its place. Changes to the value of this attribute
only take effect during an application upgrade.
The <databasestatistics> element has the following subelement.
tablestatistics Provides overrides of database‐wide statistics settings defined on the <databasestatistics> element for
a specific table. There can be multiple occurrences of the <tablestatistics> subelement on the
<databasestistics> element.
See also
• “Configuring Database Statistics Generation” on page 260
• “The Database Statistics Element” on page 261
• “The Table Statistics Database Element” on page 264
• “Using Oracle AutoTask for Statistics Generation” on page 262
The dbcp‐connection‐pool Database Configuration Element

The <database> element in file database-config.xml contains, at most, a single occurrence of subelement <dbcp-
connection-pool>. The use of the <dbcp-connection-pool> element is optional. If using PolicyCenter to manage
the connection pool rather than a JNDI data source managed by the application server, you must use this element to
configure the connection pool behavior.
Note: Guidewire implements its own version of the Apache Commons Pool, overriding specific
values to provide improved functionality.
If you experience slow performance, it is possible that the PolicyCenter server is not allocating enough database
connections. If all database connections are in use, any client attempting to connect to the server must wait until a
connection is free. By default, PolicyCenter periodically tests connections in the connection pool using a simple
query and evicts idle connections and those that fail with an exception.
The <dbcp-connection-pool> element has the following syntax. The following code sample shows required
<database>

<dbcp-connection-pool jdbc-url="string" max-idle="integer" max-total="integer" max-wait="integer"
min-evictable-idle-time="integer" num-tests-per-eviction-run="integer" password-file="string"
test-on-borrow="true|false" test-on-return="true|false" test-while-idle="true|false"
time-between-eviction-runs="integer" when-exhausted-action="block|fail|grow">
</database>
The following list describes the attributes that you can configure on the <dbcp-connection-pool> element.
Note: These attributes apply only if you use the default connection pool. If you use the server
connection pool, these settings do not apply. Configure the server connection pool instead through the
administration console provided with the server. See the Installation Guide for more information.
jdbc-url Required. Stores connection information for the database. The format of the jdbc-url value changes de‐
pending on the database type. See the Installation Guide for more information.
max-idle Maximum number of connections that can sit idle in the pool at any time. If negative, there is no limit to the
number of connections that can be idle at any given time. The default is ‐1.
max-total Maximum number of connections that the connection pool can allocate, including those in use by a client or
that are in an idle state awaiting use. A reasonable initial value for this is about 25% of the number of users
that you expect to use PolicyCenter at the same time.
If set to a negative integer, there is no limit to the number of allowed database connections. The default is ‐1.
If the number of database connections reaches the value of max-total, PolicyCenter considers the connec‐
tion pool to have no more available connections.
max-wait Maximum amount of time, in milliseconds, that the data source waits for a connection before one becomes
available in the pool to service. The default is 30000.
The value of the max-wait attribute interacts with the when-exhausted-action attribute. See that attribute
for more information.
min- Maximum time, in milliseconds, that a connection can sit idle in the pool before it is eligible for eviction due
evictable- to idle time. If a connection is idle more the specified number of milliseconds, PolicyCenter evicts the connec‐
idle-time tion from the pool. The default is 300000 milliseconds.
If this value is a non‐positive integer, PolicyCenter does not drop connections from the pool due to idle time
alone. This setting has no effect unless the value of time-between-eviction-runs is greater than 0.
During an eviction run, PolicyCenter scans the connection pool and tests the number of idle connections
equal to the value of num-tests-per-eviction-run.
num-tests- Number of idle connections that PolicyCenter tests in each eviction run. This setting has no effect unless the
per- value of time-between-eviction-runs is greater than 0. The default is 3.
eviction-run
password- Use to hide the value of the database connection password in the jdbc-url connection string. Instead of
file providing the password in the connection string, you can place the password in an external file and reference
this file from file database-config.xml. See the Installation Guide for more information.

test-on- Boolean. Whether PolicyCenter tests a connection by running a simple validation query as PolicyCenter first
borrow borrows the connection from the connection pool. If set to true, the connection pool attempts to validate
each connection before PolicyCenter uses the connection from the connection pool. If a connection fails vali‐
dation, the connection pool drops the connection and chooses a different connection to borrow. The default
is false.
PolicyCenter returns any connection used only for a query to the pool immediately after the query com‐
pletes. Thus, running a test query every time that a connection returns to the pool can potentially affect per‐
formance.
test-on- Boolean. Whether PolicyCenter tests a connection by running a simple validation query as PolicyCenter re‐
return turns the connection to the connection pool. If set to true, the connection pool attempts to validate each
connection that PolicyCenter returns from the database. The default is false.
PolicyCenter returns any connection used only for a query to the pool immediately after the query com‐
pletes. Thus, running a test query every time that a connection returns to the pool can potentially affect per‐
formance.
test-while- Boolean. Whether PolicyCenter performs validation on idle connections in the connection pool. If set to true,
idle the connection pool performs validation on idle connections. It drops connections that fail the validation test.
This attribute value has no effect unless the value of time-between-eviction-runs is greater than zero.
time- Time, in milliseconds, that PolicyCenter waits between eviction runs of idle connections in the connection
between- pool. The default is 60000.
eviction- If set to a a non‐positive integer, PolicyCenter does not launch any eviction threads.
runs
when- Specifies the behavior of the connection pool if the pool has no more connections. Set this attribute to one of
exhausted- the following values:
action • fail – If the there are no more connections available, PolicyCenter throws a NoSuchElementException
exception.
• grow – If there are no more connections available, PolicyCenter creates a new connection and returns it,
essentially making max-active meaningless.
• block – If there are no more connections available, PolicyCenter blocks connections until a new or idle
connection becomes available. If the value of max-wait is positive, then PolicyCenter blocks, at most, for
that number of milliseconds, after which PolicyCenter throws a NoSuchElementException exception. If
the value of max-wait is non‐positive, PolicyCenter blocks indefinitely.
The default is block.
The <dbcp-connection-pool> element has the following subelement.
reset-tools-params See “The reset‐tool‐params Database Configuration Element” on page 212 for more information.
See also
• “Database Parameters” on page 339
The reset‐tool‐params Database Configuration Element

The <dbcp-connection-pool> element in file database-config.xml contains, at most, a single occurrence of
subelement <reset-tool-params>. The use of the <reset-tool-params> element is optional. Use this element to
configure DBResetTool, not the DBCP connection pool. Guidewire defines the <reset-tool-params> element as a
subelement of the <dbcp-connection-pool> element because DBResetTool only works on databases defined in the
<dbcp-connection-pool> element.
The <reset-tool-params> element has the following syntax.
<database>
<dbcp-connection-pool>

</database>
The following list describes the attributes that you can configure on the <reset-tools-params> element. All of
these attributes are optional.
collation Collation value to use if creating a new H2 (QuickStart) or SQL Server database:
• H2 – Sets database collation using the Java Collation class.
• SQL Server – Sets database collation to a Microsoft Window’s or SQL collation name.
DBResetTool (dropdb) uses the value of this attribute if creating a new H2 or SQL Server database.
oracle-tnsnames (Oracle) Name of the Oracle tnsnames.ora file.
system-password (Oracle) Database system password.
system-username (Oracle) Database system username.
The <reset-tools-arams> element has no subelements.
See also
The jndi‐connection‐pool Database Configuration Element

The <database> element in file database-config.xml contains, at most, a single occurrence of subelement <jndi-
connection-pool>. The use of the <jndi-connection-pool> element is optional. However, file database-
config.xml must contain this element if you use a JNDI (Java Naming and Directory Interface) data source
managed by a JBoss, Tomcat, WebLogic, or WebSphere application server.
The <jndi-connection-pool> element has the following syntax. The following code sample shows required
<database>

<jndi-connection-pool datasource-name="string"
connections-initialized-for-application="true|false"/>
</database>
IMPORTANT If you modify the <jndi-connection-pool> element in any way, you must restart the
application server.
The following list describes the attributes that you can configure on the <jndi-connection-pool> element.
datasource-name Required. Specifies the JNDI name to assign to the data source. See the
Installation Guide for more information.
The following attribute is optional.
connections-initialized-for-application (Oracle) Boolean. Controls the number of SQL statements that PolicyCenter
executes on every connection that it borrows from an external data source.
This setting applies to the named JNDI database connection set up in this
<jndi-connection-pool> element only.
Valid values are:
• true – If configured appropriately, the data source provides
connections with certain Oracle database parameters set to their
desired values.
• false – PolicyCenter runs a set of SQL statements on each and every
database connection that it borrows from the data source.

To take advantage of this feature:
• Your PolicyCenter installation must use an Oracle database.
• You must configure the data source appropriately.
See “Configuring JNDI Connection Initialization for Oracle” on page 214 for
more information.
IMPORTANT If you set this attribute to true and do not initialize the con‐
nections properly, the application server refuses to start and logs an error
message for each incorrect setting.
The <jndi-connection-pool> element has no subelements.
See also

• “Configuring JNDI Connection Initialization for Oracle” on page 214
Configuring JNDI Connection Initialization for Oracle

Boolean attribute connections-initialized-for-application on the <jndi-connection-pool> element in
database-config.xml configures how the JNDI data source provides connections to PolicyCenter.
Attribute connections‐initialized‐for‐application Set to True
If you set the value of attribute connections-initialized-for-application to true, you must provide the
correct settings for the Oracle database parameters that the external data sources uses for connection initialization. If
you do not initialize the connections properly, the application server refuses to start and logs an error message for
each incorrect setting. Guidewire recommends that you set this attribute to true, then start the application server.
You can then determine the correct values to use for your database from the log error messages.
In general, the Oracle database parameters to use for connection initialization have the following meanings:
Oracle parameter Meaning

CURSOR_SHARING Boolean. The exact setting is dependant on your database installation.
MODULE The module name, which must include the database user name as well. For example, if your database user
name is PCPROD1, then set the module name to PolicyCenter_PCPROD1.
NLS_SORT The default sort collation for the Oracle database. The exact setting is dependent on your Guidewire instal‐
lation. See the Globalization Guide for more information.
NLS_COMP The database collation behavior. Set this value to BINARY.
To use this feature, you must configure the application server to manage the connection initialization. See “Set
Oracle Database Parameters for Connection Initialization” on page 215 for more information.
Attribute connections‐initialized‐for‐application Set to False
If you set the value of attribute connections-initialized-for-application to false, the external data source
takes no action to configure the borrowed connections before placing an item in the connection pool. This causes
PolicyCenter to run a set of SQL statements on each and every connection that it borrows from the data source.
Although the cost in time of each of these statements is very small, Guidewire applications sometimes borrow
connections at a very high rate. Thus, it is possible for the time to execute these statements to become measurable
and to become visible during performance analysis in the Guidewire Profiler.

Set Oracle Database Parameters for Connection Initialization

About this task
database-config.xml configures how the JNDI data source provides connections to PolicyCenter. To use this
feature with an external data source, you must configure your data source to initialize a set of Oracle database
parameters.
Procedure
1. Set attribute connections-initialized-for-application on the <jndi-connection-pool> element to
true.
2. Start the application server.
3. From the server log, determine the exact values to set for connection initialization.
4. Depending on your application server, set up the connection initialization as appropriate.
See “Connection Initialization for Oracle Databases” on page 215 for more information.
5. After completing your initialization configuration, restart the application server and Guidewire PolicyCenter.
Connection Initialization for Oracle Databases

If using an Oracle database, it is possible to configure an application server to initialize a data pool connection
before PolicyCenter uses that connection. This can improve performance. To use this functionality, you must set
database-config.xml to true.
WebLogic Application Servers

Use the WebLogic Administration Console to specify a SQL statement that initializes each data source connection
before WebLogic adds that connection to the pool.
If you have more than one statement to execute, you can put the statements into a SQL block. For example:
SQL BEGIN
DBMS_APPLICATION_INFO.SET_MODULE('PolicyCenter_PCPROD1', NULL);
EXECUTE IMMEDIATE 'ALTER SESSION SET NLS_SORT = BINARY_CI';
END;
Note: In the sample code, replace PolicyCenter_PCPROD1 with the actual name of the Oracle
database followed by the logon user username.
See also
Non‐WebLogic Application Servers

One way to initialize the connections appropriately is to create an Oracle logon trigger for the owner of the database
schema. The following sample code illustrates how to create a logon trigger. You must be logged on as the system
user to execute this trigger.
CREATE OR REPLACE TRIGGER

AFTER LOGON ON DATABASE
BEGIN
IF (USER = 'Guidewire schema owner')

THEN
DBMS_APPLICATION_INFO.SET_MODULE('PolicyCenter_PCPROD1', NULL);
EXECUTE IMMEDIATE 'ALTER SESSION SET NLS_SORT = BINARY_CI';
END IF;

END;
/
Note: In the sample code, replace Guidewire schema owner with Oracle database system username.
Replace PolicyCenter_PCPROD1 with the actual name of the Oracle database followed by the logon
user username.
See also
The oracle‐settings Database Configuration Element

The <database> element in file database-config.xml contains, at most, a single occurrence of subelement
<oracle-settings>. The use of the <oracle-settings> element is optional. Use this element to configure Oracle-
only database parameters.
The <oracle-settings> element has the following syntax.
<database>
<oracle-settings adaptive-optimization="OFF|REPORTING_ONLY" db-resource-mgr-cancel-sql="string"

query-rewrite="true|false" statistics-level-all="true|false"
stored-outline-category="string"/>
</database>
The following list describes the attributes that you can configure on the <oracle-settings> element. All of these
attributes are optional.
adaptive- Specifies the behavior of the Oracle Adaptive Optimization feature. Valid values are:
optimization • OFF
• REPORTING_ONLY
db-resource-mgr- Name of an Oracle Resource Consumer Group, if any.
cancel-sql
query-rewrite Boolean. Whether to enable query rewrite.

Valid values are:
• true – Enable query rewrite and use a matching materialized view.
• false – Set the Oracle query-rewrite parameter to false to disable use of Oracle materialized
views.
If not present, Guidewire does not set this value at the session level
statistics-level- Boolean. Whether to set the Oracle statistics_level parameter to ALL and enable collection of de‐
all tailed execution plan statistics.
stored-outline- Name of the stored outline category to use, if any
category
The <oracle-settings> element does not contain additional subelements.
See also

The sqlserver‐settings Database Configuration Element

<sqlserver-settings>. The use of the <sqlserver-settings> element is optional. Use this element to configure
Microsoft JDBC driver logging.
The <sqlserver-settings> element has the following syntax.
<database>
...
<sqlserver-settings jdbc-trace-file="string" jdbc-trace-level="string"
unicodecolumns="true|false"/>
...
</database>
The following list describes the attributes that you can configure on the <sqlserver-settings> element. All of
jdbc-trace-file Specifies the name of the trace file. If you do not provide a file name, this value defaults to the following:
C:\temp\msjdbctrace%u.log
PolicyCenter replaces the symbols in the file name at runtime with their meaning as listed at the follow‐
ing web site.
http://java.sun.com/j2se/1.5.0/docs/api/java/util/logging/FileHandler.html
Use the listed symbols to uniquely name the trace file.
jdbc-trace-level Valid trace level as listed at the following web site:
http://msdn.microsoft.com/en-us/library/ms378517(SQL.90).aspx?ppud=4
unicodecolumns Required if starting a new database that exclusively uses Unicode‐capable column character data types
(nvarchar, …). PolicyCenter ignores this attribute if the database does not support Unicode or if the at‐
tribute is not relevant to the new database. The default is false.
The <sqlserver-settings> element does not contain additional subelements.
See also
The upgrade Database Configuration Element

<upgrade>. The use of the <upgrade> element is optional. This element specifies various options related to database
upgrade. One important area of configuration is the degree of database parallelism to use in an Oracle database.
Database parallelism refers to the ability of an Oracle database to execute a database SQL statement such as CREATE
or INSERT using simultaneous, parallel slave processes.
The <upgrade> element has the following syntax. The following code sample shows required attributes in bold font.
<database>
<upgrade>
<mssql-db-ddl>

<mssql-filegroups admin="string" index="string" lob="string" op="string" staging="string"
typelist="string"/>



</mssql-table-ddl>
</mssql-db-ddl>
<ora-db-ddl>
typelist="string"/>
<ora-lobs caching="true|false" type="BASIC|SECURE|SECURE_COMPRESSED/>>
</ora-table-ddl>
</ora-db-ddl>

</versiontriggers>
</upgrade>
</database>
The following list describes the attributes that you can configure on the <upgrade> element. All of these attributes
are optional.
allowUnloggedOperations Boolean. Whether to disable logging of certain SQL operations during the data‐
base upgrade.
Valid values are:
• true – Run the upgrade with minimal database redo logging and enable
direct‐path INSERT operations.
• false – Run the upgrade with standard database redo logging.
Note: If you run the upgrade with attribute allowUnloggedOperations set to
true, then you need to take a full database backup after the upgrade.
collectstorageinstrumentation (Oracle) Boolean. Whether PolicyCenter collects tablespace usage and object size
data before and after the upgrade.
Valid values are:
• true – PolicyCenter collects tablespace usage and size of segments such as
tables, indexes and LOBs (large object binaries) before and after the upgrade.
You can then compare the before and after values to find the utilization
change caused by the upgrade.
• false – PolicyCenter does not collect this data.

defer-create-nonessential-indexes Boolean. Whether to defer creation of non‐essential indexes during the upgrade
process until the upgrade completes and the application server is back up. Crea‐
tion of non‐essential indexes can add significant time to the upgrade duration.
Valid values are:
• true – Defer creation of non‐essential indexes during upgrade.
• false – Do not defer creation of non‐essential indexes during upgrade.
Non‐essential indexes are:
• Performance‐related indexes that do not enforce constraints.
• Indexes on the ArchivePartition column on all entities that PolicyCenter
can archive.
If you choose to defer creation of non‐essential indexes, PolicyCenter runs the
Deferred Upgrade Tasks batch process (DeferredUpgradeTasks) as soon as the
upgrade completes and the server starts up. See “Deferred Upgrade Tasks Batch
Process” on page 106 for more information.
deferDropColumns (Oracle) Boolean. Whether to drop table columns removed during upgrade im‐
mediately or leave their removal to a later time. The database upgrade removes
some columns. For Oracle, you can configure whether the removed columns are
dropped immediately or are marked as unused. Marking a column as unused is a
faster operation than dropping the column immediately.
However, as PolicyCenter does not physically drop the removed columns from
the database, the space used by these columns is not released immediately to
the table and index segments.
Valid values are:
• true – Defer dropping removed columns until after the upgrade, possibly
during off‐peak hours of operation. The PolicyCenter database upgrade marks
the removed columns as unused instead.
• false – Drop the removed columns immediately, during the upgrade process.
degree-of-parallelism (Oracle) Controls the degree of database parallelism that Oracle uses for INSERT,
UPDATE, and DELETE database operations.
Valid values are:
• 0 – Defers to Oracle to determine the degree of database parallelism for the
operations that the attribute configures. The Oracle automatic parallel tuning
feature determines the degree based on the number of CPU processors
involved and the value set for the Oracle parameter
PARALLEL_THREADS_PER_CPU.
• 1 – Disables the parallel execution of DDL statements.
• Positive integer less than 1000 – Database parallelism, with the specified
value as the degree of parallelism.
The default is 4.
degree-parallel-ddl (Oracle) Controls the degree of database parallelism that Oracle uses to execute
DDL (Data Definition Language) statements during the database upgrade. Use to
configure the degree of database parallelism for commands such as CREATE
INDEX and the ALTER TABLE commands.
Valid values are:
• 0 – Defers to Oracle to determine to determine the degree of database
parallelism for the operations that the attribute configures. The Oracle
automatic parallel tuning feature determines the degree based on the
number of CPUs involved and the value set for the Oracle parameter
PARALLEL_THREADS_PER_CPU.
• 1 – Disables the parallel execution of DDL statements.
• Positive integer less than 1000 – Database parallelism, with the specified
value as the degree of parallelism.
The default is 4.
If you set the value of ora-parallel-dml to enable or enable_all (default),
then you need to provide a value for attribute degree-of-parallelism as well.

encryptioncommitsize Sets the commit size for rows requiring encryption. If one or more attributes use
PolicyCenter encryption, the PolicyCenter database upgrade commits batches of
encrypted values. The upgrade commits encryptioncommitsize rows at a time
in each batch.
The default value of encryptioncommitsize varies based on the database type:
• Oracle – 10000
• SQL Server – 100
Test the upgrade on a copy of your production database before attempting to up‐
grade the actual production database. If the encryption process is slow, and you
cannot attribute the slowness to SQL statements in the database, try adjusting
the encryptioncommitsize attribute. After you optimized the performance of
the encryption process, use that value of encryptioncommitsize as you upgrade
your production database.
ora-parallel-dml (Oracle) Controls database parallelism usage by Oracle in the execution of DML
(Data Manipulation Language) operations.
Valid values are:
• disable – Oracle does not execute DML statements in parallel during
upgrade.
• enable – Oracle executes DML statements in parallel during upgrade, if
configured to do so.
• enable_all – Oracle executes DML statements in parallel during upgrade in
all cases, unless turned off in the code or through configuration.
The default is enable_all.
If you set the value of ora-parallel-dml to enable or enable_all, then you
need to provide a value for attribute degree-of-parallelism as well.
Note: The value of this attribute interacts with the parallel-dml attribute on
the <versiontrigger> element. See “The versiontrigger Database Configuration
Element” on page 234 for more information.
ora-parallel-query (Oracle) Controls parallel query usage by Oracle during a database upgrade.
Valid values are:
• disable – Oracle does not use parallel queries during upgrade.
• enable – Oracle uses parallel queries during upgrade, if configured to do so.
The default is enable.
The value of this attribute interacts with the parallel-query attribute on the
<versiontrigger> element. See “The versiontrigger Database Configuration Ele‐
ment” on page 234 for more information.
sqlserverCreateIndexSortInTempDB (SQL Server) Boolean. Whether SQL Server stores temporary sort results in
tempdb. By using tempdb for sort runs, disk input and output is typically faster,
and the created indexes tend to be more contiguous. Valid values are:
• true – SQL Server stores sort results in tempdb.
• false – SQL Server stores sort results in the destination filegroup.
If you set sqlserverCreateIndexSortInTempDB to true, you must have enough
disk space available to tempdb for the sort runs, which, for the clustered index,
includes the data pages. You must also have sufficient free space in the destina‐
tion filegroup to store the final index structure, because SQL Server creates the
new index before deleting the old index.
Refer to the following web site for details on the requirements to use tempdb for
sort results.
http://msdn.microsoft.com/en-us/library/ms188281.aspx
updatestatistics (Oracle) Boolean. Whether to update table statistics during upgrade. The overall
time that it takes to upgrade the database is shorter if the database upgrade does
not update statistics.

Valid values are:

• true – Enables the upgrader to update statistics on changed objects. It also
allows the upgrader to maintain column level statistics consistent with what is
allowed in the code, data model, and configuration.
• false – Disable statistics generation during the upgrade.
If PolicyCenter does not update statistics during the upgrade:
• It reports a warning that recommends that you run the database statistics
batch process (DBStats) in incremental mode during the next maintenance
window.
• It updates the Server Tools Upgrade and Versions screen to show that the
upgrade did not update statistics.
If PolicyCenter does generate statistics during the upgrade, it updates the Upgrade
and Versions screen to report the runs of the statistics batch process, including in‐
cremental runs.
Note: Guidewire recommends that you run statistics in full mode after an up‐
grade to a major PolicyCenter version.
See the following for more information:
verifyschema Boolean. Whether PolicyCenter performs a verification of the database schema
before starting the database upgrade. This process verifies that the PolicyCenter
data model matches the physical database. This process can take some time. The
default is true.
It is possible for the verification process to take some time. Guidewire recom‐
mends that you perform this verification prior to starting the upgrade through
the use of the system_tools -verifydbschema command option. To use the
command, enter the following at a command prompt:
system_tools -password password -verifydbschema
See “System Tools Command” on page 393 for more information.
The <upgrade> element has the following subelements. Each of these elements is optional. There is, at most, a
single occurrence of each of these subelements on the <upgrade> element.
mssql-db-ddl Specifies options for SQL Server database DDL (Data Definition Language) statements. See “The mssql‐db‐
ddl Database Configuration Element” on page 221 for details.
ora-db-ddl Specifies options for Oracle database DDL (Data Definition Language) statements. See “The ora‐db‐ddl Da‐
tabase Configuration Element” on page 226 for details.
versiontriggers Specifies options for named version triggers. See “The versiontriggers Database Configuration Element” on
page 234 for details.
See also
The mssql‐db‐ddl Database Configuration Element

The <upgrade> element in file database-config.xml contains, at most, a single occurrence of subelement <mssql-
db-ddl>. The use of the <mssql-db-ddl> element is optional. Use this element to set SQL Server database DDL
(Data Definition Language) options during the creation of new objects in the database. This configuration applies to
the database at a global level.
The <mssql-db-ddl> element has the following syntax. The following code sample shows required attributes in
bold font.
<database>
<upgrade>

<mssql-db-ddl>
<mssql-filegroups op="string" admin="string" typelist="string" staging="string"
index="string" lob="string"/>
</mssql-table-ddl>
</mssql-db-ddl>
</upgrade>
</database>
There are no specific attributes on the <mssql-db-ddl> element.

The <mssql-db-ddl> element has the following subelements. Each of these elements is optional. At most, there is a
single occurrence of the <mssql-compression> and <mssql-filegroups> elements on the <mssql-db-ddl>
element. However, there can be multiple occurrences of the <mssql-table-ddl> element.
mssql- Specifies compression settings for SQL Server database tables and indexes at the global, database level.
compression See “The mssql‐compression Database Configuration Element” on page 222 for more information.
mssql- Specifies the mapping between SQL Server database filegroups and PolicyCenter logical tablespaces at the
filegroups global, database level. See “The mssql‐filegroups Database Configuration Element” on page 223 for more
information.
mssql-table- Specifies SQL Server database DDL options for a named table. These settings override values set at the glob‐
ddl al, database level. See “The mssql‐table‐ddl Database Configuration Element” on page 224 for more infor‐
mation.
See also

• “The upgrade Database Configuration Element” on page 217
The mssql‐compression Database Configuration Element

The <mssql-db-ddl> element in file database-config.xml contains, at most, a single occurrence of subelement
<mssql-compression>. The use of the <mssql-compression> element is optional. Use this element to set SQL
Server database compression options at the global, database level. See also the Installation Guide.
The <mssql-compression> element has the following syntax.
<database>
<upgrade>
<mssql-db-ddl>
</mssql-db-ddl>
</upgrade>
</database>
The following list describes the attributes that you can configure on the <mssql-compression> element.
index-compression If present, specifies the index compression setting for all indexes. Valid values are:
• NONE
• PAGE
• ROW
The default is NONE.

table-compression If present, specifies the table compression setting for all tables. Valid values are:
• NONE
• PAGE
• ROW
The <mssql-compression> element does not contain additional subelements.
See also
The mssql‐filegroups Database Configuration Element

The <mssql-db-ddl> element in file database-config.xml contains, at most, a single occurrence of subelement
<mssql-filegroups>. The use of the <mssql-filegroups> element is optional. Use this attributes to map SQL
Server filegroups to PolicyCenter logical tablespaces at the global, database level.
The use of these configuration elements applies to newly created objects only. If you map a specific table to a named
filegroup, SQL Server implements the change during table creation or table recreation only. Similarly, if you specify
a LOB filegroup for a table or all tables, only newly created tables have their LOB columns stored in the named
filegroup. The database does not store newly created LOB columns for existing tables in the named filegroup.
Note: The PolicyCenter schema verification process during server startup reports as warnings any
database tables or LOB columns that are not located in a configured filegroup. However, some of the
warnings are unavoidable due to the nature of the implementation mechanism and are purely
informational.
The <mssql-filegroups> element has the following syntax. The following code sample shows required attributes
in bold font.
<database>
<upgrade>
<mssql-db-ddl>
<mssql-filegroups op="string" admin="string" typelist="string" staging="string"
index="string" lob="string"/>
</mssql-db-ddl>
</upgrade>
</database>
The following list describes the attributes that you can configure on the <mssql-filegroups> element.
op Required. Name of a SQL Server filegroup.

admin Required. Name of a SQL Server filegroup.
typelist Required. Name of a SQL Server filegroup.
staging Required. Name of a SQL Server filegroup.

index Required. Name of a SQL Server filegroup.
lob Name of a SQL Server filegroup.
The <mssql-filegroups> element does not contain additional subelements.
See also
The mssql‐table‐ddl Database Configuration Element

The <mssql-db-ddl> element in file database-config.xml can contain any number of occurrences of subelement
<mssql-table-ddl>. The use of the <mssql-table-ddl> element is optional. Use this element to specify DDL
(Data Definition Language) options for a specific, named SQL Server database table.
The <mssql-table-ddl> element has the following syntax. The following code sample shows required attributes in
bold font.
<database>
<upgrade>
<mssql-db-ddl>
<mssql-table-filegroups lob-filegroups="string" index-filegroup="string" table-filegroup="string"/>
</mssql-table-ddl>
</mssql-db-ddl>
</upgrade>
</database>
The <mssql-table-ddl> element has the following attribute.
table-name Required. Name of the table to which these overrides apply.
The <mssql-table-ddl> element has the following subelements. Each of these elements is optional. There is, at
most, a single occurrence of the <mssql-table-compression> and <mssql-table-filegroups> elements on the
<mssql-table-ddl> element. There can be, however, multiple occurrences of the <mssql-index-ddl> element.
mssql-index-ddl Specifies DDL options for a specific index. See “The mssql‐index‐ddl Database Configuration Element”
on page 224 for more information.
mssql-table- Specifies compression for the named table. See “The mssql‐table‐compression Database Configuration
compression Element” on page 225 for more information.
mssql-table- Specifies a filegroup to associate with a table, index, or LOB. See “The mssql‐table‐filegroups Database
filegroups Configuration Element” on page 226 for more information.
See also
The mssql‐index‐ddl Database Configuration Element

The <mssql-table-ddl> element in file database-config.xml can contain any number of occurrences of
subelement <mssql-index-ddl>. The use of the <mssql-index-ddl> element is optional. Use this element to
define SQL Server database DDL options for a specific index, based on the key columns. Any value that you set at
this level overrides that same value set at the global, database level. You can create multiple <mssql-index-ddl>
elements on the parent <mssql-table-ddl> element, each of which affects a different index.
The <mssql-index-ddl> element has the following syntax. The following code sample shows required attributes in
bold font.
<database>
<upgrade>
<mssql-db-ddl>
index-filegroup="string"
key-columns="string" partition-scheme="string"/>

</mssql-table-ddl>
</mssql-db-ddl>
</upgrade>
</database>
The following list describes the attributes that you can configure on the <mssql-index-ddl> element.
key-columns Required. Comma‐delimited list of key columns, in order. Specify DESC after the column name for a de‐
scending sort order on the column.
filter-where Specifies an index filter to add after the WHERE keyword in the SQL Server CREATE INDEX ... WHERE
statement. The filter that you create must conform to standard SQL Server rules.
index-compression Specifies the compression setting for the specified index. Valid values are:
• NONE
• PAGE
• ROW
If not specified, PolicyCenter uses the SQL Server database default.
index-filegroup Name of the filegroup associated with this index. Do not use this attribute if you supply a value for the
partition-scheme attribute as the two attributes are mutually exclusive.
partition-scheme Name of a partition scheme for this index. Use of this attribute implies the use of PolicyCenter cluster‐
ing. Do not use this attribute if you supply a value for the index-filegroups attribute as the two attrib‐
utes are mutually exclusive.
The <mssql-index-ddl> element does not contain additional subelements.
See also
The mssql‐table‐compression Database Configuration Element

The <mssql-table-ddl> element in file database-config.xml contains, at most, a single occurrence of
subelement <mssql-table-compression>. The use of the <mssql-table-compression> element is optional. Any
value that you set at this level overrides that same value set at the global, database level. See also the Installation
Guide.
The <mssql-table-compression> element has the following syntax. The following code sample shows required
<database>
<upgrade>
<mssql-db-ddl>
</mssql-table-ddl>
</mssql-db-ddl>
</upgrade>
</database>
The following list describes the attributes that you can configure on the <mssql-table-compression> element. All
of these attributes are optional.
index-compression Specifies the index compression setting for the specified index. Valid values are:
• NONE
• PAGE
• ROW
If not specified, PolicyCenter uses the database default.

table-compression Specifies the table compression setting for the specified table. Valid values are:
• NONE
• PAGE
• ROW
If not specified, PolicyCenter uses the database default.
The <mssql-table-ddl> element does not contain additional subelements.
The mssql‐table‐filegroups Database Configuration Element

The <mssql-table-ddl> element in file database-config.xml contains, at most, a single occurrence of
subelement <mssql-table-filegroups>. The use of the <mssql-table-groups> element is optional. Use this
element to associate a filegroup with a table, index, or LOB. Any value that you set at this level overrides that same
value set at the global, database level.
The <mssql-table-filegroups> element has the following syntax. The following code sample shows required
<database>
<upgrade>
<mssql-db-ddl>
<mssql-table-filegroups table-filegroup="string" index-filegroup="string" lob-filegroups="string"/>
</mssql-table-ddl>
</mssql-db-ddl>
</upgrade>
</database>
The following list describes the attributes that you can configure on the <mssql-table-filegroups> element. All
of these attributes are optional. However, if you do not specify at least one of these attributes, there is no need for
this element to be present in database-config.xml.
table-filegroup Name of the filegroup to associate with this table.
index-filegroup Name of the filegroup to associate with any indexes on this table.
lob-filegroup Name of the filegroup to associate with any large object (LOB, CLOB, or spatial column).
The <mssql-table-filegroups> element does not contain additional subelements.
The ora‐db‐ddl Database Configuration Element

The <upgrade> element in file database-config.xml contains, at most, a single occurrence of subelement <ora-
db-ddl>. The use of the <ora-db-ddl> element is optional. Use this element to set Oracle database DDL (Data
Definition Language) options during the creation of new objects in the database. This configuration applies to the
database at a global level.
The <ora-db-ddl> element has the following syntax. The following code sample shows required attributes in bold
font.
<database>
<upgrade
<ora-db-ddl>
typelist="string"/>


</ora-table-ddl>
</ora-db-ddl>
</upgrade>
</database>
There are no specific attributes on the <ora-db-ddl> element.

The subelements on the <ora-db-ddl> element have the following meanings.
ora- Specifies Oracle compression settings for all tables and indexes at the global, database level. See “The ora‐
compression compression Database Configuration Element” on page 227 for more information.
ora-lobs Specifies attributes for LOB columns on all tables at the global, database level. See “The ora‐lobs Database
Configuration Element” on page 228 for more information.
ora-table-ddl Specifies DDL parameters and overrides for a specific, named Oracle database table. See “The ora‐table‐ddl
Database Configuration Element” on page 229 for more information.
tablespaces Specifies default mappings for Oracle tablespaces at a global, database level. See “The tablespaces Data‐
base Configuration Element” on page 228 for more information.
See also
The ora‐compression Database Configuration Element

The <ora-db-ddl> element in file database-config.xml contains, at most, a single occurrence of subelement
<ora-compression>. The use of the <ora-compression> element is optional. Use this element to set compression
on Oracle database indexes and tables at a global, database level.
The <ora-compression> element has the following syntax.
<database>
<upgrade>
<ora-db-ddl>
</ora-db-ddl>
</upgrade>
</database>
The following list describes the attributes that you can configure on the <ora-compression> element. All of these
attributes are optional.
index-compression Boolean. Whether to use index compression for all indexes in an Oracle database. The default is false.
table-compression Specifies table compression type for all tables in an Oracle database.
Valid values are:
• ADVANCED
• BASIC
• NONE

The <ora-compression> element does not contain additional subelements.
See also
The ora‐lobs Database Configuration Element

The <ora-db-ddl> element in file database-config.xml contains, at most, a single occurrence of subelement
<ora-lobs>. The use of the <ora-lobs> element is optional. Use this element to set options for LOB columns on
tables in an Oracle database at a global, database level.
The <ora-lobs> element has the following syntax.
<database>
<upgrade>
<ora-db-ddl>
</ora-db-ddl>
</upgrade>
</database>
The following list describes the attributes that you can configure on the <ora-lobs> element. All of these attributes
are optional.
caching Boolean. Whether to use caching for all LOB columns on a table or for the Oracle database globally. The default is
false.
type Sets LOB type globally for the database.

Valid values are:
• BASIC
• SECURE
• SECURE_COMPRESSED
The default is SECURE.
Note: SECURE and SECURE_COMPRESSED refer to the use of Oracle SecureFiles LOBs.
The <ora-lobs> element does not contain additional subelements.
See also
The tablespaces Database Configuration Element

The <ora-db-ddl> element in file database-config.xml contains a single occurrence of subelement
<tablespaces>. You must provide a <ora-tablespaces> subelement if using the <ora-db-ddl> element. Use this
element to map Oracle tablespaces to PolicyCenter logical tablespaces at a global, database level.
The <tablespaces> element has the following syntax. The following code sample shows required attributes in bold
font.
<database>
<upgrade>
<ora-db-ddl>
typelist="string"/>
</ora-db-ddl>
</upgrade>
</database>

The following list describes the attributes that you can configure on the <tablespaces> element.
The following attributes are all required.

admin Name of a PolicyCenter logical tablespace.
index Name of a PolicyCenter logical tablespace.
op Name of a PolicyCenter logical tablespace.
staging Name of a PolicyCenter logical tablespace.
typelist Name of a PolicyCenter logical tablespace.
The following attribute is optional.

lob Name of a PolicyCenter logical tablespace.
The <tablespaces> element does not contain additional subelements.
See also
The ora‐table‐ddl Database Configuration Element

The <ora-db-ddl> element in file database-config.xml can contain any number of occurrences of subelement
<ora-table-ddl>. The use of the <ora-table-ddl> element is optional. Use this element to set DDL parameters
and overrides for a specific, named table in an Oracle database.
The <ora-table-ddl> element has the following syntax. The following code sample shows required attributes in
bold font.
<database>
<upgrade>
<ora-db-ddl>
</ora-table-ddl>
</ora-db-ddl>
</upgrade>
</database>
The <ora-table-ddl> element has the following attribute.
table-name Required. Name of the table to which these overrides apply.
The subelements on the <ora-table-ddl> element have the following meanings.
ora-index-ddl Specifies options for a specific Oracle index, based on key columns. See “The ora‐index‐ddl Database
Configuration Element” on page 230 for more information.
ora-lobs Specifies options for LOB columns on a specific, named table in an Oracle database. See “The ora‐lobs
Database Configuration Element” on page 231 for more information.

ora-table- Specifies compression options on a specific, named index or table in an Oracle database. See “The ora‐
compression table‐compression Database Configuration Element” on page 231 for more information.
ora-table-date- Specifies options for date range partitioning on a specific, named table in an Oracle database. See “The
interval- ora‐table‐date‐interval‐partitioning Database Configuration Element” on page 232 for more informa‐
partitioning tion.
ora-table-hash- Specifies options for hash partitioning of a specific, named table in an Oracle database. See “The ora‐
partitioning table‐hash‐partitioning Database Configuration Element” on page 233 for more information.
ora-table- Specifies tablespace options for a specific, named table in an Oracle database. See “The ora‐table‐
tablespaces tablespaces Database Configuration Element” on page 233 for more information.
See also
The ora‐index‐ddl Database Configuration Element

The <ora-table-ddl> element in file database-config.xml can contain any number of occurrences of subelement
<ora-index-ddl>. The use of the <ora-index-ddl> element is optional. Use this element to set DDL parameters
and overrides for a specific Oracle index, based on key columns. See also the Installation Guide.
The <ora-index-ddl> element has the following syntax. The following code sample shows required attributes in
bold font.
<database>
<upgrade>
<ora-db-ddl>
</ora-table-ddl>
</ora-db-ddl>
</upgrade>
</database>
The following list describes the attributes that you can configure on the <ora-index-ddl> element.
key-columns Required. Ordered, comma‐delimited list of key columns. Specify DESC after the column name for a de‐
scending sort order on that column.
The following attributes are optional.
index-compression Specifies the index compression for this index. If you do not specify this attribute, PolicyCenter uses the
table or database default.
index-tablespaces Name of the tablespace override for the index.
The subelements on the <ora-index-ddl> element have the following meanings.
ora-index- Defines partitioning for the specified Oracle index. The <ora-index-partitioning> element has the follow‐
partitioning ing attributes:
• num-hash-partitions – The number of hash partitions to define. The default is 128.
• partitioning-type – Required if using this element. Sets the partitioning type to one of the following:
◦ LOCAL – Inherit the partitioning type from the table
◦ HASH – Use hash partitions. If you set this attribute to HASH, then you need to specify the number of
partitions to use attribute num-hash-partitions. Do not set partitioning-type to HASH if you
specify an <ora-index-range-partition> subelement.
◦ RANGE – Specify the range partitioning column list and the partition upper limits with one or more ora-
index-range-partition elements.
• range-partitioning-column-list – Optional. Use to specify the global range partitioning column list.
This attribute requires the definition of one or more ora-index-range-partitioning elements. Do not
specify the last range which is always VALUES LESS THAN (MAXVALUE). Do not use if you set attribute
partitioning-type to HASH.
The <ora-index-partitioning> element contains a single subelement:
• ora-index-range-partition – Optional. A comma‐delimited, ordered list of literal values
corresponding to the column list in the range-partitioning-column-list attribute. Use single quotes
with string values. PolicyCenter uses this value in the clause VALUES LESS THAN(value_list). Do not
use if you set attribute partitioning-type to HASH.
See also
The ora‐lobs Database Configuration Element

The <ora-table-ddl> element in file database-config.xml contains, at most, a single occurrence of subelement
<ora-lobs>. The use of the <ora-lobs> element is optional. Use this element to set attributes for LOB columns on
a specific, named table in an Oracle database. Any value that you set at this level overrides that same value set at the
global, database level.
The <ora-lobs> element has the following syntax. The following code sample shows required attributes in bold
font.
<database>
<upgrade>
<ora-db-ddl>
</ora-table-ddl>
</ora-db-ddl>
</upgrade>
</database>
The following list describes the attributes that you can configure on the <ora-lobs> element. All of these attributes
are optional.
caching Sets the LOB cache attribute for the named Oracle table. The default is false.
type Sets the LOB type for the named Oracle table. Valid values are:
• BASIC
• SECURE
• SECURE_COMPRESSED
The default is SECURE.
Note: SECURE and SECURE_COMPRESSED refer to the use of Oracle SecureFiles LOBs.
The <ora-lobs> element does not contain additional subelements.
The ora‐table‐compression Database Configuration Element

<ora-table-compression>. The use of the <ora-table-compression> element is optional. Use this element to set
compression on a specific, named index or table in an Oracle database. Any value that you set at this level overrides
that same value set at the global, database level. See also the Installation Guide.
The <ora-table-compression> element has the following syntax. The following code sample shows required
<database>
<upgrade>
<ora-db-ddl>

</ora-table-ddl>
</ora-db-ddl>
</upgrade>
</database>
The following list describes the attributes that you can configure on the <ora-table-compression> element. All of
index-compression Boolean. Whether to use index compression. Valid values are:

• true – Use compression for all indexes on this table.
• false – Do not use compression for the indexes on this table.
If you do not specify this attribute, PolicyCenter uses the database default.
table-compression Specifies table compression for this table. Valid values are:
• ADVANCED
• BASIC
• NONE
If you do not specify this attribute, PolicyCenter uses the database default.
The <ora-table-compression> element does not contain additional subelements.
The ora‐table‐date‐interval‐partitioning Database Configuration Element

<ora-table-date-interval-partitioning>. The use of the <ora-table-date-interval-partitioning>
element is optional. Use to add date range partitioning to a specific, named table in an Oracle database.
The <ora-table-date-interval-partitioning> element has the following syntax. The following code sample
shows required attributes in bold font.
<database>
<upgrade>
<ora-db-ddl>
</ora-table-ddl>
</ora-db-ddl>
</upgrade>
</database>
The following list describes the attributes that you can configure on the <ora-table-date-interval-
partitioning> element.
datecolumn Required. Name of the column to use for the date range. The column must be non‐nullable and one of the fol‐
lowing types:
• datetime
• dateonly
interval Required. The interval for each partition. Valid values are:
• DAILY
• MONTHLY
• QUARTERLY
• WEEKLY
• YEARLY
The <ora-table-date-interval-partitioning> element does not contain additional subelements.

The ora‐table‐hash‐partitioning Database Configuration Element

<ora-table-hash-partitioning>. The use of the <ora-table-hash-partioning> element is optional. Use this
element to add hash partitioning to a table in an Oracle database.
The <ora-table-hash-partitioning> element has the following syntax. The following code sample shows
required attributes in bold font.
<database>
<upgrade>
<ora-db-ddl>
</ora-table-ddl>
</ora-db-ddl>
</upgrade>
</database>
The following list describes the attributes that you can configure on the <ora-table-hash-partitioning> element.
All of these attributes are optional.
hash-column Name of the column to use for the hash function:

• For keyable entities, the default is the entity ID.
• For non‐keyable entities, you must provide a value.
num-partitions The number of hash partitions to define. The default is 128.
The <ora-table-hash-partitioning> element does not contain additional subelements.
The ora‐table‐tablespaces Database Configuration Element

<ora-table-tablespaces>. The use of the <ora-table-tablespaces> element is optional. Use this element to
provide overrides for the default table, index, and LOB tablespaces for a specific, named table in an Oracle database.
The <ora-table-tablespaces> element has the following syntax. The following code sample shows required
<database>
<upgrade>
<ora-db-ddl>
</ora-table-ddl>
</ora-db-ddl>
</upgrade>
</database>
The following list describes the attributes that you can configure on the <ora-table-tablespaces> element. All of
index-tablespace Name of the tablespace override for the specified index.
lob-tablespace Name of the tablespace override for the specified LOB column.
table-tablespace Name of the tablespace override for the specified table.
The <ora-table-tablespaces> element does not contain additional subelements.

The versiontriggers Database Configuration Element

The <upgrade> element in file database-config.xml contains, at most, a single occurrence of subelement
<versiontriggers>. The use of the <versiontriggers> element is optional. Use this element to provide overrides
for one or more named database version triggers.
The database upgrade executes a series of version triggers that make changes to the database to upgrade between
PolicyCenter versions. Usually, the default settings are sufficient. Change these settings only while investigating a
slow database upgrade.
The <versiontriggers> element has the following syntax. The following code sample shows required attributes in
bold font.
<database>
<upgrade>
</versiontriggers>
</upgrade>
</database>
The <versiontriggers> element has the following attribute, which is optional.
dbmsperfinfothreshold Specifies–for all version triggers–the threshold after which the database upgrader gathers perform‐
ance information from the database. The default is 600 (seconds).
If a version trigger takes longer than dbmsperfinfothreshold number of seconds to execute,
PolicyCenter:
• Queries the underlying database management system (DBMS).
• Builds a set of HTML pages with performance information for the interval in which the version
trigger was executing.
• Includes these HTML pages in the upgrader instrumentation for the version trigger.
You can completely turn off the collection of database snapshot instrumentation for version trig‐
gers by setting the value of the dbmsperfinfothreshold attribute to 0. If you do not have the li‐
cense for the Oracle Diagnostics Pack, you must set dbmsperfinfothreshold to 0 before running
the upgrade.
The <versiontriggers> element has the following subelement, of which there can be multiple occurrences.
versiontrigger Provides override instructions for a specific, named, version trigger. See “The versiontrigger Database Con‐
figuration Element” on page 234
See also
The versiontrigger Database Configuration Element

The <versiontriggers> element in file database-config.xml can contain any number of occurrences of
subelement <versiontrigger>. The use of the <versiontrigger> element is optional. Use this element to provide
specific override instructions for a named version trigger.
The <versiontriggers> element has the following syntax. The following code sample shows required attributes in
bold font.
<database>
<upgrade>

<versiontriggers>
</versiontriggers>
</upgrade>
</database>
The following list describes the attributes that you can configure on the <versiontrigger> element.
name Required. Case‐sensitive, fully qualified name of a version trigger.

extendedquerytracingenabled Boolean. (Oracle) Whether PolicyCenter uses extended SQL tracing (Oracle event 10046)
for the SQL statements that the version trigger executes. The default is false.
This output can be very useful if debugging certain types of performance problems. The
trace files that PolicyCenter generates exist only on the actual database computer.
PolicyCenter does not integrate this information into the upgrade instrumentation.
parallel-dml Boolean. (Oracle) Whether Oracle executes DML (Data Manipulation Language) statements
in parallel for this particular version trigger during the database upgrade.
Valid values are:
• true – Execute DML statement in parallel for inserts and updates for this version
trigger, unless the ora-parallel-dml attribute on the <upgrade> element is set to
disable.
• false – Disable parallel execution of DML statements for this version trigger, even if set
in code or if the ora-parallel-dml attribute on the <upgrade> element is set to
enable_all.
If not set, Oracle executes DML statements in parallel, if set in the code or the ora-
parallel-dml attribute on the <upgrade> element is set to enable_all (default). See
“The upgrade Database Configuration Element” on page 217 for more information.
parallel-query Boolean. (Oracle) Whether a version trigger provides a hint to the optimizer to use parallel
queries while executing SQL queries. This can be useful in improving performance as some
version triggers read large amounts of data while running their version check.
Valid values are:
• true – Execute parallel SQL queries unless the ora-parallel-query attribute on the
<upgrade> element is set to disable.
• false – Do not execute parallel SQL queries.
See “The upgrade Database Configuration Element” on page 217 for more information.
queryoptimizertracingenabled Boolean. (Oracle) Whether PolicyCenter uses query optimizer tracing (Oracle event 10053)
for the SQL statements that the version trigger executes. The default is false.
This output can be very useful if debugging certain types of performance problems. The
trace files that PolicyCenter generates exist only on the actual database computer.
PolicyCenter does not integrate this information into the upgrade instrumentation.
recordcounters Boolean. Whether to record the values of DBMS‐specific counters at the beginning and end
of each execution of the specified version trigger. Valid values are:
• true – PolicyCenter retrieves the current state of the counters from the underlying
DBMS at the beginning of execution of the version trigger. This behavior is dependent
on the value of dbmsperfinfothreshold on <versiontriggers>. if the execution time
of the version trigger exceeds the value of dbmsperfinfothreshold, PolicyCenter
retrieves the state of the counters at the end of the execution of the version trigger.
• false – PolicyCenter does not retrieve data.
PolicyCenter writes differences to the DBMS‐specific instrumentation screens of the up‐
grade instrumentation. PolicyCenter only persists these values with the upgrade instru‐
mentation if the execution time of the version trigger exceeds the configured threshold.

See also “The versiontriggers Database Configuration Element” on page 234.

updatejoinorderedhint Boolean. (Oracle) Whether to use the ORDERED hint if using a SQL UPDATE statement with a
join operation. The default is false.
updatejoinusemergehint Boolean. (Oracle) Whether to use the USE_MERGE hint if using a SQL UPDATE statement with
a join operation. The default is false.
updatejoinusenlhint Boolean. (Oracle) Whether to use the USE_NL hint if using a SQL UPDATE statement with a
join operation. The default is false.
The <version-trigger> element does not contain additional subelements.
See also

chapter 16
Database Maintenance
This topic discusses key issues for configuring and maintaining the PolicyCenter database. While PolicyCenter
automatically handles most changes to its schema, involve a database administrator in tuning and managing the
database server.
IMPORTANT The versions of third-party products that Guidewire supports for this release are subject
to change without notice. For current system and patch level requirements, visit the Guidewire
Community and search for knowledge article 1005, Supported Software Components.
See also
About the Upgrade and Versions Screen

PolicyCenter includes a Server Tools Upgrade and Versions screen that provides detailed information about each
database upgrade. The Upgrade and Versions screen includes information on the following:
• Version number of upgrade
• Status of upgrade
• Type of upgrade
• Start and time of upgrade
• Status of Deferred Upgrade Tasks batch processing
From this screen, you can drill down into more specific details about the upgrade, or download a report of the
upgrade details.
See also
Database Best Practices

Guidewire recommends the following best practices for the PolicyCenter application database.
Database Location and Synchronization

Physically locate the PolicyCenter application server and the database server in the same timezone and geographic
location. Locating the two servers in geographically distant locations increases the time that it takes the application
Database Maintenance 237
server to query the database. It is possible for the latency between the two servers to impact performance. As a
measure of this metric, Guidewire logs the time that it takes to verify access to the database at server startup.
Synchronize the application server with the database clock. The maximum time difference allowed between the
application server and database server is 29 minutes. If you use database clustering, synchronize all nodes in the
database cluster with each other.
Database Maintenance
If you need to perform any database maintenance tasks, such as applying a patch, shut down all PolicyCenter servers
that connect to the database. Restart the servers after the database maintenance is complete.
Run consistency checks on the database, especially after importing data.
Back up the database periodically to support disaster recovery options.
Monitor storage performance. If I/O (input/output) times are slower than 10ms, it indicates that there is most likely
an issue.
Monitor tablespace size allocations and disk space to ensure that PolicyCenter does not run out of space.
Update database statistics periodically so that the query optimizer selects an efficient plan for executing application
queries.
Database tables
For Oracle databases, keep the Oracle default settings as much as possible. For example, do not set a 4K block size.
Consult with Guidewire if you want to change the default Oracle settings.
Do not insert data directly into tables managed by PolicyCenter. This can cause the data distribution tool to fail and
cause other problems.
Do not add large numbers of mediumtext and CLOB columns to a table.
Do not add an index outside of PolicyCenter and not declare it in an extension file.
See also
• “Guidewire Database Direct Update Policy” on page 239

• “PolicyCenter Database Back Up” on page 240
Understanding Data Model Updates

As PolicyCenter starts, it compares system metadata (the description of the objects and tables in the config
directory) to the database to see if they match. For example, if PolicyCenter contains a new object extension added
after the last server start, the database and metadata do not match. If these two do not match, PolicyCenter attempts
to update the database to match the metadata. This type of update to the data model is different than a product
version upgrade, which includes more extensive changes to the database.
The update process calculates current checksums for all the XML files in the data model. It then compares them
with historical checksums stored in the SystemParameter entity. If the values differ, then PolicyCenter updates the
database to match the metadata. As the last step in the update, PolicyCenter updates the SystemParameter entity
with the current checksums.
If during the update PolicyCenter creates a new table, then it also generates a unique index for the table. The
TableRegistry entity stores this information. In this way, PolicyCenter guarantees uniqueness.
Before completing the startup process, PolicyCenter again verifies the data model against the physical database. If,
for some reason, the model and database disagree, PolicyCenter writes warnings to the log and, if possible, suggests
corrective actions. Take the corrective action if prompted to do so.
238 chapter 16: Database Maintenance
Note: If, for any reason, there is an interruption to the server during a database update, the server
resumes the update upon restart. PolicyCenter accomplishes this by storing the steps in the database
and marking them completed as part of the same database transaction that applies a change. This only
applies to data model updates and does not apply to product version upgrades.
See also
• “Run a Schema Verification Report” on page 239
• “About the Upgrade and Versions Screen” on page 237
• “View an Upgrade Report” on page 364
• “Understanding Guidewire Software Versioning” on page 365
Run a Schema Verification Report

Procedure
1. Ensure that PolicyCenter is running.
admin/bin
3. Enter the following command to generate a database schema verification report.
system_tools -password password -verifydbschema
Guidewire Database Direct Update Policy

PolicyCenter runs on SQL-based Relational Database Management Systems (RDBMS). You can use SQL or other
query tools directly in a read-only manner to extract or view data. Such read-only queries, depending on their scope
and how they are written, can negatively effect overall database performance even though they do not modify any
data. Therefore, Guidewire recommends that you run SQL queries in a replica or copy of their production database,
rather than the production database itself. For applications such as data warehouses and intensive reporting,
Guidewire recommends that you explore mechanisms for replicating or summarizing data into a production
reporting database for this purpose. This practice can help unexpected production performance issues due to
intensive reporting requirements or lengthy queries.
Internal application logic, embedded in PolicyCenter application code and APIs, maintains a variety of data and
metadata that relates to your application data. This might not be obvious from review of the RDBMS table structure.
Examples include:
• Calculations of summary table data for reporting
• Caching of application data in memory for faster access
• Tracking of state information associated with the underlying data, such as the processing state of an integration
message
For this reason, never use SQL to directly update the underlying RDBMS. Any such direct SQL updates can leave
the data in an inconsistent state. Guidewire can require you to restore the database to a previous state if you require
support after performing such an update query. Guidewire Support is not able to assist you with diagnosing and
correcting application issues caused by your database queries. It is your responsibility to restore PolicyCenter to a
consistent state.
WARNING Guidewire supports the built-in automatic database upgrade process only for
Guidewire InsuranceSuite products. Guidewire explicitly does not support any alternative process
that executes SQL DDL commands on the database.
If you have a legitimate need to update underlying application data, Guidewire recommends that you use Guidewire
APIs, either Java or Gosu, to perform the necessary updates. This ensures that you do not miss any critical side
effects of the updates in the process of altering the data. Using Guidewire APIs to update application data is safer
than using SQL queries with regard to consistency. However, with any programming language or API it is still
possible to update data incorrectly or in ways that do not perform well. Therefore, before using the APIs, Guidewire
strongly recommends that you review your intended updates with your Guidewire Support Partner and/or Guidewire
Professional Services team.
In the rare case in which no API exists to correct a data corruption problem, Guidewire can advise you on the SQL
queries to use to correct these problems. In these cases, the SQL queries used to update the database must be written,
or approved, by Guidewire. This process ensures that all SQL queries use correct logic and that you take all potential
side effects into account.
Do not apply any other SQL queries to modify data in a PolicyCenter database. Guidewire does not provide, nor
review, such queries for situations in which an API or supported alternate method is available.
PolicyCenter Database Back Up

PolicyCenter stores most information in the database, so the most important part of backing up the system is taking
frequent database backups. Consult the documentation for your database management system for tools and
techniques to backing up the database.
You can perform a hot (also called dynamic) or cold backup of the PolicyCenter database. You can take a hot backup
while users are still accessing PolicyCenter. Read your database documentation to understand the risks involved in
hot backups. If you plan on taking a cold backup, you must put the PolicyCenter server in maintenance mode before
performing the backup.
In the case of a complete system failure, with proper backups you can reinstall the PolicyCenter WAR or EAR file
on a new server, connecting to the same database. In the case of a database failure, you need to restart PolicyCenter
from the last database backup.
After restoring a PolicyCenter database from a backup, instruct PolicyCenter to rebuild its database statistics. See
“Understanding Database Statistics” on page 255.
Keep a Backup Copy of the PolicyCenter Configuration Files

Besides backing up the database, maintain a backup of the PolicyCenter configuration. This is especially important
for any files that you modify as part of installation or configuration of PolicyCenter. Guidewire stores all
configuration files in the PolicyCenter/modules/configuration/config directory, making it easy to keep these
files in a source control system, if desired.
Database Consistency Checks

PolicyCenter includes a number of database consistency checks that you can run on demand. These checks
determine if any unusual conditions exist in the PolicyCenter database such as orphaned child records or
inconsistencies between properties. These reports are especially useful during trial periods or for testing converted
data.
Viewing the results of consistency checks

If running the consistency checks generates any errors or issues, PolicyCenter reports the issues in the following
ways:
• In a report that you can view or download from the PolicyCenter Server Tools Consistency Checks screen.
• In the PolicyCenter application console and in the pclog.log file, if configured to do so.
See also
Recommendations for Running Database Consistency Checks

Guidewire recommends that you run the database consistency checks on a regular basis to identify potential
problems. For example, run database consistency checks:
• Before importing data into a production database
• Before and after a database upgrade
• Weekly after a new PolicyCenter deployment
• Monthly after stabilization of a new PolicyCenter deployment
• Monthly while testing imported data that you converted from a legacy system
• Monthly after you have moved the converted data to a production database
Best Practice
It is possible to trigger the execution of database consistency checks as the application server starts, by setting a
database attribute in file database-config.xml. However, Guidewire does not recommend that you use this method
of automatically running consistency checks as a general rule, especially in a production environment. The
execution of consistency checks is very resource intensive and impacts server performance significantly. Only
trigger the execution of consistency checks at server start in a development environment that contains a very small
data set.
Instead, Guidewire recommends that you execute consistency checks from the Server Tools Consistency Checks
screen in PolicyCenter.
Consistency Checks and Performance

For very large databases, running consistency checks can have a major negative effect on PolicyCenter performance.
For example, some consistency check queries use a large amount of temporary space on Oracle
For development environments with very small data sets, you can enable consistency checks to run each time the
PolicyCenter server starts. However, be aware that running consistency tests during server startup:
• Can take a large amount of time to complete
• Can impact performance severely
• Can possibly time out on large datasets
IMPORTANT Set the checker attribute on <database-config> to false under most circumstances.
Guidewire recommends that you do not set checker to true except in development environments
with very small test data sets.
Running Consistency Checks

The following table lists the various ways that you can run or trigger database consistency checks.
Ways to run consistency checks … More information

From the Consistency Checks screen “Running Consistency Checks from PolicyCenter” on page 242
From a command prompt “Running Consistency Checks from a Command Prompt” on page 242
At server start up “Configure Consistency Checks to Run at Server Start” on page 242

See also
• “Configuring the Number of Threads for Consistency Checks” on page 243
Running Consistency Checks from PolicyCenter

Guidewire recommends that you view and run consistency checks from the Server Tools Consistency Checks screen in
PolicyCenter. The Consistency Checks screen lists which consistency checks are available for specific tables. You can
also use this page to view the results of previous consistency check runs. If there are consistency check errors, the
results include SQL queries that you can use to identify records that violated the consistency check.
See also
• “Run a Consistency Check from PolicyCenter” on page 335
Running Consistency Checks from a Command Prompt

It is possible to launch database consistency checks from a command prompt. Launching consistency checks from
the command prompt is useful primarily for scheduling checks to run on a regular basis. This approach to running
consistency checks uses the system_tools -checkdbconsistency command option.
Running consistency checks using the -checkdbconsistency option can take a long time. If the connection times
out while running this command, try the following:
• Run consistency checks on fewer tables at a time.
• Increase the number of worker instance threads used by the consistency check work queue.
See also
• “Configuring the Number of Threads for Consistency Checks” on page 243
• “Run a Consistency Check Using System Tools” on page 336
Configure Consistency Checks to Run at Server Start

It is possible to set the checker attribute in file database-config.xml to automatically trigger database consistency
checks at server start.
About this task
IMPORTANT Guidewire strongly discourages the use of the checker attribute in file database-
config.xml to automatically trigger database consistency checks at server start. Database consistency
checks are very resource-intensive and can significantly degrade server performance. Only use the
checker attribute to trigger consistency checks in a development environment that contains a very
small data set.
Procedure
1. Open Guidewire Studio™ for PolicyCenter.
2. In the Project window, expand configuration→config.
a. Open file database-config.xml.
b. Add a checker attribute on the <database> element and set the attribute to true.
<database checker="true" ...>
By default, Guidewire omits this attribute in the base configuration and sets it value to false.
c. Save and close file database-config.xml.
3. Restart the application server to trigger the database consistency checks:

• If working in a development environment, restart the QuickStart server.
• If working in a production environment, create a new WAR or EAR file and deploy the file to the
production server.
See also
• “Configure Worker Threads for Consistency Checks in config.xml” on page 244
Configuring the Number of Threads for Consistency Checks

Using a larger number of threads for consistency checks can help performance as long as your server can process the
threads. Guidewire recommends starting with five threads. If you use too many threads, there is a greater chance that
current users can experience reduced performance if the database operates with a full load.
It is possible to set the number of threads to use for each consistency check in several different ways, depending on
how you configure the consistency checks to run. The following table outlines these differences.
Ways to set thread count For more information

From Consistency Checks screen “Configure Worker Threads for Consistency Checks in work‐queue.xml” on page 243
From command prompt “Configure Worker Threads for Consistency Checks in work‐queue.xml” on page 243
At server start up “Configure Worker Threads for Consistency Checks in config.xml” on page 244
See also
Configure Worker Threads for Consistency Checks in work‐queue.xml

About this task
PolicyCenter uses the work queue mechanism to run consistency checks.
Procedure
1. In the PolicyCenter Studio Project window, expand configuration→config→workqueue:
a. Open file work-queue.xml for editing.
b. Locate the block of code that contains the following workQueueClass value.
com.guidewire.pl.system.database.checker.DBConsistencyCheckWorkQueue
c. Locate the <worker> element for this code block.

d. Set the instances and batchsize attributes as appropriate for your PolicyCenter installation.
For example, the following code sets the consistency check work queue to use five worker threads, with
each worker checking out ten work items at a time.
<worker instances="5" batchsize="10"/>
2. Restart the application server for your changes to take effect:

production server.
See also
• “Configuring Work Queues” on page 92
Configure Worker Threads for Consistency Checks in config.xml

About this task
If the database connection times out while running consistency checks at server startup, Guidewire recommends that
you increase the number of worker threads available for processing the consistency checks.
Procedure
1. In Guidewire Studio™ for PolicyCenter.
2. In the Project window, expand configuration→config:
a. Open file config.xml for editing.
b. Locate the ConsistencyCheckerThreads parameter.
In the base configuration, the value of this parameter is 1.
c. Set the value of this parameter to a number that meets your business needs.
For example, set this value to five.
<param name="ConsistencyCheckerThreads" value="5"/>
3. Restart the application server for your changes to take effect:

production server.
Resize Database Columns

About this task
After the PolicyCenter database is in use, you might discover that you need to change the size of certain columns,
such as making a column name longer. PolicyCenter does not provide an automated way of doing this. However, the
following commonly used procedure provides a general outline of the steps involved making this kind of database
change.
IMPORTANT Guidewire does not support re-sizing any database columns that are part of the
PolicyCenter base configuration.
Procedure
1. Shut down PolicyCenter.
2. Alter the table and add a new temporary column that is the new size.
3. Copy all of the data from the source column to the temporary column.
4. Alter the table and drop the source column.
Depending on the database, it is possible that you need to set the data in this column to all nulls before you can
drop the column.
5. Alter the table and add the new source column that is the new size.
6. Copy the data from the temporary column to the new source column.
7. Alter the table and drop the temporary column.
8. Restart PolicyCenter.
Purging Unwanted Data

Over time, the PolicyCenter database acquires and stores ever-increasing amounts of data. Left unchecked, the
database can contain a large amount of unused and unnecessary data. Guidewire recommends that you periodically
remove this stale data to improve performance and reduce complexity during upgrade. To facilitate the removal of
unnecessary data, Guidewire provides a number of purge-related batch processes and work queues.
About Purging Activity‐related Workflow Data

Each time PolicyCenter creates an activity, it adds that activity to the following tables:
• pc_Workflow
• pc_WorkflowLog
• pc_WorkflowWorkItem
After a user completes the activity, PolicyCenter sets the workflow status to completed. PolicyCenter never uses the
pc_Workflow, pc_WorkflowLog and pc_WorkflowWorkItem table entry for that activity again. These tables grow in
size over time and can adversely affect performance and waste disk space. Excessive records in these tables also
negatively impacts the performance of the database upgrade.
Guidewire recommends that you periodically purge workflows, workflow log entries, and workflow items for
completed activities to improve database upgrade and operational performance and to recover disk space.
See also
• “Purging Workflow Data” on page 251
• “Purging Workflow Log Data” on page 252
• “Purging Work Item Set Data” on page 252
Purging Batch Process History Data

Process History Purge batch processing deletes batch process history data from the PolicyCenter ProcessHistory
table. The process uses configuration parameter BatchProcessHistoryPurgeDaysOld to determine the number of
days to retain batch process history before deletion. By default, Guidewire sets the value of
BatchProcessHistoryPurgeDaysOld to 45 days.
In the base configuration, PolicyCenter schedules the ProcessHistoryPurge batch process to run on the third day
of the month at 3:30 a.m.
It is possible to launch Process History Purge batch processing from within PolicyCenter, or, directly from a
command prompt.
PolicyCenter Navigate to the Server Tools Batch Process Info screen and run the Process History Purge batch process.
Command Launch the Purge Workflows batch process from the PolicyCenter/admin/bin directory with the following
prompt command:
maintenance_tools -password password -startprocess ProcessHistoryPurge
The PolicyCenter administration command prompt tools all require that you enter the password of an admin‐
istrative user for the tool to work. The use of a user name is optional.
See also
• “Process History Purge Batch Process” on page 116
Purging Cluster Member Data

Purge Cluster Members batch processing deletes ClusterMemberData entities from the PolicyCenter database. The
process uses Gosu class PurgeClusterMembers, in conjunction with configuration parameter
ClusterMemberPurgeDaysOld, to determine the date on which to delete a ClusterMemberData entity. By default,

Guidewire sets the value of PurgeClusterMembers to 30days.
In the base configuration, PolicyCenter schedules the PurgeClusterMember batch process to run on the first day of
each month, at 2:00 a.m.
It is possible to launch Purge Cluster Members batch processing from within PolicyCenter, or, directly from a
command prompt.
PolicyCenter Navigate to the Server Tools Batch Process Info screen and run the Purge Cluster Members batch process.
prompt command:
maintenance_tools -password password -startprocess PurgeClusterMembers
See also
• “Purge Cluster Members Batch Process” on page 117
Purging Failed Work Items

Purge Failed Work Items batch processing deletes failed work items from all work queues. The process uses Gosu
class PurgeFailedWorkItems to determine which work items to delete.
In the base configuration, PolicyCenter schedules the PurgeFailedWorkItems batch process to run on the first day
of the month, at 1:00 a.m.
It is possible to launch Purge Failed Work Items batch processing from within PolicyCenter, or, directly from a
command prompt.
PolicyCenter Navigate to the Server Tools Batch Process Info screen and run the Purge Failed Work Items batch process.
prompt command:
maintenance_tools -password password -startprocess PurgeFailedWorkItems
See also
• “Purge Failed Work Items Batch Process” on page 117
Purging Job and Policy Period Data

Purge batch processing purges jobs and prunes policy periods that meet the purge and prune criteria from the
PolicyCenter database. This process deletes jobs and other entities from the PolicyCenter database.
In the base configuration, PolicyCenter schedules the Purge work queue to run at 4:30 a.m. every night. However,
Guidewire disables Purge batch processing in the base configuration. To use this batch processing type, you must
first enable it.
It is possible to launch Purge batch processing from within PolicyCenter, or, directly from a command prompt.
PolicyCenter Navigate to the Server Tools Batch Process Info screen and run the Purge batch process.

Command Launch the Purge work queue from the PolicyCenter/admin/bin directory with the following command:
prompt maintenance_tools -password password -startprocess Purge
The PolicyCenter administration command prompt tools all require that you enter the password of an ad‐
ministrative user for the tool to work. The use of a user name is optional.
See also
• “Reset Purge Status and Check Dates Work Queue” on page 123
• “Purge Batch Process” on page 117
Purging Message History Data

Purge Message History batch processing deletes old messages from the PolicyCenter MessageHistory table.
Configuration parameter KeepCompletedMessagesForDays sets the length of time that a message remains in the
message history table before the batch process considers a message for deletion from the database. By default,
Guidewire sets the value of KeepCompletedMessagesForDays to 90 days.
In the base configuration, PolicyCenter schedules the PurgeMessageHistory batch process to run on the 20th of the
month, at 1:00 a.m.
It is possible to launch Purge Message History batch processing from within PolicyCenter, or, directly from a
command prompt.
PolicyCenter Navigate to the Server Tools Batch Process Info screen and run the Purge Message History batch process.
Command Launch the Purge Message History batch process from the PolicyCenter/admin/bin directory with the fol‐
prompt lowing command:
maintenance_tools -password password -startprocess PurgeMessageHistory
See also
• “Purge Message History Batch Process” on page 118
Purging Old Transaction ID Data

Purge Old Transaction IDs batch processing deletes SOAP header transaction IDs generated by systems external to
PolicyCenter. Guidewire does not schedule this batch process in the base configuration as the table that stores the
transaction IDs takes very little space in the database. Unless there is a constant buildup of these transaction IDs,
there is no real need to continually purge this data. In fact, if you do purge this data, it is then not possible to
determine if a new transaction is a duplicate of a transaction sent by the external system at an earlier date.
It is possible to launch Purge Old Transaction IDs batch processing from within PolicyCenter, or, directly from a
command prompt.
PolicyCenter Navigate to the Server Tools Batch Process Info screen and run the Purge Old Transaction IDs batch process.
Command Launch the Purge Old Transaction IDs batch process from the PolicyCenter/admin/bin directory with the
prompt following command:
maintenance_tools -password password -startprocess PurgeTransactionIDs

See also

• “Purge Old Transaction IDs Batch Process” on page 119
Purging Orphaned Policy Period Data

Purge Orphaned Policy Periods batch processing finds orphaned policy periods (policy periods not associated with a
specific job) and deletes them from the PolicyCenter database. The process deletes policy periods and other entities
from the PolicyCenter database.
PolicyCenter does not schedules the PurgeTransactionIDs work queue in the base configuration. You must either
run this process manually or add the work queue to file scheduler-config.xml.
It is possible to launch Purge Orphaned Policy Periods batch processing from within PolicyCenter, or, directly from
a command prompt.
PolicyCenter Navigate to the Server Tools Batch Process Info screen and run the Purge Orphaned Policy Periods batch process.
Command Launch the Purge Orphaned Policy Periods work queue from the PolicyCenter/admin/bin directory with
prompt the following command:
maintenance_tools -password password -startprocess PurgeOrphanedPolicyPeriods
See also

Purging Profiler Data

Purge Profiler Data batch processing deletes profiler data from the PolicyCenter database. This process uses the
read-only ProfilerDataPurgeBatchProcess class, in conjunction with configuration parameter
ProfilerDataPurgeDaysOld, to determine how many days to retain profiler data before deleting it. By default,
Guidewire sets the value of ProfilerDataPurgeDaysOld to 30 days.
PolicyCenter does not schedules the PurgeProfilerData batch process in the base configuration. You must either
run this process manually or add the batch process to file scheduler-config.xml.
It is possible to launch Purge Profiler Data batch processing from within PolicyCenter, or, directly from a command
prompt.
PolicyCenter Navigate to the Server Tools Batch Process Info screen and run the Purge Profiler Data batch process.
Command Launch the Purge Profiler Data batch process from the PolicyCenter/admin/bin directory with the follow‐
prompt ing command:
maintenance_tools -password password -startprocess PurgeProfilerData

See also

• “Purge Profiler Data Batch Process” on page 120
Purging Quote Clones

Purge Quote Clones batch processing deletes quote clones (copies of policy period quotes) from the PolicyCenter
database. This work queue requires that you write your own implementation of the
PolicyPeriodQuoteClonePlugin plugin interface and register it in the Plugins registry.
PolicyCenter does not schedules the PurgeQuoteClones work queue in the base configuration. You must either run
this process manually or add the work queue to file scheduler-config.xml.
It is possible to launch Purge Quote Clones batch processing from within PolicyCenter, or, directly from a command
prompt.
PolicyCenter Navigate to the Server Tools Batch Process Info screen and run the Purge Quote Clones batch process.
Command Launch the Purge Quote Clones work queue from the PolicyCenter/admin/bin directory with the following
prompt command:
maintenance_tools -password password -startprocess PurgeQuoteClones
The PolicyCenter administration command prompt tools all require that you enter the password of an ad‐
ministrative user for the tool to work. The use of a user name is optional.
See also

• “Purge Quote Clones Work Queue” on page 120
Purging Rate Book Export Data

Purge Rate Book Export Result batch processing removes Excel and XML files associated with
RateBookExportResult objects from the PolicyCenter database. Configuration parameter
RateBookExportResultAgeForPurging sets the length of time to retain these files before deletion. By default,
Guidewire sets the value of RateBookExportResultAgeForPurging to 60 days.
In the base configuration, PolicyCenter schedules the PurgeRateBookExportResult work queue to run every day at
3:00 a.m.
It is possible to launch Purge Rate Book Export Result batch processing from within PolicyCenter, or, directly from
a command prompt.
PolicyCenter Navigate to the Server Tools Batch Process Info screen and run the Purge Rate Book Export Result batch process.
Command Launch the Purge Rate Book Export Result batch process from the PolicyCenter/admin/bin directory with
maintenance_tools -password password -startprocess PurgeRateBookExportResult

See also
• “Purge Rate Book Export Result Work Queue” on page 120

Purging Rating Worksheets Data

Purge Rating Worksheets batch processing removes WorksheetContainer objects from the PolicyCenter database.
Configuration parameter RatingWorksheetContainerAgeForPurging sets the minimum number days after the
closure of a job before the WorksheetContainer associated with its policy is eligible for deletion from the database.
By default, Guidewire sets the value of RatingWorksheetContainerAgeForPurging to90 days.
PolicyCenter does not schedules the PurgeWorksheets work queue in the base configuration. You must either run
this process manually or add the work queue to file scheduler-config.xml.
It is possible to launch Purge Rating Worksheets batch processing from within PolicyCenter, or, directly from a
command prompt.
PolicyCenter Navigate to the Server Tools Batch Process Info screen and run the Purge Rating Worksheets batch process.
Command Launch the Purge Rating Worksheets work queue from the PolicyCenter/admin/bin directory with the fol‐
maintenance_tools -password password -startprocess PurgeWorksheets
See also
• “Purge Rating Worksheets Work Queue” on page 121

Purging Temporary Risk Assessment Data

Purge Risk Assessment Temporary Store batch processing deletes temporary objects created for risk assessment
from the PolicyCenter database. Configuration parameter PurgeRiskAssessmentTempStoreDays specifies the
length of time to retain the objects before deletion. By default, Guidewire sets the value of
PurgeRiskAssessmentTempStoreDays to 30 days.
In the base configuration, PolicyCenter schedules the PurgeRiskAssessmentTempStore work queue to run every
Monday at 1:00 a.m.
It is possible to launch Purge Risk Assessment Temporary Store batch processing from within PolicyCenter, or,
directly from a command prompt.
PolicyCenter Navigate to the Server Tools Batch Process Info screen and run the Purge Risk Assessment Temporary Store batch
process.
Command Launch the Purge Risk Assessment Temporary Store batch process from the PolicyCenter/admin/bin direc‐
prompt tory with the following command:
maintenance_tools -password password -startprocess PurgeRiskAssessmentTempStore

See also
• “Purge Risk Assessment Temporary Store Work Queue” on page 121
Purging Temporary Policy Period Data

Purge Temporary Policy Periods batch processing removes PolicyPeriod objects with a status of Temporary from
the PolicyCenter database. Whenever you start a new policy transaction (job) or create a new policy revision,
PolicyCenter creates a PolicyPeriod object. For a short amount of time during object initialization, the policy
period has a status of Temporary. If an error occurs during initialization of the job or policy period, it is possible for
the policy period to remain in the PolicyCenter database with a status of Temporary.
Guidewire disables Purge Temporary Policy Periods batch processing in the base configuration. To enable this batch
process, set configuration parameter PurgeTemporaryPolicyPeriodsEnabled to true.
Configuration parameter PurgeTemporaryPolicyPeriodsAfterDays specifies the length of time to retain the
temporary object before deletion. By default, Guidewire sets the value of
PurgeTemporaryPolicyPeriodsAfterDays to 14 days.
PolicyCenter does not schedules the PurgeTemporaryPolicyPeriods work queue in the base configuration. You
must either run this process manually or add the work queue to file scheduler-config.xml.
It is possible to launch Purge Temporary Policy Periods work batch processing from within PolicyCenter, or,
directly from a command prompt.
PolicyCenter Navigate to the Server Tools Batch Process Info screen and run the Purge Temporary Policy Periods batch process.
Command Launch the Purge Temporary Policy Periods work queue from the PolicyCenter/admin/bin directory with
maintenance_tools -password password -startprocess PurgeTemporaryPolicyPeriods
See also
• “Purge Temporary Policy Periods Work Queue” on page 121
Purging Workflow Data

Purge Workflow batch processing deletes any completed workflows, after resetting any referenced workflows. This
process uses Gosu class PurgeWorkflow, in conjunction with configuration parameter WorkflowPurgeDaysOld, to
determine the number of days to retain workflow data before deletion. By default, Guidewire sets the value of
WorkflowPurgeDaysOld to 60 days, the number of days since the last update to the workflow (the date the workflow
completed).
In the base configuration, PolicyCenter schedules the PurgeWorkflows batch process to run on the first day of each
month, at 1:30 a.m.
It is possible to launch the Purge Workflow batch process from within PolicyCenter, or, directly from a command
prompt.
PolicyCenter Navigate to the Server Tools Batch Process Info screen and run the Purge Workflow batch process.
prompt command:
maintenance_tools -password password -startprocess PurgeWorkflows
See also
• “Purge Workflow Batch Process” on page 122

Purging Workflow Log Data

Purge Workflow Logs batch processing deletes completed workflow logs. It does not remove workflow records,
only workflow log records. The process uses Gosu class PurgeWorkflowLogs, in conjunction with configuration
parameter WorkflowLogPurgeDaysOld, to determine the number of days to retain the workflow logs before deletion.
By default, Guidewire sets the value of WorkflowLogPurgeDaysOld to 30 days.
In the base configuration, PolicyCenter schedules the PurgeWorkflowLogs batch process to run on the first day of
It is possible to launch the Purge Workflow Logs batch process from within PolicyCenter, or, directly from a
command prompt.
PolicyCenter Navigate to the Server Tools Batch Process Info screen and run the Purge Workflow Logs batch process.
Command Launch the Purge Workflow Logs batch process from the PolicyCenter/admin/bin directory with the fol‐
maintenance_tools -password password -startprocess PurgeWorkflowlogs
See also
• “Purge Workflow Logs Batch Process” on page 122

Purging Work Item Set Data

Work Item Set Purge batch processing deletes work item sets from the database. The process uses Gosu class
WorkItemSetPurge, in conjunction with configuration parameter BatchProcessHistoryPurgeDaysOld, to specify
the number of days to retain work item sets before deletion.
In the base configuration, PolicyCenter schedules the WorkItemSetPurge batch process to run on the second day of
It is possible to launch the Work Item Set Purge batch process from within PolicyCenter, or, directly from a
command prompt.
PolicyCenter Navigate to the Server Tools Batch Process Info screen and run the Work Item Set Purge batch process.
Command Launch the Purge Workflow Logs batch process from the PolicyCenter/admin/bin directory with the fol‐
maintenance_tools -password password -startprocess WorkItemSetPurge

See also
• “Work Item Set Purge Batch Process” on page 126
Purging Work Queue Instrumentation Data

Work Queue Instrumentation Purge batch processing deletes instrumentation data for work queues from the
PolicyCenter database. The process uses Gosu class WorkQueueInstrumentationPurge, in conjunction with
configuration parameter InstrumentedWorkerInfoPurgeDaysOld, to determine how long to retain work queue
instrumentation before deletion. By default, Guidewire sets the value of InstrumentedWorkerInfoPurgeDaysOld to
45 days.
In the base configuration, Guidewire schedules the WorkQueueInstrumentationPurge batch process to run on the
second day of the month, at 2:30 a.m.
It is possible to launch the Work Queue Instrumentation Purge batch process from within PolicyCenter, or, directly
from a command prompt.
PolicyCenter Navigate to the Server Tools Batch Process Info screen and run the Work Queue Instrumentation Purge batch proc‐
ess.
prompt command:
maintenance_tools -password password -startprocess WorkQueueInstrumentationPurge
See also
• “Work Queue Instrumentation Purge Batch Process” on page 126


chapter 17
Database Statistics Generation
This topic discusses database statistics, metadata that describe the underlying database.
Understanding Database Statistics

Database statistics are metadata that describe the underlying database. For example, database statistics store row
counts in a table, the distribution of data in the table, and much more. A database management system uses statistics
to determine query plans to optimize performance.
PolicyCenter provides database statistics generation designed specifically for how the PolicyCenter application and
data model interact with the physical database.
Use the Server Tools Database Statistics screen to determine if any statistics are out of date. Guidewire recommends
that you archive database statistics as standard practice. This ensures that you have a record of the database history
that you can review, if it becomes necessary.
IMPORTANT Have your database administrator (DBA) review the database statistics with you.
Database Statistics Generation for Oracle Databases

There are several different ways in which to generate database statistics in Guidewire PolicyCenter if using an
Oracle database:
• Use the Oracle Autotask infrastructure to manage the task of gathering database table statistics.
• Run PolicyCenter batch processing DBStats periodically to collect database table statics.
You enable the use of each method by setting the value of the useoraclestatspreferences attribute on the
<tablestatistics> element in file database-config.xml.
Statistics useoraclestatspreferences Description For more information

gathering
Oracle AutoTask true Disable DBStats batch processing and “Using Oracle AutoTask for
use Oracle AutoTask to manage the Statistics Generation” on page
collection of database statistics. 262
DBStats batch false (Default) Disable Oracle AutoTask and “Database Statistics Work
processing use DBStats to manage the collection of Queue” on page 105
database statistics.
Database Statistics Generation 255

Note: A change in the value of useoraclestatspreferences takes effect only during an application
upgrade.
Disable the automatic generation of database statistics using Oracle by doing one of the following:
• Disable the Oracle AutoTask “auto optimizer stats collection” automated task.
• Set the AUTOSTATS_TARGET preference to ORACLE. This action ensures that the automated task gathers
statistics for the Oracle Dictionary only.
If using DBStats batch processing to manage the collection of database statistics:
• Do not execute Oracle dbms_stats manually.
• Manually execute, or schedule, DBStats batch processing.
See also Disable Automatic Database Statistics Generation by Oracle in the System Administration Guide.
Guidewire recommendations for Oracle database installations

• Guidewire recommends that Oracle implementations only update database statistics during quiet periods, such as
weekends or nights, so that these updates do not occur while PolicyCenter is under heavy load. By default,
updating statistics on a table or index invalidates existing query plans related to that table or index.
• Guidewire recommends that Oracle implementations use the NO_INVALIDATE => AUTO_INVALIDATE option
while updating database statistics. This is the default option. This option is also what the Guidewire Database
Statistics batch process uses, unless the configuration parameter DiscardQueryPlansDuringStatsUpdateBatch
is set to true.
Setting NO_INVALIDATE => FALSE to force immediate invalidation of query plans has a high likelihood of
causing issues with concurrent batch updates. Using AUTO_INVALIDATE greatly reduces this risk. Ideally, set the
_optimizer_invalidation_period parameter to a low value (a few minutes) to reduce the time window during
which Oracle might invalidate a plan.
Disable Automatic Database Statistics Generation by Oracle

About this task
If you do not intend to use the Oracle AutoTask infrastructure to manage the collection of database table statistics,
then disable the generation of database statistics by Oracle before you install Guidewire PolicyCenter. If Oracle
database statistics is enabled after the installation of PolicyCenter, do the following:
Procedure
1. Disable Oracle database statistics generation.
2. Delete the schema statistics.
3. Gather full database statistics using the Guidewire DBStats batch process.
Database Statistics Generation for SQL Server Databases

Guidewire recommends that you only update database statistics during quiet periods, such as weekends or nights, so
that these updates do not occur while PolicyCenter is under heavy load. By default, updating statistics on a table or
index invalidates existing query plans related to that table or index.
Guidewire requires that you ensure the following options are set to true on the SQL Server database used for
PolicyCenter:
• Auto Create Statistics
• Auto Update Statistics
Updates to Database Statistics

Updating database statistics can take a long time on a large database. Collect full statistics only if there are
significant changes to data such as after a major upgrade. Guidewire recommends that you collect full statistics if
256 chapter 17: Database Statistics Generation
using the zone_import command or if there are performance problems. Under standard operating conditions, you
generally need to gather incremental database statistics only on a frequent basis.
If you encounter performance problems or degradation related to the database, check the Database Statistics screen on
the Info Pages section of the Server Tools. If this screen shows suspicious or inaccurate statistics, update database
statistics. If the data change is high, consider using a daily schedule for updating incremental statistics. Use the daily
schedule to provide statistics on tables that have had a configurable percentage of data changed since the last
statistics process was run.
Understanding Database Statistics Batch Processing

PolicyCenter uses Database Statistics batch processing DBStats to generate database statistics. The database
statistics batch process is resource-intensive. To prevent an accidental start of this process, you cannot start Database
Statistics from the Server Tools Batch Process Info screen. Instead, you must start the process manually from a
command prompt.
IMPORTANT Consult with your Database Administrator before starting the database statistics process.
Some PolicyCenter batch processes use work or scratch tables to store intermediate calculations. Other batch
processes populate denormalized tables that PolicyCenter uses internally for performance reasons. These processes
can update database statistics on the scratch tables and denormalized tables during their execution.
Automatic Generation of Database Statistics During Upgrade

PolicyCenter automatically updates specific database statistics during an upgrade, in conjunction with selected batch
processes, or during the zone_import process.
For database upgrades, PolicyCenter updates database statistics for objects that the upgrade process changes
significantly. For optimum performance, generate full database statistics during the next maintenance window after
performing a major upgrade.
Managing Database Statistics using System Tools

You can use the system_tools command options to explicitly update database statistics or to generate the SQL
statements to update statistics. It is possible to update database statics fully or incrementally.
Full database sta‐ Generates database statistics for every table in the PolicyCenter database.
tistics
Incremental data‐ Generates database statistics for tables for which the change in the table data caused by inserts and dele‐
base statistics tes exceeds a certain percentage threshold. You specify this threshold through the
incrementalupdatethresholdpercent attribute on the <databasestatistics> element in file
database-config.xml. The default is 10 percent.
It is possible to pause the database statistics updating process, just as you can with other work queues. Use the
Server Tools Work Queue Info page to pause an in-progress work queue.
See also

Perform a Full Database Statistics Update

Procedure
1. Ensure that the PolicyCenter is running.

2. In a command prompt, navigate to the following location in the PolicyCenter installation directory:
admin/bin
3. Enter the following command to update statistics for all tables.
system_tools -password password -updatestatistics description false
Update Statistics on Tables that Exceed a Change Threshold

About this task
This process does not update statistics on any table that contains locked statistics.
Procedure

admin/bin
3. Enter the following command to update statistics for tables exceeding the change threshold.
system_tools -password password -updatestatistics description true
Check the Database Statistics Updating Process

Procedure

admin/bin
3. Enter the following command to check on the state of the process that updates database statistics.
system_tools -password password -getupdatestatsstate

Cancel the Database Statistics Updating Process

About this task
If you receive the following error on submitting a new database statistics job, it is possible that the Database
Statistics process terminated abruptly:
The Database Statistics process is already in progress
If you receive this error message, you need to manually cancel the process.
Procedure
admin/bin
3. Enter the following command to cancel the process that updates database statistics.
system_tools -password password -cancelupdatestats
Generate Statistics SQL for All Tables

About this task
You can use the results purely as a reference, or you can edit the statements and execute them outside of
PolicyCenter.
Procedure
admin/bin
3. Enter the following command to generate database statistic SQL statements for all tables.
system_tools -password password -getdbstatisticsstatements
Result
PolicyCenter groups the output statements by table.
Generate Statistics SQL for Tables that Exceed a Threshold

Procedure
admin/bin

3. Enter the following command to generate database statistic SQL statements for tables exceeding the change
threshold.
system_tools -password password -getincrementaldbstatisticsstatements
Result
PolicyCenter groups the output statements by table.
Configuring Database Statistics Generation

You control which database statistics statements PolicyCenter generates by configuring the database connection in
the database-config.xml file. You control the number of threads that PolicyCenter uses for the work queue used in
generating database statistics through configuration of file work-queue.xml.
See also
Configuring the Number of Threads for Statistics Generation

Note: This topic is relevant only if using DBStats batch processing to manage the collection of
database statistics.
PolicyCenter uses the work queue mechanism for statistics generation. You can therefore configure work queue and
worker attributes for statistics generation in the work-queue.xml file. Access this file in Guidewire Studio at
configuration→config→workqueue. Set parameters for the work queue with workQueueClass set to
com.guidewire.pl.system.database.dbstatistics.DBStatisticsWorkItemWorkQueue. For example:
<work-queue
workQueueClass="com.guidewire.pl.system.database.dbstatistics.DBStatisticsWorkItemWorkQueue"
progressinterval="86400000">
<worker instances="5" batchsize="10"/>
</work-queue>
In this configuration, the statistics generation work queue uses five worker threads and each worker checks out ten
work items at a time.
If you edit work-queue.xml, you must rebuild and redeploy PolicyCenter.
See also
Statistics and the Database Configuration File

File database-config.xml contains a single <databasestatistics> element, which has the following format.
<databasestatistics samplingpercentage="integer" databasedegree="integer"

incrementalupdatethresholdpercent="integer">
<tablestatistics name="string" samplingpercentage="integer" databasedegree="integer"
action="delete|keep|update">
<histogramstatistics name="string" numbuckets="integer" />
...
</tablestatistics>
...

In the following example, an Oracle database connection shows the use of these parameters:
<databasestatistics samplingpercentage="0" databasedegree="4" incrementalupdatethresholdpercent ="15">

<tablestatistics name="pc_table1" samplingpercentage="0" databasedegree="4">
<histogramstatistics name="scheduledsenddate" numbuckets="254"/>
</tablestatistics>
<tablestatistics name="pc_table2" action="delete" />
The previous example configures the following database statistic generation behavior:
• Collects statistics on all PolicyCenter tables using the automatic sampling size and with a degree of parallelism of
4.
• Samples table pc_table1 using Oracle AUTO_SAMPLE_SIZE, which is the recommended value.
• Uses a parallel degree of 4.
• Defines a histogram with 254 buckets (time slots) on pc_check.scheduledsenddate. Guidewire requires that
you provide a value for this attribute.
• Deletes statistics on pc_table2 due to the attribute action="delete".
See also
The Database Statistics Element

You use the <databasestatistics> element in database-config.xml to specify database statistic parameters that
override the database defaults specified on the database. This element has the following attributes.
databasedegree On Oracle, this attribute controls the degree of parallelism for each individual state‐
ment. The default is 1. PolicyCenter uses the value of this attribute for all state‐
ments.
SQL Server ignores the databasedegree attribute.
incrementalupdatethresholdpercent Specifies the percentage of table data that must have changed since the last statis‐
tics process for the incremental statistics generation batch process to update statis‐
tics for the table.
The default is 10.
numappserverthreads On both Oracle and SQL Server, the numappserverthreads attribute controls the
number of threads that PolicyCenter uses to update database statistics for staging
tables during import only. Command prompt tool table_import launches this im‐
port.
The value defaults to 1. If the value is greater than 1, then the PolicyCenter server
assigns a table at a time to each thread as the thread becomes available. Each
thread executes all of the database statistics statements for its assigned table.
For all other statistics generation operations, set the number of threads by specify‐
ing the number of workers for the database statistics work queue. Set the
instances attribute on the <workers> subelement of the <work-queue> element
for the database statistics work queue. This element has
workQueueClass="com.guidewire.pl.system.database.dbstatistics.DBStati
sticsWorkItemWorkQueue".
The default is 1.
samplingpercentage On Oracle, this attribute controls the value of the estimate_percent parameter in
the dbms_stats.gather_table_stats() SQL statements. You can set
samplingpercentage to an integer from 1 to 100 to directly set the
estimate_percent value. However, Guidewire recommends highly that you set the
samplingpercentage value to 0 to set estimate_percent to AUTO_SAMPLE_SIZE.
The default value is 0.
On SQL Server, the samplingpercentage attributes controls the value of the WITH
FULLSCAN/SAMPLE PERCENT clause in the UPDATE STATISTICS statements. A value
of 100, the default, translates into WITH FULLSCAN, as does a value of 0.
The default is 0.
useoraclestatspreferences On Oracle, this attribute sets the database statistics preferences to be able to use
the Oracle Autotask infrastructure instead of the DBStats batch process from
PolicyCenter. The default is false, which requires that you disable the Autotask
and schedule DBStats batch processing in its place. Changes to the value of this
attribute only take effect during an application upgrade.
The values you set for these attributes apply to all the tables in the database. You can fine tune these values and set
specific values on individual tables by using the <tablestatistics> subelement. Setting values on a specific table
overrides the values set on the database for just that table.
See also
• “Table Import Command” on page 399
Using Oracle AutoTask for Statistics Generation

The purpose of the useoraclestatspreferences attribute on the <databasestatistics> element in file
database-config.xml is twofold:
• To set the table statistics preferences according to those set for Oracle in file database-config.
• To have the Oracle AutoTask infrastructure manage the task of collecting table statistics, based on preferences
that Guidewire sets in the base configuration.
To set Oracle to collect table statistics, set this attribute value to true, for example:
<database name="ExampleDatabase" dbtype="oracle" env="oracle" printcommands="true">

<databasestatistics databasedegree="4" useoraclestatspreferences="true">
...
Any change to this attribute value takes effect only during an upgrade, either a full upgrade or a rolling
(configuration) upgrade. To force PolicyCenter to recognize the change without an application upgrade, increment
the application metadata version and restart the application server.
After you set this attribute to true, the next application upgrade does the following:
• It clears all existing preferences for table statistics.
• It resets the preferences for table statistics to those currently defined for Oracle table statistics in database-
config.xml.
• It creates a new tab named Oracle Statistics Preferences in the Server Tools Info Pages→Database Catalog Statistics
Information screen.
To confirm that the application upgrade set the database statistics preferences, review the Oracle Statistics Preferences
tab and verify the preferences.
Enabling Oracle AutoTask

After you set this parameter to true for the first time, do the following:
• Ensure that the statistics target setting has the default value of AUTO.
• Ensure that you enable Oracle AutoTask for statistics collection.
The Oracle DBA can run the following SQL commands to set these value back to their default.
EXEC dbms_stats.set_global_prefs('AUTOSTATS_TARGET','AUTO');
EXEC dbms_auto_task_admin.enable(client_name => 'auto optimizer stats collection', operation => NULL,
window_name => NULL);

Gathering new statistics

If you enabled the database before you set useoraclestatspreferences to true, you need to delete the existing
schema statistics and gather new statics. The Oracle DBA can use the following SQL commands for this task.
EXEC dbms_stats.delete_schema_stats('PCUSER');
EXEC DBMS_STATS.GATHER_SCHEMA_STATS(ownname=>'PCUSER', options=>'GATHER');
Note: Replace PCUSER with the actual PolicyCenter database user.
After the application upgrade

After you perform an application upgrade with useoraclestatspreferences set to true, there is no longer any
need to run DBStats batch processing. At this point, Oracle AutoTask automatically handles statistics collection
during the maintenance windows defined for the database. To disable DBStats batch processing, remove its schedule
from scheduler-config.xml.
Future application upgrades

For future application upgrades, Guidewire recommends that you consider setting the updatestatistics attribute
on the <upgrade> element in database-config.xml to true, for example:
<upgrade degree-parallel-ddl="6" degree-of-parallelism="6" ora-parallel-dml="enable_all"

updatestatistics="true">
Setting the updatestatistics attribute to true allows the PolicyCenter upgrader to create any additional
histograms required by the new application version.
For more information, refer to the following Oracle documentation:
• Oracle Database Administrator’s Guide, "Managing Automated Database Maintenance Tasks"
• Oracle Database SQL Tuning Guide, "Managing Optimizer Statistics: Basic Topics"
Resetting useoraclestatspreferences
Any change to the useoraclestatspreferences attribute in database-config.xml (from false to true or from
true to false) takes effect only after an upgrade, either a full upgrade or a rolling (configuration) upgrade.
However, if you reset this attribute from true to false, PolicyCenter throws an exception during the next upgrade
and prevents the upgrade from continuing due to locked table statistics in the Oracle database. Review the details of
the exception provided in the server log to determine which table statistics need to be unlocked. See “Revert to
DBStats Batch Processing for Database Statistics” on page 263 for information on how to unlock the table statistics.
See also
• “The Oracle Statistics Preferences Tab” on page 345
Revert to DBStats Batch Processing for Database Statistics

You must perform a sequence of manual steps if you want to revert to using the table statistics preferences set in
database-config.xml rather than using Oracle AutoTask infrastructure to manage the task of collecting table
statistics. Any change to the useoraclestatspreferences attribute in database-config.xml (from false to true
or from true to false) takes effect only after an upgrade, which can be either a full database upgrade or a rolling
configuration update. However, if you attempt to reset this attribute from true to false, PolicyCenter throws an
exception during the next upgrade and prevents the upgrade from continuing due to locked statistics in certain
Oracle database tables. Review the exception log detail to determine which tables statistics need to be unlocked.
Procedure
1. Reset attribute useoraclestatspreferences to false in file database-config.
2. Start the application server in upgrade mode.
The upgrade fails due to locked table statistics in the Oracle database.
3. Review the server log for which table statistics need to be unlocked.
Search for text that is similar to the following:
com.guidewire.pl.system.exception.UpgradeException: The following tables have locked statistics...
4. Unlock the specified table statistics.

The following example code illustrates what SQL commands the BillingCenter database DBA needs to
execute.
SET serverout on timing on

DECLARE
CURSOR locked_tables_cur IS SELECT table_name
FROM user_tab_statistics
WHERE STATTYPE_LOCKED IS NOT NULL;
str VARCHAR2(256);
BEGIN
FOR locked_tables IN locked_tables_cur
LOOP
str:= 'EXEC DBMS_STATS.UNLOCK_TABLE_STATS(USER, ''' || locked_tables.table_name|| ''')' ;
dbms_output.put_line(str);
EXECUTE IMMEDIATE str;
END LOOP;
END;
/
This step is mandatory. Otherwise, the PolicyCenter upgrade fails.

5. Delete the current table statistics preferences. The following example code illustrates what SQL commands
the BillingCenter database DBA needs to execute.
EXEC DBMS_STATS.DELETE_SCHEMA_PREFS('PCUSER', 'NO_INVALIDATE');

EXEC DBMS_STATS.DELETE_SCHEMA_PREFS('PCUSER', 'CASCADE');
EXEC DBMS_STATS.DELETE_SCHEMA_PREFS('PCUSER', 'STALE_PERCENT');
EXEC DBMS_STATS.DELETE_SCHEMA_PREFS('PCUSER', 'ESTIMATE_PERCENT');
EXEC DBMS_STATS.DELETE_SCHEMA_PREFS('PCUSER', 'DEGREE');
EXEC DBMS_STATS.DELETE_SCHEMA_PREFS('PCUSER', 'METHOD_OPT');
Note: Replace PCUSER with the actual PolicyCenter database user.

6. Restart the application server in upgrade mode.
7. After the application upgrade completes, schedule the execution of DBStats batch processing to collect
database table statistics.
The Table Statistics Database Element

The <databasestatistics> element in database-config.xml has one subelement, <tablestatistics>. You use
this element to override database-wide statistics settings defined on the <databasestatistics> element for a
specific table. You can override the databasedegree (Oracle only), samplingpercentage attributes, and statistic
gathering behavior of PolicyCenter. You must provide a name parameter to identify the table for which you want to
set values:
<tablestatistics name="string" samplingpercentage="integer" databasedegree="integer"

action="update|delete|keep|force"/>
By default, PolicyCenter on Oracle does not generate statistics on any table used for processing work items.
PolicyCenter deletes any existing statistics on these tables whenever it updates statistics. You can override this
behavior by using the action attribute of the <tablestatistics> element. You can set the action attribute to one
of the following values:
delete Delete the statistics on the table. This value does nothing in SQL Server.

force Update statistics for this table while running incremental statistics, regardless of the value of attribute
incrementalupdatethreshold on the <databasestatistics> element.
keep Keep the existing statistics. PolicyCenter does not update statistics for any table for which the user explicitly specifies
keep as the value for the action attribute. This value affects any type of database.
update Update the statistics for the table:

• Full statistics update ‐ Always update
• Incremental statistics update ‐ Update only if the change in table data caused by inserts and deletes exceeds the
value of attribute incrementalupdatethreshold on the databasestatistics> element.
The default value is update.

The <tablestatistics> element is optional. If you do not specify a <tablestatistics> element for a table,
PolicyCenter uses the database-wide statistics defined on the <databasestatistics> element. If you do specify a
<tablestatistics> element, you must also supply a value for the action attribute.
See also
• “The Histogram Statistics Database Element” on page 265
The Histogram Statistics Database Element

The <tablestatistics> element in database-config.xml has several subelements, one of which is the
<histogramstatistics> element. Use this element to specify a column-specific value for the databasedegree
attribute (ignored on SQL Server) and the samplingpercentage attributes. By default, PolicyCenter issues a single
dbms_stats.gather_table_stats(... 'FOR COLUMNS ...') statement for all columns of interest in the table,
including:
• All columns that are the first key column of an index. (Oracle only).
• The retired column, if present.
• The subtype column, if present.
• All columns that have the createhistogram attribute set to true. (Guidewire sets this value internally.)
If you specify non-default values for either the databasedegree or the samplingpercentage on a particular
column, PolicyCenter issues a separate statement for those values alone.
The <histogramstatistics> element has the following format:
<histogramstatistics name="string" numbuckets="integer" />
The name attribute specifies a column name. The numbuckets attribute controls the maximum number of buckets for
the specified histogram. Guidewire requires that you provide a value for this attribute. The default value for the
number of buckets is 254 for the retired and subtype columns. For all other columns, PolicyCenter uses 75, the
database default.
Notes
• For performance reasons, PolicyCenter does not currently create a histogram on publicid columns. These
columns are rarely, if ever, referenced in a WHERE clause.
• Also for performance reasons, PolicyCenter tries to combine as many columns as possible into a single
statement. Certain tabs in the Database Catalog Statistics page display a
dbms_stats.gather_table_stats(...'FOR COLUMNS ...') statement with only the associated column for
each histogram, regardless of the parameter values. This enables you to specify the most granular statement if a
given histogram is out of date.
See also

chapter 18
Importing and Exporting

Administrative Data
Guidewire categorizes certain types of application data as administration data. (Guidewire frequently shortens this
tem to just admin data.) For example, activity patterns are administration data. This topic provides information
related to importing administrative data into, and exporting data from, Guidewire PolicyCenter. While users enter
much of the information into PolicyCenter directly, at times, it is more convenient or necessary to enter information
in bulk.
Ways to Import Administrative Data

It is possible to import administrative data into Guidewire PolicyCenter by using the following mechanisms.
Mechanism How Description

PolicyCenter Administration tab Import and export administrative data files in XML format by using functionality available
from the PolicyCenterAdministration tab. The import functionality provides warnings on col‐
lisions between incoming data and existing data and provides a mechanism for you to re‐
solve the collisions.
For more information, see:
• “Importing and Exporting Administrative Data from PolicyCenter” on page 278
File import import_tools Use the import_tools command prompt tool to import one or more CSV or XML files
ImportToolsAPI containing administrative data. Use this command to import administrative data only.
Guidewire does not support using the import_tools command to import other types of
data.
It is also possible to use the ImportToolsAPI web service to import one or more XML files
containing administrative data.
For more information, see:
• “Import Tools Command” on page 388
• “Using Tools to Import Administrative Data” on page 277
File load import directory Load data from CSV files contained in the following directory:
configuration→config→import→gen
At initial database upgrade, starting from an empty database, PolicyCenter loads the ad‐
ministrative data contained in the files in this directory. These files include the default
activity patterns, non‐renewal question templates, and roles and role privileges.
Importing and Exporting Administrative Data 267

Mechanism How Description

You can add additional files to this directory for load at initial server startup. For details,
see “About the import Directory” on page 268.
IMPORTANT PolicyCenter loads files from the import→gen directory only if the database is
empty, during the database upgrade at initial server start.
File load Security directory Load security zone data from file security-zones.xml, contained in the following direc‐
tory:
configuration→config→Security
For information on the syntax to use with this XML‐formatted file, see “Importing Security
Zone Data” on page 280.
IMPORTANT PolicyCenter loads data from file security-zones.xml only if the database is
empty, during the database upgrade at initial server start.
Staging tables table_import Use the table_import command prompt tool to import administrative data from staging
tables into PolicyCenter database tables. For details, see “Table Import Command” on
page 399.
PolicyCenter supports bulk data import from staging tables for loading zone data only.
IMPORTANT Never modify PolicyCenter operational database tables directly with SQL commands.
About the import Directory

After the initial PolicyCenter installation, the database is empty of data. The first time that you start the PolicyCenter
application server after installation, PolicyCenter upgrades the database and populates it with data. As part of this
initial upgrade, PolicyCenter automatically loads the following administrative data:
• Root group of the user and group tree
• Initial security zone
• Default activity patterns
• Default roles and role permissions (privileges)
PolicyCenter can also import default business rules on initial server startup, depending on the value of configuration
parameter BizRulesImportBootstrapRules.
Administration Data Import at Initial Server Startup

During the initial server startup with an empty database, PolicyCenter uses the data defined in the following files to
load the default activity patterns, roles, and role permissions:
• activity-patterns.csv
• roleprivileges.csv
• roles.csv
These files exist in the following location in Guidewire Studio:
configuration→config →import →gen
It is possible to modify but not remove these files. However, at a bare minimum, these files must define the super
user role (superuser) and its privileges in the base configuration. Otherwise, it is not possible to start PolicyCenter
the first time.
Security Zone Data Import at Initial Server Startup

During the initial server startup with an empty database, PolicyCenter loads the security zone data defined in file
security-zones.xml. The use of this file is optional. You cannot use this mechanism to load security zone data
after the initial server startup to populate the database with data.
File security-zones.xml exists in the following directory:
268 chapter 18: Importing and Exporting Administrative Data
configuration→config →Security
See also
• “Importing Security Zone Data” on page 280
Business Rules Import at Initial Server Startup

During the initial server startup with an empty database, it is possible for PolicyCenter to automatically load
business rules defined in the following location in PolicyCenter Studio:
configuration→config →import →bizrules
Configuration parameter BizRulesImportBootstrapRules controls this behavior:
• If the value of BizRulesImportBootstrapRules parameter is true, PolicyCenter imports the business rules
defined in this location.
• If the value of BizRulesImportBootstrapRules is false, PolicyCenter ignores any business rules in that
location.
IMPORTANT Guidewire recommends that you set the value of this configuration parameter to true in
a non-production test environment only.
See also
• “Automatic Import of Business Rules at Server Startup” on page 315
Controlling the Import of Data During Server Startup

During server startup, PolicyCenter examines file importfiles.txt to determine if it lists additional administrative
data files to load. This file exists in the following location in PolicyCenter Studio:
configuration→config →import →gen
In the base configuration, file importfiles.txt contains the name of one file to load, that of
nonRenewalExplanationPatterns.csv.
If you want to load additional administrative data at initial database upgrade, do the following:
1. Place the appropriately formatted CSV files containing the data in the gen folder in Studio.
The file format must be CSV.
2. Add the filename to the list of files in importfiles.txt.
Any file named in importfiles.txt must exist in the gen folder or the database upgrade fails.
PolicyCenter loads the data in sequential order of the files listed in file importfiles.txt.
About Adding Admin Data after Initial Server Startup

In general, it is not possible to import additional administrative data contained in the files in the import→gen
directory after the initial database upgrade. PolicyCenter loads the administrative data contained in the gen folder
only if starting from an empty (non-upgraded) database. Thereafter, PolicyCenter ignores the files in the gen
directory.
There is one exception. It is possible to add the data in file roleprivileges.csv in the gen folder using an
import_tools command option. This command option adds the privileges associated with each PolicyCenter role in
the roleprivileges.csv file to those that already exist in the database.
See also
• “Add Additional Roles and Privileges after Initial Server Startup” on page 270
Add Additional Roles and Privileges after Initial Server Startup

Procedure
admin/bin
4. Run the following command to import the data in file roleprivileges.csv in the import→gen folder:
import_tools -password password -privileges
See also
• “Character Set Encoding for File Import” on page 270
Character Set Encoding for File Import

In the base configuration, if you do not supply a value for the import_tools -charset option, PolicyCenter
assumes a character set encoding of UTF-8 for the import file. It is possible that the file that you want to import
contains characters not recognized by the import tools default character encoding.
If this is the case, then you need to set the character set encoding to a more appropriate value. For example, for a file
that contains single-byte data as accented characters or characters with umlauts, a -charset value of ISO-8859-1 is
possibly more appropriate.
See also
Maintain Data Integrity During Administrative Data Import

About this task
Some PolicyCenter administrative data has dependencies on business roles. For example, PolicyCenter associates
roles with groups. Therefore, if you export administrative data from one system into another, you must also export
all PolicyCenter roles from the old system and import the roles into the new system.
Procedure
1. Export the PolicyCenter roles.
2. Export the administration data.
3. Import the roles into the new system.
4. Import the administration data into the new system.
Administrative Data and the PolicyCenter Data Model

To import or export data from PolicyCenter, you must understand its data model. In particular, you must understand
how PolicyCenter uses each user interface field and how that use maps to the database. To build this understanding,
review the PolicyCenter Data Dictionary, accessible in the following location in the PolicyCenter installation
directory:
build/dictionary/data/index.html
The Data Dictionary describes the structure of business objects stored persistently by PolicyCenter, and the
dictionary defines the properties and foreign key references for each object. If you change the data model, regenerate
the PolicyCenter Data Dictionary by using the following command:
gwb genDataDictionary
See also
Public ID Prefix
Each entity that you import into PolicyCenter requires a unique public ID. This is separate from the system ID that
PolicyCenter assigns internally and uses for most system processing. Foreign key references between related objects
use this public ID.
Typically, a company imports data from multiple external sources. If you do import data from multiple sources, use
a naming convention to generate public IDs for external sources. For example, if you import from two systems
(Adminsystem and Salessystem), each source can have a contact entity with ID=5432. Thus, Guidewire
recommends that you use the following ID format so that the IDs do not register as duplicates:
origin:ID
By using this format, the contact from the first system comes in as adminsystem:5432 and the contact from the
second system comes in as salessystem:5432. Thus, there is no risk of duplicate IDs. There is also the benefit of
knowing from which system the record originated.
Public IDs need to be unique only within objects of the same type. For example, all policy objects must have a
different public ID. However, an account and a policy with the same public ID do not conflict with each other.
Public IDs cannot exceed 20 characters in length.
Support for Unique Public IDs in a Development Environment

During development and testing of a PolicyCenter configuration, multiple developers can create administrative data
in their own PolicyCenter installations. Administrative data includes users, groups, roles and so forth. You combine
this data into a single production database at the end of the development cycle.
If you ever want to transfer data from one database to another with the export and import utilities, the two databases
must have different public ID prefixes. Set the public ID prefix by modifying the value of the following parameter
in config.xml:
<param name="PublicIDPrefix" value="pc"/>
PolicyCenter appends this public ID prefix to each administrative entity created in an individual PolicyCenter
configuration. Thus, the public ID format becomes the following:
{PublicIDPrefix}:{ID}
If you have multiple developers, you must coordinate the public ID prefix so that no public ID prefixes overlap. For
example, each engineer can use their initials or computer names as a public id prefix.
You can use the same prefix for multiple development and testing databases if you do not ever transfer data between
them.

Constructing a CSV File for Import

You can only import data for an entity type that already exists in the Guidewire data model. Each CSV-formatted file
you import must have a heading that defines what the file contains and how to interpret it. The following example
illustrates a very simple import file:
1: ADDRESS
2: type,data-set,entityid,addresstype,addressline1,createuser
3: Address,0,ab:1001,home,1253 Paloma Ave.,import_tools
4: Address,0,ab:1002,business,325 S. Lake Ave.,import_tools
The import_tools command distinguishes between two types of information in an import file:
• Heading information
• Data information.
PolicyCenter treats any line that contains the string entityid as a heading.
PolicyCenter considers as data any line:
• Without an entityid string
• With comma delimited values
• With a value in its third comma-delimited field
In the example shown, the import_tools command treats line 2 as a heading and lines 3 – 4 as data. The
import_tools command ignores line 1. If the command encounters a data line before a heading line, it returns an
error.
IMPORTANT Guidewire supports using the import_tools command to import administrative data
only.
Constructing a Heading Line in CSV‐formatted Files

Every CSV-formatted file that you import into PolicyCenter must have a heading line. The heading line initializes
the import for a particular entity object in the data model. A heading consists of comma-separated fields. The first
three fields must be, in order:
• type
• data-set
• entityid
All subsequent fields must refer to columns, typelists, foreign keys, and joinarrays on the entity.
For example, a file importing into the Address entity has a heading that appears as:
type,data-set,entityid,addresstype,addressline1,createuser
The following fields follow the three required fields (type, data-set, and entityid):
• addresstype – Represents a typelist
• addressline1 – Represents a column
You do not have to specify all fields in the entity within the import file. You must specify at least the required fields.
You can determine which fields PolicyCenter requires by viewing the entity description in the PolicyCenter Data
Dictionary.
Use lowercase to specify fields, including arrays. In this example, specify AddressLine1 in the data model as
addressline1 in the import file.
To specify a foreign key, use the foreign key name without the concluding ID. In this example of a Person import:
1: type,data-set,entityid,firstname,lastname,primaryaddress,workphone,primaryphone,
taxid,vendortype,specialtytype
2: Person,0,demo_sample:1,Ray,Newton,demo_sample:4000,818-446-1206,work,,,
3: Person,0,demo_sample:2,Stan,Newton,demo_sample:4002,818-446-1206,work,,,
4: Person,0,demo_sample:3,Harry,Shapiro,demo_sample:1004,818-252-2546,work,,,

5: Person,0,demo_sample:4,Bo,Simpson,demo_sample:1003,619-275-2346,work,,,
6: Person,0,demo_sample:5,Jane,Collins,demo_sample:4003,213-457-6378,work,,,
7: Person,0,demo_sample:6,John,Dempsey,demo_sample:1006,213-475-9465,work,,,
The primaryaddress field is a foreign key to the Address entity. It appears as PrimaryAddressID in the
PolicyCenter Data Dictionary but as primaryaddress in the import data.
If you specify a field in the heading that is not a recognizable column, typelist, foreign key or array, the import
program silently ignores the column and any associated data. In the following example, the import ignores the %$zed
heading field and the somedata value in line 3:
1: ADDRESS
2: type,data-set,entityid,addresstype,addressline1,createuser, %$zed
3: Address,0,ab:1001,home,1253 Paloma Ave.,import_tools, somedata
4: Address,0,ab:1002,business,325 S. Lake Ave.,import_tools,
Constructing Data Lines in CSV‐formatted Files

Every CSV-formatted file that you import into PolicyCenter must have a heading line, followed by one or more data
lines. The import_tool identifies data lines by scanning the file for lines with the following characteristics:
• The line does not contain the three required heading fields.
• The line contains comma-delimited values.
• The line contains a third field that is non-empty.
Each data line represents a single instance of a data model entity.
Data Line ‐ Field 1
The first field in any data line must be an entity name or an entity subtype name.
1: Policy
2: type,data-set,entityid,account,corepolicynumber,policytype,producttype,productversion,
systemofrecorddate,,,,,,,,,,,
3: Policy,0,ds:1,ds:1,34-123436-CORE,wc,wc_workerscomp,1,1/1/2002,,,,,,,,,,,
4: Policy,0,ds:2,ds:1,25-123436-CORE,bop,bop_businessowners,1,1/1/2002,,,,,,,,,,,
5: Policy,0,ds:3,ds:3,54-123456-CORE,personalauto,pa_personalauto,1,1/1/2002,,,,,,,,,,,
8: Policy,0,ds:6,ds:1,20-123436-CORE,businessauto,ba_businessauto,1,1/1/2002,,,,,,,,,,,
9: Policy,0,ds:7,ds:1,50-123436-CORE,umbrella,u_umbrella,1,1/1/2002,,,,,,,,,,,
...
In lines 3 - 9, the entity name Policy appears in the first field as required. The capitalization of an entity or subtype
name must be identical to that used in the PolicyCenter Data Dictionary. For example, to create a RevisionAnswer
data line the entry name would be invalid if you specified it as revisionanswer.
The second field in a data line is the value of the highest-numbered data-set of which the imported object is part.
2: type,data-set,entityid,account,corepolicynumber,policytype,producttype,productversion,
systemofrecorddate,,,,,,,,,,,
3: Policy,0,ds:1,ds:1,34-123436-CORE,wc,wc_workerscomp,1,1/1/2002,,,,,,,,,,,
PolicyCenter orders data-sets by inclusion. Thus, data-set 0 is a subset of data-set 1 and data-set 1 is a subset of data-
set 2, and so forth. It is possible to request a particular data-set while converting CSV to XML. By default,
PolicyCenter requests data-set 10240. PolicyCenter assumes that data-set 10240 includes every data-set that it is
possible to create.
You can leave the second field blank, in which case PolicyCenter always includes this object in the import
regardless of the requested data-set.

The third field in any data line must be the public ID for that particular data object. This field is mandatory. For
example, ds:2 is the public ID of the Policy on line 4.
Foreign Key and Column Data

The import_tools command imports both column and typelist data values from the CSV file. In the previous
example, the policytype column has a value of wc in line 3 and a value of bop in line 4. You represent foreign key
data by a string in one of two formats:
publicID
or
entity_id:identity_source
If there is more than one : (colon), the import ignores everything after the second : (colon).
1 ADDRESS
2 type,data-set,entityid,addresstype,addressline1,createuser
3 Address,0,ab:1001,home,1253 Paloma Ave.,import_tools
4 Address,0,ab:1002,business,325 S. Lake Ave.,import_tools
6
7 PERSON
8 type,data-set,entityid,firstname,lastname,primaryaddress,inaddressbook,loadrelatedcontacts,
referred,contactaddresses
9 Person,0,ab:2001,John,Foo,ab:1002,true,true,true,ContactAddress|address[ab:10001,ab:1002]
10 Person,0,ab:2002,Paul,Bar,ab:1002,false,false,false,ContactAddress|address[ab:10001]
11 Person,0,ab:2003,David,Goo,,false,true,false,,
In the previous example, the primaryaddress on line 9 is a foreign key to the Address specified on line 4.
If PolicyCenter cannot resolve a foreign key reference and does not require the foreign key, PolicyCenter imports
the data, sets the foreign key field to null, and reports an error. If PolicyCenter does require the foreign key, then
PolicyCenter reports an error and does not import that data.
Simple Array Data

You specify simple array data, referencing a single foreign key by using the following format:
arraykey|foreignkey[publicID,publicID,...]
In the PERSON example (line 9), the arraykey value is the array key on the parent entity (Person). The foreignkey
is the foreign key name of the array without the ID. ContactAddress is the array key and address is the foreign key
name. The public ID values [publicID,publicID,...] correspond to public IDs that are referenced by the foreign
key.
In this format, the arraykey is optional. However, you might want to retain it for readability.
Complex Array Data

You might need to specify more complex arrays that have a mixture of data types. If you specify arrays that contain
a mixture of columns, foreign key data, or typelists, use a different format. The basic format of these complex array
entries appear as follows:
[ [array_entry];[array_entry]; ...]
Enclose each array_entry in brackets. Separate multiple entries with semicolons. Enclose all completed entries in a
second set of brackets. Each array_entry is made up of comma-separated [type|value] pairs as follows:
[[[type|value],[type|value]];[[type|value],[type|value]]]
The type is the name of a column, typelist, or foreign key, as in a heading line. The value is the column value,
typelist typecode, or a foreign key. In the following sample, there are three array_entry specifications, the first
and last array_entry specifications appear in bold:
Group
type,data-set,entityid,users
Group,0,demo_sample:27,[[[user|demo_sample:101],
[loadfactor|50],[loadfactortype|loadfactorview]];[[user|demo_sample:102],
[loadfactor|100],[loadfactortype|loadfactoradmin]];
[[user|demo_sample:103],[loadfactor|50],[loadfactortype|loadfactorview]]]
Construct an XML File for Import

About this task
To create a properly formatted XML file of administrative data for import, Guidewire recommends that you perform
the following sequence of steps:
Procedure
1. First, export the current administrative data as an XML file by using the functionality available on the Export
Data screen available to PolicyCenter administrators. This screen provides the ability to export various types of
administrative data in XML format.
2. Modify the file to suit your business needs, carefully preserving the XML formatting for administrative data.
3. Regenerate the XSD files by using the following command in the PolicyCenter installation directory:
gwb genImportAdminDataXsd
It is important to regenerate the XSD files every time that you modify the data model.
4. Re-import the modified file by using either the import_tools command or the Export Data screen available to
PolicyCenter administrators. This screen provides the ability to import administrative data in XML format into
PolicyCenter.
See also
• “Importing and Exporting Administrative Data from PolicyCenter” on page 278
• “Constructing the XML for the Administrative Data Import File” on page 275
Constructing the XML for the Administrative Data Import File

The PolicyCenter/build/xsd/pc_import.xsd file defines the XML schema used for import and export of
administrative data. This file further references information in two other XSD files in the same directory:
• pc_entities.xsd
• pc_typelists.xsd
You can use any schema-aware XML editor to help format information properly according to these XSD definitions.
You generate these XSD files by entering the following command in a command prompt open in the PolicyCenter
installation folder:
Regenerate the XSD files any time you modify the data model. These files are likely to change as you configure the
data model.
The following XML example shows the default activity pattern 30 Day Diary from the PolicyCenter administrative
export file.
<?xml version="1.0"?>
<import xmlns="http://guidewire.com/pc/exim/import" version="p5.86.a12.309.46"
usePeriodicFlushes="true">
<ActivityPattern public-id="sample_pattern:19">

<ActivityClass>task</ActivityClass>
<AutomatedOnly>false</AutomatedOnly>
<Category>reminder</Category>
<Code>30_day_diary</Code>
<Command/>
<Description/>
<Description_L10N_ARRAY/>
<DocumentTemplate/>
<EmailTemplate/>
<EscBusCalLocPath/>
<EscalationBusCalTag/>
<EscalationDays/>
<EscalationHours/>
<EscalationInclDays/>
<EscalationStartPt/>
<Mandatory>false</Mandatory>
<PatternLevel>All</PatternLevel>
<Priority>normal</Priority>
<Recurring>true</Recurring>
<ShortSubject/>
<ShortSubject_L10N_ARRAY/>
<Subject>30 day diary</Subject>
<Subject_L10N_ARRAY/>
<TargetBusCalLocPath/>
<TargetBusCalTag/>
<TargetDays>30</TargetDays>
<TargetHours/>
<TargetIncludeDays>elapsed</TargetIncludeDays>
<TargetStartPoint>activitycreation</TargetStartPoint>
<Type>general</Type>
</ActivityPattern>
</import>
You can:
• Modify any existing entry in the administrative export file and re-import the file.
• Add additional entries by using the existing entries as a model and re-import the file.
Foreign Key References

Within an XML-formatted administrative data import file, it is common to have references between objects in the
file. For example, a user object can reference the group of which it is a member. As the group definition is elsewhere
in the XML file, or perhaps the definition exists elsewhere, the user definition refers to this group with a foreign key.
The foreign key is the object’s public ID. For example, the XML file can contain:
<User publicID="demo_sample:100"> … </User>

…
<Group publicID="demo_sample:200">
…
<Users>
<GroupUser>
<User publicID="demo_sample:100" />
…
</GroupUser>
</Users>
</Group>
In this example, the user demo_sample:100 is a member of group demo_sample:200.

It is possible to reference an item within a single XML file before you define the item. For example, you can refer to
supervisor of a user before you define the supervisor in the file. This enables you to define all groups first.
PolicyCenter reports errors only if a referenced object still does not exist after reading the entire file.
Validating the Import XSD File

It is important to regenerate the PolicyCenter XSD files after you modify the data model. These files are likely to
change as you configure the data model. You generate these XSD files by opening a command prompt in the
PolicyCenter installation directory and entering the following command:

You can validate the XML of your import file against a pc_import.xsd file by using the following code:
uses java.io.File
uses java.io.FileInputStream
uses javax.xml.validation.SchemaFactory
uses javax.xml.XMLConstants
uses javax.xml.parsers.SAXParserFactory
uses org.xml.sax.helpers.DefaultHandler
uses gw.testharness.TestBase
// Provide the correct directory of the pc_import.xsd

var schemaFile = new File("pc_import.xsd")
TestBase.assertTrue(schemaFile + " Should exists", schemaFile.exists());
var factory = SchemaFactory.newInstance(XMLConstants.W3C_XML_SCHEMA_NS_URI);
var schema = factory.newSchema(schemaFile)
var spf = SAXParserFactory.newInstance()
spf.setSchema(schema)
spf.Validating = true
spf.NamespaceAware = true
var parser = spf.newSAXParser();
var fis = new FileInputStream("myImportFile.xml")
parser.parse(fis, new DefaultHandler())
Using Tools to Import Administrative Data

Guidewire provides the import_tools command for importing new or updated administrative data into existing
tables in the PolicyCenter database. PolicyCenter reads the import data from a comma-separated values (CSV) file
or an XML file.
IMPORTANT PolicyCenter supports this command for importing administrative data, but not for
importing other types of data.
PolicyCenter uses the public ID of each object in the data import to determine if an object with that public ID
already exists in the database. See “Administrative Data and the PolicyCenter Data Model” on page 270 for a
discussion of public IDs
During import, if PolicyCenter finds a match in entity public IDs, it does the following:
• If there is no difference between the import record and the database record, PolicyCenter ignores the import
record.
• If there are differences between the two records, PolicyCenter overwrites any existing database record values
with the values from the import file. PolicyCenter does not throw concurrent data change exception if the
imported records overwrite existing records in the database.
• If there are null entries for a record in the import file, PolicyCenter nulls out those values in the record in the
database.
only.
See also
Import Administrative Data Using Import Tools

Before you begin
The MaximumFileUploadSize parameter in file config.xml must be larger than the size of any file that you attempt
to import. The MaximumFileUploadSize parameter value is in megabytes (MB). In the base configuration, the
default value of MaximumFileUploadSize is 20 MB.
Procedure
1. Create a CSV or XML file describing the data, by using one of the following methods:
• Create the XML or CSV file manually.
• Export the current administrative data as an XML file from PolicyCenter and modify the file.
2. Import the CSV or XML file by using the import_tools command:
import_tools -password password -import fileName
The -import option requires that you provide the name of the file to import (fileName). There are a number
of other options that you can set as well.
See also
• “Constructing a CSV File for Import” on page 272
• “Construct an XML File for Import” on page 275
Importing and Exporting Administrative Data from PolicyCenter

You can import and export administrative data directly from PolicyCenter by using functionality that is available to
PolicyCenter administrators. It is only possible to import and export XML-formatted administrative data through the
PolicyCenter interface.
About Importing PolicyCenter Administrative Data

You can import administrative data directly from within PolicyCenter by using the Import Data screen available to
PolicyCenter administrators. To access this screen, navigate to the following location in PolicyCenter:
Administration→Utilities→Import Data
You can import XML-formatted administrative data only.
During the import, PolicyCenter looks for existing data objects that have public IDs that match those of the data
objects being imported. PolicyCenter notifies you if it determines there are public ID matches. You can then chose to
do one of the following:
• Overwrite all existing data with the imported data
• Discard updates to any existing data and keep the existing data
• Interactively resolve each data conflict on a case-by-case basis
Importing Arrays
PolicyCenter handles arrays differently depending on whether it is importing an owned array or a virtual array:
• Owned array – If an entity owns the array, PolicyCenter notifies you of the differences between the imported
data and any existing data. However, you do not have the choice of resolving the array elements. PolicyCenter
only gives you the option to delete the current array and replace all of the contents of the imported array.
• Virtual array – It is not possible replace the contents of a virtual array with those of an imported array as it is
not possible to delete a virtual array.
Import Administrative Data into PolicyCenter

Procedure
1. Log into Guidewire PolicyCenter as a user with the viewadmin and soapadmin permissions.
2. Select the Administration tab.
3. In the left-hand navigation pane, expand Utilities→Import Data.

4. Click Browse... to search for the XML file containing data to import.
5. Click Finish to import data from the file.
Next steps
During an import, PolicyCenter does not run validation rules. However PolicyCenter does run pre-update rules. For
this reason, run user exception and group exception batch processing after you import administrative data.
About Exporting PolicyCenter Administrative Data

You can export administrative data directly from within PolicyCenter by using the Export Data screen available to
PolicyCenter administrators. To access this screen, navigate to the following location in PolicyCenter:
Administration→Utilities→Export Data
It is possible to export the following data sets independently:
• Admin
• Policy Form Patterns
• Policy Holds
• Roles
IMPORTANT If a particular data set is not on the export list, you cannot export it by using this
function.
During export or import of users and groups, PolicyCenter also exports or imports any entities referred to by any
User or UserRole object through a foreign key or array.
You can export XML-formatted administrative data only.
Exporting user data

PolicyCenter handles the export of user data differently between development and production systems.
Server mode User export data

Development mode The export file contains usernames and the associated passwords in hash code format.
Production mode The export process strips both username and password information from the export file.
Notes
1. During export or import of users and groups, PolicyCenter also exports or imports any entities referenced by a
User or UserRole object through a foreign key or array.
2. It is not possible to import a previously exported data file from a production environment back into
PolicyCenter as it does not contain the required username/password data.
Export Administrative Data from PolicyCenter

Procedure
1. Log into Guidewire PolicyCenter as a user with the viewadmin and soapadmin permissions.
3. In the left-hand navigation pane, expand Utilities→Export Data.
4. Select the data set to export.
5. Click Export to download the XML file.
Understanding Roles and Permissions

A permission (or privilege) is a granular task or ability to see or do something within PolicyCenter. A role is a
named collection of permissions, and, typically, maps to a job function or job title.
PolicyCenter stores role information in file roles.csv and permission information in file roleprivileges.csv.
Within Guidewire Studio Project window, these two files exist in the following location:
configuration→config→import→gen
PolicyCenter loads the contents of these two files into the database upon initial database upgrade, at first server
startup after installation. See “About the import Directory” on page 268 for details on how PolicyCenter works with
the files in the gen directory.
See also
• “About Adding Admin Data after Initial Server Startup” on page 269
Role Definitions
File roles.csv contains a list of PolicyCenter roles, along with a human-readable name and description for each
role. Within this file, set the name and description fields to whatever is useful in uniquely identifying the role.
PolicyCenter reads the file, starting with the first row that contains the entityid identifier and imports the data into
the database.
The following code samples are examples of role definition entries:
Roles,
type,data-set,entityid,description,name,carrierinternalrole,roletype
Role,0,superuser,Superuser with full permissions,Superuser,true,user
Role,0,underwriter_supervisor,Base permissions for a supervisor,Underwriting Supervisor,true,user
Role,0,underwriter,Permissions for underwriter,Underwriter,true,user
,,,,,,
Role Permission Definitions

File roleprivileges.csv contains the mappings that link roles to a set of permissions. PolicyCenter reads the file
starting with the first row that contains the entityid identifier and imports the data into the database.
The following code samples are examples of permission definition entries:
type,data-set,entityid,permission,role
RolePrivilege,0,sample_data:2,abcreate,k
RolePrivilege,0,sample_data:3,abdelete,audit_examiner
RolePrivilege,0,sample_data:4,abedit,audit_examiner
RolePrivilege,0,sample_data:5,abview,audit_examiner
RolePrivilege,0,sample_data:6,anytagcreate,audit_examiner
,,,,
Each row in file roleprivileges.csv maps a single permission to a role. Each role has multiple permissions and
thus multiple rows. For example, the abcreate entry grants permission to create a contact to the audit_examiner
role.
The PolicyCenter Security Dictionary provides a full list of role permission, along with a brief description of each. It
also provides a list of the correspondences between roles and permissions.
Importing Security Zone Data

In the base default configuration, Guidewire provides a single security zone named Default Security Zone.
PolicyCenter imports this default security zone as part of the administration data import into an empty database at
initial server startup.
It is possible to add additional security zones to PolicyCenter in the following ways.
File security- Define additional security zones in file security-zones.xml. PolicyCenter loads this data as part of the
zones.xml administration data load into an empty database at initial server startup. See “Understanding the Security
Zones File” on page 281 for more information.
Export/import of Import new security zone data using the export and import functionality accessible in the PolicyCenterAd-
data ministration tab. See “Import Security Zones” on page 282 for more information.
Understanding the Security Zones File

Besides importing the default security zone at initial server startup, PolicyCenter also loads any additional security
zone data defined in file security-zones.xml. The use of this file is optional. You cannot use this mechanism to
load security zone data after the initial server startup populates the database with data.
You access file security-config.xml in Guidewire Studio™ by navigating in the Project window to
configuration→config→Security.
In the base configuration, file security-zones.xml provides only the top-level <import> XML element. To be
useful, you must add one or more <SecurityZone> subelements. There is no limit to the number of
<SecurityZone> elements that can exist in security-zone.xml.
The <SecurityZone> element has the following syntax.
<SecurityZone public-id="… ">

<Description>Some meaningful description…</Description>
<Name>Some meaningful name…</Name>
</SecurityZone>
The attributes and subelements of <SecurityZone> have the following meanings.

SecurityZone public-id Yes Internal identifier for the security zone. Guidewire recommends that you make the val‐
ue of this attribute meaningful for your organization. See “Public ID Prefix” on page
271 for more information on public IDs.
Name Yes External‐facing name for the security zone. This name is visible in the PolicyCenter user
interface.
Description No Description of the security zone, which, if provided, is externally facing in the
PolicyCenter user interface.
See also
• For information on PolicyCenter import of administration data at initial server startup, see “About the import
Directory” on page 268.
View the XSD for Security Zones

About this task
There is no separate XSD for security-zones.xml. Instead, view the relevant XSD information in the generated
pc_entities.xsd file.
Procedure
1. Open a command prompt and navigate to the PolicyCenter installation directory
3. In the file system, navigate to the following location in the PolicyCenter installation directory:
build/xsd
4. Open file pc_entities.xsd.

5. Search for SecurityZoneXSDType.
Import Security Zones

Procedure
1. Within Guidewire PolicyCenter, navigate to the following location:
Administration→Utilities→Export Data
2. Select Admin from the Data to Export drop-down list, then click Export.
3. Open the resulting file (admin.xml) for editing.
4. Review the file for existing security zones (<SecurityZone public-id="...">).
Pay particular attention to the public-id value for each existing security zone. Any security zone that you
add to this file must have a unique public-id value. See “Public ID Prefix” on page 271 for more information
on public IDs.
5. Add unique entries for each security zone to import.
For example:
<SecurityZone public-id="pc:236">
<Description>Some meaningful description…</Description>
<Name>Some meaningful name…</Name>
</SecurityZone>
This example sets the public-id value to pc:236. Chose a public ID value that makes business sense for your
organization.
6. After saving the file, navigate to the following location in PolicyCenter:
Administration→Utilities→Import Data
7. Browse to find your modified file and click Next.
If there are conflicts between the administrative data in the import file and the existing administrative data,
PolicyCenter provides a mechanism for conflict resolution. You can chose to do one of the following:
• Overwrite all existing data with the imported data
• Discard updates to any existing data and keep the existing data
• Interactively resolve each data conflict on a case-by-case basis
8. After resolving all data conflicts, click Finish.
Result
You can now use the updated set of security zones in PolicyCenter without server restart.
See also
• “Construct an XML File for Import” on page 275
• “About Exporting PolicyCenter Administrative Data” on page 279
• “About Importing PolicyCenter Administrative Data ” on page 278
About Importing Zone Data

PolicyCenter provides a collection of zone data files for various localities with small sets of zone data that you can
load for development and testing purposes. The zone data files are in the following location in the Project window of
Guidewire Studio™.
configuration→config→geodata
In the geodata folder, PolicyCenter stores the zone information in country-specific zone-config.xml files, with each
file in its own specific country folder. For example, the zone-config.xml file that configures address-related
information in Australia is in the following location in the Studio Project window.
configuration→config→geodata→AU
Guidewire provides the US-Locations.txt and similar files for testing purposes to support autofill and
autocomplete when users enter addresses. This data is provided on an as-is basis regarding data content. The
provided zone data files are not complete and might not include recent changes.
Also, the formatting of individual data items in these files might not conform to your internal standards or the
standards of third-party vendors that you use. For example, the names of streets and cities are formatted with mixed
case letters but your standards may require all upper case letters.
The US-Locations.txt file contains information that does not conform to United States Postal Service (USPS)
standards for bulk mailings. You can edit the US-Locations.txt file to conform to your particular address
standards, and then import that version of the file.
Importing a Zone Data File

To import zone data, use the zone_import command. The zone_import command imports data in CSV format from
specified files into database staging tables for zone data. It is only possible to import zone data for a single country
at a time. The zone data files that you import must contain zone data for a single country only. To load zone data for
multiple countries, use the command multiple times with different, country-specific zone data files each time.
After you use the zone_import command to import zone data, set the PolicyCenter server run level to MAINTENANCE.
Then, move the data from the staging tables to the production tables by using the table_import command. See
“Table Import Command” on page 399 for more information on the table_import command. Finally, set the
PolicyCenter server run level back to MULTIUSER.
Guidewire expects that you import address zone data upon first installing PolicyCenter, and then at infrequent
intervals thereafter as you receive data updates.
Import a Zone Data File
Procedure
1. Start the PolicyCenter server.
2. Clear the zone data staging tables.
If you have multiple countries defined, you can include the -country countryCode option to clear staging
zone data only for the country you want to load:
zone_import -password password -clearstaging [-country countrycode]
3. Load the zone data file into the staging tables:
zone_import -password password -country countrycode -import filename
4. Clear existing zone data in production. Perform this step if zone data already exists for the country whose data
you intend to load.
zone_import -password password [-country countrycode] -clearproduction
5. Set the PolicyCenter server run level to MAINTENANCE:
system_tools -password password -maintenance
6. Load zone data from the staging tables into production:
table_import -server url -password password -integritycheckandload -zonedataonly
7. Set the server run level back to MULTIUSER:
system_tools -password password -multiuser
See also
• For information on using the zone_import command, see “Zone Import Command” on page 403.
• For information on using the table_import command, see “Table Import Command” on page 399.
• For information on importing zone data and database staging tables generally, see the Integration Guide.
• For information on the web service ZoneImportAPI that also imports zone data, see the Integration Guide.
Importing Custom Zone Data Files

You can create your own zone data files, in CSV format. The import tool uses configuration information from zone-
config.xml files for specific countries to determine which data fields to import and what each field represents for
each country. PolicyCenter stores the zone-config.xml files for base-configuration countries in country-specific
folders. Navigate to the following location in the Studio Project window to view these files:
configuration→config→geodata
For example, Guidewire provides zone configuration data for France in file zone-config.xml in the following
location:
configuration→config→geodata→FR
Each country-specific zone-config.xml file must have a single top-level <Zones> element for that country.
Underneath each <Zones> element are <Zone> elements that define the zone fields to import from zone data files for
that country. For each field, the fileColumn attribute indicates the position of the field within lines of the files.
The following example XML code from a zone-config.xml file defines the fields to import from zone data files for
the United States. The example code specifies for United States zone data files that the third field in the comma-
separated values of each line of corresponds to a city.
<Zones countryCode="US">
...
<Zone code="city" fileColumn="3" granularity="2">
...
See also
• For complete information about zone-config.xml, including a description of its XML elements and attributes,
see the Globalization Guide.

chapter 19
Free‐text Batch Load Command
The free-text batch load command loads the Guidewire Solr Extension, a full-text search engine, with index
documents for all policies in your PolicyCenter application database. Index documents are XML documents that
contain a subset of the information from policies in PolicyCenter. The Guidewire Solr Extension indexes the
documents after it receives them from the free-text batch load command.
Note: The free-text batch load command runs on the host on which the Guidewire Solr Extension
resides. The command is located in the /opt/gwsolr/pc/solr/policy_active/conf directory, not
the PolicyCenter/admin/bin directory.
See also
When to Run the Free‐text Batch Load Command

Generally, PolicyCenter updates the Guidewire Solr Extension whenever someone changes a policy and that change
affects index documents stored there. In response, the Guidewire Solr Extension incrementally indexes the changed
information. Occasionally, you need to load the Guidewire Solr Extension and have it build new indexes based on
the newly loaded information.
IMPORTANT Users must not perform free-text searches while the free-text batch load command runs.
Otherwise, the search results will be incomplete. Set the EnableDisplayBasicSearchTab script
parameter to false to temporarily hide the free-text search user interface.
Run the free-text batch load command whenever any of the following occur:
• Policies are bulk loaded into the PolicyCenter application database.
• Indexes become corrupted in the Guidewire Solr Extension, as reported by the Guidewire Solr Extension web
application.
• Changes are made to the metadata definitions of entities and relationships in the policy graph, and these changes
affect index documents stored in the Guidewire Solr Extension.
• Changes are made to attribute definitions in schema.xml.
• Changes to the mapping (policy-search-config.xml or custom mappers) that affect already indexed periods.
• Your instance of PolicyCenter is upgraded to a later version, and the upgrade changes metadata definitions for
index documents stored in the Guidewire Solr Extension.
Do not run the free-text batch load command if you configured free-text search for embedded operation. Whenever
the Guidewire Solr Extension runs in embedded mode, use the Internal Tools Free Text Search screen to execute the
Free‐text Batch Load Command 285
batch load command instead. If you run the batch load command with embedded operation, PolicyCenter limits the
number of index items to 10,000.
See also
• “Free-text Search” on page 382
Prerequisites for Running the Free‐text Batch Load Command

Before you run the free-text batch load command for the first time, you must modify the following files:
• data-config.xml – Specifies for the batch load command the location of the collated and compiled index
documents for the Guidewire Solr Extension to load. It also specifies the fields the index documents contain.
• batchload.sh/batchload.bat – Specifies the batchload-config-databaseBrand.xml configuration file to
use for your database brand.
• batchload-config-databaseBrand.xml – Specifies the SQL Select statement that extracts policies from the
PolicyCenter application database. Specifies the URL for the Guidewire Solr Extension. Specifies a working
directory, and specifies a sort binary.
Each time before you run the free-text batch load command, you must do all of the following if PolicyCenter is
running:
• Suspend the PCSolrMessageTransport message destination from the Event Messages page on the Administration
tab.
Suspending the message destination prevents PolicyCenter from sending updated index documents to the
Guidewire Solr Extension if users modify policies while the free-text batch load command runs.
• Set the EnableDisplayBasicSearchTab script parameter to false from the Script Parameters page on the
Administration tab.
Setting the script parameter to false prevents users from accessing the Basic tab to perform free-text searches
while the free-text batch load command runs.
Guidewire recommends that you take a backup of the current index before running each free-text batch load
command. This backup provides a snapshot of the index that you can restore if the batch load does not complete
successfully. For information about how to perform a backup, see the Solr documentation on the following web site:
https://archive.apache.org/dist/lucene/solr/ref-guide/apache-solr-ref-guide-4.10.pdf
Run the Free‐text Batch Load Command

About this task
You run the free-text batch load command on the host on which the Guidewire Solr Extension resides. Run the
command only if you configured free-text search for external operation.
Procedure
1. In PolicyCenter, do the following:
a. Suspend the PCSolrMessageTransport message destination.
b. Set the EnableDisplayBasicSearchTab script parameter to false.
2. Shut down and restart the Guidewire Solr Extension.
Shutting down the Guidewire Solr Extension forces it to pick up any changes to data-config.xml.
3. Navigate to the following location in the Solr installation:
/opt/gwsolr/pc/solr/policy_active/conf
4. Run the batchload command.

286 chapter 19: Free‐text Batch Load Command
5. After the command finishes, examine the status response to verify that your load succeeded.
A problem-free load gives the same positive number for Total Rows Fetched and Total Documents Processed.
6. In PolicyCenter, do the following:
a. Resume the PCSolrMessageTransport message destination.
b. Set the EnableDisplayBasicSearchTab script parameter to true.
Recovering from Solr Batch Load Errors

The free-text batch load command queries the PolicyCenter database for data and then locally processes the
information intensively on disk. The command eventually produces an XML file with index documents ready for the
Guidewire Solr Extension. In the last step, the command tells the Guidewire Solr Extension to load the index
documents.
The batch load command can fail in the last step and end without loading any index documents. For example, an
error in file data-config.xml can cause the load to fail. In such cases, you do not need to run the entire batch load
command again. Instead, you can invoke the Guidewire Solr Extension directly to complete its portion of the batch
load process by using the following URL:
http://hostName:8983/pc-gwsolr/pc_policy_active/dataimport?command=full-import&entity=policy
Clean‐up Tasks after Running the Free‐text Batch Load Command

Each time after you run the free-text batch load command, you must do all of the following:
• Resume the PCSolrMessageTransport message destination from the Event Messages page on the Administration
tab.
Resuming the message destination lets PolicyCenter send updated policy data to the Guidewire Solr Extension
for incremental indexing.
• Set the EnableDisplayBasicSearchTab script parameter to true from the Script Parameters page on the
Administration tab.
Setting the script parameter to true lets users access the Basic tab to perform free-text searches.
• Consider deleting files from the workDir directory.
Free‐text Batch Load Command and Native SQL

The free-text batch load command extracts all policy data from the PolicyCenter application database by using
native SQL. The SQL Select statement that the batch load command uses is defined in configuration files for
specific database brands. These configuration files are located on the host on which the Guidewire Solr Extension
resides, in the following directory:
opt/gwsolr/pc/solr/policy_active/conf
The configuration files that contain native SQL are:
• For H2 databases – batchload-config-h2.xml, suitable only for development
• For Oracle databases – batchload-config-oracle.xml, suitable for development or production
• For SQL Server databases – batchload-config-sqlserver.xml, suitable for development or production
See also
Free‐text Batch Load Command 287

288 chapter 19: Free‐text Batch Load Command

chapter 20
Production Data Fix Tool
This topic describes a tightly constrained system for updating data on a running production server other than through
PCF pages or web services. Guidewire calls this mechanism the Production Data Fix tool.
WARNING Only use the Production Data Fix tool under extraordinary conditions, with great
caution, and upon advice of Guidewire Support. Before registering a data change on a production
server, register and run the data change on a development server. Guidewire recommends multiple
people review and test the code and the results before attempting the data change on a production
server.
Overview of the Production Data Fix Tool

In typical conditions, PolicyCenter data changes in the database using the following techniques:
• Users change data through the user interface, defined by PCF pages
• External systems change data through specific integrations exposed as web services
There may be a need to change production data in a way that had not been predicted enough to define PCF pages or
web services for the situation. In typical cases, you can write a new web service or other integration to satisfy your
integration need. However, in rare cases there may not be an opportunity to bring your production server down for
this improvement to the application.
PolicyCenter provides a tightly constrained system for updating data on a running production server. Guidewire calls
this mechanism the Production Data Fix Tool.
server.
Separation of Permissions
To decrease security risks, the Production Data Fix tool separates its action into separate tasks, each of which has
different permissions and entry points.
Permission Code Description

Execute a data change admindatachangeexec Permission to execute the data change code.
Production Data Fix Tool 289

Permission Code Description

Register a data change wsdatachangeedit Permission to register a data change Gosu program.
Use one of the following methods to register data change code:
• The data_change command prompt tool
• The DataChangeAPI WS‐I web service
View a data change admindatachangeview Permission to view the Data Change screen.
By having multiple different paths and multiple different roles, there is no single point of attack.
IMPORTANT Guidewire recommends that you force separation of responsibilities into two different
PolicyCenter users. Give each user either the wsdatachangeedit permission (to register data change
code) or the admindatachangeexecd permission (to execute the code), but not both permissions.
Preserving Results
ClaimCenter captures the results of script execution. This increases accountability and makes debugging easier.
Replay Prevention
To prevent replay attacks, the Production Data Fix tool runs each registered script a maximum of one time. If you
need to run it again, you must first re-register the script and create a new change control reference.
The Data Change Screen
Administrative users with the admindatachangeview permission can view a special PolicyCenter Data Change
administration screen that displays information about data change operations. To access this page, navigate to the
following location in Guidewire PolicyCenter:
Administration→Utilities→Data Change
See also
• “Writing Gosu Data Change Code” on page 291

• “Registering Data Change Code” on page 291
• “Run Gosu Data Change Code” on page 293
Typical Use of the Production Data Fix Tool

There are several steps in using the Production Data Fix tool. The following list describes these steps.
Task More information

Writing and testing the Gosu database code “Writing Gosu Data Change Code” on page 291
Registering the Gosu database code, either through the use of a system tool “Registering Data Change Code” on page 291
or through web services
Running the Gosu database code “Run Gosu Data Change Code” on page 293
See also
• “Logging Data Change Operations” on page 293
290 chapter 20: Production Data Fix Tool

Writing Gosu Data Change Code

The first stop in using the Production Data Fix tool is to write Gosu code that does the following:
• Correctly and safely makes only necessary changes to the production data
• Persists the changes to the database
WARNING Carefully test your data change code. Guidewire strongly recommends that multiple
people review and approve the code for safety and correctness before proceeding.
To persist changes to the database, use the gw.Transaction.Transaction class and its method runWithNewBundle.
You pass the method a block that runs code. If the block does not throw an exception, PolicyCenter persists any data
changes from your Gosu block to the database. If the block throws an exception, no changes persist to the database.
Design your data change code to minimize the number of entity instances you change. Too many changes in entity
data increases the chance of memory issues or concurrent data exceptions.
Save your Gosu code to a local file that ends in .gsp.
See also
• Gosu Reference Guide
Registering Data Change Code

There are two ways to register your data change code
• Through the data_change command prompt tool
• Through the DataChangeAPI web service
The data change registration details vary between these two variants.
In all cases, before proceeding you must have:
• Data change code in the form of a Gosu script that you have already tested in your development environment.
• A human-readable description for your data change.
• A unique reference ID that you create to represent this data change.
IMPORTANT If you need to re-run a successful data change, you must first re-register the script with a
new reference ID. This is requirement preserves the integrity of the results log.
See also
Register a Data Change Using a Command Prompt

About this task
The command prompt tool data_change works by calling the DataChangeAPI web service on a running
PolicyCenter server.
WARNING Before registering a data change on a production server, register and run the data
change on a development server first. Guidewire recommends multiple people review and test the
code and the results before attempting the data change on a production server.
Procedure
2. Open a command prompt and navigate to the following location in the PolicyCenter installation:
admin/bin

data_change –description DESCRIP –edit REFID –gosu PATH –server SERVERURL –user USER –password PW
For example:
data_change –description "Fix Employee Name"

–edit REFID_1234 –gosu c:\PolicyCenter\datachange\gosudatachange_REFID1234.gsp
–server http://TESTINGSERVER:8180/pc –user su –password gw
Result
The script outputs results such as:
Running data_change.gsp
Connecting as su to URL http://localhost:8180/pc/ws/gw/webservice/systemTools/DataChangeAPI
Edit change ref=REFID1234 publicId=cc:1
See also
• “Data Change Command Tool Reference” on page 294
Register a Data Change Using a Web Service

About this task
WARNING Before registering a data change on a production server, register and run the data
change on a development server first. Guidewire recommends multiple people review and test the
code and the results before attempting the data change on a production server.
Procedure
1. Call the DataChangeAPI web service method updateDataChangeGosu.
2. Pass the method the following arguments as String objects:
• The reference ID
• A human-readable description
• The Gosu code to run
For example:
var gosuScript = "gw.transaction.Transaction.runWithNewBundle(\ bundle -> {

print(""DATA CHANGE!"") })"
var publicID = datachangeAPI.updateDataChangeGosu("REFID_1234",

"Fix for Issue 1234 regarding missing Employee ID", gosuScript)
Result
The method call returns the public ID of the new DataChange entity instance.
See also
• “Register a Data Change Using a Command Prompt” on page 291
Run Gosu Data Change Code

Before you begin
Before executing data change code, confirm that someone created and registered a data change. You must know the
reference ID for the registered data change.
About this task
server, register and run the data change on a development server first. Guidewire recommends
multiple people review and test the code and the results before attempting the data change on a
production server.
Procedure
1. Log into your PolicyCenter development environment as an administrator with the admindatachangeview
permission.
3. Expand Utilities→Data Change.
4. In the list of data changes, use the Reference column to find the data change request by its reference ID.
5. Click on the data change row in that list.
The screen shows the Gosu code for that data change.
6. Review the Gosu code to confirm it is what you expect.
7. Click Execute.
During code execution, PolicyCenter does not display the results immediately in the Result pane. Instead, the
status of Executing shows in the list of data changes.
8. After a few seconds, click Reload on the screen to view the current status and results.
• If the change is successful, PolicyCenter confirms the success in a message that uses your reference ID:
• If the change was not successful, PolicyCenter shows any compile errors or exceptions in the user interface
in the Result pane.
9. Confirm your changes in the database and check your logging results from the change.
10. If the data change appears safe in your development environment, carefully register and run the data change
on the production server.
See also
Logging Data Change Operations

Use data change Gosu APIs to configure logging within data change code. These APIs generate logging information
that users can see in human-readable output in data change user interface.
Logging Entity Field‐level Changes

To log entity field-level changes, call DataChange.util.setDetailResultWriting(bundle). The logging
information includes information about added objects, deleted objects, and field-level changes on every object. For
updated properties, the logging information includes each field value before the change and after the change.
Logging Arbitrary Text Data

To log arbitrary text data, call DataChange.util.ResultsWriter. That property returns an appender, which is an
object that implements the interface java.lang.Appendable. That object has several method signatures of the
append method. The simplest method signature takes a CharSequence object, such as a standard String object.
Example Logging Code
The following sample code uses the setDetailResultWriting method and the ResultsWriter property to create
log messages.
gw.transaction.Transaction.runWithNewBundle(\ bundle -> {
// For demonstration, get a User object and make minor data change to the first name
var u = gw.api.database.Query.make(User).select().first()
bundle.add(u)
u.Contact.FirstName = u.Contact.FirstName + "SUFFIX"
// Determine what you want to write to the data change log

var msg = "For PublicID '${u.PublicID}' User.DisplayName is now '${u.DisplayName}'!"
// To log arbitrary text in Data Change UI, get a results writer (type is java.lang.Appender)
var rw = DataChange.util.ResultsWriter
rw.append("Add arbitrary log message here\n")
// enable detailed logging of each property value before and after our change
DataChange.util.setDetailResultWriting(bundle)
// for testing in Studio Scratchpad, also print to standard console

// print("To console: " + msg)
})
To test and debug your code in Studio Scratchpad, you may want to print to the console using the standard print
statement. Also, add one more argument to the runWithNewBundle method to represent a user name. For example,
pass the String value "su" to create your writable bundle as the super user.
Data Change Command Tool Reference

this mechanism the Production Data Fix tool. The Production Data Fix tool uses either the data_change command
option, or, the DataChangeAPI web service to make changes to the production data.
Because the data_change command allows arbitrary execution of data, Guidewire strongly recommends that you
carefully control the ability to create and run code on a production server. The user who runs this command must
have permission wsdatachangeedit.
server.
data_change -help
data_change -password password [-server url] [-user user] {
-edit refID -gosu filepath [-description desc] |
-discard refID |
-status refID |
-result refID }
See also
• For a description of how and when to use the data_change command to change data on a running production
server, see “Production Data Fix Tool” on page 289.
• For a description of how to use the DataChangeAPI web service, see “Data Change Web Service Reference” on
page 295.
• For specifics of the data_change command options, see “Data Change Options” on page 387.
Data Change Web Service Reference

To register a data change, or to check status on a data change operation, use the following methods on the
DataChangeAPI web service.
DataChangeAPI method Description
updateDataChangeGosu Use to register a data change.

Pass the method the following arguments as String objects:
• The reference ID
• A human‐readable description
• The Gosu code to run
The method returns the public ID of the new DataChange entity instance.
Note the following:
• If the data change succeeds with no compile errors, you cannot edit the data change Gosu code.
You must re‐register the script with a new reference ID to make changes to the code.
• If the data change was never run, or had compile errors, you can update (edit) and rerun Gosu
code with the same reference ID.
discardDataChange Use to discard a data change that you already registered.
Pass the method a data change reference ID as a String object. You cannot discard a data change
that was already run.
getDataChangeResult Use to review the result of a data change that you already registered.
Pass the method a data change reference ID. The method returns a String object that represents
the results in the DataChange entity instance. If running the data change code generated parse er‐
rors, the results include the errors.
getDataChangeStatus Use to view the status of a data change that you already registered.
Pass the method a data change reference ID. The method returns a DataChangeStatus typecode.
Valid values include:
• Open
• Discarded
• Executing
• Failed
• Completed
See also
• “Overview of the Production Data Fix Tool” on page 289
• “Typical Use of the Production Data Fix Tool” on page 290


part 6
Business Rules Administration

chapter 21
Administering Business Rules
This topic provides information relevant to administering PolicyCenter business rules.
See also
• “Business Rules Import and Export” on page 309
Business Rules in Guidewire PolicyCenter

PolicyCenter provides a management tool that business analysts and rule administrators use to do the following:
• Add, modify, and delete underwriting business rules
• Manage the process of testing and deploying business rules
• Manage the import and export of business rules between development, test, and production PolicyCenter servers
You access and manage business rules for underwriting within Guidewire PolicyCenter in the following location:
Administration→Business Settings → Business Rules→Underwriting Rules
PolicyCenter business rules, managed from and within PolicyCenter, are distinct from Gosu rules, managed through
PolicyCenter Studio.
See also
•
• Rules Guide
Business Rule Roles and Permissions

In the base configuration, PolicyCenter provides the following permissions for use with underwriting business rules.
Permission Code Provides ability to…

Approve Rules uwruleapprove Promote a business rule to Approved status.
Deploy Rules uwruledeploy Deploy approved business rules.
Edit Rules uwruleedit Perform the following business rule edit operations:
• Create, clone, edit, delete
• Revert a rule to a specific rule version
• Enable or disable a business rule
Administering Business Rules 299
Permission Code Provides ability to…

• Promote a business rule to Staged status
Import Rules uwruleimport Import business rules into PolicyCenter.
View Rules uwruleview View the Business Rules screens.
PolicyCenter provides the following user role for use with business rules.
Rules Admin All permissions related to underwriting business rules
Business Rule Configuration Parameters

PolicyCenter provides the following parameters in file config.xml to manage the deployment and configuration of
business rules:
• BizRulesDeploymentEnabled
• BizRulesDeploymentId
• BizRulesImportBootstrapRules
• BizRulesLeafSearchNumOfHops
• BulkLoadRuleHeadByIdCacheEnabled
• PreloadBizRulesBeansCacheEnabled
Production Servers
If you are running PolicyCenter business rules on a production server, you must set BizRulesDeploymentEnabled
to true and provide a value for BizRulesDeploymentId. See “Business Rule Production Server Configuration” on
page 300 for more information.
Performance Considerations
See “Business Rule Performance” on page 301 for information on how to use the
BulkLoadRuleHeadByIdCacheEnabled and PreloadBizRulesBeansCacheEnabled configuration parameters to
improve the performance of the business rules.
See also
• For information on the individual configuration parameters see the Configuration Guide.
Business Rule Production Server Configuration

If you use PolicyCenter business rules in a production environment, you must do the following:
• Set configuration parameter BizRulesDeploymentEnabled to true.
• Set configuration parameter BizRulesDeploymentId to the server ID of the production server importing the
business rules.
Note: If you are running multiple production clusters, the BizRulesDeploymentID value for each
separate cluster must be unique.
It is possible to set the ID of the application server as you start the server (best practice) or through the <registry>
element in file config.xml. Set the value of the server ID to the value of BizRulesDeploymentId.
See also
300 chapter 21: Administering Business Rules
Business Rule Performance

Executing a business rule often involves making multiple queries to the database. A large number of database
queries can affect business rule performance significantly. Guidewire provides the following business rule
parameters to use in configuring a caching strategy that improves business rule performance.
Parameter Cache Description Possible values

BulkLoadRuleHeadByIdCacheEnabled Rule Parameter determines whether • true ‐ PolicyCenter uses a
Head PolicyCenter populates the Rule Head single query to populate the
cache cache with a single query to the Rule Head cache.
database or multiple queries. • false ‐ PolicyCenter uses
The Rule Head cache stores the Rule‐ multiple queries to populate
Head IDs for all business rules. the Rule Head cache.
PreloadBizRulesBeansCacheEnabled Rule Parameter determines whether • true ‐ PolicyCenter uses a
Entities PolicyCenter populates the Rule single query to preload the Rule
cache Entities cache with a single query to Entities cache with business
the database or multiple queries. rule entities from the database.
The Rule Entities cache stores the en‐ • false ‐ PolicyCenter uses
tities that comprise the entire rule multiple queries during rule
graph for all rule versions that can execution to fetch the
possibly execute in the application. necessary business rule entities.
IMPORTANT Using a single query to bulk load a cache significantly reduces the number of database
queries needed during rule execution and the total time needed to run the queries.
Preloading the Rule Head Cache at Server Start

It is possible to preload the Rule Head cache at server start. Guidewire enables this functionality by default in the
base configuration by adding the following entry to file preload.txt:
gw.api.bizrules.startup.BizRulesPreloadActions#preloadRulesBeansDataIntoCache
To access this file, navigate to the following location in PolicyCenter Studio.
configuration→config→startup
To disable this functionality, either remove or comment out this line of code in the file.
Preloading the Rule Entities Cache at Server Start

It is possible to preload the Rule Entities cache during server start. However, Guidewire disables this functionality
by default in the base configuration. To enable this process, add the following method invocation to file
preload.txt:
gw.api.bizrules.startup.BizRulesPreloadActions#preloadRulesVersionsCache
To access this file, navigate to the following location in PolicyCenter Studio.
configuration→config→startup
Business Rule Versioning

PolicyCenter associates a version number and a status with each individual business rule. You see this information in
the Business Rules screen, in the Version field. PolicyCenter manages the information in the Version field, you cannot
change it yourself. If you have multiple versions of a rule, the Version field becomes a drop-down list, from which
you can select a specific version of a rule to view.
Each version number starts at 0 and increments by 1 for each deployed version of the business rules. If you edit a
business rule, PolicyCenter adds a plus sign (+) to the version number of the rule that you are editing, for example,
0+. Along with the version number, each business rule shows its state in parenthesis next to the version number, for
example 0+ (Draft). Each version of a business rule is always in one of the following states:
• Draft
• Staged
• Approved
At the creation of a new rule version, PolicyCenter generates a work-in-progress rule with an initial status of Draft.
This draft rule contains the same data values as the parent rule version from which it was made.
The work-in-progress rule must go through the three states of Draft, Staged, and Approved before it is possible to
deploy the new rule version. After these three stages are complete, and after you deploy the rule, PolicyCenter
assigns an updated version number to the rule. This version of the rule becomes the latest running version of the rule
that PolicyCenter evaluates at runtime.
The version of a rule that PolicyCenter evaluates at runtime contains the word Evaluated after the version number.
The Evaluated version of a rule is always the latest iteration of a rule that can run in a specific environment.
Rule versioning in multi‐cluster production environments

PolicyCenter associates the BizRulesDeploymentId value (name) with a rule during its deployment. Thus, if you
deploy the rule in one production cluster, the name appears along with the rule version number in other production
clusters to identify that the rule was deployed in a particular production cluster. For example, if the value of
BizRulesDeploymentId is GW100-A, the rule version appears as 1.GW100-A in all other production clusters.
Rules for Deleting a Business Rule Version

PolicyCenter associates a version number and a status (state) with each individual business rule. Whether it is
possible to delete any given version of a rule depends on the state of the business rule. The following table lists the
rules for deleting a rule version.
Rule version state Rule deletion action

Deployed It is not possible delete a rule version after its deployment
Approved or Staged Clicking Delete deletes all rule versions down to the last deployed rule version.
Draft – Direct parent deployed Clicking Delete Draft deletes all rule versions down to the last deployed rule version.
Draft – Rule previously staged or Clicking Delete Draft deletes the immediate draft version of the rule. After deleting the draft
approved version of the rule, it then becomes possible to delete the staged and approved versions of
the rule.
Business Rule Deployment

In a production PolicyCenter cluster, Guidewire requires that you deploy a rule before PolicyCenter can evaluate the
business logic of the rule. Thus, to deploy a business rule is to make that version of the business rule active within
PolicyCenter.
Before deployment, PolicyCenter creates a work-in-progress version of the rule. At rule deployment, PolicyCenter
assigns a version number to the rule.
A rule must go through the following states before you can deploy the rule:
• Draft
• Staged
• Approved
A rule must be in the approved state before it is possible to deploy the rule. Approving a rule does not deploy the
rule. You must actively deploy a rule before PolicyCenter evaluates that rule at runtime.
Rule deployment is distinct from rule import. On completing a rule import, you must deploy any new approved rule
versions before PolicyCenter can evaluate the rules.
In general, Guidewire recommends that you use separate cluster environments for each of the following in working
with PolicyCenter business rules:
• Development
• Test
• Production
Typically, you deploy a rule version in a production environment only.
IMPORTANT In a production environment, you must set configuration parameter

BizRulesDeploymentEnabled to true and provide a value for configuration parameter
BizRulesDeploymentId. If you have multiple PolicyCenter production clusters, the
BizRulesDeploymentId value for each cluster must be unique.
See also
• “Business Rule Versioning” on page 301
• “Business Rule State” on page 303
Business Rule State

At initial creation, all business rules are in the Draft state. As you begin the process of reviewing, evaluating, and
updating a business rules, its state can vary and progress.
The Guidewire PolicyCenter business rules have the following sequence of states.
Draft PolicyCenter assigns a status of Draft after you save the initial version of the rule. A rule reverts to Draft status
whenever the rule undergoes any type of editing. It is not possible export a rule in the Draft state.
Staged You manually move a rule in the Draft state to the Staged state, after you complete rule editing. Typically, this is the
point in the rule lifecycle that you export the rule from the development environment and import the rule into the
testing environment.
Approved You manually move a rule from the Staged state to the Approved state, usually in the testing environment after you
complete the rule evaluation phase. Typically, this is the point in the rule lifecycle that you export the rule from the
testing environment and import it into the development environment.
Deployed You manually move a rule from the Approved state to the Deployed state. Typically, this occurs after you import the
rule into the production environment.
The following diagram describes the various states associated with a business rule, as it is created, edited, and
deployed in development and production environments.

Business Rules States
Development Testing Production
Test, Reject Rule

Staged Approved
Draft
Create/Edit Rule
Test, Import Rule Deploy

Import
Promote to
Rule
Test, Approved
Promote to Edit
Business Staged
Analyst
Approved Deployed
Export Back to
Implementation Staged Development
Specialist Export Rule and
Testing*
Staged Approved
Export Business Rule Business Rule
Rule
* Export deployed rules from Production and import back to Development and Testing to keep all environments in sync.
Business Rule Lifecycle

Whenever you deploy a rule in the production server, Guidewire recommends that you export the rule from the
production server and import it back to the development and testing servers. Always make changes on the
development server and move the rule to testing then to production. If you always export the deployed rule from
production back to development, the version number and rule status on development stays synchronized with the
production server.
For example, you create a new rule; the version is 0+. The version stays the same as you move it from development
to testing, and then to production. Whenever you deploy the rule on the production server, the version number
increases to 1. You export the rule from production back to the development server. Now on both development and
production servers, the rule version is 1. On the development, you make changes to version 1. Whenever you import
the rule through testing to production and deploy it, the rule version becomes 2.
The following table shows how rule version and status change as you export and import it from development,
testing, and production servers.
Action Development Testing Production

1. Create a new rule. 0+ (Draft)
2. Edit rule and save it. Repeat several times. 0+ (Draft)
3. Rule is ready for testing. Promote to Staged status. 0+ (Staged)

Action Development Testing Production

4. Export rule and import to testing server. 0+ (Staged)
5. Rule passes testing. Promote to Approved status. 0+ (Approved)
6. Export rule and import to production server. 0+ (Approved)
7. Deploy the rule. The version number increases from 0 to 1. 1 (Evaluated)
8. Export rule and import back to development and testing servers. 1 1
9. Edit rule. 1+ (Draft)
10. Ready for testing, promote to Staged. 1+ (Staged)
11. Export and import to testing server. 1+ (Staged)
12. Testing passed. Promoted to Approved. 1+ (Approved)
13. Export and import to production. 1+ (Approved)
14. Deploy. The version number increases from 1 to 2. 2 (Evaluated)
15. Export rule and import back to development and testing servers. 2 2
Guidewire does not require that you move the rule back to development before making changes. However, if you do
not export a deployed rule back to the development server, the version number on the development server does not
correspond to the version on production. For example, if you never export the rule back to development, the version
number stays at 0+ on the development server, and increases each time you deploy it on production.
The following table shows rule version and status change for the same rule, except that you do not export the rule
from production to development.
Action Develop‐ Testing Production

ment
1. Create a new rule. 0+ (Draft)
2. Edit rule and save it. Repeat several times. 0+ (Draft)
3. Rule is ready for testing. Promote to Staged status. 0+ (Staged)
4. Export rule and import to testing server. 0+ (Staged)
5. Rule passes testing. Promote to Approved status. 0+ (Ap‐
proved)
6. Export rule and import to production server. 0+ (Ap‐
proved)
7. Deploy the rule. The version number increases from 0 to 1. 1 (Evaluat‐
ed)
8. Rule requires updates. On development server, edit rule and save. Repeat sever‐ 0+ (Draft)
al times.
9. Ready for testing, promote to Staged. 0+ (Staged)
10. Export and import to testing server. 0+ (Staged)
11. Testing passed. Promoted to Approved. 0+ (Ap‐
proved)
12. Export and import to production. The version number increases because the 1+ (Ap‐
rule version currently on the production server is 1 (Evaluated). The rule being proved)
imported has changes built on top of that version, so the version becomes 1+. If
changes to the rule have also been made on production server, you must re‐
solve the conflict manually on import.

Action Develop‐ Testing Production

ment
13. Deploy. The version number increases from 1 to 2. 2 (Evaluat‐
ed)
Business rule logging

PolicyCenter writes audit information to the console log and to the PolicyCenter log file whenever a rule evaluation
triggers a business rule. Unless specified otherwise, PolicyCenter writes the business rule information at the log
INFO level for the following logging categories and subcategories:
BizRules
BizRules.Audit
BizRules.Autocomplete
BizRules.Compiler
BizRules.ContextHelp
BizRules.Export
BizRules.Import
BizRules.UI
BizRule.Validation
The business rule information that PolicyCenter logs includes the following:
• The rule context
• The rule ID
• Whether the rule condition evaluates to true or false
• The rule action parameters as a list of name-value pairs, which is empty if the condition evaluates to false
Default logging level

The default logging level for business rules is INFO. It is possible for the standard operation of the business rules to
generate a large amount of entries in the business rules logging file. You can reduce the amount and frequency of
business rule logging by setting the logging level to a higher threshold, for example, WARN.
To change the default logging level for the business rules, add the following entry to file logging.properties:
log4j.category.BizRules=WARN
See also
• “Application Logging” on page 25
• “Set Log Level” on page 329
Invalid Business Rules

Invalid Rules during Server Start
If business rules exist in the database, PolicyCenter validates these rules during the start of the application server. If
PolicyCenter finds an invalid rule, it generates an error message in the application log similar to the following:
ERROR BizRules.Validation Rule Test Rule has validation error: Could not resolve symbol
for : SomeActivity
Guidewire recommends that you review the application log after starting the application server to determine if
PolicyCenter reports any rule as invalid.
Invalid Rules during Rule Execution

How PolicyCenter treats an invalid rule depends on the rule state and the server mode.
Rule state Server mode Action

Draft Non‐Production PolicyCenter skips the execution of the invalid rule and continues with rule
execution.
Staged, Approved,or Non‐Production PolicyCenter generates an exception and stops rule execution.
Deployed
Deployed Production PolicyCenter generates an exception and stops rule execution.


chapter 22
Business Rules Import and Export
You transfer rules between different server environments by exporting the rules from one server environment and
importing the rules into the other server environment. Typically, one or more appointed Rules Administrators
manage the export and import of business rules between different PolicyCenter server environments.
Overview of Business Rule Export and Import

The following flow is a typical business rule development lifecycle:
• Create and develop the draft business rules on a development server.
• Export the business rules to a testing server that replicates the production environment and stage the rules.
• After testing, approve the rules on the staging server.
• Export the business rules to the actual production server and deploy the rules there.
You use the import and export of business rules to move rules between the multiple servers.
During import, PolicyCenter only brings in rules that are new or have changes on top of existing versions. The
version number increases whenever you deploy the rule.
Whenever importing rules from one server environment to another, PolicyCenter takes one of the following actions
for each rule:
PolicyCenter Action
Imports the rule PolicyCenter imports the rule if any of the following are true:
• The rule for import is new and does not exist in the importing server environment.
• The rule for import is an update to an existing rule in the importing server environment.
Does not import PolicyCenter does not import the rule into the importing server environment if any of the following are
the rule true:
• If there is no difference between the existing rule and the rule for import.
• If the existing rule is already an update of the rule for import.
Raises a conflict PolicyCenter raises a conflict if it cannot automatically decide whether to import the rule. This can happen
for example, if the existing rule and the rule for import both have updates that conflict.
As a concrete example, suppose that you update the rule on the testing server. Then you independently
update the rule on the development server. You export the rule from the development server and import
it into the testing server. The testing server raises a conflict while trying to import the rule. Within
PolicyCenter, you must choose whether to keep the existing rule or take the importing rule.
This type of rule conflict can occur also if you update or customize a default business rule and Guidewire
later provides an updated version of that rule in an upgrade. In that case, you must choose either to keep
one or the other of the updated rules or manage the merge of the two rules.
Business Rules Import and Export 309

Rule export does not export utility functions, context objects, or other changes to the business rules plugin. You
must propagate changes to Gosu functions or context objects on the development server to the testing and
production servers.
About Rule Version Conflicts

During the rule import operation, PolicyCenter raises a rule conflict issue if there is a mismatch in rule versions for
any given rule between:
• The existing rule version as defined in the application database
• The importing rule version
The Rule Administrator resolves all rule import conflicts manually through the following business rules screens:
• Import/Export Status screen
• Complete Import screen
• Compare Rules screen
To access the Import/Export Status screen, navigate to the following location in Guidewire PolicyCenter:
Administration→Business Settings→Business Rules→Import/Export Status
The other screens open from the Import/Export Status screen.
IMPORTANT The user who is responsible for resolving rule conflicts, either in a production
environment or non-production environment, must have the necessary rule edit permission.
Business Rule Import Failure

If, for some reason, PolicyCenter detects an error with a rule import operation, it sets the Status value for that import
operation to Failure in the Import/Export Status screen. Click the Failure link to open the Rule Import/Export Failure Reason
screen for more information on the reason for the import failure.
Source and Target Data Models and Rule Export and Import
The business rules data model in the source and target systems must match. Otherwise, the rule import into the target
system fails. Ensure that the business rules data model of the source and target systems is the same before exporting
any business rule. If necessary:
• Modify the business rules data model of one system so that it matches the business rules data model of the other
system.
• Regenerate the business rules export file.
• Repeat the business rules import operation.
Export Business Rules from Guidewire PolicyCenter

About this task
It is possible to export business rules from PolicyCenter in multiple ways from the Underwriting Rules screen. To
export a rules, it must be in one of the following states: Staged, Approved, or Deployed. It is not possible export a
rule in the Draft state.
Procedure
2. Navigate to the following screen:
Administration→Business Settings→Business Rules→Underwriting Rules
3. From the More drop-down list, chose one of the following options:
310 chapter 22: Business Rules Import and Export
Export Se- Exports only the selected business rules. PolicyCenter exports the selected rules from the currently active
lected screen only.
Export All Exports all business rules visible on all pages. PolicyCenter does not export rules hidden because of filters.
Nor does PolicyCenter export rules in the Draft state that have never been deployed. If a rule is in the Draft
state and has one or more previously deployed versions, PolicyCenter exports the last deployed version.
Selecting one of these options opens the Import/Export Status screen.
See also
• “Import Business Rules into Guidewire PolicyCenter” on page 311
• “The Import/Export Status Screen” on page 311
Import Business Rules into Guidewire PolicyCenter

Before you begin
You can only import a business rule into Guidewire PolicyCenter that was previously exported from Guidewire
PolicyCenter.
Procedure
2. Navigate to the following administrative screen:
Administration→Business Settings→Business Rules
3. Choose one of the following options:
Import rules from… Import option

Underwriting Rules screen From the More drop‐down list, select Import from File.
Import/Export Status screen Click Import from File.
Selecting one of these options opens the Import/Export Status screen.
See also
• “Export Business Rules from Guidewire PolicyCenter” on page 310
• “The Import/Export Status Screen” on page 311
• “About the Business Rules Export File” on page 312
The Import/Export Status Screen

After you begin the import or export of business rules, PolicyCenter opens the Import/Export Status screen
automatically. This screen shows progress and status information for each import or export session initiated by any
PolicyCenter rule administrator. To update the information on this screen, click Refresh.
You can also initiate a rule import operation from this screen by clicking Import from File.
The Imports Table

The Imports summary table contains information about each rule import operation, including the following:
• The user who initiated the import operation
• The start time of the import operation
• The source file for the import operation
• The status of the import operation

• A Review button if there are no rule conflicts in this import operation
• A Complete Import button if this import operation generated rule conflicts between the import and existing rules
The Exports Table

The Exports summary table contains information about each rule export operation, including the following:
• The user who initiated the export operation
• The start time of the export operation
• The number of rules in the export file
• A Cancel button that is visible only during the export operation
• A Download button that becomes visible after the export operation completes
See also
• “About the Business Rules Export File” on page 312
• “The Review Import/Complete Import Screens” on page 312
About the Business Rules Export File

The Import/Export Status business rule screen shows the progress of a business rules export operation. After export
completes, click Download to save exported rules to the file system.
PolicyCenter writes the rules for export to a file with a .gwrules extension. This file is in binary format. It is not
possible to view or edit this file outside of Guidewire PolicyCenter. The primary use for this file is to provide the
ability to export business rules from one PolicyCenter server environment and import those rules into another
PolicyCenter server environment.
See also
• “Export Business Rules from Guidewire PolicyCenter” on page 310
• “Import Business Rules into Guidewire PolicyCenter” on page 311
• “The Compare Rules Screen” on page 314
The Review Import/Complete Import Screens

The business rules Import/Export Status screen displays information about the import of each business rule import
operation. The functionality changes on the screen depending on whether a rule conflict exists in an import
operation.
Rule con‐ Functionality

flicts?
No If there are no issues to resolve for a given rule import operation, PolicyCenter displays a Review button in the
Imports table for that rule. Clicking Review opens the Review Import screen.
Yes If there are pending rule conflicts to resolve for a given rule import operation, PolicyCenter displays a Complete
Import button in the Imports table for that rule. Clicking Complete Import opens the Complete Import screen.
The Review Import and Complete Import screens are identical except that the functionality changes depending on
whether you are reviewing a rule import operation or resolving a rule import conflict.
The Review Import / Complete Import screen consists of the following distinct sections:
• “The Rule Import Disposition Table” on page 313
• “The Rule Import Manage Synchronization Table” on page 313
To access the business rules Import/Export Status screen, navigate to the following location within Guidewire
PolicyCenter:
The Rule Import Disposition Table

Clicking Review or Complete Import in the business rules Export Status screen opens either the Review Import or the
Complete Import screen for that rule import operation. These two screens are identical in layout. At the top of the
screen is Disposition table that provides summary information about the rules in the rule import file. The following
list describes the various columns in the table.
Column Description
Outstanding The subcategories under Outstanding have the following meanings:
• New Rule – Number of rules in the import file that have no equivalent on the importing server.
• New Version – Number of rules in the import file that are newer versions of rules on the importing server.
• Rule Deployment – Number of rules being imported that have a deployed version.
• Version Conflict – Number of rules for which there are version conflicts between the importing rules and
the existing rules on the importing server.
Note: If there are no rule conflicts for a given import operation, all subcategories for that operation display
zero (0).
Imported Number of rules PolicyCenter imported from the import file.
Discarded Number of rules discarded by user action.
Applied Edited Number of rules that the Rule Administrator edited and saved.
No Change Number of rules that require no additional action. These rules exist in the importing server already and do
not require updating. For example, the rule in the import file is an earlier version of the rule version on the
importing server.
The Rule Import Manage Synchronization Table

Clicking Review or Complete Import in the Import/Export Status screen opens either the Review Import or the Complete
Import screen for that rule import operation. These two screens are identical in layout. At the bottom of the screen is
Manage Synchronization table that provides summary information about each rule in the rule import operation.
If you undertake an action to resolve a rule import conflict, you must save and apply your change. See “Resolve
Rule Import Conflicts” on page 315 for details.
The following list describes the various columns in the table.
Column Description
Rule Name Name of the rule. PolicyCenter adds a warning next to the rule name if the importing rule name duplicates the
name of a differently configured rule on the importing server.
Status Current status of the rule. Some examples are:
• New Rule – This rule does not exist on the importing server.
• New Rule Version – This version of the rule does not exist on the importing server.
• No Change – The existing rule is a more updated version of the importing rule.
• Rule Deployment – The existing version of the rule needs deployment to match the rule version of the importing
rule.
• Versions conflict – There are conflicts between the existing and importing versions of the rule that you must
manage manually.
Depending on the rule status, it is possible for the list of possible actions in the Available Actions column to change.
Available If there is no import conflicts with the given rule, there are no actions available in this column.
Actions
Column Description
If there is a conflict between the importing and existing rule versions, the following actions are available:
• Existing Version – Keep the existing rule on the importing server and discard the imported rule.
• Importing Version – Keep the importing rule version and discard the existing rule version on the importing server.
• Compare – Click to open the Compare Rules screen. Use this screen to compare the two rules side‐by‐side to view
their similarities and differences and determine the course of action to take.
Existing The version of the rule that exists in the importing server. Click to access a read‐only view of the rule.
Version
Importing The version of the rule in the import file. Click to access a read‐only view of the rule.
Version
In this table, there is an additional column with no heading just to the left of the Rule Name column. An icon (with a
gear) in this column indicates that the PolicyCenter business rules editor does not manage this rule. As the rule is
external to the editor, it is not possible to edit the rule in most respects.
Business Rule Validation Errors

If there is at least one rule validation error in the business rules Complete Import screen, PolicyCenter inserts an
exclamation icon (!) to the immediate left of the name of the invalid rule. If you hover the cursor over the icon,
PolicyCenter shows additional information in the tooltip message.
PolicyCenter shows … to the immediate left of the rule name if the validation work queue has not yet run and there
are pending validations.
The Compare Rules Screen

PolicyCenter opens the Compare Rules screen if you click a Compare link in Complete Import screen. The Compare link
appears if there is a conflict between an existing rule and a rule that you are importing.
The Compare Rules screen creates a side-by-side table view to make it easier to compare the details of the two rules.
PolicyCenter highlights the fields that are different between the two rule definitions.
After you review each of the rules, choose one of the following actions.
Keep Exist- Chose to retain the currently existing rule and discard the importing rule. If you select this option, PolicyCenter
ing Version reopens the Complete Import screen with the Existing Version radio button selected for the rule under Available Actions.
Replace with Chose to accept the importing rule and discard the existing rule. If you select this option, PolicyCenter reopens
Importing the Complete Import screen with the Importing Version radio button selected for the rule under Available Actions.
Version
Edit Select one of the following from the Edit drop‐down list to modify either the existing or importing rule:
• Existing Version
• Importing Version
If you choose to edit a rule version, PolicyCenter opens the rule in edit mode. Editing a rule creates a new Draft
version of the rule.
If you save the edited rule, it becomes the resolved rule version for import completion. After making this update,
you can no longer select either the importing or existing rule in the Complete Import screen. The status of the rule in
the Complete Import screen changes to Edited Version. If you select the rule and apply your change, the Applied Edited
field increments by one (1).
PolicyCenter disables the edit functionality if the existing rule version is in Draft mode.
See also
• “Resolve Rule Import Conflicts” on page 315
Resolve Rule Import Conflicts

About this task
To resolve rule version conflicts that occurs during a rule import process, perform the following sequence of steps.
IMPORTANT The user who is responsible for resolving rule conflicts, either in a production
environment or non-production environment, must have the necessary rule edit permission.
Procedure
1. Navigate to the following location in Guidewire PolicyCenter:
2. For each rule import that has a status other than Completed, click Complete Import.
The Complete Import screen opens.
3. Review each rule that shows an import conflict in the Complete Import screen. Do one of the following:
a. Click the Existing Version or the Importing Version link to see the details for that rule version.
b. Click the Compare link to open the Compare Rules screen in which you can view the two rule versions in a
side-by-side table.
4. (Optional) After reviewing the two rules in the Compare Rules screen, do one of the following:
• Click Keep Existing Version to accept the existing rule on the import server with no change.
• Click Replace with Importing Version to accept the import rule with no change.
• Select an Edit option to open an editable view of either the existing or importing version of the rule.
• Note: PolicyCenter disables the edit functionality if the existing rule version is in Draft mode.
5. (Optional) Make your edits the Compare Rules screen, then click Save Edited Version.
If you edit either of the rule versions in the Compare Rules screen, that version of the rule becomes the resolved
version of the rule. Thereafter, you cannot ever use the importing or existing version of the rule to resolve the
rule conflict.
PolicyCenter reopens the Complete Import screen.
6. In the Complete Import screen, select the resolution type by selecting a radio button in the Available Actions cell
for the rule of interest. Your choices are:
• Use the existing rule version.
• Use the importing rule version.
• Use the new, edited, version of the rule.
7. Select the rule for update by selecting the checkbox to the left of the rule name.
8. Save and apply your changes by clicking one of the following function buttons:
Import Selected Enabled if you select one or more rules. It must be possible to import these rules.
Discard Selected Enabled if you select one or more rules.
Import All Remaining Enabled if there are no remaining unresolved conflicts. It must be possible to import these rules.
Discard All Remaining Enabled if you select one or more rules.
9. Repeat these steps as necessary to resolve all rule conflicts in the rule import operation.
See also
• “The Compare Rules Screen” on page 314
Automatic Import of Business Rules at Server Startup

After the initial PolicyCenter installation, the database is empty of data. The first time that you start the PolicyCenter
application server after installation, PolicyCenter upgrades the database and populates it with data. As part of this
initial database upgrade, it is possible to automatically load data from the following Studio Project folder into the
PolicyCenter database:
configuration→config→import→bizrules
Configuration parameter BizRulesImportBootstrapRules in config.xml controls whether PolicyCenter imports
the business rules in folder bizrules automatically at initial server start, if you have an empty database.
IMPORTANT Guidewire recommends that you set this parameter to true in an environment in which
you plan to do initial rule review or rule development only.
See also
• “Business Rules Import at Initial Server Startup” on page 269
Import PolicyCenter Business Rules at Server Startup

About this task
It is possible to automatically import any business rules that exist in the PolicyCenter bizrules directory at initial
server startup.
Procedure
1. Set configuration parameter BizRulesImportBootstrapRules in config.xml to true.
2. Place the business rules to import in the following location in PolicyCenter Studio:
configuration→config→import→bizrules
3. With an empty database, start the PolicyCenter application server.
PolicyCenter populates the database with the business rules in the bizrules directory.
4. Reset configuration parameter BizRulesImportBootstrapRules to false.
See also
• “About the import Directory” on page 268
Correct Duplicate BizRuleDeploymentId Values

About this task
If you attempt to import a set of business rules that has a BizRulesDeploymentId value that is the same as the
current PolicyCenter cluster, the rule import fails. This happens, for example, if a development cluster has the same
BizRulesDeploymentId value as the production cluster into which you are importing the files. To avoid this
scenario, Guidewire recommends that each PolicyCenter sever cluster have its own unique value for
BizRulesDeploymentId.
Note: You set the value of BizRulesDeploymentId in config.xml.
Correcting this issue is a multi-step process:
Procedure
1. Temporarily, start the application server with the value of BizRulesDeploymentEnabled set to false.
In this case, PolicyCenter ignores the value of BizRulesDeploymentId and import validation does not stop
the import of the file.
2. After the rule import succeeds, restart the application server with the value of BizRulesDeploymentEnabled
set to true.
See also
Business Rule Import after Application Upgrade

About this task
It is possible for Guidewire to provide new rule versions for base configuration business rules in the PolicyCenter
upgrade package. If this is the case, you need to import any Guidewire updates to these business rules into your
installation after you complete the upgrade. If you have altered the base configuration rules in your installation, you
also need to resolve any discrepancies between the imported (source) and existing (target) business rules.
Procedure
1. Complete the upgrade of your PolicyCenter application.
2. Point an application instance to an empty database.
3. Set the value of configuration parameter BizRuleImportBootstrapRules in file config.xml to true for that
application instance.
IMPORTANT Guidewire recommends that you set the value of this parameter back to false
after you complete the following steps.
4. Start the application server.

This action causes PolicyCenter to load the base configuration business rules into the database. See “Business
Rules Import at Initial Server Startup” on page 269 for more information.
5. Within PolicyCenter, navigate to the following location:
Administration→ Business Settings→ Business Rules→ Underwriting Rules
6. From the More drop-down list, select Export All.
7. Navigate to the Import/Export Status screen:
a. Click Download next to the export rule file that you want to download.
b. Save the file (*.gwrules) to a local file location.
8. Point an application instance to the upgraded application database.
9. Within PolicyCenter, navigate to the Import/Export Status screen:
a. Click Import from File.
b. Use the Browse functionality to identify the file that you want to upload, then click Import.
After import completes, PolicyCenter shows either a Review or Complete Import button for your import. If there
are no conflicts between the source and target rules, PolicyCenter shows a Review button. If there are conflicts
between the two sets of rules, PolicyCenter shows a Complete Import button. Clicking one of these buttons
opens either the Review Import or the Complete Import screen.
10. If there are conflicts between the rule versions, you must resolve the rule conflicts before continuing.
See “Resolve Rule Import Conflicts” on page 315 for information on how to use the Complete Import screen to
resolve rule import conflicts.


part 7
Administration Tools
chapter 23
Server Tools
Guidewire provides server tools to assist you with certain server and database administration tasks.
See also
• “Internal Tools” on page 381
Accessing the Server Tools

You must have the internaltools permission to access the Server Tools screens. In the base configuration, the
Superuser role has this permission by default. If the value of parameter EnableInternalDebugTools in config.xml
is true and the server is running in development mode, all users have access to the Server Tools screens.
To access the Server Tools screens, press ALT+SHIFT+T on any PolicyCenter screen after you log into the application.
See also
Batch Process Info

Use the Server Tools Batch Process Info screen to view information about PolicyCenter batch processes, including
writer threads for work queues. Use the drop-down on the Batch Process Info screen to filter the batch process list.
This drop-down contains the following options.
Any Shows all batch processing types.

Schedulable Shows batch prossing types marked as Schedulable in the BatchProcesstype typelist. You define a schedule for
batch processes, including writer processes for work queues, in file scheduler-config.xml.
Runnable Shows batch procesing types marked as APIRunable in the BatchProcesstype typelist. It is possible to run these
types of batch processes using APIs.
From this screen, you can perform the following actions.
Refresh Click to update the information in the Processes table.

Download Click to download some of the information on this screen as a HTML report. See “Download a Batch Process
History Report” on page 323 for details.
Server Tools 321

Suspend Sched- Stops the batch processing scheduler that triggers the execution of the batch processes. After stopping the
uler scheduled, you can restart it using this same button.
The Batch Process Info screen contains the following areas or tabs that contain batch process information.
Processes Lists all the available batch processing types along with information about each individual batch process. It is also
possible to run certain actions on a selected batch processing type from this screen. See “Processes Table Columns”
on page 322 for more information.
Chart Shows the execution time in seconds and the number of operations performed by the batch process over time in a
graphical format.
History Shows records of past runs of the selected batch process in tabular form.
Note: It is possible to run writers for work queues either from the Work Queue info screen or from the
Batch Process Info screen.
See also
• “Administering Batch Processing” on page 81
• “Processes Table Columns” on page 322
• “Chart and History Tabs” on page 323
Processes Table Columns

The Processes area of the Server Tools Batch Process Info screen contains a number of columns. The column labels
have the following meaning:
Column Description
Batch Process Name of the batch processing type.
Description Description of the batch processing type.
Action Actions that you can perform on the selected batch processing type. These actions include:
• Run – Runs a batch process. The Run button is active for all batch process types that belong to the
BatchProcessTypeUsage category UIRunnable.
• Stop – Stops an actively running batch process.
• Download History – Downloads a batch process history report for the selected batch process in HTML format.
See “Download a Batch Process History Report” on page 323 for more information.
You cannot start multiple runs of non‐exclusive custom batch processes from the Batch Process Info screen. In‐
stead, you must use the maintenance_tools command to start multiple runs of non‐exclusive custom batch
processes.
Last Run Date on which this batch processing type last run.
Last Run Status Completion status from the last run of this batch processing type. This field shows as Failed or Interrupted if the
batch process OperationsFailedReasons is non‐null.
Next Scheduled Date of the next scheduled run for this batch processing type.
Run
Schedule Scheduling actions that you can perform on the selected batch processing type. These actions include:
• Stop – Disables the scheduled runs for the selected batch processing type.
• Start – Enables the scheduled runs for the selected batch processing type.
Cron-S M H DOM Column header stands for seconds, minutes, hours, days of month, month, and day of week.
M DOW
322 chapter 23: Server Tools

See also
• See “Work Queues and Batch Processes, a Reference” on page 98 to determine if it is possible to run multiple
instances of a batch processing type.
• See “Maintenance Tools Command” on page 390 for details of how to start multiple batch processes of the same
type.
• See the Integration Guide for a discussion of the meaning of the Exclusive property on a batch process.
Chart and History Tabs

The Server Tools Batch Process Info screen contains Chart and History tabs at the bottom of the screen. Select a batch
processing type to view its chart or history.
The Chart tab shows the execution time in seconds and the number of operations performed by the batch process
over time in a graphical format. The History tab includes a table of records of past runs of the selected batch process
in tabular form.
The History table includes the following information:
Column Description
Start Requested The time at which PolicyCenter received the request to start the process.
Started The time that a batch process started.

Completed The time that a batch process completed.
Scheduled Yes indicates that the start request was the result of the regular execution of a schedule. No indicates that a
user made the request manually.
Server Server on which the start request was made. The server on which the request was made is not necessarily the
server on which PolicyCenter executed the batch process.
Description Optional description.
Ops For batch processes that are work queue writers, the value of Ops (operations) is the number of work items
that the work queue processed. This number includes work items that failed.
Failed For batch processes that are work queue writers, the value of Failed is a counter that PolicyCenter increments
each time work item processing encounters an exception.
It is possible for a work queue to attempt to process a failed work item multiple times. Therefore, the Failed and
Ops numbers do not necessarily match the total number of work items.
Failure Reason For batch processes that are work queue writers, Failure Reason is the reason that a work item failed processing.
See also
Download a Batch Process History Report

Procedure
1. Navigate to the Server Tools Batch Process Info page.
2. Select a batch process in the Processes table.
3. Click Download History in the Actions column of the row for that particular batch process.
4. Select the date rage for the records that you want to download.
5. Click Complete Download.
6. Select the location for the local file download and click OK.
7. Unzip the download file into a local directory.
8. Find and double-click index.html to open the report.
Server Tools 323
See also

Work Queue Info

Use the Server Tools Work Queue Info screen to control and view information associated with work queues. From this
screen, you can track work queues as they process information. Each work queue has both a writer and one or more
workers. You can run the writer to add a batch to the work queue, and you can monitor the progress of the workers
processing the batch.
The Work Queue Info screen contains multiple tables that contain work queue information. These tables include:
• “Work Queue Table Columns” on page 324 table
• “Item Statistics Tabs and Columns” on page 325 table
Besides the work queue information available on the Work Queue Info screen, it is possible to download a number of
specialized reports from this screen as well.
Note: It is possible to run writers for work queues either from the Work Queue info screen or from the
Batch Process Info screen.
See also

• “Work Queue Table Columns” on page 324
• “Item Statistics Tabs and Columns” on page 325
• “Work Queue Reports” on page 327
Work Queue Table Columns

The Work Queue area of the Server Tools Work Queue Info screen includes a table containing information and
functionality related to work queues. The table column headings have the following meaning:
Column Description
Work Queue Name of the work queue.
Available Number of work items available for processing.
Checked Out Number of work items checked out by workers.
Failed Number of work items that failed during processing.
Executors Running Number of workers processing the work queue.
Cluster-wide State Work queue status, for example, started or stopped.

Writer Status Status of the writers.
Actions Actions that you can perform on the selected work queue. These actions include:
• Run Writer – Launches the writer to write work items for the work queue.
• Notify – Wakes workers by notifying the executor that there are items to process.
• Stop – Stops the selected work queue.
• Restart – Restarts the selected work queue.
• Download History – Downloads the historical instrumentation data for the work queue, in CSV format.

See also
• See “Worker Task Management” on page 94 for a discussion of the executor function.
• See “The Work Queue History Report” on page 328 for more information on the work queue history report.
Item Statistics Tabs and Columns

The Item Statistics area of the Server Tools Work Queue Info screen provides information on work items related to the
work queue selected in Work Queue table. This region has three tabs the contain a number of tables.
Tab For more information

By Writers “The By Writers Tab” on page 325
By Executors “The By Executors Tab” on page 325
Work Item “The Work Items Tab” on page 326
The By Writers Tab

The By Writers tab on the Server tools Work Queue Info screen contains item counts generated by the writer associated
with the work queue selected in the Work Queue table. Each row represents one wake period for the writer. These
table column headers have the following meanings.
Column Description
Process ID ID for the writer process.
Item Creation Time Time at which the writer woke and began writing work items. The first item in the table for a queue has a
creation time that matches the queue’s current Last Execution Time for the Writer value.
Server Name of the server running the work queue.
Scheduled Yes indicates that the start request was the result of the regular execution of a schedule. No indicates that a
user made the request manually.
Number of Items Total work items in the queue regardless of status.
Worker End Time Time at which the last worker completed the last item in the work queue.
Execution Time Time, in minutes, since the start of the process (the execution time so far).
Available Total number of available items in the queue.
Checked Out Number of items checked out by workers for processing.
Succeeded Number of items that completed successfully.
Failed Number of items that failed.
The By Executors Tab

The By Executors tab on the Server Tools Work Queue Info screen lists active executors. The table column headers have
the following meanings:
Column Description
Hostname Server on which the executor is running.
Max. Number of Workers Maximum number of workers available to the executor.
Processed Items Number of items processed by the workers.

Exceptions Number of exceptions encountered during processing.
Server Tools 325

Column Description
Failed items Number of work items that failed processing.
Status Status of the executor, for example, running.
Started Timestamp of the start of the executor.
Up For Length of time since the start of the executor.
Under the By Executors tab is a By tasks tab. The By tasks columns have the following meanings:
Column Description
ID The unique identifier of the task.
Writer The identifier of the writer process.
Success Whether the worker succeeded in the processing of the work items.
Checked out items The number of work items the worker checked out.
Processed items The number of work items the worker processed.
Exceptions The number of exceptions, if any, encountered during item processing.
Orphans Reclaimed The number of orphaned work items the worker adopted for processing.
Failed items The number of work items that failed.

Skipped items The number of items that the process skipped.
Started Timestamp of the start of the task.
Ended Timestamp of the end of completion of the task.
Active Whether the task is still active.
Consecutive Errors If processing resulted in exceptions, the number of consecutive work items found that resulted in excep‐
tions during processing by the worker.
The Work Items Tab

The Work Items tab on the Server Tools Work Queue Info screen lists work items. The table column headers have the
following meanings.
Column Description
ID Unique identifier of the work item.
Create time Timestamp of the work item creation time.
Update time Timestamp of the last update to the work item.
Available at Timestamp of when the work item is available processing. This value is null for failed work items.
Server Server that processed the work item.

Writer Writer that created the work item.
Attempts How many attempts a worker has made to process the item.
Activity ID ID of the activity involved. This field is only visible if you select the Activity Escalation work queue.
Subject Subject of the activity involved. This field is only visible if you select the Activity Escalation work queue.
PolicyCenter shows data on this screen during the active execution of database checks only.

Work Queue Reports

Use the Server Tools Work Queue Info screen to generate and download multiple report types of work queue data. You
can use the data from these reports to calculate different kinds of secondary data, for example, work queue
efficiency.
The Work Queue Report

The Work Queue report available from the Server Tools Work Queue Info screen contains the following linked HTML
files:
• Files that include a summary of the work queues
• Files that include detailed information for specific work queues
The report provides data for each worker by thread and by host, such as:
• How long since the start of the worker
• How many items the worker processed
• Throughput (items processed per minute of execution) per thread and per host
• Start and end times for the worker
• Last wake up time
• Items processed for each thread (cumulative, average, and maximum)
• Uptime (cumulative, average, and maximum)
• Execution time (cumulative, average, and maximum)
For each report, you can specify the following:
• The maximum number of writer runs, executors, and batches for each worker
• The number of hours for which to generate item distribution data in the report
Note: Internally, Guidewire sets the allowable maximum number of writer runs to show in the report
at 150.
The Work Queue Runs View

The Work Queue report available from the Server Tools Work Queue Info screen includes a view called Work Queue
Runs. This view shows work queue statistics organized by writer run, including how long it took the writer to
produce work items and how long the workers processed those items. The view can be confusing if the writer is run
repeatedly before finishing the work items produced by the previous run.
See also
• “Download a Work Queue Report” on page 327
Download a Work Queue Report

Procedure
1. Navigate to the Server Tools Work Queue Info page.
2. Click Download on the top-level menu.
3. Specify the parameters for the report.
6. Unzip the download file and double-click index.html to open the Work Queue report.
7. Click the Work Queue Runs link at the bottom of the Work Queue report to open the Queue Runs view.
Server Tools 327
See also
• “The Work Queue Report” on page 327
The Work Queue Raw Data Report

The Work Queue Raw Data report available from the Server Tools Work Queue Info screen is a set of CSV-formatted
files. There is one CSV file for each work queue. The files contain time-sliced raw data from the ProcessHistory
and InstrumentedWorkerTask tables.
You can perform analysis of this data in third-party tools such as Microsoft Excel.
See also
• “The Work Queue Raw Data Report” on page 328
Download the Work Queue Raw Data Report

Procedure
2. Click Download Raw Data on the top-level menu.
3. Select the date range for the records that you want to download.
6. Unzip the download file and open the work queue instrumentation reports individually.
See also
The Work Queue History Report

The Server Tools Work Queue Info screen includes the ability to download information on the history of a particular
work queue. The report, in CSV format, includes the following information:
• Process ID
• Writer start and end times
• Duration
• Failures by workers
You can clear the instrumentation data for all work queues by running the Work Queue Instrumentation Purge
process.
See also
• “Download the Work Queue History Report” on page 328
• “Work Queue Instrumentation Purge Batch Process” on page 126
Download the Work Queue History Report

Procedure
2. Select a work queue in the Work Queue table.
3. Click Download History in the Actions column of the row for that particular work queue.
4. Select the date range for the records that you want to download.
7. Unzip the download file and open the work queue history report.
See also
• “The Work Queue History Report” on page 328
About Work Queue Efficiency

The Server Tools Work Queue Info screen provides a wide variety of data, along with types of multiple work queue
reports. It is possible to use this information to generate additional types of data. For example, the Work Queue
report contains information on the processing time for each item (Item Processing Time). Using this data, you can
calculate the efficiency of a work queue by dividing the work item processing time by the total active time of a
worker.
Set Log Level

Use the Server Tools Set Log Level screen to set a specific logging level for the different logging categories. The
logging levels that you specify in the Set Log Level screen persist until you change them or restart the server.
The Logger drop-down presents the following types of logging categories:
Type Description
Guidewire de‐ Guidewire provides a number of default PolicyCenter logging categories. You can also see the list of
fault Guidewire logging categories by running the system_tools command from a command prompt and adding
the -loggercats option.
Guidewire inter‐ Guidewire provides a number of logging categories that apply to Guidewire internal classes. These logging
nal classes categories start with com.guidewire.*. These logging categories generally look like a fully qualified class
path.
Third‐party soft‐ Guidewire applications integrate with certain types of third‐party software. The manufactures of this soft‐
ware ware provide their own logging categories. These logging categories start with org.*, for example,
org.apache.* or org.eclipse.*. These logging categories generally look like a fully qualified class path.
Making Logging Level Changes Permanent

To make your log level settings permanent, you must edit file logging.properties and set the log level there.
Access this file from the Project window in Guidewire Studio™ by navigating to configuration→config→logging, and
then opening logging.properties.
In the default PolicyCenter configuration, file logging.properties does not contain a unique logger definition for
each Guidewire logger. To add additional loggers to this file, see “Adding Loggers to the Logging Properties File”
on page 27.
Note: Some log4j loggers do not appear on the Set Log Level screen until actually used. This is a
standard log4j behavior.
Setting Logging Levels Through System Tools

It is also possible to set a logging level through the following system tools command.
system_tools -updatelogginglevel logger level
Server Tools 329

The PolicyCenter administration command prompt tools all require that you enter the password of an administrative
user for the tool to work. The use of a user name is optional.
To use this command, you must supply the name of a specific logger category (logger) and the new logging level
(level) for that logger. Use the system_tools -loggercats command option to see a list of valid PolicyCenter
logger categories.
You must refresh the Server Tools Set Log Level screen after using the system_tools command to see your changes
reflected in that screen.
See also
• “The Logging Properties File” on page 25
View Logs
The Server Tools View Logs screen contains the following items:
Field Purpose
Log File Use to select the log file to view.
Filter Use to display the log file lines that contain the specified word or words.
Max Lines to Display Use to set the maximum number of lines to show on the screen.
By default, PolicyCenter writes log files to tmp/gwlogs/PolicyCenter/logs/. To specify a different location in

which the View Logs screen checks for log files, specify a different value for property guidewire.logDirectory in
file logging.properties.
See also
• “The Log Files Directory and the View Logs Screen” on page 26
Info Pages
The Server Tools Info Pages provide information to help manage a PolicyCenter server and database. Guidewire
intends these screens for use by Guidewire Support, Integration Engineers, Database Administrators, and System
Administrators to diagnose existing and potential database-related performance problems. You can also use these
screens to review the results of a load operation.
Configuration
The Server Tools Configuration screen lists the values of the configuration parameters in your PolicyCenter
environment. This screen also includes a Download button. Click Download to download a copy of the following
configuration files:
• config.xml
• messaging-config.xml
• scheduler-config.xml
• work-queue.xml
You can find these files in the config folder within the downloaded ZIP file. The ZIP file also includes a current
directory, which includes the in-memory state of config.xml and work-queue.xml parameters on the server.
Guidewire makes the in-memory state available because it is possible to change certain configuration parameters
using a web service or JMX APIs after server startup.
Archive Info
Use the Server Tools Archive Info screen to view information about any archiving processing taken by PolicyCenter.
Note: the value of configuration parameter ArchiveEnabled must be true before you can view the
Archive Info screen.
On the Archive Info screen, you can take the following actions.
Refresh Click to update the information on the Archive Info screen. Clicking the Refresh button also refreshes the
IArchiveSource plugin.
Download Click to download archive information to an HTML report. This report is a summary of the information
shown on the Archive Info screen.
View Progress Click to open the Work Queue Info screen. From this screen you can view the progress of the Archive work
queue. You can also start an unscheduled run of the archive work queue from the Work Queue Info screen. For
more information, see “Work Queue Info” on page 324.
Export Upgrade In- Click to download and save a .dat file that contains information on the database version for a specific ar‐
fo chive operation.
Import Upgrade In- Click to upload a .dat file that you previously exported. You must browse for a file (click Browse) to upload
fo before the Import Upgrade Info button becomes active.
Browse... Click to open a standard file picker dialog.
The Archive Info screen contains the following tables.
Overview The Overview table includes the following:

• The number of entities that PolicyCenter archived
• The number of items that the archiving process skipped or excluded
• The number of items that failed the archiving process
The screen information divides the excluded items into the following categories:
• Items that PolicyCenter excluded from archiving due to business logic
• Items that PolicyCenter excluded from archiving due to a failure
Click Reset to set the total number of items excluded to zero.
Archive The Archive Source Information table indicates the last refresh time of the Archive Info screen and the
Source In- IArchiveSource plugin. The Archive Source Information also shows the availability of the store, retrieve and delete
formation services. PolicyCenter derives these values from the following variables in Gosu class ArchiveSource:
• storeStatus
• retrieveStatus
• deleteStatus
Possible values for these services include:
• Available – The service is available.
• Failure – The last archive, restore, or delete operation failed.
• Manually – A user has manually flagged the service as unavailable.
• Not configured – The service is not yet configured.
• Not enabled – Archiving is not yet enabled.
• Not started – Archiving is not yet started.
• Queue – The service is not available. PolicyCenter, however, allows the queueing of user requests.
Archive As its name suggests, the Archive Summary by Datamodel Version table shows archiving information organized by data
Summary by model version. For each database version, the table shows the following:
Data Model • The number of entities PolicyCenter succeeded in archiving.
Version • The number entities that PolicyCenter did not archive due to business logic.
• The number of entities that PolicyCenter did not succeed in archiving due to a failure.
Each excluded category has a Reset button to set the count of excluded entities back to zero.
Server Tools 331
Click a specific version to view additional archive information for that data model version. See “The Archive Sum‐
mary by Data Model Version Table” on page 332 for more information.
See also
• “Configuring Archive Logging Operations” on page 37
The Archive Summary by Data Model Version Table

The Server Tools Archive Info screen contains an Archive Summary by Data Model Version table. This area organizes
archive data by data model version. Clicking a specific data model version link opens an Archive Summary for Data
Model (n, n, n, n, n) screen.
On the Archive Summary screen, you can do the following.
Reset counts Click Reset to set the value of Excluded or Failed items back to zero.
Filter the information by Set a value for Begin Time and End Time, then click View.
time period
Review archived items Select the Archived tab to view a list of items archived with this data model version and for the
specified time period.
Review skipped, excluded, Select the appropriate tab to view the reason that archiving process skipped or excluded an item
and failed items from archiving, or the reason the items failed the archiving process. For each reason, you can click
Reset All Items to reset the count to zero.
Domain Graph Info

The Server Tools Domain Graph Info screen includes the following tabs:
• Graphs
• Warnings
IMPORTANT Guidewire strongly recommends that you review the Warnings tab for possible issues
every time you change the data model.
The Graphs Tab

The Graphs tab of the Server Tools Domain Graph Info screen contains the following items:
• A DOT formatted version of the archive domain graph. The DOT format is a means of showing object
relationships in plain text.
• A Download button, which if clicked, generates a ZIP file that contains the text version of the archive domain
graph and the means to view the graph visually.
The Warnings Tab

The Warnings tab of the Server Tools Domain Graph Info screen shows non-fatal violations of the archive domain graph
that PolicyCenter detected while starting the server. PolicyCenter provides warnings for these situations rather than
preventing the server from starting because it is possible to prevent the erroneous situation through business logic.
The server also performs other graph checks while starting. If these checks fail, the server does not start. Because
the server does not start, you cannot use the Archive Graph Info screen to view errors detected by these checks. Instead,
the server reports the error in the application log and prints the archive domain graph in DOT notation.
See also
Consistency Checks
Use the Server Tools Consistency Checks screen to view and run consistency checks on the PolicyCenter database.
The screen consists of two tabs:
• Run consistency checks
• View consistency checks definitions
The Server Tools Consistency Checks screen executes the Database Consistency Check batch process. See “Database
Consistency Check Work Queue” on page 105 for more information.
Note: If the server running a database consistency check fails for some reason, PolicyCenter provides
for a graceful recovery. After the server becomes operational again, PolicyCenter starts the check at
the place it last left off and completes the check.
The Run Consistency Checks Tab

The Run Consistency Checks tab on the Server Tools Consistency Checks Info screen manages the process of running
database consistency check batch processing. From this screen, you can:
• Start, stop, pause, restart, or cancel consistency check batch processing.
• Set the tables or table groups on which to run the consistency checks.
• Set the number of workers to use in the batch processing.
• Set the type of consistency checks to run.
• Review the results of each consistency check batch processing in tabular format.
The Run Consistency Check tab contains the following items.
Item Action
Download all se- Click to download the consistency check results for all selected consistency check runs in the results table. See
lected “Run a Consistency Check from PolicyCenter” on page 335 for information on how to run a consistency check.
Delete Click to delete the results of prior consistency check runs for those rows with a check mark in the results ta‐
ble.
Refresh Click to update the information on the screen while processing is active.
Run Consistency Click to submit a batch processing job to perform one or more database consistency checks. This action exe‐
Checks cutes the Database Consistency Check batch process. See “Database Consistency Check Work Queue” on page
105 for more information.
This button is not available if the Database Consistency Check work queue is not active,
See also
Pause/Resume Click to pause a currently executing batch processing job. During a pause in processing, PolicyCenter changes
Consistency the button label to Resume. Clicking the button resumes processing execution. If this button is not visible, click
Checks the Refresh button.
Cancel Consis- Click to cancel all currently executing batch processes. PolicyCenter displays this button only if there is a cur‐
tency Checks rently executing consistency check.
All tables Click to select All tables (default) to run consistency check against all database tables.
Specify tables If you select Specify tables, PolicyCenter opens a table picker from which you can select the database tables
Specify table against which the consistency checks run. You must select at least one table.
groups If you select Specify table groups, PolicyCenter opens a table groups picker from which you can select the table
groups against which the consistency checks run. You must select at least table group.
Change Click to change the number of workers used to execute the consistency check. Enter a positive integer value. If
you change the number of workers in an active work queue, PolicyCenter stops the existing work queue and
restarts it with the new number of workers.
This button is not available if the Database Consistency Check work queue is not active,
Check all types? Click to select Specify Types to see the list of available check types from which you can make a selection of
Server Tools 333

The Consistency Checks Results Table

After completing a consistency check, the results table columns have the following meanings.
Column Description
Download Click the Download icon to download a ConsistencyCheckRundate.zip file that contains the set of database re‐
ports generated by this consistency check run.
Download Er- Click the Download Errors icon to download a ConsistencyCheckRunErrorsOnlyRunDate.zip file that contains
rors the consistency check log file and stack trace. PolicyCenter displays this column only if there are SQL errors in the
database consistency check.
See “Correct a Consistency Check SQL Failure” on page 336 for more information.
View Click the View icon to open a pop‐up from which you can view the same reports contained in the
ConsistencyCheckRundate.zip file, after you supply your user credentials. PolicyCenter displays this column in
test and development mode only. The column is not visible in production mode.
Delete Click the Delete icon to remove a consistency check row from the table.
Rerun Rerun a consistency check that has a SQL error. PolicyCenter displays this button only if there are SQL errors in
the database consistency check. See “Correct a Consistency Check SQL Failure” on page 336 for more informa‐
tion.
Description List of tables against which the consistency checks ran.
With Errors Number of errors encountered by the consistency checks run.
Total Checks Total number of consistency checks that ran.
Not started Number of consistency checks that have not yet started in the current consistency checks run. Click the Refresh
button to update this data during a currently executing check run.
In progress Number of actively executing consistency checks at any given moment.
Finished Number of consistency checks that completed in this run.
Start time Time at which this set of consistency checks started.
End time Time at which this set of consistency checks ended.
Duration Length of time that this set of consistency checks took to run.
Version Database version, listing (in order):
• Application major version
• Application minor version
• Platform major version
• Platform minor version
• Data model version
See “Understanding Guidewire Software Versioning” on page 365.
ID Identifier (ID) of the stored results of this consistency check run.
If the consistency checks results table lists multiple check runs, use the check box next to a table row to select a
consistency check run for further action.
See also
• “The View Consistency Checks Definitions Tab” on page 335
• “Correct a Consistency Check SQL Failure” on page 336

The View Consistency Checks Definitions Tab

The View consistency checks definitions tab on the Server Tools Consistency Checks Info screen provides information on
the consistency checks available for each database table. In this tab, you can undertake the following actions.
Action Description
Download Click Download to download a ZIP file that contains a set of linked HTML files that describe the consistency
consistency checks that Guidewire provides in the PolicyCenter base configuration.
check infor‐
mation
Search by ta‐ Search by table name to find the consistency checks related to a specified table. Most consistency checks op‐
ble name erate on the specified table, but some checks, such as typelist table checks, operate on other tables as well.
To search, enter a complete or partial table name in the Table name fragment field and click Search. The results of
the search show in a table that lists the table name, the consistency check name, and a description of the
consistency check.
To clear the results of the search, click Reset.
Filter by check Filter the list of consistency check types to see a list of all tables for which that consistency check is available.
type
View SQL Review the SQL query used to generate a given database consistency check. First, select a consistency check,
query then:
• Select the Command tab at the bottom of the screen to view the SQL command of the consistency check.
The SQL command retrieves a count of rows that violate the consistency check.
• Select the Query to identify rows tab at the bottom of the screen (if available) to view the SQL query used to
identify rows that violate the consistency check. SQL queries to identify rows that violate consistency
checks are not available for all check types.
See also
• “The Run Consistency Checks Tab” on page 333
• “Correct a Consistency Check SQL Failure” on page 336
Run a Consistency Check from PolicyCenter

Procedure
1. Navigate to the Server Tools Consistency Checks screen.
2. Select the Run Consistency Checks tab.
3. (Optional) Specify any, or all, of the following items:
• Description that PolicyCenter prepends to the standard description of the tables and checks in the
consistency check reports.
• Tables or table groups on which to run the consistency checks.
• Number of workers to use in executing the consistency check.
• Type of consistency check to run.
4. Click Run Consistency Checks.
5. After the batch process completes, select one of the following in the summary table:
Download arrow Downloads a Zip file that contains the full set of consistency check reports.
Download Errors arrow Downloads a Zip file that contains only the consistency checks that contain SQL errors.
View icon Opens a pop‐up window from which you can view the full set of consistency check reports.
6. If you downloaded the consistency check file to your local system, unzip the file into its own directory.
Server Tools 335
7. Locate the index.html file and double-click it to open it in a browser.

8. In file index.html, do the following:
• Click a table name to view all consistency checks related to that table.
• Click a check name to view all tables that the consistency check runs against.
See also
Run a Consistency Check Using System Tools

Procedure
2. Navigate to the following location in the PolicyCenter installation:
admin/bin
3. Run the system tools command with the -checkdbconsistency option:
system_tools -password password -checkdbconsistency [...]
You must supply a value for password. You can optionally supply additional parameters for the -
checkdbconsistency option.
See also
Correct a Consistency Check SQL Failure

Before you begin
If a database consistency check fails due to a SQL error, PolicyCenter adds the following icons to the Server Tools
Consistency Checks screen:
• Download Errors
• Rerun
Procedure
1. In the PolicyCenter Server Tools Consistency Checks screen, click Download Errors next to the check that
generated the error.
2. Open the error report:
a. Click the number in the With Errors column.
b. On the details report that opens, click Details.
c. Review the SQL query that generated the error.
d. Correct any identified errors.
The listed table name indicates the object against which the consistency check ran. The check name
indicates which consistency check actually generated the error.
3. In the PolicyCenter Server Tools Consistency Checks screen, click Rerun next to the consistency check that
generated the SQL error.
See also
Consistency Check Frequency

Guidewire recommends that you run all database consistency checks periodically during implementation and during
the stabilization period before going into live production. Running the consistency checks during the implementation
and stabilization phases frequently helps to discover issues with application configuration.
Upgrade
Guidewire recommends that you run all consistency checks before and after a database upgrade. Running these
checks before and after a database upgrade helps to verify the validity of the data and identify potential issues.
See also
Database Table Info

The Server Tools Database Table Info screen provides a way to access information about the tables used to store the
PolicyCenter data model. The screen contains the following buttons:
Verify Click to have PolicyCenter compare the database schema with the schema defined in the data
model metadata files. PolicyCenter shows any errors that it finds on the screen.
Download Database Schema Click to download the results of the database schema verification. If there were no errors, the re‐
Verification Errors port is empty.
Download Database Table Info Click to download a ZIP file containing a number reports that document each table in the
PolicyCenter data model.
See also
• “Database Table Info Reports” on page 337
• “Understanding the Database Table Info Reports” on page 338
• “View the Database Table Info Reports” on page 338
Database Table Info Reports

The Server Tools→Info Pages→Database Table Info screen provides the following reports.
Screen Description
All Tables Provides information about all PolicyCenter tables.
Guidewire Version Lists schema version and build information for PolicyCenter.
Indexes by Table Lists the indexes on a table and provides information about the associated key col‐
umns.
Spatial Indexes Provides information about the spatial indexes.
Server Tools 337

Screen Description
Primary Key Constraints by Table Lists the primary key constraints on tables and provides information about the fields
that reference the keys.
Foreign Key Constraints by Table Lists the foreign key constraints on tables and provides information about the tables
referenced by the keys.
Typekey Columns by Typelist Lists the referencing typekey columns for each typelist.
Number of columns and min/max row lengths Displays the number of columns and categories of columns and the minimum and
maximum row length in each table. Overly large row lengths in a database can lead to
inefficiencies in data queries.
Possibly Redundant Backing FK Indexes Lists foreign key indexes that may be redundant, including information about whether
the index is unique and if it is an extension.
Indexes with Shared Prefixes Lists indexes that share multiple leading key columns. It is possible to use this infor‐
mation to find redundant indexes.
Indexes with the Same Key Columns Lists indexes that have the same key columns.
Indexes without a Description Lists indexes that do not have a description.
Indexed Views Lists any indexed views and the view definitions.
Event Paths from Tables to Listening Objects Shows paths from event‐generating entities to non‐event‐generating entities. Each
row in the table contains a non‐event‐generating entity E, along with one of the paths
from an event‐generating entity to E. PolicyCenter uses each path to generate a query
to find the event‐generating entity instances that reference an instance of the non‐
event‐generating entity.
Event Paths to Listening Tables Shows the same paths as those in the Event Paths from Tables to Listening Objects report,
but the entities in the second column are the event‐generating entities. Each entry in
the table contains an event‐generating entity E, along with a path from E to a non‐
event‐generating entity.
Instrumentation Queries Lists the queries that PolicyCenter executes against the database while building the
data the comprises the download reports.
PolicyCenter copies table-specific information from each report category to the individual report for the respective
table, excepting the All Tables and Instrumentation Queries reports.
Understanding the Database Table Info Reports

The report information that you download from the Server Tools Info Pages→Database Table Info screen consists of the
following categories of information:
• Summary reports with links to detailed table reports
• Copies of specific configuration files.
Configuration Files
The download ZIP file contains several directories of configuration files:
• The config directory contains a number of PolicyCenter configuration files as defined at server startup.
• The current directory contains the in-memory state of the batch-process-config.xml, config.xml, and
work-queue.xml parameters on the server. Guidewire makes the in-memory state available because it is possible
to change certain configuration parameters using a web service or JMX APIs after server startup.
View the Database Table Info Reports

Procedure
2. Navigate to the Server Tools Info Pages Database Table Info screen.
3. Click Download Database Table Info.
4. Save the download ZIP file.
5. Unzip the ZIP to a local directory.
6. Find and double-click file index.html to view a table of contents for the downloaded reports in a browser
window.
7. Do one of the following:
• Click a report type to open that report.
• Click config_files on the table of contents page to access copies of certain metadata configuration files.
Database Parameters
The Server Tools Database Parameters screen displays information about the database configuration. You can view the
information on-screen, or you can download a set of linked HTML reports that contain the same information.
To download and access the HTML reports, click Download Database Parameters Info, unzip the resulting file, and click
index.html.
To view the information on-screen, select a view type from the View drop-down list. Depending on the database type,
the View list contains the following items.
View Lists
Database and Driver Versions of the database and its associated driver.
Database Connection Pool Settings Connection pool settings as configured in database-config.xml if using
PolicyCenter to manage the connection pool. See “The dbcp‐connection‐pool
Database Configuration Element” on page 210 for more information on these
parameters.
If you use the application server to manage the connection pool, then this
screen does not show connection pool parameters. Instead, tune the connec‐
tion pool by using the Administrative Console of the application server.
Database Connection Properties Properties related to the database connection.
Guidewire Database Config Guidewire‐specific database configuration parameters. The table lists the
<database> element attributes specified in file database-config.xml or lists
the default value if file database-config.xml does not specify a value.
Guidewire Database Config Statistics Settings Guidewire‐specific database configuration parameters related to statistics gath‐
ering.
Guidewire Database Upgrade Configuration Guidewire‐specific database configuration parameters related to upgrade as de‐
fined by the <upgrade> element in file database-config.xml.
Guidewire Version Information about this specific version of PolicyCenter.
Linguistic Search Options Options defined for linguistic searching in PolicyCenter. See the Globalization
Guide for more information.
Linguistic Search Oracle Functions and Java Source Oracle functions, and Java source, for linguistic searching.
Oracle DB Options and Features Oracle database options.

Oracle DBA Auto Tasks Automated tasks defined for the Oracle database.
Oracle DBA Scheduler Jobs Scheduled jobs for the Oracle database.
Oracle NLS Instance Params Oracle database NLS (National Language Support) options.
Oracle NLS Session Params Oracle database NLS (National Language Support) options.
Oracle Patch History Patches applied to this Oracle database.
Oracle Permanent NLS DB Params Oracle database NLS (National Language Support) options.
Server Tools 339

View Lists
Oracle Registry Oracle registry values.
Oracle Session Init Params Initial parameters of the Oracle session for the current connection with
PolicyCenter.
Oracle SGA Summary Info Information about the Oracle System Global Area (SGA) shared memory.
Oracle State of Current Instance Information on the current Oracle instance connected to PolicyCenter.
Oracle System Statistics Information on Oracle database statistics.
Queries Executed to Build Download SQL queries used to generate the information in the HTML download reports.
SQL Server Database Options Options set on the SQL Server database, such as auto create statistics and
auto update statistics, the recovery model, collation, and so forth.
SQL Server Server Global Server Settings Global server settings for the SQL Server instance.
SQL Server Server Instance Attributes and Values Attributes and values for the SQL Server instance connected to PolicyCenter.
SQL Server Session Properties Properties of the SQL Server session for the current connection with
PolicyCenter.
Summary of Queries Executed to Build Download Number of queries used, and the execution time, to generate the information
in the HTML download reports.
Database Storage
The Server Tools Database Storage Information screen provides information about the space and memory taken up by
the database on the PolicyCenter server. You can both view and download database storage information from this
screen.
The following tables lists the filtering options for the database storage information:
Option Available Description

Include Estimation of Compression Oracle If checked, you need to also select the compression level for estimating savings.
Savings for Tables and Indexes SQL Server
Select Compression Level for Esti- Oracle Only available if you chose to include estimation of compression savings in the
mating Savings SQL Server database storage information. Options include:
• Oracle – Advanced
• SQL Server – Screen, Row
Collect Index Physical Stats for all SQL Server If you select Yes, then PolicyCenter includes statistics on all tables in the database.
tables If you select Specify tables, PolicyCenter includes statistics only on the database ta‐
bles that you select.
Select Mode for Collecting Index SQL Server Options include:
Physical Stats • None
• Detailed
• Limited
• Sampled
Be aware that if you select the Detailed option without selecting individual tables
the report can take a long time to generate.
After setting the desired filtering options, chose one of the following:
• To see the database storage information on the current PolicyCenter screen, click Display Database Storage Info.
• To download the database storage information to view later, click Download Database Storage Info.
After you click Display Database Storage Info, the screen shows information at the bottom of the screen, along with a
drop-down that you can use to filter the information.
Oracle Database Options

To filter the information in the Server Tools Database Storage Information screen, select one of the following options
from the Storage Set to Display drop-down filter:
Storage Set to Dis‐ Description

play
Guidewire Version Guidewire product‐specific information such as application and platform version.
Indexes Alloc Space Space allocation of each index on a table, in megabytes.
Oracle LOBs Alloc Space allocation information for Large Object Blocks (LOBs), sorted by LOB name.
Space
Oracle User LOBs Space allocation information similar to that shown for Oracle LOBs Alloc Space, sorted by table name.
Queries Executed to List of SQL queries used to generate the data. The data includes the SQL used to generate the query and
Build Download other information such as the number of rows returned the time it took for the query to run.
Summary of Queries Simple summary of the number of queries involved in generating the data and the total database time
Executed to Build that the queries took to execute.
Download
Table Alloc Space + Size in megabytes for each table in the database.
Estimated Advanced The information that you see depends on the filter options that you set:
Compression Settings • If you select the Include Estimation of Compression Savings for Tables and Indexes option, PolicyCenter
modifies the storage set name to indicate that fact and provides the requested information.
• If you do not select the Include Estimation of Compression Savings for Tables and Indexes option, you see only
Table Alloc Space as the storage set name. PolicyCenter does not show compression estimation savings
information.
Tablespace Space Size of each tablespace in megabytes, along with information on how much of the space is in use or is
free space.
Tablespaces List of each tablespace in the database along with related information.
User Indexes Information on each user index in the database, selectable by index name.
User Tables Information on each user table in the database, selectable by table name.
SQL Server Database Options

To filter the information in the Server Tools Database Storage Information screen, select one of the following options
from the Storage Set to Display drop-down filter:

play
Data Spaces Lists the filegroups taken up by the data and the amount of space taken and allocat‐
ed by PolicyCenter.
Database Space Details about the amount of disk spaced taken up by the database.
Guidewire Version Guidewire product‐specific information such as application and platform version.
See “Understanding Guidewire Software Versioning” on page 365.
Index Physical Statistics + Estimated Database Lists the statistics about indexes on the physical table. Includes information such as
Compression Savings for Page Level Compo- minimum, average, and maximum record size. To change which tables and indexes
nents the Index Physical Statistics filter shows, select a new index.
Server Tools 341


play
The information that you see depends on the filter options that you set:
• If you select the Include Estimation of Compression Savings for Tables and Indexes
option, PolicyCenter modifies the storage set name to indicate that fact and
provides the requested information.
• If you do not select the Include Estimation of Compression Savings for Tables and
Indexes option, you see only Index Physical Statistics as the storage set name.
PolicyCenter does not show compression estimation savings information.
Index Usage Stats Information on the usage of an index on a table, selectable by table name.
Indexes with High Fragmentation Display average percentage of fragmentation for indexes in the database.
Queries Executed to Build Download List of SQL queries used to generate the data. The data includes the SQL used to
generate the query and other information such as the number of rows returned the
time it took for the query to run.
Summary of Queries Executed to Build Down- Simple summary of the number of queries involved in generating the data and the
load total database time that the queries took to execute.
Tables and Indexes Lists paging and allocation type of a table and its indexes. To change which table the
Tables and Indexes filter shows, select a new table.
TempDB Summary Shows paging information for the database.
Data Distribution
Use the Server Tools Data Distribution screen to a run batch processing job that generates data on the distribution of
various items in the database. You can then view this information on-screen or download a set of reports that details
this information.
There are multiple categories of data distribution reports.
Compari- The report shows data for the various items selected for inclusion in any of the comparison reports. It also shows
son re- the individual row count for the data, as of that date. The intent of this report is to provide a way to visualize data
ports growth. The download ZIP file contains an HTML report plus separate CSV reports. The HTML report contains dis‐
tinct columns representing the data from each report used for comparison.
Combined The report combines information from multiple runs. The intent of the report is to provide a way to create smaller
reports reports that require less generation time and then combine the information into one report. The download ZIP file
contains an HTML report that contains information on each of the previous reports combined into this report plus
tables that contain the combined data.
It is also possible to start the data distribution batch process (DataDistribution) directly from the command
prompt by using a maintenance_tools command option.
• “Generate and View a Data Distribution Report” on page 342
• “Download Comparison and Combined Data Distribution Reports” on page 343
Generate and View a Data Distribution Report

Procedure
1. Navigate to the Server Tools Info Pages and select Data Distribution.
2. Select from the available options listed under Data Distribution Batch Job Parameters.
3. Click Submit Data Distribution Batch Job.
4. After the batch job completes, select one of the following in the summary table:
Download arrow Downloads an DataDistribution.zip file that contains the set of database reports.
View icon Open a pop‐up window from which you can view the same reports contained in the
DataDistribution.zip file, after you supply your user credentials.
5. If downloaded the report to your local system, unzip the download Zip file into its own directory.
7. Use the links on the screen to navigate through the distribution reports.
See also

• “Download Comparison and Combined Data Distribution Reports” on page 343
Download Comparison and Combined Data Distribution Reports
Before you begin
To create either a comparison or a combined data distribution report, two or more data distribution reports and their
output must exist in the database.
Procedure
1. Navigate to the Server Tools Info Pages→Data Distribution screen.

2. In the summary table, select (check) at least two of the generated reports.
PolicyCenter enables the following download buttons:
• Download Comparison Zip File
• Download Combined Zip File
3. Click the appropriate button.
Database Statistics
The Server Tools Database Catalog Statistics Information screen provides reports about out-of-date statistics in the
database indexes, histograms, staging tables, and PolicyCenter application tables. This screen is not available with
the development-only QuickStart database.
The Database Statistics screen contains the following tabs.
Tab Description
Database Statistics Use to generate and download database statistics reports for the entire database or for specific tables.
Information See “Generate and Download a Database Statistics Report” on page 344.
Execution History Use to view information about individual database statistics reports and take action with respect to each
report. See “Working with Database Statistics Reports” on page 344.
Oracle Statistics Use to work with Oracle database preferences for database table statistics. This tab is only available after
Preferences you set the useoraclestatspreferences attribute to true and perform an application upgrade.
See “Using Oracle AutoTask for Statistics Generation” on page 262 for more information.
Database Statistics Generation

Batch processing type Database Statistics (DBStats) generates the data available on Database Catalog Statistics
Information screen.
Server Tools 343
Development Mode
In development mode, it is possible to run Database Statistics batch processing in any of the following ways:
• From a command prompt, using the -updatestatistics option of the system_tools command
• From the Execution History tab of the Server Tools Database Statistics screen
Production Mode
In production mode, it is possible to run Database Statistics batch processing in the following ways only:
• From a command prompt, using the -updatestatistics option of the system_tools command.
Oracle AutoTask
For Oracle databases, it is possible to use the Oracle Autotask infrastructure to manage the collection of database
table statistics. To use Oracle AutoTask, do the following:
• Disable any scheduled runs of DBStats batch processing.
• Set attribute useoraclestatspreferences attribute on the <databasestatistics> element in file database-
config.xml to true.
You must use either Oracle AutoTask or DBStats batch processing to manage the collection of database statistics.
Do not attempt to use both methods simultaneously.
See also
• “Generate and Download a Database Statistics Report” on page 344
• “Working with Database Statistics Reports” on page 344
Working with Database Statistics Reports

To access the Database Catalog Statistics Information screens, click Info Pages→Database Statistics in the left-hand
navigation pane of the Server Tools screen. Depending on the database type, you can access the following tabs from
this screen:
Tab More information

Database Statistics Info “Generate and Download a Database Statistics Report” on page 344
Execution History “The Execution History Tab” on page 345
Oracle Statistics Preferences “The Oracle Statistics Preferences Tab” on page 345
Generate and Download a Database Statistics Report

About this task
Use the Server Tools Database Statistics screen to generate database statistics about how the PolicyCenter application
and data model interact with the physical database.
Procedure
2. Navigate to the Server Tools Info Pages→Database Catalog Statistics Information screen.
3. Select one of the following values for the View database catalogs statistics on all tables option:
Yes View statistics data for all database tables. This selection can take a few seconds to complete, depending on the
size of your database.
No View statistics data for the selected tables only. If you select this option and do not select any tables, PolicyCenter
shows statistics data from the database metadata only.
4. Click Download.
5. Save the report ZIP file to a local directory.
6. Unzip the file.
7. Double-click file index.html to open the HTML report summary.
8. Use the report links to navigate through the various database statistics reports.
See also
• “Generate and Download a Database Statistics Report” on page 344
• “Working with Database Statistics Reports” on page 344
The Execution History Tab

The Execution History tab of the Server Tools Database Statistics screen lists information about each run of the batch
process that updates database statics. This information includes the following:
• Date and time of the start and stop of the database statistics generation
• Type of database statistics generation, either full or incremental
• Description of the particular database statistics generation
From this screen, you can also perform the following actions.
Action Description
Refresh Update the contents of the summary table.
Run Incremental Statis- Generate incremental database statistics. This action updates database statistics for tables exceeding
tics the change threshold only. This functionality is available in development mode only.
Run Full Statistics Generate full database statistics. This functionality is available in development mode only.
Download Click the download arrow to download the data from this report.
Delete Click the trash can icon to delete this data row from the summary table.
The Oracle Statistics Preferences Tab

The Oracle Statistics Preferences tab presents information about the table statistics preferences set for the Oracle
database. To view this information, navigate to the following location in Guidewire PolicyCenter:
Server Tools Info Pages→Database Statistics screen, Oracle Statistics Preferences tab
To be able to view the Oracle Statistics Preferences tab:
• The database is Oracle.
• Attribute useoraclestatspreferences on the databasestatistics element in file database-config.xml is
set to true.
Note: You must perform an upgrade (either full or rolling) after setting attribute
useoraclestatspreferences to true in order for the change to the useoraclestatspreferences
attribute to become effective.
Actual and Configured Table Preferences

To configure database statistics in BillingCenter, one sets statistics options in file database-configl.xml, either at
the global database level, or, at the individual table level. To be clear, you set statistics options in database-
configl.xml, which BillingCenter then uses to set the actual table statistics preferences during the upgrade.
Server Tools 345
It is also possible to set table statistics preferences directly in the Oracle database by executing the following
command:
DBMS_STATS.SET_TABLE_PREFS
If you set table statistics preferences directly, the actual table statistic preferences no longer match the configured
table statistics preferences. During a database upgrade, BillingCenter ignores any actual table statistics preferences
and sets table statistics preferences to those defined in the new database-config.xml file. The information
provided on the Oracle Statistics Preferences tab provides a mechanism to capture direct changes made to the table
statistics preferences that are lost during a database upgrade
Button Actions
The following list describes the actions of the individual button in the Oracle Statistics Preferences tab.
Button Action
Refresh Refreshes the table information on the screen.
Download Downloads the database-config table statistics options as a JSON file.
Reapply Config Reapplies the table statistics options configured in file database-config.xml to the database,
discarding any preferences set outside the application. As you initiate this process,
BillingCenter inserts an entry into the application log and then inserts another entry at the
end of the process.
Generate Config to Match Actual Generates an XML file that details the table statistics preferences actually in use in the
database.
Dev Mode only (Drop‐down) Selects that environment in which to generate the table statistics.
You can also filter the list of tables using the drop-down at the right of the tab.
Oracle Statspack
The Server Tools Oracle Statspack screen provides a means to download HTML reports based on any two Oracle
database statspack snapshots from the same instance start-up time. You create statspack snapshots by using a tool
such as SQL*Plus or SQL Developer. Also, Oracle provides a script called spauto.sql that you can modify and run
to automate statspack snapshot gathering.
The Server Tools Oracle Statspack screen is available only if the database server is Oracle. For PolicyCenter to display
statspack information, you must also:
• Install the statspack package in your Oracle database (spcreate.sql).
• Grant SELECT privileges on all the PERFTEST tables to the PolicyCenter database user.
If you do not install the statspack or fail to grant the correct permission, PolicyCenter displays an error message on
the screen. You cannot select any snapshots on the screen either. Refer to the Oracle documentation for statspack
installation instructions.
Guidewire Recommendations
Guidewire recommends the following with regards to Oracle statspack snapshots:
• Collect statspack snapshots at regular intervals if you do not have a license for Oracle AWR.
• Collect the snapshots at level 7 or higher to collect execution plan and segment statistics.
• Regularly purge statspack snapshots older than a certain period.
Guidewire recommends that you use the Guidewire AWR tool, rather than Oracle Statspack, if you have a license
for the following:
• Oracle Diagnostics package
• Tuning package
If you do not have these licenses, Guidewire recommends that you acquire the two licenses.
See also
• “Oracle AWR” on page 347
Oracle AWR
The Server Tools Oracle AWR Information screen is available only if the database server is Oracle. Use the Oracle AWR
Information screen to generate a set of Guidewire performance reports using AWR snapshots that you define in the
database. The Guidewire AWR reports provide a view of the database activity that contains more detailed
information than the Oracle Standard AWR report available with the Oracle database.
The Guidewire AWR reports provide the following additional information:
• Messaging analysis
• Concurrent batch processes
• Distributed worker activity
• Database statistics
The Guidewire AWR reports require that you have a license for the following:
• Oracle Diagnostics package
• Tuning package, if you select either Probe in Memory SQL Monitoring or Probe on Disk SQL Monitoring
Refer to the Oracle documentation for details.
See also
• “Oracle Statspack” on page 346
• “Download an Oracle AWR Unused Indexes Report” on page 349
Generate Guidewire AWR Reports

About this task
Use the Server Tools Oracle AWR Information screen to generate a set of performance reports using AWR snapshots
that you define in the database.
Procedure
2. Navigate to the Server Tools Info Pages→Oracle AWR Information screen.
3. Set the options for the report.
See “Automatic Generation of Oracle Standard AWR Reports” on page 348 for a discussion of the Include
native Oracle report option.
4. Select two database snapshots from the list at the bottom of the screen.
Each snapshot must share the same Oracle instance startup time.
5. Click Generate Perf Report.
6. After PolicyCenter completes generating the report, select one of the following:
Download arrow Downloads an AWRReport.zip file that contains the set of database reports.
View icon Open a pop‐up window from which you can view the same reports contained in the AWRReport.zip
file, after you supply your user credentials.
Generate Guidewire AWR Reports Using System Tools

Before you begin
Generating a Guidewire AWR report using system tools requires that you know the ID of two database snapshots.
Server Tools 347
Procedure
2. Navigate to the following location in the PolicyCenter installation:
admin/bin
3. Enter the following command to generate a list of database snapshot IDs:
system_tools -password password -oraListSnaps numSnaps
You must enter a value for password. You must limit the list of snapshots by entering a value for numSnaps.
4. Enter the following command to generate the Guidewire AWR report:
system_tools -password password -oraPerfReport beginSnapshotID endSnapshotID
You must enter the IDs of two database snapshots.
Result
The system_tools -oraPerfReport command option reports the process ID of the process generating the
performance report. You can check on the status of this process using the -processstatus option of the
maintenance_tools command.
See also
Generate Oracle Standard AWR Reports Using a SQL Script

About this task
Guidewire provides a SQL script that you can execute against the database server to generate the Oracle Standard
AWR reports.
Procedure
2. Navigate to the Server Tools Info Pages→Oracle AWR Information screen.
3. Generate and download an AWRReport.zip file as described in “Generate Guidewire AWR Reports” on page
347.
4. Extract the contents of the AWRReport.zip file to a local directory.
5. Navigate to the following directory:
AWRInfo→sqlscripts
6. Execute the following script against the database server:
oraclescripts.sql
Executing this script creates an Oracle Standard AWR report.
Automatic Generation of Oracle Standard AWR Reports

Guidewire provides the means to generate an Oracle Standard AWR report automatically as you generate a
Guidewire AWR report. To do so, ensure that you include report option Include native Oracle AWR report in the list of
options against which to run the report. The PolicyCenter base configuration sets (checks) this option by default.
To generate a report with this option, the Oracle database user must have EXECUTE privilege for database package
DBMS_WORKLOAD_REPOSITORY. If the database user does not have the required privilege, PolicyCenter does not
generate a report and prints a message to the application log.
If the required privilege does not exist for the database user, do one of the following:
• Request that the database administrator grant the required privilege and rerun the report from Oracle AWR screen.
• Manually execute script oraclescripts.sql from the database server.
See also
• “Generate Oracle Standard AWR Reports Using a SQL Script” on page 348
Oracle AWR Unused Indexes Information

The Server Tools Oracle AWR Unused Indexes Information screen provides the means to generate and download a report
that contains information on the following types of indexes:
• Indexes that have no logical or physical reads
• Indexes that are not found in query plans
The Oracle AWR Unused Indexes Information screen is available only if the database server is Oracle.
See also
Download an Oracle AWR Unused Indexes Report

About this task
The Server Tools Oracle AWR Unused Indexes Information report contains information on indexes that have no logical or
physical reads or indexes that are not found in query plans.
Procedure
1. Log into Guidewire PolicyCenter using an administrative command.
2. Navigate to the Server Tools Info Pages→Oracle AWR Unused Indexes Information screen.
3. Set the options for the report.
4. Select two database snapshots from the list at the bottom of the screen.
Guidewire recommends that you select snapshots that have a wide range.
5. Click Download.
6. In the download dialog, set the download location.
7. Click OK to download and save the report.
8. Open the downloaded ZIP file and click index.html to open the index to the linked set of report files.
Oracle Outlines
The Server Tools Oracle Outlines screen is only available if the database server is Oracle. The screen contains
information on any Oracle outlines defined in the PolicyCenter Oracle database. Oracle defines an outline as a
collection of hints associated with a specific SQL statement, used to provide SQL execution plan stability. Consult
the Oracle documentation on how to write and use Oracle Outlines.
The Oracle Outlines summary table contains the following information:
Name Name of the Oracle outline. Click to open the Outline Details screen. This screen contains the SQL hints used to
define the outline.
Category Optional name used to group stored outlines.
Server Tools 349

Used Whether the Oracle database has ever used this particular outline.
Time Stamp Date and time of the last use of this outline.
Version Oracle database version.

SQL SQL query text that creates the database outline.
Signature Unique identifier for this outline.
Compatible Whether the stored outline is compatible with this database version.
Enabled Whether Oracle enables this outline in the database.
Note: Oracle AWR contains a column that indicates if PolicyCenter used an Oracle Outline for a
given SQL statement.
See also
• “View an Oracle Outlines Report” on page 350
View an Oracle Outlines Report

About this task
The Server tools Oracle Outlines screen contains information on any Oracle Outlines defined in the Oracle database.
Procedure
1. Log into Guidewire PolicyCenter using an administrative command.
2. Navigate to the Server Tools Info Pages→Oracle Outlines screen.
3. Click Download.
4. In the download dialog, set the download location.
5. Click OK to download and save the report.
6. Open the downloaded ZIP file and click Outlines.html to open the report.
See also
• “Oracle Outlines” on page 349
SQL Server DMV Snapshot

The Server Tools SQL Server DMV Snapshot screen is available only if the database server is SQL Server. Use this
screen to generate and download performance reports using SQL Server Dynamic Management Views.
It is also possible to generate a SQL Server DMV report using the -mssqlPerRpt option of the system_tools
command.
See also
Generate SQL Server DMV Snapshot Report

Procedure
1. Log into Guidewire PolicyCenter using an administrative report.
2. In the left-hand navigation pane, navigate to Server tools Info Pages→SQL Server DMV Snapshot.
3. (Optional) Select Include Database Statistics if you want to include database statistics information in the report.
4. Click Generate Perf Report.

This action launches an internal batch process that gathers performance data and creates the report.
5. After the batch process completes, select one of the following in the summary table:
Download arrow Downloads an DMVReport.zip file that contains the set of database reports.
View icon Open a pop‐up window from which you can view the same reports contained in the DMVReport.zip
file.
6. After you download the DMVReport.zip file, unzip the file into its own directory.
7. Locate file index.html and double-click it to open it in a browser.
8. Use the links on the screen to navigate through the distribution reports.
Microsoft JDBC Driver Logging

The Server Tools Microsoft JDBC Driver Logging screen is available only if the database server is SQL Server. Use this
screen to specify the following:
• The logging level for the Microsoft JDBC driver. Be cautious setting the logging level, as detailed logging can
slow the system significantly.
• The logging format, either Simple, for a more readable format, or XML for XML output, usually parsed by another
system.
• The log file location, adding special components that PolicyCenter replaces at runtime. For example, use %u to
append a unique number to each log to avoid conflicts.
If you have already enabled logging through database-config.xml or from previous use of this screen, use this
screen to make additional changes to reset the logging level. After clicking Set Logging Level, PolicyCenter flushes
and closes any existing logging files before beginning the new trace. If you choose OFF as the logging level,
PolicyCenter disables any current logging as well as flushing and closing any existing logging files.
The ability to control logging of the Microsoft JDBC driver through PolicyCenter only works if using the internal
connection pool, not if using an external JNDI data source connection pool.
Note: Using this screen is a better option if tracing a particular operation, in order to minimize system
impact and size of the trace file.
See also
Load History
The Server Tools Load History Information screen displays information about specific PolicyCenter database
operations. For example, loading data into the staging tables impacts the loader history information on this screen.
On this screen:
• Click Refresh to reload and update the table data.
• Click Edit to make the Description field for each load history writable.
The load history summary table contains the following information:
Download Click the Download arrow to download a LoadHistoryInfo.zip file that contains a set of HTML reports. The re‐
ports consist of a summary table and a set of links to individual reports that load different views of the database
operation data onto the screen. These are the same reports that are available by clicking the View icon.
View Click the View icon to view the on‐screen Load History Detail report for the selected database operation. The detail
view consists of a summary table and a set of tabs that load different views of the database operation data onto
the screen. These are the same reports that are available for download by clicking the Download icon.
Delete Click the trash can icon to remove the data for this database operation.
Server Tools 351

Load Opera- Type of database operation. This can be, for example, any of the following:
tion Type • Database table load operations
• Staging table clearing operations
• Database statistics generation operations
Start Time Start date and time of the database operation.
End time Completion date and time of the database operation.
Duration Length of time, in seconds, to complete the operation.
Error Count Number of reported errors for this database operation..
Calling User Name of user who initiated this database operation.
Description Click Edit to make the Description field for each database operation row writable. Click Update after entering text to
save your work and update the field.
See also
• “View a Load History Report” on page 352
• “The Load History Detail Report” on page 352
View a Load History Report

Procedure
2. Navigate to the Server Tools Load History Information screen.
3. For the database load operation that interests you, select one of the following in the summary table:
Download arrow Downloads a LoadHistoryInfo.zip file that contains the set of database reports.
View icon Opens a new screen that displays the Load History Detail report. See “The Load History Detail Report” on
page 352 for details.
4. If you downloaded the report file, unzip the file into its own directory.
6. Use the links on the screen to navigate through the linked reports.
The Load History Detail Report

The Server Tools Load History detail report contains a summary table for the selected database operation. As with the
main Load History Information table, the report summary view lists information about the type of database operation,
the user who initiated the operation, and similar information.
Underneath the report summary table is a set of tabs, each of which loads a different view of the load operation data
onto the screen. The following list describes these tabs.
Parameters Lists the values of the configuration parameters used in generating the data for the database operations reports.
See “Load History Detail Report ‐ Parameters tab” on page 353 for more information.
Steps Lists the individual steps in the data load operation. Click a step to view detail data about that step. See “Load
History Detail Report ‐ Steps tab” on page 353 for more information.
Row Counts Lists information about database tables impacted by the data load. Use the information on this screen to quickly
assess whether the amount of data loaded by operation was the amount that you expected the operation to
load.
Integrity Lists the integrity checks that PolicyCenter ran against each affected database table before the data load opera‐
Checks tion.

Inserts Lists the results of the SQL INSERT_INTO queries that PolicyCenter ran against the affected database tables. This
tab lists all INSERT_INTO run for this database load. In contrast, the Steps tab lists only the INSERT_INTO queries
that run for each particular step.
Callbacks Lists the operations that PolicyCenter executes before and after a staging table load operation. This tab lists all
callback operations for this database load. In contrast, the Steps tab lists only the callback operations that exe‐
cute for each particular step.
Statistics Lists the SQL commands used to generate database statistics.
Commands
Load History Detail Report ‐ Parameters tab

The Server Tools Load History Detail report contains a summary table for the selected database operation. Directly
underneath this table is a Parameters tab. Selecting this tab opens a table that lists the database parameters that affect
the results of the Load History Detail report. The table shows the current values of these parameters.
Setting Parameters for the Load History Report

The following table describes how to set the database parameter values that affect the results of the load history
detail report.
Parameter How to set

allowRefsToExistingNonAdminRows Use the table_import -allreferencesallowed command prompt option.
clearErrorTable Use the table_import -clearerror command prompt option.

parallelism Modify file database-config.xml.
populateExclusionTable Use the table_import -populateexclusion command prompt option.
updateDBStatisticsWithEstimates Use the table_import -estimateorastats command prompt option.
Load History Detail Report ‐ Steps tab

The Server Tools Load History Detail report contains a summary table for the selected database operation. Directly
underneath this table is a Steps tab. Selecting this tab opens a table that lists the steps involved in the operation in
this particular load operation.
Steps and Ops Link

If a step contains multiple sub-steps, that step becomes a clickable link that opens a Steps and Ops detail screen for
that step. The Steps and Ops detail screen contains a summary table that describes the substeps involved in this step
operation. It can also contain an additional table depending on the type of step. For example, both callback and
INSERT_INTO operations generate an additional table of database information.
Be aware that the Table Name column can contain the name of an actual database table or the name of a database
callback. If it is a callback, then the associated Callbacks table contains specific information about that callback
operation.
Load Integrity Checks

The Server Tools Load Integrity Checks screen reports on the SQL integrity checks that run as a database load
operation executes. For each integrity check, the screen lists the SQL query that the check performs, along with a
description of the integrity check.
This screen has two tabs.
View by Staging Ta- This tab lists the set of integrity checks that PolicyCenter executes against a specific staging table. Use the
ble Staging Table drop‐down to select the table of interest.
Server Tools 353

View by Load Error This tab lists the set of tables against which PolicyCenter runs a particular integrity check. Use the Load
type Error Type drop‐down to select a specific integrity check.
In both tabs, you are able to set the value of Allow Non Admin References to true. If you do, PolicyCenter checks
foreign key references to administrative tables on load, such as users and groups. In the base configuration,
Guidewire disables these references by default.
See also
Load Errors
The Server Tools Load Errors screen displays errors generated by failed integrity checks. You can use this screen to
drill down through a table name to the specific error generated by a load operation. Errors relate to a particular
staging table row. For each error, the Load Errors screen shows:
• The table
• The row number
• The logical unit of work ID (LUWID)
• The error message
• The data integrity check query that failed.
In some cases, PolicyCenter cannot identify or store a single LUWID for the error.
Runtime Environment Info

The Server Tools Runtime Environment Info screen lists information about the runtime environment for PolicyCenter.
This information includes Guidewire platform and PolicyCenter build information, system properties, and
environment variables.
Safe Persisting Order

The Server Tools Safe Persisting Order screen provides information on the following:
• The order in which PolicyCenter writes object data to the database during a bundle commit.
• The order in which PolicyCenter runs the Preupdate rules on a set of objects.
The screen contains a table with columns that show the following:
Column Description
Entity Entity name
Order Entity ranking order
Table Entity database table name
Preupdate Whether the entity triggers a Preupdate rule set
To see only those objects affected by Preupdate rules, select With rules only from the drop-down filter.
Note: PolicyCenter does not use Preupdate rules.
Bundle Commits to the Database

PolicyCenter uses the defined object ordering in every bundle write to the database for insert, write, or delete
operations. Thus, during a commit to the database of a bundle containing multiple objects, PolicyCenter writes the
objects to the database in the order listed in the table Order column.
Preupdate Rules
In running the rules in the Preupdate rule set, PolicyCenter first computes the set of objects on which to run the
Preupdate rules. PolicyCenter then runs the Preupdate rules for this set of objects as determined by the order listed in
the table Order column.
Loaded Gosu Classes

Guidewire no longer supports the Server Tools Loaded Gosu Classes screen.
Serialization Info
The Server Tools Serialization Info screen (under Info Pages) shows, for any specific server in the cluster, the entire set
of Java objects (classes) deserialized by that server instance. This screen contains an optional filter (Including listed in
the serialization whitelist classes) that filters the list of classes:
• Checking this box means that the list of class names includes the names of all Java classes encountered and
deserialized by the local server. This list includes the names of classes that exist in the serialization white
(permitted) list as well.
• Un-checking this box means that the list of class names includes only the names of classes encountered and
deserialized that are not on the serialization white list. Guidewire recommends that you add these classes to the
serialization whitelist. After you complete your whitelisting of Java classes, the class listing will be empty.
Enabling Object Deserialization
Configuration parameter SerializationWhitelistEnabled in config.xml determines whether PolicyCenter

permits only those Java classes placed on a serialization whitelist to be deserialized. Before you enable the use of the
whitelist, ensure that you first add any additional types needed due to customizations. Primarily, this is due to
creating batch processes that accept custom objects as arguments. Such objects must be serialized if the batch
process is invoked from a server other than the server with the batch role. If you do not add these objects to the
whitelist, it is possible that the related services that uses them to not function properly.
See also
The Black and White Serialization Lists

In Guidewire Studio, you can access the serialization black list and white lists in the following location:
configuration→config→security
To blacklist a Java class, add an entry in serialization-blacklist.lst in that folder. To whitelist a Java class,
add an entry in serialization-whitelist.lst in that folder.
In making entries in these files, use the following syntax:
• Place a # symbol at the beginning of a line to indicate that the line is a comment.
• Use a separate line for each class or package name, for example, gw.api.myPackage.*.
• Do not place the * separator in the middle of a class or package name. For example, do not do the following:
#Incorrect example
gw.api.*.myClass
• Use blank lines and leading spaces as desired to enhance readability of the file.
Server Tools 355

Management Beans
Note: The Management Beans screen is accessible to users with the soapadmin permission only.
The Server Tools Management Beans screen lists PolicyCenter management beans, which are PolicyCenter objects that
represent different resources. Click the name of a resource to open the Guidewire Managed Bean Properties screen for
the selected resource.
The Bean Properties screen most often contains a MBean Property table that lists the properties associated with the
selected resource bean. Property values are either read-only or editable. Guidewire marks the editable properties
with a small blue triangle in the upper left-hand corner of the Value field. Click anywhere in an editable field to make
that field editable. After modifying a field, you can either save your work or cancel your changes by clicking Save or
Cancel.
A few Bean Properties screens also contain an Operation table that lists available operations associated with this
resource bean. Click Execute to execute a selected operation. The Result field shows the result of the operation.
Startable Services
The Server Tools Startable Services screen contains summary information on all of the startable plugin services in the
PolicyCenter cluster. A startable plugin is a special type of PolicyCenter plugin. At runtime, PolicyCenter creates a
background process, or service, for each startable plugin. The summary table provides the following information for
each service.
Name Name of the plugin service.

Status Current status of the service.
Host One of the following:

• Name of the host on which the plugin service is running.
• Distributed, if cluster configuration allows the startable plugin services to run on all cluster members.
Action Available actions, either starting or stopping a selected service.
See also
Cluster Members and Components

The Server Tools Cluster Members screens provide information on the server cluster installation. The screen consists
of the following areas:
• This Application Instance
• Application Server Instances
• Components tab
• History tab
From this screen you can also schedule (or cancel) a planned shutdown of a specific server in the PolicyCenter
cluster.
See also
• “The Cluster Components Screen” on page 360

• “Download a Server Component Report” on page 362
• “Schedule a Planned Cluster Member Shutdown” on page 362
Cluster Members – This Application Server Instance

The Server Tools Cluster Members (This Application Server Instance) screen provides the following information about the
local server instance from which you access the screen.
Host Name of the machine on which this PolicyCenter server instance is running.
Server ID Name for this server instance. You specify the server ID by either adding an entry for this cluster member in the
<registry> element in config.xml, or, by setting a JVM option at server startup. If you do not specify a server
ID, PolicyCenter uses the host (machine) name as the server ID.
UUID Universally Unique ID for this server machine. PolicyCenter randomly generates a UUID for each machine at each
machine startup.
Server Roles List of server roles assigned to this PolicyCenter server.
See also
About Planned Server Shutdowns

The Server Tools Cluster Members screen provides the ability to start a planned or scheduled shutdown of a specific
server in the PolicyCenter cluster. This action affects only the server on which you start or stop the scheduled
shutdown. The Actions column of the Application Server Instances table contains a button that toggles between Start
planned shutdown and Cancel planned shutdown, depending on whether a server shutdown is currently scheduled.
Scheduling a Server Shutdown

Clicking Start planned shutdown opens the Schedule Planned Shutdown screen. From this screen, you can set the details
of the planned server shutdown. In this screen, you need to set the following items:
Banner text Choose the text of the banner message to show on the PolicyCenter screen to warn
users of the server shutdown. The banner message occurs immediately as you
schedule the shutdown and contains a countdown to the time of the scheduled
shutdown. You can choose from several listed messages or create your own custom
message.
Shutdown date and time Set the date and time of the server shutdown.
Action to take with respect to running Decide how to handle any currently running batch processes
batch processes
After you initiate a scheduled shutdown, PolicyCenter does the following on the affected server:
• It does not start any new batch processes.
• It manages the stopping of any currently running batch processes depending on the setting for Terminate Batch
Processes field.
• It requests that all work queues and message destinations stop processing immediately.
• It completes the transmission of any current messages without starting any new message transmissions.
• It completes the processing of any current work items without beginning work on any new work items.
After you click OK in the Schedule Planned Shutdown confirmation dialog, PolicyCenter reopens the Cluster Members
screen. In this screen, you now see the following:
• The Actions entry for the affected server contains a Cancel planned shutdown button.
• The Planned Shutdown entry for the affected server provides information about the planned shutdown.
• The screen contains a banner with the chosen shutdown message and a countdown timer to the server shutdown
date and time.
Server Tools 357
Shutdown Message Details

You have several choices for shutdown messages in the Schedule Planned Shutdown screen. If you do not elect to
create a custom message, you can select one of several default messages. PolicyCenter stores these messages as
display keys in file display.properties.
Web.TabBar.SystemAlertBar.PlannedShutdown.RollingUp Rolling upgrade in progress, please save your work and log out
gradeMessage to redirect to a new server.
Web.TabBar.SystemAlertBar.PlannedShutdown.ScaleInMe Please save your work and log out to re‐direct to the new
ssage server.
After you initiate a planned shutdown, PolicyCenter stores only the current time and the shutdown time in the
database. It stores other information, such as the shutdown message itself, in a single-threaded atomic reference in
memory. PolicyCenter clears this message reference under the following conditions:
• At the restart of the shutdown PolicyCenter server
• At the cancellation of the planned server shutdown
Terminating Running Batch Processes

The Schedule Planned Shutdown screen contains a Terminate Batch Processes checkbox. PolicyCenter terminates the
running batch processes differently depending how you set this field.
Terminate Batch Processes Result

Checked PolicyCenter sets a flag on all currently running batch processes (including workqueue writers)
assigned to this server to terminate their operation. To the extent that the batch process logic
permits, the process attempts to stop as gracefully and as safely as possibly.
Unchecked PolicyCenter does not instruct the currently running batch processes on this server to terminate
prematurely. All batch processes assigned to this server continue until completed, regardless of
how long it takes to complete the work.
In either case, PolicyCenter does not show a Planned Shutdown status of ready on the Cluster Members screen until all
the affected processes have stopped on the server.
Note: Batch processes run as non-daemon threads. As such, it is not possible for the server to perform
a graceful shutdown if any batch processes are still running on the server at the time of the shutdown.
Using System Tools to Schedule a Shutdown

It is also possible to initiate a scheduled shutdown of a server using either the SystemToolsAPI web service or its
associated system_tools command prompt option. For example, the system_tools command uses the following
syntax:
system_tools -user user -password password -scheduleshutdown serverId [-
terminatebatchprocesses -shutdowndelay minutes]
If you include the -terminatebatchprocess command option, the shutdown process is the same as if you checked
the Terminate Batch Processes checkbox in the Schedule Planned Shutdown screen.
For more information on the use of the system_tools command, see “System Tools Command” on page 393. For
information on the use of the SystemToolsAPI web service, see the PolicyCenter Integration Guide.
Cluster Members – Application Server Instances

The Server Tools Cluster Members (Application Server Instances) table provides the information about individual server
instances recognized by the cluster. To update this information, click Refresh Cluster Members. A review of this
information is one way to determine if a cluster member has become unreachable, and therefore, missing from the
list.
This screen provides the following information.
Server ID Name for each PolicyCenter server in the cluster.

Status Status of each server instance n the cluster.
Host Name of the machine on which each PolicyCenter server is running.
User Sessions Number of user sessions active on each server instance in the cluster.
Run Level Run level for each server instance in the cluster.
Version Version information for each application installation. PolicyCenter provides the version information in the fol‐
lowing format:
Guidewire build.customer build (n,n,n,n,n)
These numbers have the following meaning:
• Guidewire build – Guidewire application version, for example, 9.0.0 or 9.0.1.
• Customer build– Custom label, defined in file customer-version.properties. If not defined, this field is
empty.
• (n,n,n,n,n) – Numbers that have the following meaning:
◦ Platform major version
◦ Platform minor version
◦ Application major version
◦ Application minor version
◦ Data model version
See “Understanding Guidewire Software Versioning” on page 365 for more information.
Server Roles Specific roles assigned to each server instance in the cluster.
Server Started Date and time at which this server instance started.
Connection Date and time this cluster member first started.
Started
Last Update Date and time of the last update on this server instance.
Planned Shut- Date and time of any planned shutdown of a cluster member.
down
Actions Action button to manage a planned shutdown of a cluster server instance. The button label is either Start Plan-
ned Shutdown or Cancel Planned Shutdown.
You can also initiate a server shutdown using the administrative system_toools command option -
scheduleshutdown. See “System Tools Command” on page 393 for details.
See also
• “The Cluster Components Screen” on page 360

• “Schedule a Planned Cluster Member Shutdown” on page 362
Cluster Members – Components Tab

The Server Tools Cluster Members – Components tab provides information on the components running on the server
that you select in the Cluster Members (Application Server Instances) table. These components, with the exception of the
work queues, are all singleton components. You can view all components, or, filter the list to see only those
components of a specific type.
The Components table provides the following information about the components running on the selected server.
Type Type of component, one of the following:

• Batch Process
• Message Destination
• Startable Service
• System
• Work Queue
Server Tools 359

Name Name of component. This name is the work queue name, or, the message destination name, for example.
Started Date and start time for each component.
State State of the component, for example, Assigned.
Retry Date and time of the deadline in which to complete the failover process.
Failover If a lease manager detects the expiration of a component lease owned by another cluster member, the first lease
manager starts a failover process for this lease. This failover process is not instantaneous. It is possible that during
the failover process the cluster member performing the failover itself might fail, or lose database connection, or
encounter other potential problems.
To be able to handle this situation gracefully, the cluster member starting the failover gives itself a deadline to com‐
plete the failover process. If the current failover process does not finish by the specified timestamp, some other
cluster member needs to start the failover process for the same component lease.
See “Automatic Failover of a Component Lease” on page 158 for a description of how PolicyCenter calculates this
timestamp.
You can view similar information to that of the Components table on the Cluster Components screen.
Cluster Members – History Tab

The Server Tools Cluster Members – History tab provides information on the history of the server that you select in the
Cluster Members (Application Server Instances) table.
The History table provides the following information for the selected server.
Host Name of the machine on which the PolicyCenter server is running.

UUID Universally Unique ID for this server machine. PolicyCenter randomly generates a UUID for each machine at
each machine startup.
Env Environment for this PolicyCenter server.
Last Run Level Run level for this PolicyCenter server at the point it stopped
Server Roles Specific roles assigned to this server instance.

Server Started Date and time at which this server started.
Server Stopped Date and time at which this server stopped.
The Cluster Components Screen

The Server Tools Cluster Components screen provides the following information.
Type Type of component to show in the table, one of the following:

• All types
• Batch Process
• Message Destination
• Startable Service
Name Name of component. This name is the work queue name, or, the message destination name, for example.
State State of the component, for example, Assigned, meaning that there is a server that has the lease to run this
component.
Start Request- Date and time that the server received the component start request.
ed
Started Date and start time of the component.

Owner Server ID of the PolicyCenter server on which the component runs.

Lease Expira- Date and time at which the component lease expires. See “Cluster Member Shutdown” on page 143 for more
tion information on component leasing.
Stopped Stop date and time of the server instance on which the component is running, due to one of the following rea‐
sons:
• Server instance stopped as intended
• Server instance failed due to lost power or network connection and automatic recovery failed or is still in
progress
Actions PolicyCenter shows a Complete Failover button if a component lease expires and a custom failover plugin imple‐
mentation instructs PolicyCenter not to perform an automatic failover.
See “Cluster Member Shutdown” on page 143 for more information.
Terminate Re- Date and time at which the server instance received a terminate request for this component.
quested
Transfer Re- Date and time at which the server instance received a request to transfer this component to another server in‐
quested stance.
Transfer Tar- Server ID of the server to which the transfer request was made.
get
Retry Failover Date and time of the deadline in which to complete the failover process.
If a lease manager detects the expiration of a component lease owned by another cluster member, the first
lease manager starts a failover process for this lease. This failover process is not instantaneous. It is possible that
during the failover process the cluster member performing the failover itself might fail, or lose database connec‐
tion, or encounter other potential problems.
To be able to handle this situation gracefully, the cluster member starting the failover gives itself a deadline to
complete the failover process. If the current failover process does not finish by the specified timestamp, some
other cluster member needs to start the failover process for the same component lease.
See “Automatic Failover of a Component Lease” on page 158 for a description of how PolicyCenter calculates
this timestamp.
You can view similar information to that of the Components table on the Cluster Members screen, the Components tab.
Available Actions on this Screen

The following list describes the actions that you can take in Cluster Members screen.
Action Description
Download cluster server re‐ Click Download to download an HTML report of the cluster components and component history.
port See “Download a Server Component Report” on page 362 for details.
Filter by component type Select a component type from the Types drop‐down list to filter the information by a specific
component type.
Filter by component state Select a component state from the State drop‐down list to filter the information by a specific
component state.
Filter by component Click Filter by Component to open a Select Components screen. In this screen, you can select individ‐
ual components of all available types to show in the component table.
Refresh the component in‐ Click Refresh to update the server component information in the table.
formation
Review component history Click the name of a component to open a component history detail screen.
detail
Server Tools 361

Download a Server Component Report

Procedure
2. Navigate to the Server Tools Cluster→Components screen.
3. Click Download to open the Specify Record Limits for this Report screen.
4. Enter the maximum number of days to include in the generated component history report.
5. Click Complete Download to generate a ZIP file that contains the actual report.
6. Save the download file to your local machine.
7. Extract the Zip file to a local directory.
8. Search for and double-click file index.html to open the report in a web browser.
9. Click the Component History link for the component history report.
Schedule a Planned Cluster Member Shutdown

About this task
You schedule a planned shutdown of a cluster member from the Server Tools Cluster Members screen.
Note: It is also possible to initiate a planned server shutdown using the administrative system_toools
-scheduleshutdown command option. See “System Tools Command” on page 393 for details.
Procedure
1. Navigate to the Server Tools Cluster Members screen.
2. Find the table row that corresponds to the cluster member for which you want to schedule a shutdown.
3. If necessary, scroll the screen all the way to the right until you see the Actions column.
4. Click Start Planned Shutdown.
5. In the Schedule Planned Shutdown screen, select the message that you want to show to the application user about
the planned shutdown.
Select one of the provided messages, or, enter custom text in the provided field.
6. Select the date and time of the planned shutdown.
7. Decide how you want PolicyCenter to handle the currently running batch processes (including work queue
writers) on the server.
See “About Planned Server Shutdowns” on page 357 for more information on the option.
8. Click Schedule Shutdown.
Result
After you schedule the shutdown, the Cluster Members screen shows the following for your selected cluster member:
• The Planned Shutdown column shows the date and time that you initiated the planned shutdown. It also shows the
date and time of the actual planned shutdown.
• The Actions column shows a Cancel Planned Shutdown button. To cancel a planned server shutdown, click Cancel
Planned Shutdown.
All PolicyCenter screens, for all users logged into the affected cluster member, display a banner indicating the date
and time of the planned shutdown and your selected message.
Upgrade and Versions

The Server Tools Upgrade and Versions screen displays information about any upgrade process that is run on this
server.
Note: Some functionality on the Upgrade and Versions screens is visible only if configuration parameter
ClusteringEnabled is set to true in config.xml.
Guidewire divides this screen into the following areas:
• A row of buttons that control upgrade functionality, not all of which are visible at all times.
• A summary table that contains upgrade information for this particular application server.
The Upgrade and Versions Screen Summary Table
The upgrade summary table contains the following information.
Version Version information for each application installation. PolicyCenter provides the version information in the fol‐
lowing format:
Guidewire build.customer build (n, n, n, n, n)
• Guidewire build – Guidewire application version, 9.0.0, for example.
• Customer build– Custom label, defined in file customer-version.properties. If not defined, this field is
empty.
• (n, n, n, n, n) – Numbers that have the following meaning:
◦ Platform major version
◦ Platform minor version
◦ Application major version
◦ Application minor version
◦ Data model version
See “Understanding Guidewire Software Versioning” on page 365 for more information.
Status Status of each upgrade, for example, new schema or upgraded.
Type Type of each upgrade, for example, install or full.
Start Time Date and time of the beginning of the upgrade.
End Time Date and time of the completion of the upgrade.
Duration Length of time that the upgrade process took. Time is shown as both the total number of seconds involved and
the time for just the database upgrade portion of the upgrade.
Deferred Up- Execution status of the Deferred Upgrade Tasks batch process. See “Deferred Upgrade Tasks Batch Process” on
grade Tasks page 106 for details.
Status
View Details Click the View icon to open a pop‐up from which you can view the same reports contained in the
UpgradeInfo.zip file.
See “View an Upgrade Report” on page 364 for more information.
Download De- Click the Download arrow to download an UpgradeInfo.zip file that contains a set of HTML reports describing
tails various aspects of the upgrade process.
See “View an Upgrade Report” on page 364 for more information.
Remove Detail Click the trash can icon to remove this row of data.
Data
The Upgrade and Versions Screen Buttons
The top of the Upgrade and Versions screen contains a number of buttons that initiate specific upgrade actions. Not all
buttons are visible at all times. Individual buttons become visible depending on various aspects of the cluster and its
members.
The Upgrade and Versions screen contains the following buttons that initiate upgrade-related actions.
Refresh Click to update the information on this screen.
Server Tools 363

Start Rolling Upgrade Click to open a Start Rolling Upgrade screen from which you can initiate a rolling upgrade for this
cluster. Before you can start the upgrade, you must actively affirm that you have taken certain pre‐
upgrade steps by selecting the check box next to each pre‐upgrade step.
See “Performing a Rolling Upgrade” on page 172 for details.
Rolling Upgrade Complete Click to notify PolicyCenter that the in‐progress rolling upgrade is now complete. The rolling up‐
grade is complete after you have restarted all cluster members with the new target configuration.
Without this notification, PolicyCenter prevents cluster members with the old configuration from
starting up again.
Initiate Backout Click to attempt to stop an in‐progress rolling upgrade and initiate a backout of the new target
configuration. The attempt to stop and backout the rolling upgrade may or may not succeed, de‐
pending at what point in the rolling upgrade that the request to stop occurs.
If a backout of the rolling upgrade is possible, PolicyCenter opens a series of confirmation screens
to assist you in correcting issues related to backing out the new configuration.
Start Full Upgrade Click to open a Start Full Upgrade screen. Clicking Yes on this screen sets a flag in the database that
you intend to start a full database upgrade. You must set this database flag before deploying your
new configuration to members of the cluster.
If you do not set this flag, PolicyCenter refuses to start and refuses to alter the database. After you
complete the full upgrade, PolicyCenter deletes this flag from the database. You must set the flag
again before starting a new full upgrade.
Note: You can also initiate a full PolicyCenter upgrade using the administrative system_toools
command option -startfullupgrade. See “System Tools Command” on page 393 for details.
Cancel Full Upgrade Click to attempt to cancel an in‐progress full upgrade. If PolicyCenter determines that it is safe to
cancel the upgrade, it does so.
See also
• “Review Profiler Upgrade Information” on page 375
The Upgrade Report

The Server Tools Upgrade and Version screen contains the means to view details of a specific upgrade or download the
data for local viewing. The downloadable upgrade information contains the following set of linked reports:
Upgrade Instance Lists information on various upgrade statistics.

It is also to possible to download Guidewire Profiler data by clicking the Raw Profiler Data link at the bottom of
the Upgrade Instance screen.
Database Parame- Lists information on various database parameters, including the database connection pool settings, the da‐
ters tabase configuration settings, and similar information.
See also
• “Understanding Data Model Updates” on page 238
View an Upgrade Report

Procedure
1. Navigate to the Server Tools Upgrade and Versions screen.
2. For the upgrade process that interests you, click one of the following:
Download Details arrow Downloads a Zip file that contains various types of upgrade information.
View Details icon Open a pop‐up window from which you can view the upgrade reports.

3. If you downloaded the report file, unzip the file into its own directory.
4. Locate file index.html and double-click it to open it in a browser.
5. Use the links on the screen to navigate through the linked reports.
Understanding Guidewire Software Versioning

Guidewire application versioning is a way to label a particular snapshot of a Guidewire application. It is a string
label that Guidewire applies to its software. You see this version number primarily in Server Tools, in the Cluster
Members screen and the Upgrade and Versions screens.
Guidewire manages software versioning through internal classes that provide the following information:
• Guidewire application name and version
• Guidewire module name and version
PolicyCenter writes version information to the application log at PolicyCenter start.
Application Version Numbering

Both the Server Tools Cluster Members screen and the Upgrade and Versions screens show the PolicyCenter application
version as the following sequence of numbers:
A.B (a,b,c,d,e)
Version # Meaning
A Guidewire application version plus the application release build version, 9.0.0.905, for example.
B Custom version label
a Platform major version
b Platform minor version
c Application major version
d Application minor version
e Data model version number
The following example illustrates these concepts.

9.0.0.905.20161017 (6,22,11,45,175)
Notice that:
• 9.0.0.905 is the application version number (9.0.0) plus the application release build version number (905).
• 20161017 is a custom label that you define in file customer-version.properties. In this case, the label
indicates the date of a PolicyCenter configuration upgrade. If not defined, this field is empty.
• 175 is an upgrade version number that you define in file extensions.properties.
WARNING In a production environment, Guidewire requires that you increment the data model
version number whenever you make changes to the data model, before you restart the application
server. Otherwise, unpredictable results can occur. Use of the extensions.properties file in a
development environment is optional.
See also
• “File customer-version.properties” on page 366
• “Create a Custom Version Label File” on page 366
• “Deploy a Custom Version Label File on Tomcat” on page 366
Server Tools 365
File customer‐version.properties
Use file customer-version.properties to define a custom build version number. If you create this file and
populate it correctly, PolicyCenter appends your custom version number to the Guidewire software version label.
Most commonly, you use a custom build number to label and track a configuration deployment that you undertake as
a rolling upgrade.
File customer-version.properties contains the following single property:
customer.build
The following example illustrates how to use this property:
customer.build=20160914-PCF-upgrade
PolicyCenter does not contain file customer-version.properties in the base configuration. Instead, you must
create this file and place it in a location that PolicyCenter recognizes.
See also
Create a Custom Version Label File

About this task
Guidewire recommends that you place any customer-version.properties file that you create in the res directory.
Procedure
1. Open Guidewire Studio™.
2. In the Studio Project window, expand configuration→res.
3. Select res and right-click to open the context menu.
4. Select New→File.
5. Name the file customer-version.properties.
6. Enter a value for customer.build, for example:
customer.build=20160914-PCF-upgrade
7. Save your work.
See also
Deploy a Custom Version Label File on Tomcat

Before you begin
Before starting this procedure, ensure that you complete all steps in “Create a Custom Version Label File” on page
366.
Procedure
1. Create a Tomcat WAR file:
a. Open a command prompt in the PolicyCenter installation directory.
b. Execute the following command:
gwb warTomcatDBcp

PolicyCenter adds file customer-version.properties to the following JAR file in the generated WAR
file:
WEB-INF/lib/configuration.jar
2. Deploy the WAR file to the application server.
Result
After starting the server, the Server Tools Upgrade and Versions screen shows the custom version label, appended to
the standard application version number.
See also
Cache Info
The Server Tools Cache Info screen provides information in both table and chart form of PolicyCenter server cache
information. Guidewire recommends that you use this information to help you monitor how well the cache is
performing.
See also
•
• “Server Cache Tuning Parameters” on page 67
The Cache Summary View

The Server Tools Cache Summary view on the Cache Info screen includes the following information:
Max Cache Space (KB) Maximum amount of space to allot to the global cache.
Stale Time (mins) Maximum time allowed for an object to be in the cache without a database refresh.
Page Loaded at Date and time of the last refresh of the data on this screen.
Available Actions on this Screen

From the Cache Info screen, you can do the following:
Edit Click Edit to enter different values for Maximum Cache Space and Stale Time. If you change these parameters from the
Cache Summary view, the values you specify apply only to the PolicyCenter server to which you are connected. If
you restart the server, your changes are lost. For your changes to persist, edit the their values in file config.xml.
Refresh Click Refresh to update the information shown on this screen. This action also updates the date and time values
shown for the Page Loaded at field.
Download Click Download to download a CSV‐formatted file containing detailed cache information. See also “Understanding
the Cache Data Report” on page 369.
Clear Glob- Click Clear Global Cache to clear the cache of all entities. Note, however that the cache always contains some ob‐
al Cache jects to support an active server.
Cache Summary Graphs

The Cache Summary screen provides the following graphs.
Server Tools 367
Graph Description
Cache Size The memory used by the cache over time.
Hits and Misses (Stacked) The number of cache hits (an object was found in the cache) and misses (object was not found
in the cache) and the miss percentage.
Type of Cache Misses The number of cache misses caused by PolicyCenter evicting an object because the cache was
full and the number of missed caused by PolicyCenter evicting an object due to reaping.
This graph is not visible if configuration parameter GlobalCacheDetailedStats in
config.xml is set to false.
Evict Information Information about cache evictions over time, including:

• Number of times no entry was found to evict when cache was full
• Number of evictions within active time when cache was full
• Number of evictions when cache was full
• Number of evictions due to reaping
This graph is not visible if configuration parameter GlobalCacheDetailedStats in
config.xml is set to false.
Current Age Distribution The number of objects of various ages in the cache.
Current Cache Contents for Age All The percentage of types of objects present in the cache for all ages.
See also
The Historical Performance View

Note: For the Historical Performance screen to be visible, configuration parameter
GlobalCacheDetailedStats in file config.xml must be set to true.
The Server Tools Historical Performance view on the Cache Info screen provides the following graphs:
Graph Description
Space Retained Memory used by the cache over the past couple days. The time shown is a much lon‐
ger period than the cache size graph on the Cache Summary tab, which only displays
the past 15 minutes.
In this case, the x‐axis represents the average values for each time period during each
of the past eight days. This is to allow for comparison of cache behaviors against hour‐
ly trends.
Hits and Misses (stacked) Number of hits (object was found in the cache) and misses (object was not found in
the cache) and the miss percentage over the past day and past seven days.
Miss % The percentage of cache read attempts in which the object was not found in the
cache over the past day and the past seven days.
Number of Misses because item was evicted The number of misses over the past day and the past seven days due to PolicyCenter
when cache was full having evicted an object from the cache because the cache was full.
The Cache Details View

The Server Tools Cache Details view on the Cache Info screen includes graphs of the following:
Graph Description
Age Distribution by time A number of graphs that show the age distribution of objects in the cache. The Cache Details tab
shows age distributions for zero to 30 minutes ago.
Current Cache Contents by age A number of graphs that show the percentage of types of objects in the cache over time.

Understanding the Cache Data Report

Clicking Download on the Server Tools Cache Info screen downloads a CSV-formatted file containing information on
the current state of the server cache.
In looking at the cache data provided in the downloaded report, Guidewire recommends that you first calculate the
cache miss ratio around the time of the performance degradation. The cache miss ratio is the ratio of (misses) /
(misses + hits).
The cache miss ratio is a useful metric in that it normalizes the cache values. For example, suppose that you have the
following hit and miss cache values that you use to calculate each individual cache miss ratio.
Timestamp # cache misses # cache hits Cache miss ratio

Time 1 12345 2345 12345 / (12345 + 2345) = 0.8404
Time 2 1234 23456 1234 / (1234 + 23456) = 0.4998
The calculation of the cache miss ratio enables you to compare the data in way that is not possible by simply
examining the raw data.
High Miss Ratio

If the miss ratio is high, it is very likely that the performance issue involves the cache. Thus, you need to look at the
reasons for the cache misses, especially the following:
• Number of misses because item was evicted when cache was full
• Number of misses because item was evicted due to reaping
If the number of misses is high due to the cache being full, you can try increasing the size of the cache and
observing the result. Keep in mind, however, that the cache uses the heap. Thus, increasing the size of the cache can
potentially reduce the amount of heap memory available for other computations.
If the number of cache misses is high due to reaping, you can try increasing the reaping time and observing the
result. Another approach would be to examine ways to reduce the time since the last access of the cache objects. For
example, how long does it take to reuse objects that have already been fetched from the database? Is it possible to
hold onto the objects without having to fetch the objects again from the database?
Cache Misses and Hits

The cache miss count and the cache miss ratio are probably the most important data points in the download cache
data report. Essentially, the more cache hits the better, as it means the cache was able to service that many requests.
Each cache miss is expensive. It means that the looked-for object was not in the cache. Regardless of the hit count, a
high miss count can indicate a problem with the cache. Keep in mind, however, that the miss count is relative. A
cache miss can occur simply because it is the first time a request is made to the cache for an object. Thus, the object
is not in the cache. A cache miss can also occur as the sought-after object may have already been evicted from the
cache. The downloaded cache report provides data on the various reasons for object eviction from the cache.
Using the assumption that there is a constant cost for each cache miss, it is possible to compare the miss counts
directly. The cost of a cache hit is essentially zero, or very, very small compared with the cost of a cache miss. A
cache miss may take a 10 millisecond trip to the database to access the data whereas a cache hit is essentially zero
time.
Guidewire Profiler
The Server Tools Guidewire Profiler screen provides access to a set of tools that are useful in gathering and analyzing
information on the runtime behavior and performance of Guidewire PolicyCenter. The Profiler records the time
spent in specific processing areas done by the application code, as well as the configured rules and PCF screens. Use
of this information can help narrow down issues to the potentially problematic components such as PCF screens,
rules, code sections, or workflows.
Server Tools 369
Profiler code is useful, for example, to record the following types of operations and information:
• SQL statements that PolicyCenter is executing
• Parameters passed to those SQL statements
• Row counts
• Name of the currently executing rule
Guidewire recommends that you exercise care in using this feature. Storing too much information can cause the
Profiler screen to become too cluttered, require more space for storage and, for long-running processes, hold on to too
much memory at runtime.
Note: Guidewire Profiler does not collect memory usage statistics. You can use a third-party tool to
gather memory usage and garbage collection information.
See also
• “Server Memory Management” on page 69
Guidewire Profiler Concepts

There are several concepts that are important to an understanding of how to configure application profiling. The
following list describes these concepts.
Concept Description More information

Entry point Refers to the type of interactions that you select to profile. “Profiler Entry Points” on page 370
Profiler tag Alias for a piece of code to profile in the Guidewire application. “Profiler Tags” on page 376
Profiler frame Contains information on a specific invocation of profiled code. “Profiler Frames” on page 376
Profiler stack Stores profiling information for a specific thread. “Profiler Stacks” on page 376
See also
• “Guidewire Profiler” on page 369

• “Guidewire Profiler Analysis” on page 373
Profiler Entry Points

The Server Tools Guidewire Profiler collects profile data based on the type of interaction with the application that
you request to be profiled. Guidewire refers to the different types of possible interactions as entry points, with the
entry point type indicating the type of request or action that initiated application processing. The following list
describes the various entry point types that you can access from the Guidewire Profiler Configuration screen.
Entry point Description

Web Application user interface in which you can perform tasks and actions. Clicking around in the Guidewire
application in a browser potentially causes rules, web services, messages, and similar items to trigger.
See “Web Session Profiling” on page 371 for more information.
Batch process Batch processes wake up at regular intervals, and can execute large queries. See “Administering Batch
Processing” on page 81.
Work queue Long running processes that pick up work to be done from a queue. These processes are typically dis‐
tributed across several servers. See “Administering Batch Processing” on page 81.
Message destination Process that delivers messages to one or more destinations. See the Integration Guide.
Web service SOAP requests received by PolicyCenter. See the Integration Guide.

Entry point Description

Startable service Plugin implementations that PolicyCenter initializes at server startup and de‐initializes at server shut‐
down. For example, it is possible to register a plugin implementation as a listener on a JMS queue. See
the Integration Guide.
Except for Web entry points, Guidewire Profiler stores configuration information in the database. This information
is visible to all PolicyCenter servers in the cluster. For a change in configuration to a batch process, work queue, or
message destination to take effect, you need to restart that batch process, work queue, or message destination. Any
change in configuration can take some time to propagate through the cluster. Also, it may take up to the cache stale
time for a configuration change to become visible.
The next time profiling starts for a given entry point, Guidewire Profiler checks whether profiling is enabled for that
entry point. If profiling is enabled, Guidewire Profiler records the profiling data in the form of a profiler stack. The
Profiler records multiple stacks if the initial thread spawns more threads and the developer profiles the spawned
threads. Except for Web profiling, PolicyCenter persists this data database, making it possible to retrieve the data
later.
See also
• “Web Session Profiling” on page 371

• “PolicyCenter Application Profiling” on page 372
• “Ways to View a Guidewire Profiler Analysis Reports” on page 374
Web Session Profiling

After you enable web profiling in Guidewire Profiler, PolicyCenter records all subsequent round-trips to the server
as a separate profiler stack. Unlike other types of entry points, PolicyCenter does not persist the stacks from Web
requests to the database. Instead, PolicyCenter stores the stacks from web requests in the user session.
In using web profiling, Guidewire recommends the following:
• Enable the web profiler only as absolutely needed as it degrades performance.
• Start the profiler from the PolicyCenter application screen that you want to profile.
• Log out of the application after you have analyzed or downloaded the profiler data to free the memory used by
the profiler.
It is only possible to enable profiling for web entry points for the current session.
Enable Web Profiling
Procedure
1. Navigate to the Server Tools Guidewire Profiler screen.

2. Select the profiling options that you want to use for this profiling session.
3. Restart the associated batch process, work queue, or message destination, if appropriate.
4. Click Enable Web Profiling for this Session.
This action returns you to the application screen from which you started.
5. Exercise the application screens that you want to profile.
6. Press ALT+SHIFT+P to return to Guidewire Profiler.
7. Click Disable Web Profiling to end the current Web profiling session.
This action generates the Web Profiler screen.
Server Tools 371
See also
• To understand the various web profile tracing options, see “Profiler Trace Options” on page 372.
• To understand the Web Profiler screen, see “Guidewire Profiler Analysis” on page 373.
• To download and save this snapshot of application profiling data, see “View Uploaded Profiler Reports” on page
375.
PolicyCenter Application Profiling

You enable application profiling in the Server Tools Guidewire Profiler→Configuration screen. On the Configuration
screen, directly below the Web Profiler area, is an area labeled Entry Point Configuration. This area contains a set of
buttons that you click to enable the other types of entry points profiling. These buttons are:
• Enable Profiling For Batch Process
• Enable Profiling For Work Queue
• Enable Profiling For Message Destination
• Enable Profiling For Web Service
• Enable Profiling For Startable Service
Clicking the button for an entry point opens a secondary screen in which you select the specific item to profile.
For example, to enable application profiling for message destinations:
• Click Enable Profiling for Message Destinations. This action opens a screen in which you can select the specific
message destination to profile.
• Select a specific message destination and click OK to return to the Configuration screen.
PolicyCenter then adds a message destination row to the Entry Point Configuration table, with a different column for
each profile trace option. If a trace option is available for the chosen entry point, PolicyCenter places an icon in the
table cell for that option. Initially, PolicyCenter disables all of the available trace options, marking the table cell with
a red X. It is necessary to click the icon to enable the option. This action changes the icon to a green check mark.
For each entry point row, PolicyCenter also adds an Edit Configuration button at the right-hand end of the row. This
button provides an alternative way to enable tracing options, as well as a way to disable tracing for this entry point.
If you disable an application profile entry type, PolicyCenter removes the row for that entry point from the Entry
Point Configuration table.
See also
• “Profiler Entry Points” on page 370
• “Profiler Trace Options” on page 372
Profiler Trace Options

It is possible to enable different types of trace options in the Server Tools Guidewire Profiler→Configuration screen for
the various types of entry points. These options are quite expensive to compute. Guidewire recommends that you
narrow down your performance issues first before trying these options.
The exact number and type of tracking options available for profiling depend on the following:
• The application server’s operating system
• The application server’s database type
• The chosen entry point
To enable a specific profiling option, select or check the box next it. The following list describes the types trace
options that are available in Guidewire Profiler.
Trace option Description

Individual Stacks Available for work queue entry points only. If checked, this tracing option stores one stack for
each work item and does not roll up the data.
Use caution with the Individual Stacks option as the amount of data to store could be unbound‐
ed. The Individual Stacks option is recommended for use if there are only a few items in the
queue.
Stack Trace Tracking Captures the Java stack trace and the PCF trace at the point at which a query executes.
Use caution with this tracing option as it can be very performance intensive. As a general rule,
Guidewire recommends that you do not enable this tracing option unless there are very specif‐
ic reasons to use it.
Query Optimizer Tracing (Oracle) This trace indicates how the database arrived at a specific execution path. This creates
a trace file on the database server.
High Resolution Clock (Microsoft Windows) This tracing option uses the Microsoft Windows 100 nanosecond clock.
Extended Query Tracing This trace tracks the parse‐execute‐fetch actions of a SQL statement. If executing against an
Oracle database, this tracing option also tracks any wait times associated with the SQL state‐
ment, including what the cause of the wait.
This stack trace creates a trace file on the database server.
Diff DBMS Instrumentation Coun- (Oracle) Enable this option to capture the DBMS counters at the beginning of the profiling ses‐
ters sion and to include analysis of the differences in the DBMS‐specific report.
DBMS Instrumentation Capture (Oracle) The Profiler generates a DBMS report if an action exceeds the threshold value set by
Threshold for each Action (millis) this option. This tracing option is available only if you first enable Diff DBMS Instrumentation Coun-
ters tracing. Click Edit Configuration and select Edit DBMS Instrumentation Capture Threshold (millis) to
edit the default value of 0.
See also
• “Profiler Entry Points” on page 370
• “PolicyCenter Application Profiling” on page 372
Guidewire Profiler Analysis

After you disable the current application profiling session, PolicyCenter generates a Profiler Analysis screen, or, in the
case of Web profiling, a Web Profiler screen.
The Profiler Analysis screen of the Guidewire Profiler contains the following regions:
• Profiler Source – Lists each profiler run in the current user session. To see the results for a particular run, select
it from the Name list.
• Profiler Results – Provides a means to view different views of the profiler data for the selected run. Selecting a
different view type updates the screen data to reflect your choice.
View Type Result

Stack Queries Lists the queries that were executed by each stack. The first list shows the stacks in the profiled session. The
second shows the queries executed in that stack. More details can be obtained by clicking on each tab.
Aggregated Quer- Lists all queries executed as part of the profiled session and some statistics about them, including number of
ies times executed, average time, and more.
Search by Query Searches the session for a query. This view enables you to determine the source of a particular query. Paste in
a query, for example from the AWR report, and click Search.
Elapsed Lists each frame in chronological order within its stack along with the time in seconds between when
PolicyCenter pushed and then popped the frame.
Server Tools 373

View Type Result

Chrono Lists each frame in chronological order within its stack along with the time in seconds between PolicyCenter
creating the stack and pushing the frame. See the Rules Guide.
Group Frames Lists frames in each stack aggregated by tag and presented in order of total aggregate time on the stack.
Group Stacks Presents data similar to that presented in the Group Frames view, except Guidewire Profiler aggregates the
frames across all stacks in the session instead of by stack.
Stacks Grouped Lists the stacks in the profiling session grouped by name, along with timing information on each stack group.
Rule Execution Lists rules that fired during the session. If no rules fired during the session, the Profiler Result region contains
the message “No profiler stacks found”.
See also

• “View Uploaded Profiler Reports” on page 375
• Rules Guide
Save a Profiler Analysis Report

About this task
This procedure creates a .gwprof file that you can download and store outside of Guidewire Profiler. To view the
contents of this file, you must upload the .gwprof file back into Guidewire Profiler.
Procedure
1. Navigate to the Server Tools Guidewire Profiler→Profiler Analysis→Profiler Analysis screen.

2. Select the profiler entry point, for example, Batch Processes.
3. In the Profiler Source area at the top of the screen, select the report that interests you.
4. In the Profiler Result area, select a value from the View Type drop-down list.
5. Click Download.
6. In the dialog that opens, select Save File and click OK.
See also
• “View Uploaded Profiler Reports” on page 375
Ways to View a Guidewire Profiler Analysis Reports

It is possible to view Guidewire Profiler analysis reports in any of the following ways.
View type More information

View current report in Profiler Analysis screen “Guidewire Profiler Analysis” on page 373
Search for historical report and view in Profiler Analysis screen “View Historical Profiler Reports” on page 375
Upload historical report and view in Profiler Analysis screen “View Uploaded Profiler Reports” on page 375
See also

View Historical Profiler Reports

Procedure
1. Navigate to the Server Tools Guidewire Profiler→Profiler Analysis→By Time Range screen.
2. Click Search.
The screen shows two calendar date pickers.
3. Enter a start and stop date in which to search for existing profiler analysis reports.
PolicyCenter prints the results of the search to the screen:
• If PolicyCenter finds no existing profiler reports that occurred within the specified time frame, it prints a
message to that effect.
• If PolicyCenter does find one or more profiler reports that occurred within the specified time frame, it lists
the reports.
View Uploaded Profiler Reports

Before you begin
The file to upload must be a file with a .gwprof extension downloaded previously from Guidewire PolicyCenter.
Procedure
1. Navigate to the Server Tools Guidewire Profiler→Profiler Analysis→Saved File screen.
2. In the Restore Snapshot field, browse to find a .gwprof file for upload.
3. Click OK.
Result
The saved profiler data loads in the Saved File screen. which then becomes the Profiler Analysis screen. If you navigate
away from this screen, Guidewire Profiler deletes the uploaded data from the screen.
See also
• “Save a Profiler Analysis Report” on page 374
Review Profiler Upgrade Information

About this task
The Guidewire Profiler Upgrade screen provides profiler analysis of upgrade operation on the PolicyCenter server.
The Upgrade screen provides information on the database upgrade at initial server start and all subsequent upgrade
operations.
Procedure
1. Navigate to the Server Tools Guidewire Profiler→Profiler Analysis→Upgrade screen.
See also
Server Tools 375
Profiler Tags
Profiler tags represents sections of code that Guidewire Profiler can profile. A profiler tag is an alias for a piece of
code in the Guidewire application for which you want to gather performance information.
The code represents profiler tags by instances of the gw.api.profiler.ProfilerTag class. The constructor for the
ProfilerTag takes a String parameter defining the ProfilerTag name.
Always create a static final ProfilerTag object and preserve it. If you attempt to create more than one instance of
the same ProfilerTag object, PolicyCenter generates a warning message in the application log that is similar to the
following:
WARN Duplicate tag name: tag_name
Profiler Frames
A profiler frame contains information corresponding to a specific invocation of profiled code, such as its start and
finish times.
In each Profiler session:
• Whenever the code calls push on the profiler stack, Guidewire Profiler creates a profiler frame and pushes the
frame onto the stack.
• Whenever the code calls pop on the profiler stack, Guidewire Profiler removes the profiler frame from the stack.
The Profiler continues to stores the frame information, however, so as to make the information available for
future examination.
The code represents Profiler frames by instances of gw.api.profiler.ProfilerFrame.
See also
• “Understanding Properties and Counters on a Frame” on page 377
Profiler Stacks
A profiler stack stores profiling information for a specific thread. A profiler stack implements the standard push and
pop functionality of a stack. The push and pop actions correspond to the beginning and end, respectively, of a piece
of code represented by a profiler tag. Thus, at any time, the current contents of the profiler stack reflect all profiler
tags whose code PolicyCenter is currently executing. The code represents Profiler stacks by instances of
gw.api.profiler.ProfilerStack.
If a profiler stack has been initialized for the current thread, the call to Profiler.push(ProfilerTag.MYTAG)
pushes a new frame with tag MYTAG on to that profiler stack. Otherwise, the call has no effect.
Similarly, Profiler.pop(frame) is just a pass-through to calling pop on the profiler stack of the current thread.
Using Custom Profile Tags with Guidewire Profiler

It is possible to profile custom code in the Server Tools Guidewire Profiler by creating your own custom profile tags.
To define a custom Profiler tag, create a globally-accessible Gosu class using the Java-based PCProfilerTag class
as a model. Guidewire recommends that you add all custom profiler tags to this Gosu class.
The following sample code illustrates how to create a custom Gosu class for custom profiler tags.
package gw.profiler
uses gw.api.profiler.ProfilerTag
class MyProfilerTags {
public static final var MY_TEST_TAG1 = new ProfilerTag("MyTestProfiler1")
//..
private construct() {

// Do not instantiate}
}
}
To profile a block of custom code, use the following pattern to push and pop profiling information onto the profiler
stack.
uses gw.api.profile.Profiler
...
ProfilerFrame frame = Profiler.push(ProfilerTag.MY_TEST_TAG1)

try {
// CODE TO PROFILE
} finally {
Profiler.pop(frame)
}
See also
• “Understanding Properties and Counters on a Frame” on page 377
Profiling Spawned Threads

Some processes spread their workload across multiple threads. If you want to use Guidewire Profiler to profile these
threads, use the following pattern:
gw.api.profiler.Profiler.createPotentiallyProfiledRunnable(ProfilerTag entryPointTag,
String entryPointDetail, GWRunnable block)
This generates a new Runnable object that executes the given block. This Runnable object profiles the block if the
calling thread is also being profiled. If this is the case:
• The Profiler associates the stack for that thread with the stack of the calling thread.
• The Profiler persists that thread along with the stack of the calling thread.
See the Javadoc for the Profiler.createPotentiallyProfiledRunnable method for more details.
Understanding Properties and Counters on a Frame

Guidewire profiler frames can hold custom properties and counters that provide more information about system
events. Consider the following example:
uses gw.api.profiler.Profiler
uses gw.api.profiler.ProfilerTag
public class MyProfilerTags {
public static var paramValue : String

public static var ctrValue : Integer
var frame = Profiler.push(MyProfilerTags.myProfilerTag)
public function profileCode() {
public static var myProfilerTag: ProfilerTag = new ProfilerTag("MY_TAG")
try {
//Some paramter value, set by method code...

var param = "some parameter value"
//Some counter value, set by method code

var ctr = 5
frame.setPropertyValue("PARAMETER", param)
frame.setCounterValue("COUNTER", ctr)
} finally {
Server Tools 377

Profiler.pop(frame)
After the sample code pops a profiler frame off the stack, the frame contains information about the calculated values
of PARAMETER and COUNTER. The Server Tools Profiler Analysis screen then shows these values as well.
See also
• “Using Custom Profile Tags with Guidewire Profiler” on page 376
Understanding the Guidewire Profiler API

You can use the ProfilerAPI web service to configure the profiler from an external system. In addition to enabling
and disabling profiling for the various entry points, you can enable Web profiling on all subsequent sessions. This
API does not provide the ability to profile a specific session or to profile active sessions.
See also
Profiler‐related Configuration Parameters

Guidewire provides several profiler-related configuration parameters in config.xml. The first two configuration
parameters in the list refer to SQL Server databases only.
Parameter More information

IdentifyQueryBuilderViaComments Configuration Guide
IdentifyORMLayerViaComments Configuration Guide

ProfilerDataPurgeDaysOld Configuration Guide
See also
• Gosu Reference Guide
Product Model Info

The Server Tools Product Model Info screen contains a single Reload Availability button. If you click this button,
PolicyCenter attempts to reload availability data from the directory specified in the
ExternalProductModelDirectory parameter defined in config.xml.
Note: Guidewire disable the Reload Availability button during a rolling upgrade. The button does not
become active until a user clicks Rolling Upgrade Complete in the Server Tools Upgrade and Versions
screen.
The server attempts to synchronize the lookup entities and the existing product model availability data with the
XML files stored in the external lookup directory:
• If the reload is successful, PolicyCenter displays an informational message.
• If reload is not successful, PolicyCenter displays an error message.
In either case, you can check the server log files for details of the reload operations or problems that occurred.
To access the Product Model Info screen:

• The server must be in development mode.
• The EnableInternalDebugTools parameter in config.xml must be set to true.
• The logon user role must have the View ProductModelInfo tools screen permission. The code for this permission is
toolsProductModelInfoview. In the base configuration, only the Superuser role has this permission.
See also
Server Tools 379


chapter 24
Internal Tools
Guidewire provides internal tools to assist you with certain administrative tasks.
WARNING Guidewire does not support the use the tools found in the Internal Tools screens.
Guidewire provides these tools for use during development only. Guidewire does not support the
use of these tools in a production environment. Use these tools at your own risk.
See also
• “Server Tools” on page 321
Accessing the Internal Tools

The server must be in development mode and you must have the internaltools permission to access the Internal
Tools screens. In the base configuration, the Superuser role has this permission by default.
To access the Internal Tools screens, press ALT+SHIFT+T on any PolicyCenter screen after you log into the application
as a user with the necessary role permission.
See also
Reload
The Reload screen is useful while you develop a configuration. From this screen you can reload key configuration
files into a running PolicyCenter installation. You can choose from the following options:
Option Description
Reload PCF Files Verifies and reloads all PCF files. If there are errors in the PCF files, PolicyCenter writes the errors to the
log.
Verify All PCF Files Verifies the PCF files without reloading them.
Reload Web Templates Reloads the entire PolicyCenter user interface including the config/web/templates directory.
Reload Workflow Engine Reloads the Workflow engine.
Reload Display Names Reloads label definitions only from the display_languageCode.properties file for the locale.
Internal Tools 381

Testing System Clock

The system clock plugin, TestingClock, enables you to get and set the current system time in PolicyCenter. This
non-production tool is useful during the testing phase. Use this tool to move the system time forward as necessary to
determine if a process completes correctly. You cannot set the system time to a time before the current time.
IMPORTANT Use the system clock plugin for testing purposes only. You can only adjust the system
clock if the PolicyCenter server is in development or test mode.
You must implement and configure the TestingClock plugin to use this tool.
See also
• See the Integration Guide for implementation details.
PC Sample Data
The PC Sample Data screen is for loading sample data into PolicyCenter for development purposes only. Guidewire
does not support this tool for a production environment.
See also
Free‐text Search
The Free-text Search screen helps you manage the Guidewire Solr Extension, a full-text search engine, during
development. The screen is an alternative to the free-text batch load command. The Free-text Search screen provides
one operation to drop the indexes and another operation to load and index data. The free-text batch load command
performs both operations in a single command.
During development, use The Free-text Search screen instead of the free-text batch load command to avoid the
installation and setup procedure required to use the command.
To access the Free-text Search screen, you must run the PolicyCenter application in development mode. The
application hides the screen whenever you run the application in production mode.
The Free-text Search screen provides the following buttons:
• Sync Policy Index – Provides the functionality similar to the free-text batch load command. It extracts policy data
from the application database and sends it to the Guidewire Solr Extension for indexing. Click this button after
you click the Drop Policy Index button.
• Drop Policy Index – Drops the policy indexes from the Guidewire Solr Extension. Click this button before you click
the Sync Policy Index button.
• Run Consistency Check – Confirms that the policy index data in the Guidewire Solr Extension matches the policy
data in the application database. Click this button after you click the Sync Policy Index button or after you run the
free-text batch load command.
You can access the Free-text Search screen if all the following are true:
• The server is in development mode.
• The EnableInternalDebugTools parameter in config.xml is true.
• The FreeTextSearchEnabled parameter in config.xml is true.
See also
• “Run the Free-text Batch Load Command” on page 286
382 chapter 24: Internal Tools
Archiving Test
PolicyCenter provides an Internal Tools Archiving Test screen that you can use to test archiving on a job or policy
term. Use this tool during development to see the effect of Archive Policy Terms batch processing on a selected job
or policy term. From this screen, you can do the following:
Action Description
Archive term by policy transaction number Enter the job ID of the policy term that you want to archive. For validation:
• If you chose to skip validation, PolicyCenter archives the policy term based on the
current configuration, including the IPCArchviginPlugin implementation.
• If you do not skip validation, PolicyCenter archives the policy term regardless of
whether there are open jobs or whether the server time has reached the date set by
PolicyTerm.NextArchiveCheckDate.
Clicking Archive does not run Archive Policy Term batch processing. Instead, the tool
reads the current configuration parameter values and runs the IPCArchivingPlugin
plugin implementation.
Archive term by policy number and term Enter the policy and term number of the policy term that you want to archive. This ac‐
tion behaves in a similar fashion to that of Archive term by policy transaction number.
Flush other work queues PolicyCenter cannot archive a policy term if that policy term is associated with one or
more workflows. To more effectively archive policy terms, Guidewire recommends that
you run this command to clean up workflows before archiving.
This command runs the following batch processing types:
• Purge Workflow
• Purge Workflow Logs
• Workflow
Run archiving batch process Click Run to start the Archive Policy Term batch processing. This action is the same as
running Archive Policy Term batch processing from the Server Tools→Work Queue Info
screen.
See also
• “Archive Policy Terms Work Queue” on page 101
• “Purge Workflow Batch Process” on page 122
• “Purge Workflow Logs Batch Process” on page 122
• “Workflow Work Queue” on page 126
Internal Tools 383

384 chapter 24: Internal Tools

chapter 25
Command Prompt Tools
PolicyCenter includes a number of administrative tools as command prompt tools that you can use for help with
administrative tasks on your PolicyCenter server.
Note: For tools that build PolicyCenter, see the Installation Guide.
Administration Tools Overview

PolicyCenter provides a set of tools that you can use to perform administrative tasks directly from a command
prompt. Typically, these commands are meant to run on an administrator’s workstation. The tools are all found in the
PolicyCenter/admin/bin directory, unless otherwise noted. These tools all execute against a running PolicyCenter
instance.
There are *.bat and *.sh versions of each administration tool to support installations on Windows and UNIX
systems, respectively. You can only use these tools if the PolicyCenter server is actively running.
User Credentials
The PolicyCenter administration command prompt tools all require that you enter the password of an administrative
user for the tool to work. The use of a user name is optional.
User Name
PolicyCenter does not require the use of a user name for an administration command. The use of a user name is
optional:
• If you do supply a value for the -user parameter, it must be the user name of a user with administrative
privileges.
• If you do not supply a value for the -user parameter, the command defaults to user su (in the base configuration)
and you must supply the password for that user.
Password
PolicyCenter expressly requires the use of a password of a user with administrative privileges for an administration
command. If you do not supply a value for the -password parameter, the execution of the command fails
Masking User Credentials

In the administration command, it is possible mask the input of the user name and password. To do so, insert a dash
in the command in place of the actual user name or password, for example:
system_tools -ping -user - -password -
Command Prompt Tools 385
Before PolicyCenter executes the tool script, it prompts you to enter the missing information. If you chose to mask
the user name and password information:
• PolicyCenter does not echo the entered user name or password back to the screen.
• The operating system, either Windows or Linux, does not store the plain text information in the command prompt
buffer and command history.
Accessing Administration Tool Help

To access help for any of the PolicyCenter command prompt tools, enter -help after the tool name. For example,
enter the following at the command prompt to generate a list of command options with a description of each option
for the import_tools administrative tool:
import_tools -help
Administrative Tool Command Syntax

The administrative tools command descriptions use the following command syntax.
tool_name Bold font indicates that this is the actual command name, for example, import_tools.
-option All command options start with a minus sign (-). Command options are either mandatory or optional. See the
following discussion.
| An upright bar indicates a Boolean OR. For example, A | B | C means A or B or C.
{ ... } A set of curly braces indicates a set of mutually exclusive choices. You must one chose (and only one) item from a
set of choices. For example, { A | B | C } indicates you must choose either A or B or C, but not more than of
one the listed options.
arguments Specifies the arguments required by a tool option such as a file name or directory, for example, import_tools ...
-import file.
... A series of dots after the argument indicates that you can enter multiple items of the same type. For example, -
import file ... indicates that you can enter multiple file names (file) after the -import argument.
[ ... ] A set of square brackets indicates that the argument is optional. For example, [-user] indicates that the com‐
mand permits you to set a user value (-user), but does not require that you set this value.
In contrast, an argument not enclosed in square brackets indicates that an argument is mandatory. For example,
for all the administrative commands, the -password argument is mandatory. Thus, the command syntax does not
surround the -password argument by square brackets as the argument is mandatory.
PolicyCenter Command Prompt Tools Summary

The following list provides a summary of the PolicyCenter administration tools available from a command prompt.
Tool Description
“Data Change Command” Provides a mechanism for making changes to code on a running production server.
on page 387
WARNING Only use the data_change command under extraordinary conditions, with
great caution, and upon advice of Guidewire Support. Before registering a data change on
a production server, register and run the data change on a development server. Guidewire
recommends multiple people review and test the code and the results before attempting
the data change on a production server.
“Import Tools Command” Set of utilities for loading XML‐formatted data into PolicyCenter.
on page 388
“Maintenance Tools Com‐ Set of utilities for performing maintenance operations on the server (for example, running esca‐
mand” on page 390 lation/exception rules, calculating statistics, and more.)
386 chapter 25: Command Prompt Tools

Tool Description
“Messaging Tools Com‐ Provides a set of utilities for managing integration event messages (for example, retrying a mes‐
mand” on page 391 sage, skipping a message, purging the message table, and more).
“System Tools Command” Provides a set of utilities for controlling the server (for example, pinging the server, bringing the
on page 393 server in and out of maintenance mode, updating database statistics, and more.)
“Table Import Command” Used for importing tables into the database.
on page 399
“Template Tools Command” Helps in converting between template versions.
on page 401
“Workflow Tools Com‐ Allows you to manage user workflows in the system.
mand” on page 402
“Zone Import Command” Loads zone data from a file to a staging table.
on page 403
Data Change Command

this mechanism the Production Data Fix tool. The Production Data Fix tool uses either the data_change command
option, or, the DataChangeAPI web service to make changes to the production data.
Because the data_change command allows arbitrary execution of data, Guidewire strongly recommends that you
carefully control the ability to create and run code on a production server. The user who runs this command must
have permission wsdatachangeedit.
server.
data_change -help
data_change -password password [-server url] [-user user] {
-edit refID -gosu filepath [-description desc] |
-discard refID |
-status refID |
-result refID }
See also
• For a description of how and when to use the data_change command to change data on a running production
server, see “Production Data Fix Tool” on page 289.
• For a description of how to use the DataChangeAPI web service, see “Data Change Web Service Reference” on
page 295.
Data Change Options

You can use any of the following options with the data_change command. You must always supply the -password
option.

server.
Option Description
-description desc Human‐readable description (desc) of the change. Include this option with the edit option. For
testing, the description is optional. For production use, include the description. Put quotes around
the description to permit space characters in the description.
-discard refID Instruction to discard a data change that you already registered. You must supply a data change ref‐
erence ID (refID). You cannot discard a data change that was already run.
-edit refID Instruction to create a new data change or edit an existing data change. You must supply a unique
reference ID (refID) for this data change.
If the data change succeeded with no compile errors, you cannot edit it. You must re‐register the
script with a new reference ID.
If the data change was never run, or had compile errors, you can update (edit) the Gosu code with
the same reference ID.
If you use the edit option, you must:
• Include the -gosu option to include your Gosu data change code
• Include the -description argument to provide a description
-gosu filepath Full path name (filepath) to a Gosu script. You must include this option with the edit argument.
You can use a full path name, or a relative path that is relative to the current working directory.
-password password Password (password) to use to connect to the server. PolicyCenter requires the password.
-result refID Result of a data change that you already registered. You must supply a data change reference ID
(refID). If a user attempted to run it and there were parse errors, the results include the errors.
-server url Specifies the PolicyCenter host server URL. Include the port number and web application name, for
example:
http://servername:8180/pc
-status refID Status of a data change that you already registered. You must supply a data change reference ID
(refID). This option prints the status of the data change, which is one of the following:
• Open
• Discarded
• Executing
• Failed
• Completed
-user user User (user) to use to run this process.
Import Tools Command

import_tools -help
import_tools -password password [-server url] [-user user] {
-import filename1, filename2 ... [-charset charset] [-dataset dataset]
[-ignore_all_errors] [-ignore_null_violations]
[ { -output_csv filename | -output_xml filename } ] | -privileges }
The import_tools command imports new or updated data into existing tables in the PolicyCenter database. You can
only import data for valid entities or their subtypes. PolicyCenter supports this command for importing
administrative data but not for importing other data into PolicyCenter.

Note: PolicyCenter does not fire any events related to the data you add or modify through this
command. PolicyCenter does not throw concurrent data change exceptions if the imported records
overwrite existing records in the database.
Data that you import into PolicyCenter through the use of import_tools is immediately available. You do not need
to restart the PolicyCenter server for the changes to take effect.
only.
IMPORTANT The MaximumFileUploadSize parameter in config.xml must exceed the size of any file
that you attempt to import. The MaximumFileUploadSize parameter value is in megabytes (MB). The
base configuration default value of MaximumFileUploadSize is 20 MB.
See also
• “Ways to Import Administrative Data” on page 267
• “About the import Directory” on page 268
• “Using Tools to Import Administrative Data” on page 277
Import Tools Options

You can use any of the following options with the import_tools command. You must always supply the -password
option.
Option Description
-charset charset Character set encoding (charset) for the files to import.If this option is null,
PolicyCenter sets the default character encoding to UTF‐8. See also “Character Set
Encoding for File Import” on page 270.
-dataset integer Integer value (integer) representing the dataset to import from a CSV‐formatted
file, for example:
RolePrivilege,0,default_data:3,abdelete,audit_examiner
PolicyCenter orders datasets by inclusion. The number of the smallest dataset is al‐
ways 0. Thus, dataset 0 is a subset of dataset 1, and dataset 1 is a subset of dataset 2,
and so forth.
To import all data, set this value to ‐1.
-ignore_all_errors Causes the tool to ignore any errors in a CSV‐formatted input file.
-ignore_null_violations Causes the tool to ignore violations of null constraints in a CSV‐formatted input file.
-import filename1, filename2, ... Imports administrative data from one or more CSV (comma‐separated values data)
files or XML files.
It is possible to provide a list of file names in a separate file. To do so, create a file
that contains a comma‐separated list of files names. Prefix an @ character to the
name of the list file, for example:
-import @files.lst
To convert data using the -output_csv or -output_xml options, provide only a sin‐
gle file name.
-output_csv filename If used with the -import option, outputs comma‐separated values to the specified
file and then stops processing. PolicyCenter imports no data into the server. Use this
option to convert XML input files to CSV‐formatted output files.
-output_xml filename If used with the -import option, outputs XML to the specified file and then stops
processing. PolicyCenter imports no data into the server. Use this option to convert
CSV input files to XML‐formatted output files.

Option Description
-password password Password to use to connect to the server. PolicyCenter requires the password value.
-privileges Adds the role privileges contained in file roleprivileges.csv in the Guidewire Stu‐
dio™ modules/configuration/config/import/gen folder to those roles that al‐
ready exist in the database.
See also “About Adding Admin Data after Initial Server Startup” on page 269.
-server url Specifies the PolicyCenter host server URL. Include the port number and web appli‐
cation name, for example:
http://servername:8180/pc
-user user User to use to run this process.
Maintenance Tools Command

maintenance_tools -help
maintenance_tools -password password [-server url] [-user user] {
-processstatus process |
-startprocess process [-args arg1, arg2, ...] |
-terminateprocess process |
-whenstats }
The maintenance_tools command starts, terminates, or retrieves the status of a PolicyCenter process. For a list of
processes that the maintenance_tools command can start, see “The Work Queue Scheduler” on page 89.
Maintenance Tools Options

You can use any of the following options with the maintenance_tools command. You must always supply the -
password option.
Option Description
-args arg1 arg2 ... Arguments to use while starting a process. Use only with -startprocess.
If you have multiple arguments, separate each one with a space. The command does not vali‐
date the provided arguments.
To use arguments with custom batch processes, see the Integration Guide, especially the fol‐
lowing method:
ProcessesPlugin.createBatchProcess(type, args)
-processstatus process Returns the status of a batch process. For the process value, specify a valid process name or
a process ID.
For work queues, this option returns the status of the writer process. It does not check
whether additional work items remain in the work queue. Thus, it is possible for the process
status to report completion after the writer finishes adding items to the work queue while
the work queue contains unprocessed work items.
-server url Specifies the PolicyCenter host server URL. Include the port number and web application
name, for example:
http://serverName:8180/pc
-startprocess process - Starts a new batch process. For the process value, specify a valid process code. See also -
args... args.
For a list of batch process codes, including work queue writer processes, see “Work Queues
and Batch Processes, a Reference” on page 98.
-terminateprocess process Terminates a batch process. For the process value, specify a valid process name or a process
ID.

Option Description
It is not possible to terminate single phase processes using this option. Single phase process‐
es run in a single transaction. Thus, there is no convenient place to terminate the process.
See “Work Queues and Batch Processes, a Reference” on page 98 to determine if it is possi‐
ble to terminate a process.
-whenstats Reports the last time PolicyCenter calculated statistics on the server.
Messaging Tools Command

messaging_tools -help
messaging_tools -password password [-server url] [-user user] {
-config destinationID |
-purge date |
-restart -destination destinationID [-wait wait] [-retries retries] [-initial initial]
[-backoff backoff] [-poll poll] [-threads threads] [-chunk chunk] |
-resume destination destinationID |
-resync -destination destinationID -account -accountID |
-retry messageID |
-retrydest destinationID |
-skip messageID |
-statistics destinationID |
-suspend destinationID }
You use the messaging_tools command to manage a message destination from the command prompt. To do so,
you must know the message destination ID. The person who creates the message destination assigns this ID. You
create and configure message environments and destinations in file messaging-config.xml. Access messaging-
config.xml in Guidewire Studio at the following location:
configuration→config→Messaging
Messaging Tools Options

You can use any of the following options with the messaging_tools command. You must always supply the -
password option.
Op‐ Description
tion
-account accountID Use to specify the account ID (accountID) of the account to re‐synchronize. See -resync.
-config - Returns the configuration for a message destination.
destination destinationID
- Specifies a message destination (destinationID).

-purge date Deletes completed messages that are older than a specified date. The purge tool deletes mes‐
sages in Acked, ErrorCleared, Skipped or ErrorRetried state with send time before the
specified date. The date format is mm/dd/YYYY.
If the purge tool succeeds in removing these messages without error, it reports Message
table purged.
Since the number and size of messages can be very large, periodically use this command op‐
tion to purge old messages to avoid the database from growing unnecessarily.

Op‐ Description
tion
-restart - Restarts the messaging destination with new configuration settings:

destination destinationID • destinationID – The destination ID of the destination to restart.
-wait wait • wait – the number of seconds to wait for the shutdown before forcing it.
-retries retries • retries – The number of automatic retries to attempt before suspending the messaging
-initial initial destination.
-backoff backoff • initial – The amount of time in milliseconds after a retryable error to retry sending a
-poll poll message.
-threads threads • backoff – The amount to increase the time between retries, specified as a multiplier of
-chunk chunk the time previously attempted. For example, if the last retry time attempted was 5
minutes, and backoff is set to 2, PolicyCenter attempts the next retry in 10 minutes.
• poll – Each messaging destination pulls messages from the database (from the send
queue) in batches of messages on the batch server. The application does not query again
until pollinterval amount of time passes. After the current round of sending, the
messaging destination sleeps for the reminder of the poll interval. If the current round of
sending takes longer than the poll interval, then the thread does not sleep at all and
continues to the next round of querying and sending. See the Integration Guide for details
on how the polling interval works. If your performance issues primarily relate to many
messages for each primary object for each destination, then the polling interval is the most
important messaging performance setting.
• threads – To send messages associated with a primary object, PolicyCenter can create
multiple sender threads for each messaging destination to distribute the workload. These
are threads that actually call the messaging plugins to send the messages. Use the -
threads option to configure the number of sender threads for safe‐ordered messages.
PolicyCenter ignores this setting for non‐safe‐ordered messages, as PolicyCenter uses one
thread for each destination for these types of messages. If your performance issues
primarily relate to many messages but few messages per claim for each destination, then
this is the most important messaging performance setting. For more information, see the
Integration Guide.
• chunk – number of messages to read in a chunk.
-resume - Resumes the operation of the specified message destination.
-resync - Re‐synchronizes an account with specified ID against a specific message destination. Use -
destination destinationID destination and -account to specify the destination and policy.
-account
accountID
-retry messageID Attempts to resend a message that failed. The message must be a candidate for retrying. A
message is a candidate for retry if the error at the destination system is temporary and the
message destination does not have an automatic retry mechanism. For instance, the message
is a candidate for retry if the destination contains a locked record and refuses the update.
-retrydest destinationID Retries all retryable messages for a message destination.
-server url Specifies the PolicyCenter host server URL. Include the port number and web application
name, for example:
-skip messageID Skips a message with the specified ID. If you mark a message as skipped, then PolicyCenter
stops trying to resend the message. After you skip a message, you cannot retry it.
-statistics destinationID Prints the statistics for the specified destination.
-suspend destinationID Suspends a message destination. Use this command option if the destination system is going
to be shut down or to halt sending while PolicyCenter processes a daily batch file.

System Tools Command

system_tools -help
system_tools -password password [-server url] [-user user] {
-cancelshutdown serverId |
-cancelupdatestats |
-checkdbconsistency [-tableselection tblSelection -checkTypeSelection checkTypeSelection] |
-completefailedfailover type componentId |
-components |
-daemons |
-dbcatstats [regularTables stagingTables typelistTables] |
-getdbccstate |
-getdbstatisticsstatements |
-getincrementaldbstatisticsstatements |
-getPerfReport ID |
-getupdatestatsstate |
-listPerfReports [number] |
-loggercats |
-maintenance |
-mssqlPerfRpt [numTopQueries numHotObjects collectStatistics] |
-multiuser |
-nodefailed serverId [-evenifincluster -filepath filepath] |
-nodes |
-oraListSnaps numstats |
-oraPerfReport beginSnapshotID endSnapshotID probeVDollarTables |
-ping |
-recalcchecksums |
-reloadloggingconfig |
-requestcomponenttransfer type componentId targetOwner |
-scheduleshutdown serverId [-terminatebatchprocesses | -shutdowndelay minutes] |
-sessioninfo |
-startfullupgrade |
-updatelogginglevel loggername logginglevel |
-updatestatistics description update |
-verifyconfig filepath |
-verifydbschema |
-version }
See also

System Tools Options

You can use any of the following options with the system_tools command. You must always supply the -password
option.
Option Description
-cancelshutdown serverId Cancels the planned shutdown of the server specified by serverId.
-cancelupdatestats Cancels the process that is updating database statistics if running. Use the fol‐
lowing option to verify the process state:
-getupdatestatsstate
-checkdbconsistency Checks the consistency of data in the database. The -checkdbconsistency op‐
tion runs consistency checks as an asynchronous batch process. PolicyCenter
provides this option so that you can schedule consistency checks to run during
a time period of low activity on the database server.
The -checkdbconsistency option has two optional arguments:
• tblSelection
• checkTypeSelection
Option Description
Specify the tblSelection argument as one of the following:
• all – Run consistency checks on all tables.
• tableName – The name of a single table on which to run checks.
• tg.tableGroupName – The name of a table group. Use the <database>
element in database-config.xml to define table groups. For more
information, see the Installation Guide.
• @fileName – The name of a file name that contains one or more valid table
names or table group names entered in comma‐separated values (CSV)
format. Prefix table group names with tg., such as tg.MyTableGroup. You
can combine table groups and individual table names in the same file.
Specify the checkTypeSelection argument as one of the following:
• all – Run all consistency checks on the specified tables.
• checkName – The typecode of a single consistency check to run.
• @fileName – The name of a file with one or more valid consistency check
names entered in comma‐separated values (CSV) format.
If you specify one optional argument, you must specify both.
To run consistency checks from PolicyCenter, use the Server Tools Consistency
Checks→Info Page screen, described in “Consistency Checks” on page 333.
For more information, see “Database Consistency Checks” on page 240.
-completefailedfailover type Manually completes component failover for a failed component. You must sup‐
componentId ply the component type (type) and component ID (componentId).
-components Provides information about the components that exist on each PolicyCenter
server in the cluster. The report contains the following information for each
component:
• Component type
• Component code
• Component state
• Component start date and time
• Server ID of the server instance on which the component exists
• Component ID
The report information is similar, but not identical, to the cluster information
available from the Server Tools Cluster Members and Cluster Components screens.
See “Cluster Members and Components” on page 356 for information on these
screens.
-daemons Sets the server to the daemons run level. For information about the various run
levels, see “Server Run Levels” on page 55.
-dbcatstats Used with no arguments, the option returns a ZIP file of database catalog statis‐
tics info for all the tables in the database.
The -dbcatstats option takes the following optional arguments:
• regularTables
• stagingTables
• typelistTables
This option, used with three arguments, returns a ZIP file of database catalog
statistics info for the specified tables.
Specify each of the arguments as one of the following:
• all / none – Select all or none of the tables of this type
• <tableName> – The name of a single table of this type
• @<fileName> – The name of a file with one or more valid table names of
this type entered in comma‐separated values (CSV) format.
For example, -dbcatstats none none all returns database catalog statistics
information for all the typelist tables. You must specify either no arguments or
all three arguments if you use this command option.
You can specify the target destination for the database catalog statistics ZIP file
by adding the –filepath filepath option. If you do not provide a path,
PolicyCenter uses the current directory.

Option Description
This process can take a long time, and it is possible for the connection to time
out. If the connection times out while running this command option, try reduc‐
ing the number of tables on which to gather statistics by using the arguments
listed previously.
For information about configuring database statistics generation, see “Under‐
standing Database Statistics” on page 255.
-evenifincluster [-filepath filepath] Consider the cluster member as failed even if it is still in the cluster. Use only
with the -nodefailed option.
The -filepath parameter sets the location (filepath) for an optional report.
IMPORTANT This command option overrides an important safety check on the
server. Use this command option in certain defined circumstances only. See -
nodefailed for details.
-getdbccstate Returns the status of any currently executing database consistency checks,
Processing or Completed, for example.
-getdbstatisticsstatements Retrieves the list of SQL statements to update database statistics and prints the
list to the console. See “Understanding Database Statistics” on page 255.
-getincrementaldbstatisticsstatements Retrieves the list of SQL statements to update database statistics for tables ex‐
ceeding the change threshold. Prints the list to the console.
The incrementalupdatethresholdpercent attribute of the
<databasestatistics> element in database-config.xml defines the change
threshold. See “Understanding Database Statistics” on page 255.
-getPerfReport ID Downloads the performance report with the specified ID. You can retrieve a list
of available performance report IDs by running the -listPerfReports com‐
mand option.
-getupdatestatsstate Returns the state of the process running the statistics update.
-listPerfReports number Lists IDs and other information for available database performance reports. You
can specify an optional integer (number) to specify the number of available
downloads to list, ordered starting with the most recent. If unspecified or 0,
this command lists all available downloads.
The list shows the ID of the report and the status, indicating if the performance
report batch job succeeded, failed, or is still running. The list also includes the
start and end times of the batch job and the description of the batch run.
You can use the ID of the performance report to download the report with the
-getPerfReport ID option.
-loggercats Displays the available logging categories.

-maintenance Sets the server to the maintenance run level. For information about the various
run levels, see “Server Run Levels” on page 55.
-mssqlPerfRpt numTopQueries Generates a SQL Server DMV (Dynamic Management Views) performance re‐
numHotObjects port using the MSDMReport batch process. This command option has the follow‐
collectStatistics ing arguments:
• numTopQueries
• numHotObjects
• collectStatistics
Replace numTopQueries and numHotObjects with integer values for the num‐
ber of top queries and hot objects to report.
Replace collectStatistics with true or false to specify whether
PolicyCenter gathers database statistics while generating the DMV report.
You must specify all three arguments or none. If you do not specify any argu‐
ments, PolicyCenter uses defaults of 400 top queries, 400 hot objects, and does
collect statistics.
-multiuser Sets the server to the multiuser run level. For information about the various
run levels, see “Server Run Levels” on page 55.

Option Description
-nodefailed serverId Releases any tasks owned by the PolicyCenter server specified by serverId.
Only use this command option if the server referenced by serverId has already
been stopped or otherwise shutdown.
See also the -evenifincluster option.
IMPORTANT You must ensure that the server referenced by serverId is actual‐
ly stopped if using the -evenifincluster option. PolicyCenter does not pre‐
vent you from using this option if the server is still running. However, this op‐
tion overrides an important safety check on the server. It can produce unex‐
pected and possibly negative results if the server is running.
Use the -evenifincluster option only if both of the following are true:
• The server in question is no longer running.
• The standard operation of the -nodefailed command failed due to the
server retaining its cluster membership.
-nodes Provides information about each PolicyCenter server in the cluster. The report
contains the following information on each cluster member:
• ID of this cluster server
• Whether the server instance is actively in the cluster
• Server run level
• Time the server instance started
• Time at which PolicyCenter last updated the server instance
• Number of user sessions active on the server instance
• Whether a planned shutdown is in progress
• Time of the planned shutdown
• Whether background tasks are still active on the server
The report information is similar, but not identical, to the cluster information
available from the Server Tools Cluster Members screen. See “Cluster Members
and Components” on page 356 for information on that screen.
-oraListSnaps numSnaps Lists numSnaps number of available Oracle AWR snapshot IDs, starting with the
most recent snapshot. You can generate performance reports using the -
oraPerfReport option with these available beginning and ending snapshot IDs.
-oraPerfReport beginSnapshotID Generates a Guidewire AWR performance report using the OraAWRReport batch
endSnapshotID process. This command option has the following arguments:
probeVDollarTables • beginSnapshotID
• endSnapshotID
• probeVDollarTables
Specify the beginning and ending snapshot IDs and whether to probe VDollar
tables. The two snapshots must share the same Oracle instance startup time.
The third argument can also specify a file by prefixing the file name with an @
sign, for example, @filename.properties.
Optionally, you can prefix the file name with the path to the file, if the file is not
in the current directory. This file is a standard properties file with the following
property names (default value in parenthesis):
• probleVDollarTables (false)
• capturePeekedBindVariables (false)
• searchQueriesMultipleHistoricPlans (false)
• searchQueriesBeginSnapOnly (true)
• searchQueriesEndSnapOnly (true)
• includeInstrumentationMetadata (false)
• outputRawData (false)
• includeDatabaseStatistics (true)
• probeSqlMonitor (true)
• capturePeakedBindVariablesFromAWR (false)
• genCallsToAshScripts (false)

Option Description
You must spell and capitalize each property as shown or PolicyCenter ignores
the property. If you specify a property, you must set value of that property to
either true or false. If you do not specify a property, PolicyCenter uses the de‐
fault value for that property.
The -oraPerfReport option reports the process ID of the process generating
the performance report. You can check on the status of this process using the
following command:
maintenance_tools -processstatus processID
View the performance report on the Info Page. See “Oracle AWR” on page 347.
-password password Password (password) to use to connect to the server. PolicyCenter requires the
password.
-ping Pings the server to check if its active. The returned message indicates the serv‐
er run level. The possible responses are:
• MULTIUSER
• DAEMONS
• MAINTENANCE
• STARTING
For information about functionality available at various run levels, see “Server
Modes” on page 53.
-recalcchecksums Recalculates file checksums that PolicyCenter uses for clustered configuration
verification.
-reloadloggingconfig Directs the server to reload the logging configuration file.
-requestcomponenttransfer type Requests transfer of ownership of a component of the specified type (type) and
componentId ID (componentId) to the specified PolicyCenter server (targetOwner).
targetOwner Use the -components command option to determine the component informa‐
tion to enter.
The -requestcomponenttransfer command option fails if the component can‐
not be successfully stopped or the current owning server is unable to process
the request.
-scheduleshutdown serverId Schedules the planned shutdown of the server specified by serverId. Use with
[-terminatebatchprocesses the following optional options:
-shutdowndelay minutes] • -terminatebatchprocesses – Determines how the shutdown process
handles the batch processes (including work queue writers) currently
running on the server. See “About Planned Server Shutdowns” on page 357
for information on this command option.
• -shutdowndelay minutes – Server shuts down in the number of minutes
specified by minutes.
It is also possible to schedule a server shutdown in the following ways:
• From the Server Tools Cluster Members screen. See “Cluster Members and
Components” on page 356 for information.
• Through the SystemToolsAPI web service. See the PolicyCenter Integration
Guide for information.
-shutdowndelay minutes Server shuts down in the number of minutes specified by minutes. Use with
the -scheduleshutdown option only.
-server url Specifies the PolicyCenter host server URL. Include the port number and web
application name, for example:
-sessioninfo Returns the session information of the server.

-startfullupgrade Starts the process of full application upgrade for the entire cluster of server
members. You must shut down all cluster members and deploy a new WAR/EAR
file to each one before attempting the upgrade.

Option Description
See the PolicyCenter upgrade documentation for details of the PolicyCenter ap‐
plication upgrade process.
See the Upgrade Guide for more information.
-terminatebatchprocesses Immediately terminates all running batch processes on the specified server. Use
with the -scheduleshutdown option only.
-updatelogginglevel logger level Sets the logging level of logger with the given name. For the root logger, specify
RootLogger for the logger name.
-updatestatistics description update Launches the Update Statistics batch process to update database statistics.
It is possible to specify an optional text description (description) for this batch
process execution. PolicyCenter shows the text of the description on the Execu-
tion History tab of the Database Statistics info page.
Specify one of the following values for update:
• true – Update database statistics for tables exceeding the change threshold
only. Guidewire defines this change threshold through the
incrementalupdatethresholdpercent attribute of the
<databasestatistics> element in database-config.xml.
• false – Generate full database statistics.
IMPORTANT Updating database statistics can take a long time on a large data‐
base. Only collect statistics if there are significant changes to data, such as after
a major upgrade, after using the zone_import command, or if there are per‐
formance issues.
See also
-verifyconfig filepath Compares the following two server configurations:
1. The new or target server configuration contained in a WAR/EAR file
pointed to by the filepath parameter.
2. The existing server configuration of the cluster member on which you
run the system_tools command.
The tool provides an on‐screen report that contains information about the fea‐
sibility of a configuration upgrade for the server instances in the cluster. For ex‐
ample, the tool provides the following types of information:
• Configurations are different – Requires a full PolicyCenter server upgrade.
• Configurations are identical – No upgrade is necessary.
• Configurations are compatible – Guidewire permits a rolling upgrade.
If a rolling update is not possible, the command lists the incompatible or miss‐
ing files.
If a rolling update is in progress, there are two possible configurations active in
the cluster. Each individual server instance is using either the source configura‐
tion or the target configuration.
The -verifyconfig command option checks for both configurations on the
server instances on which you run the command and reports which of the con‐
figurations is active on this cluster member. If neither configuration is active,
the command reports that a rolling update is in progress and that it is not possi‐
ble to verify the configuration at this time.
-verifydbschema Verifies that the data model matches the underlying physical database.
-version Returns the running server version, the database schema version, and configu‐
ration version.

Table Import Command

table_import -help
table_import -password password [-server url] [-user user] {
-clearerror |
-clearexclusion |
-clearstaging |
-deleteexcluded [-batch] |
-encryptstagingtbls [-batch] |
-getLoadHistoryReport reportID [-filepath filepath] |
-integritycheck [-allreferencesallowed] [-batch] [-clearerror]
[-numthreadsintegritychecking num ] [-populateexclusion] |
-integritycheckandload [-allreferencesallowed] [-batch] [-clearerror] [-estimateorastats]
[-numthreadsintegritychecking num ] [-populateexclusion] [-zonedataonly] |
-listLoadHistoryReports numReports |
-populateexclusion [-batch] |
-updatedatebasestatistics [-batch] [-integritycheckandload] }
The table_import command loads data from staging tables into PolicyCenter. Most of the options for this
command require the server to be at the MAINTENANCE run level. Before you use those command options, use the
system_tools -maintenance command option to set the server run level to MAINTENANCE. Use the system_tools
-multiuser command option to set the server run level to MULTIUSER after the table import command completes.
It is not possible to use the system_tools -terminateprocess command option to terminate the table_import
command.
IMPORTANT PolicyCenter supports bulk data import from staging tables for loading zone data only.
For more information, see “Zone Import Command” on page 403.
See also
• “Load History” on page 351
Table Import Options

You can use any of the following options with the table_import command. You must always supply the -password
option.
Option Description
-allreferencesallowed Allows references to existing non‐administrative rows in all operational tables.
This option only applies with the following command options:
-integritycheck
-integritycheckandload
This option corresponds to the Boolean parameter allowRefsToExistingNonAdminRows
used by the integrity check methods of the TableImportAPI web service. Guidewire rec‐
ommends that you use this option or the equivalent API parameter, set to true only if
absolutely necessary.
This option can cause performance degradation during the check and load process, which
would noticeably slow down the loading of staging table.
See also “The Load History Detail Report” on page 352.

Option Description
-batch Runs the table_import command in a batch process. This option only applies with the
following command options:
-deleteexcluded
-encryptstagingtbls
-integritycheck
-populateexclusion
-updatedatabasestatistics
You can run table_import in a batch process from any node in a PolicyCenter cluster.
However, table import batch processing must run physically on a server designated as a
batch server. Therefore, in running the command, provide the URL of a batch server and
also provide the user credentials for that batch server.
Note: The -batch option does not wait until the started process completes before
returning. Instead, it returns immediately and prints the ID of the started process (PID).
The process caller is responsible for waiting for the process to complete before taking
further action.
-clearerror Clears the error table.
See also “The Load History Detail Report” on page 352.
-clearexclusion Clears the exclusion table.
-clearstaging Clears the staging tables. Requires the server to be at the MAINTENANCE run level.
-deleteexcluded Deletes rows from staging tables based on contents of exclusion table.
-encryptstagingtbls Do not use. While this command option is available, Guidewire does not support this
command option in PolicyCenter.
-estimateorastats Executes queries for row counts on production tables and sets the database statistics. If
you do not use this option, the import command uses information in database statistics
to report approximate row counts. Use the -estimateorastats option only to load pro‐
duction tables that are empty or have very few rows. Used with the -
integritycheckandload command option.
This option only applies with the following command option:
• -integritycheck
• -integritycheckandload
This parameter corresponds to the Boolean parameter
updateDBStatisticsWithEstimates used by the integrity check methods of the
TableImportAPI web service.
This command option applies to Oracle databases only.
-filepath filepath Path to target directory in which to download a report. Use with the -
getLoadHistoryReport command option.
-getLoadHistoryReport Downloads a compressed Zip version of the load history report as specified by the value
reportID of reportID. Does not require the server to be at the MAINTENANCE run level. Use the -
listLoadHistoryReports option to determine the ID to use. Use the optional -filepath
parameter to specify the target directory for the download.
-integritycheck Validates the contents of the staging tables. You can optionally specify:
-allreferencesallowed
-clearerror
-numthreadsintegritychecking
-populateexclusion

Option Description
-integritycheckandload Validates the contents of the staging tables and populate operational tables. You can op‐
tionally specify one of the following command options as well:
-allreferencesallowed
-clearerror
-estimateorastats
-numthreadsintegritychecking
-populateexclusion
-zonedataonly
-listLoadHistoryReports Lists the most recent load history reports. Optional parameter numReports is the number
[numReports] of reports to list:
• If you supply a positive integer for numReports, then PolicyCenter lists that number of
most recent reports.
• If you do not supply a value for numReports, then PolicyCenter lists all available
reports.
Does not require the server to be at the MAINTENANCE run level.
-messagesinks sinks, ... Deprecated. This option does nothing.
-numthreadsintegritychecking Specifies the number of threads that PolicyCenter is to use in running database table in‐
num tegrity checks. The value of num has the following meaning:
• Not specified – PolicyCenter assumes the number of threads to be one, no
multithreading.
• 1 – No multithreading, the default.
• 2 ‐ 100 – PolicyCenter runs the database integrity checks with the number of specified
threads.
• > 100 – PolicyCenter throws an exception.
This option only applies with the following command options:
-integritycheck
-populateexclusion Populate the exclusion table with rows to exclude.
See “The Load History Detail Report” on page 352.
-server url Specifies a PolicyCenter server URL. Include the port number and web application name,
for example:
If running the table import command in a batch process, see -batch for more informa‐
tion.
-updatedatabasestatistics Updates the database statistics on the staging tables. Run the table import command
with this option after populating the staging tables, but before using the -
integritycheck or -integritycheckandload options.
See “The Load History Detail Report” on page 352.
-user user Specifies the user to use to run this process.
-zonedataonly Sets the import to load zone data only. Used with the -integritycheckandload com‐
mand option.
Template Tools Command

template_tools -help
template_tools -password password [-server url] [-user user] {
-convert_dir directory |
-convert_file filename [working_dir directory] |
-import_dir objectsfile fieldsfile directory [working_dir directory] |
-import_files objectsfile fieldsfile outfile |

-list_templates |
-validate_all |
-validate_template templateID }
The template_tools command contains options to list, manage, and validate document templates.
See also
Template Tools Options

You can use any of the following options with the template_tools command. You must always supply the -
password option.
Option Description
-convert_dir directory Converts all templates in the specified directory to the new format.
-convert_file filename Converts the specified template to the new format.
-import_dir objectsfile fieldsfile Imports context objects and form fields from the provided CSV‐formatted files into
directory all the templates in the specified directory. This option has the following argu‐
ments:
• objectsfile – File containing the context objects for import, in CSV format.
• fieldsfile – File containing the fields for import, in CSV format.
• directory– Directory that contains the templates to update.
-import_files objectsfile Imports context objects and form fields from the provided CSV‐formatted files into
fieldsfile outfile the specified template descriptor file (outfile). This option has the following argu‐
ments:
• objectsfile – File containing the context objects for import, in CSV format.
• fieldsfile – File containing the fields for import, in CSV format.
• outfile – Template descriptor file to update.
-list_templates Lists all of the templates available for validation.
-password password Password (password) to use to connect to the server. PolicyCenter requires the
password.
-server url Specifies the PolicyCenter host server URL. Include the port number and web appli‐
cation name, for example:

-validate_all Validates all the templates in a similar manner to -validate_template.
-validate_template templateID Validates a single template. Validates that the given template descriptor
(templateID) is in a valid format, and that all template descriptor context objects
and form fields are valid given the current data model.
-working_dir directory Specifies a directory for use as the root (working directory) for relative paths.
Workflow Tools Command

workflow_tools -help
workflow_tools -password password [-server url] [-user user] {
-complete workflowID |
-resume workflowID |
-resume_all |
-suspend workflowID }

You can also control workflows using the WorkflowAPI web service. See the Integration Guide.
Workflow Tools Options

You can use any of the following options with the workflow_tools command. You must always supply the -
password option.
Option Description
-complete workflowID Completes running workflow for the specified workflow (workflowID).
-resume workflowID Resume named workflow (workflowID) in the error or suspended state.
-resume_all Resume all workflows in the error or suspended state.
-server url Specifies the PolicyCenter host server URL. Include the port number and web application name, for
example:
-suspend workflowID Suspend the named workflow (workflowID).

Zone Import Command

zone_import -help
zone_import -password password [-server url] [-user user] {
-import filename -country country [-clearstaging] [-charset charset] |
-clearproduction [-country country] |
-clearstaging [-country country] }
The zone_import command imports data in CSV format from specified files into database staging tables for zone
data. It is only possible to import zone data for a single country at a time. The zone data files that you import must
contain zone data for a single country only. To load zone data for multiple countries, use the command multiple
times with different, country-specific zone data files each time.
Guidewire expects that you import address zone data upon first installing PolicyCenter, and then at infrequent
intervals thereafter as you receive data updates.
See also
• For a discussion of zone data, importing a zone data file, and working with custom zone data files, see “About
Importing Zone Data” on page 282.
• For more information on importing zone data and database staging tables generally, see the Integration Guide.
• For information on the web service ZoneImportAPI that also imports zone data, see the Integration Guide.
Zone Import Options

You can use any of the following options with the zone_import command. You must always supply the -password
option.
Option Description
-charset charset Character set encoding of the zone data file. The default is UTF‐8.
-clearproduction Clears zone data from the production tables. Optionally, specify the -country option to clear data
for only one country.

Option Description
-clearstaging Clears zone data from the staging tables. Optionally, specify the -country option to clear data for
only one country.
-country countrycode Used with -import, -clearproduction, and -clearstaging command options:
• If used with the -import option, -country specifies the country of the zone data in the import
file.
• If used with either the -clearproduction or -clearstaging options, -country specifies the
country of the zone data to clear from the tables.
-import filename Imports zone data from the specified file (filename). You must set a value for the -country op‐
tion.
If you include the optional -clearstaging option, PolicyCenter clears the data in the staging ta‐
bles for the specified country before importing the data from the import file.
-server url Specifies the PolicyCenter host server URL. Include the port number and web application name,
for example:

Admin Guide

Uploaded by

Copyright:

Available Formats

Admin Guide

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Admin Guide

Uploaded by

Copyright:

Available Formats

Guidewire PolicyCenter™

System Administration Guide

Product Name: Guidewire PolicyCenter

About PolicyCenter documentation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16

2 Application Logging. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .25

4 PolicyCenter Server Administration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .51

Cache Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .64

6 Administering Batch Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .81

Archive Policy Terms Work Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

Retire Activities Batch Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124

8 Working with a PolicyCenter Cluster . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147

9 Working with Component Leases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157

Automatic Failover of a Component Lease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158

10 Deploying Configuration Changes in a Clustered Environment . . . . . . . . . . . . . . . . . . . . . . . . 165

12 Securing Access to PolicyCenter Objects. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185

13 Securing Access to Notes and Documents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193

14 Securing Access to Accounts, Jobs, and Policy Periods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197

16 Database Maintenance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237

Purging Work Item Set Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252

17 Database Statistics Generation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255

18 Importing and Exporting Administrative Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267

Importing Custom Zone Data Files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284

19 Free-text Batch Load Command. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285

20 Production Data Fix Tool . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289

22 Business Rules Import and Export . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309

Business Rule Import after Application Upgrade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317

The Cluster Components Screen. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 360

24 Internal Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 381

25 Command Prompt Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 385

Template Tools Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 401

About PolicyCenter documentation

The following table lists the documents in PolicyCenter documentation:

16 About PolicyCenter documentation

Conventions in this document

bold Highlights important sections of code in for (i=0, i<someArray.length(), i++) {

18 About PolicyCenter documentation

PolicyCenter Default System Users

User Default Owner

IMPORTANT Do not make direct assignments to user defaultowner.

User PolicyChange Daemon

Deleting User policychange_daemon

User Renewal Daemon

Deleting User renewal_daemon

User Super User

User su and the Command Line Tools

User System User

Temporary System Users

Change the Unrestricted User

<param name="UnrestrictedUserName" value="userName"/>

Minimum and Maximum Password Length

<param name="MinPasswordLength" value="6"/>

24 chapter 1: Managing Users

Overview of Application Logging

The Logging Properties File

The Logging Directory Path

Changes to the Logging Level

Logging and the env Environment Property

The Log Files Directory and the View Logs Screen

Adding Loggers to the Logging Properties File

Make an Example Logger in the Logging Properties File Active