Troubleshooting Guide: IBM DB2 Universal Database

® ®
IBM DB2 Universal Database IBM
Troubleshooting Guide
GC09-2850-01
® ®
IBM DB2 Universal Database IBM
Troubleshooting Guide
GC09-2850-01
Before using this information and the product it supports, be sure to read the general information under Notices.
This document contains proprietary information of IBM. It is provided under a license agreement and is protected
by copyright law. The information contained in this publication does not include any product warranties, and any
statements provided in this manual should not be interpreted as such.
You can order IBM publications online or through your local IBM representative.
v To order publications online, go to the IBM Publications Center at www.ibm.com/shop/publications/order
v To find your local IBM representative, go to the IBM Directory of Worldwide Contacts at
www.ibm.com/planetwide
To order DB2 publications from DB2 Marketing and Sales in the United States or Canada, call 1-800-IBM-4YOU
(426-4968).
When you send information to IBM, you grant IBM a nonexclusive right to use or distribute the information in any
way it believes appropriate without incurring any obligation to you.
© Copyright International Business Machines Corporation 1993, 2000. All rights reserved.
US Government Users Restricted Rights – Use, duplication or disclosure restricted by GSA ADP Schedule Contract
with IBM Corp.
Contents
Chapter 1. Troubleshooting overview . . 1 Interpreting diagnostic log file entries . . . 58
7 Introduction to problem determination . . . . . 1 Interpreting the db2diag.log file informational
Identifying the version and service level of your record . . . . . . . . . . . . . . 60
product . . . . . . . . . . . . . . . . 3 Interpreting an SQL structure in the
DB2 server process model . . . . . . . . . . 4 db2diag.log file . . . . . . . . . . . 61
Proactive monitoring tools . . . . . . . . . . 7 Dump files . . . . . . . . . . . . . 61
Snapshot and event monitors . . . . . . . . 8 Trap files . . . . . . . . . . . . . . 62
Setting up snapshot monitoring . . . . . . 9 Formatting trap files (Windows) . . . . . 63
Creating an event monitor . . . . . . . . 9 Analyzing trap files . . . . . . . . . 64
Activating an event monitor . . . . . . . 9 Common signals and exceptions that cause
Formatting an event monitor . . . . . . 10 trap file generation . . . . . . . . . . 64
Basic monitoring commands . . . . . . . 10 Analyzing the stack trace back . . . . . . 65
Configuring health indicators for health monitors 11 Platform specific error logs . . . . . . . . 66
Health monitor output. . . . . . . . . . 12 Operating system errors . . . . . . . . 67
Interpreting event monitor and snapshot output 13 Setting up the system error log (UNIX) . . . 68
Diagnosing the source of the problem . . . . . 13 System core files (UNIX) . . . . . . . . 69
Diagnostic tools . . . . . . . . . . . . 13 Accessing system core file information (UNIX) 69
7 db2diag - Analyzing db2diag.log files . . . 13 Example of the dbx Command . . . . . 70
db2greg - Displaying and altering the Global Accessing event logs (Windows) . . . . . 70
Registry (UNIX) . . . . . . . . . . . 14 Exporting event logs (Windows) . . . . . 70
db2look - Mimicking databases . . . . . . 14 Accessing the Dr. Watson log file (Windows) 71
db2nstck and db2_call_stack - Generating 7 Combining DB2 and OS Diagnostics . . . . . 71
EDU call stacks . . . . . . . . . . . 18
9 db2pd - Monitoring and troubleshooting DB2 19 Chapter 2. Specific troubleshooting
db2support - Collecting environment techniques . . . . . . . . . . . . . 75
information . . . . . . . . . . . . 35 Installation issues . . . . . . . . . . . . 75
Platform specific tools . . . . . . . . . . 38 Location of installation error logs . . . . . . 75
Diagnostic tools (Windows) . . . . . . . 38 Installation methods for DB2 UDB (Windows and
Diagnostic tools (UNIX) . . . . . . . . 38 UNIX) . . . . . . . . . . . . . . . 76
Troubleshooting commands (AIX) . . . . 39 Interpreting Windows installation error logs . . 78
Troubleshooting commands (UNIX) . . . 39 Interpreting UNIX installation error logs. . . . 82
Commands for DB2 in a partitioned Tracing installation problems . . . . . . . 86
database environment (UNIX) . . . . . 40 Troubleshooting a DB2 Information Integrator
Determining active process status . . . . . 41 installation . . . . . . . . . . . . . 86
Trace files . . . . . . . . . . . . . . 43 Instance creation and update issues . . . . . . 86
Basic trace diagnostics . . . . . . . . . 43 Determining the cause of instance creation
DB2 trace files . . . . . . . . . . . 44 problems . . . . . . . . . . . . . . 86
DB2 trace memory buffers . . . . . . 46 Identifying instance update issues (UNIX) . . . 87
Dumping a trace file . . . . . . . . 46 Avoiding fixpak upgrade problems . . . . . 90
Formatting a DB2 trace file . . . . . . 47 License issues . . . . . . . . . . . . . 90
DRDA trace files. . . . . . . . . . . 48 Troubleshooting licensing problems . . . . . 91
JDBC trace files . . . . . . . . . . . 50 Migration issues . . . . . . . . . . . . . 92
Enabling JDBC traces . . . . . . . . 50 Migration recommendations . . . . . . . . 92
Enabling DB2 Universal JDBC traces . . . 51 Migration restrictions . . . . . . . . . . 94
Obtaining traces of applications that use 7 Reverse FixPak upgrade restrictions . . . . . 95
CLI-based Legacy Type 2 JDBC Driver . . 52 Migrating databases . . . . . . . . . . 95
Obtaining traces of applications that use Application Development issues . . . . . . . 96
the DB2 Universal JDBC Driver. . . . . 53 Debugging and Optimizing an Application . . . 97
First failure data capture . . . . . . . . . 53 Proper application error handling . . . . . . 97
First failure data capture locations . . . . . 54 Displaying the contents of a bind file using the
Setting the error capture level for the db2bfd tool . . . . . . . . . . . . . 99
administration notification log file . . . . . 55 Resolving compilation errors . . . . . . . 100
Interpreting administration notification log file Avoiding linker errors . . . . . . . . . 101
entries . . . . . . . . . . . . . . 55 Troubleshooting suspended or looping
Setting the diagnostic log file error capture applications . . . . . . . . . . . . . 102
level . . . . . . . . . . . . . . . 58
© Copyright IBM Corp. 1993, 2000 iii

Error-Handling Considerations in partitioned Error logging for warehouse programs and
database environments . . . . . . . . . 102 transformers . . . . . . . . . . . . . 124
DB2 Development Center issues . . . . . . . 102 Tracing errors created by the Apply program 125
Enabling the Next push button in a wizard . . 103 Start error trace files . . . . . . . . . . 125
Correcting a user ID or password . . . . . 103 Information Catalog Center issues . . . . . . 126
Resolving routines that fail to build on another Finding help for information catalog center
connection . . . . . . . . . . . . . 104 issues . . . . . . . . . . . . . . . 126
Supporting null input values to Java stored Recovering Information Catalog Center
procedures . . . . . . . . . . . . . 104 components and data . . . . . . . . . . 126
Testing null input values to SQL user-defined DB2 Cube Views issues . . . . . . . . . . 126
functions . . . . . . . . . . . . . . 104 Database connection behavior using the OLAP
Diagnosing SQL routines that fail to build . . . 105 Center . . . . . . . . . . . . . . . 126
Correcting color schemes that make text DB2 Cube Views backward compatibility with
unreadable . . . . . . . . . . . . . 105 the OLAP Center and metadata API . . . . . 127
Diagnosing SQL routine build failures with the Improving queries issued to federated data
DSNTPSMP build utility . . . . . . . . 106 sources . . . . . . . . . . . . . . 127
Diagnosing Java routine build failures with the Connectivity issues . . . . . . . . . . . 128
DSNTJSPP build utility . . . . . . . . . 106 Problem determination process for connectivity 129
Diagnosing routine build failures on DB2 UDB Determining if communication paths are
for OS/390 or DB2 UDB for z/OS servers, error operational . . . . . . . . . . . . . 129
code -471 . . . . . . . . . . . . . . 107 Verifying the initial connection . . . . . . 130
Setting display for the pop-up information for Determining intermittent and persistent
fields and controls . . . . . . . . . . . 107 problems . . . . . . . . . . . . . . 130
Diagnosing system connection exceptions when Determining the location of the communication
trying to establish multiple JDBC connections breakdown . . . . . . . . . . . . . 131
(AIX) . . . . . . . . . . . . . . . 108 Testing connectivity using the PCT tool -
Maintaining correct formatting for stored example . . . . . . . . . . . . . . 132
procedure source code (DB2 for iSeries) . . . 109 Configuring PCT on the server . . . . . 132
Backup and recovery issues . . . . . . . . 109 Starting the PCT listener on the server . . . 133
Verifying backup images using the db2ckbkp Configuring PCT on the client . . . . . . 134
command. . . . . . . . . . . . . . 110 Starting a PCT communication test on the
History file analysis . . . . . . . . . . 110 client . . . . . . . . . . . . . . 134
DB2 Data Links Manager issues . . . . . . . 111 Creating a loopback connection - example . . . 136
1 Setting up a backup strategy for DB2 Data Links Confirming TCP/IP configuration for a DB2
1 Manager . . . . . . . . . . . . . . 111 instance . . . . . . . . . . . . . . 138
File system backup and restore utilities . . . . 112 Verifying the TCP/IP setup on the server . . 138
1 Obtaining Data Links Manager environment Starting DB2 communication listeners on the
1 information using the dump utility server . . . . . . . . . . . . . . 139
1 (dlfm_dump) . . . . . . . . . . . . 113 Verifying the TCP/IP setup on the client . . 140
1 DB2 Data Links logging manager . . . . . . 114 Confirming NetBIOS configuration for a DB2
DB2 Data Links server failure . . . . . . . 114 instance . . . . . . . . . . . . . . 141
DB2 Data Links failure and recovery overview 115 Verifying the NetBIOS setup on the server 141
| Determining the file system directories that are Starting DB2 communication listeners on the
| needed to restore to the current point in time . . 116 server: . . . . . . . . . . . . . . 142
Reconciling database tables with DB2 Data Verifying the NetBIOS setup on the client 142
Links file date using the db2_recon_aid utility . 117 Testing a database connection using the
DB2 Data Links Manager recovery scenarios . . 119 Configuration Assistant . . . . . . . . . 143
Business intelligence issues . . . . . . . . . 122 Common connectivity problems . . . . . . 143
DB2 Data Warehouse Center logging . . . . 122 Initial connection mainframe or midrange
Viewing DB2 Data Warehouse Center runtime server is not successful . . . . . . . . 143
errors . . . . . . . . . . . . . . . 123 Problems encountered after an initial
Viewing DB2 Data Warehouse Center build-time connection . . . . . . . . . . . . 143
errors . . . . . . . . . . . . . . . 123 Common DB2 Connect problems . . . . . 144
Viewing log entries in the Data Warehouse Database crashes . . . . . . . . . . . . 144
Center . . . . . . . . . . . . . . . 123 db2dart and INSPECT–Database inspection tools 145
DB2 Data Warehouse Center component trace Bringing an OFFLINE table space ONLINE . . 151
data . . . . . . . . . . . . . . . 123 Data movement issues . . . . . . . . . . 151
Running warehouse agents as a user process Loading data . . . . . . . . . . . . 151
(Windows) . . . . . . . . . . . . . 123 Message files from IMPORT, EXPORT, and
Running a Data Warehouse Center component LOAD . . . . . . . . . . . . . . . 153
trace . . . . . . . . . . . . . . . 124 Extender issues . . . . . . . . . . . . . 153
iv DB2 Troubleshooting Guide

DB2 Spatial and Geodetic Extenders . . . . . 153 Problem determination particular to multi-partition
How to interpret DB2 Spatial Extender environments . . . . . . . . . . . . . 192
messages . . . . . . . . . . . . . 153 Common problems in multi-partition
DB2 Spatial Extender stored procedure environments . . . . . . . . . . . . 193
output parameters . . . . . . . . . . 154 Partitioned instance set-up . . . . . . . 193
DB2 Spatial Extender function messages . . 156 Log full situation during redistribution 195
DB2 Spatial Extender CLP messages . . . . 158 Dropping a database partition . . . . . 196
DB2 Control Center messages . . . . . . 159 SQL1035 (database in use) when taking
Tracing DB2 Spatial Extender problems with offline backups in parallel . . . . . . 196
the db2trc command . . . . . . . . . 160
DB2 XML Extender . . . . . . . . . . 161 Chapter 3. Searching knowledge
Starting the trace for XML Extender . . . . 161 bases . . . . . . . . . . . . . . . 199
Stopping the trace . . . . . . . . . . 162
DB2 troubleshooting resources. . . . . . . . 199
XML Extender UDF return codes . . . . . 163
XML Extender stored procedure return codes 163
| SQLSTATE codes and associated message Chapter 4. Getting fixes . . . . . . . 201
| numbers for XML Extender . . . . . . . 163 Release Notes . . . . . . . . . . . . . 201
XML Extender messages . . . . . . . . 168 How to effectively search for known problems . . 201
Locking issues . . . . . . . . . . . . . 182 Authorized Program Analysis Reports (APARs)
Tracking deadlocks . . . . . . . . . . 183 and fix packs . . . . . . . . . . . . . 202
Tracking lock waits and timeouts . . . . . . 183
Analyzing a lock snapshot . . . . . . . . 185 Notices . . . . . . . . . . . . . . 205
Identifying the owner of a lock via db2pd . . . 187 Trademarks . . . . . . . . . . . . . . 207
Performance issues . . . . . . . . . . . 187
Data To Collect For Optimizer Problems . . . 190 Index . . . . . . . . . . . . . . . 209
Analyzing access plans . . . . . . . . . 190
Tuning practices . . . . . . . . . . . 192
Contacting IBM . . . . . . . . . . 211
Sort overflows . . . . . . . . . . . 192
Product information . . . . . . . . . . . 211
DB2 Connect performance troubleshooting . . 192
Contents v
vi DB2 Troubleshooting Guide
Chapter 1. Troubleshooting overview
7 Introduction to problem determination
7 The first step in good problem analysis is to describe the problem completely.
7 Without a problem description, you will not know where to start investigating the
7 cause of the problem. This step includes asking yourself such basic questions as:
7 v What are the symptoms?
7 v Where is the problem happening?
7 v When does the problem happen?
7 v Under which conditions does the problem happen?
7 v Is the problem reproducible?
7 Answering these and other questions will lead to a good description to most
7 problems, and is the best way to start down the path of problem resolution.
7 What is the problem?:
7 When starting to describe a problem, the most obvious question is ″What is the
7 problem?″ This may seem like a straightforward question; however it can be
7 broken down into several other questions to create a more descriptive picture of
7 the problem. These questions can include:
7 v Who or what is reporting the problem?
7 v What are the error codes and error messages?
7 v What are the symptoms?
7 v How does it fail? For example: loop, hang, crash, performance degradation,
7 incorrect result.
7 v Can the problem be reproduced?
7 v What is the business impact?
7 The following problem scenario demonstrates how and when you might ask these
7 questions.
7 What is the environment and configuration?:
7 Determining where the problem originates is not always easy, but is one of the
7 most important steps in resolving a problem. Many layers of technology almost
7 always exist between the reporting and failing components. Networks, disks, and
7 drivers are only a few components to be considered when you are investigating
7 problems.
7 v Is the problem platform specific, or common to multiple platforms?
7 v Is the current enviroment and configuration supported?
7 v Is the application running local on the database server or on a remote server?
7 v Is there a gateway involved?
7 v Does the database reside on individual disks, or on a RAID array?
© Copyright IBM Corp. 1993, 2000 1

7 These types of questions will help you isolate the problem layer, and are necessary
7 to determine the problem source. Remember that just because one layer is
7 reporting a problem, it does not always mean the root cause exists there.
7 Part of identifying where a problem is occurring is understanding the environment

7 in which it exists. You should always take some time to completely describe the
7 problem environment, including the operating system, its version, all
7 corresponding software and versions, and hardware information. Confirm you are
7 running within an environment that is a supported configuration, as many
7 problems can be explained by discovering software levels that are not meant to run
7 together, or have not been fully tested together.
7 When does the problem happen?:
7 Developing a detailed time line of events leading up to a failure is another

7 necessary step in problem analysis, especially for those cases that are one-time
7 occurrences. You can most easily do this by working backwards --start at the time
7 an error was reported (as exact as possible, even down to milliseconds), and work
7 backwards through available logs and information. Usually you only have to look
7 as far as the first suspicious event that you find in any diagnostic log, however this
7 is not always easy to do and will only come with practice. Knowing when to stop
7 is especially difficult when there are multiple layers of technology each with its
7 own diagnostic information.
7 v Does the problem only happen at a certain time of day or night?
7 v How often does this happen?
7 v What sequence of events lead up to the time the problem is reported?
7 v Does the problem happen after an environment change such as upgrading
7 existing or installing new software or hardware?
7 Responding to questions like this will help you create a detailed time line of
7 events, and will provide you with a frame of reference in which to investigate.
7 Under which conditions does the problem happen?:
7 Knowing what else is running at the time of a problem is important for any
7 complete problem description. If a problem occurs in a certain environment or
7 under certain conditions, that can be a key indicator of the problem cause.
7 v Does the problem always occur when performing the same task?
7 v Does a certain sequence of events need to occur for the problem to surface?
7 v Do other applications fail at the same time?
7 Answering these types of questions will help you explain the environment in
7 which the problem occurs, and correlate any dependencies. Remember that just
7 because multiple problems may have occurred around the same time, it does not
7 necessarily mean that they are always related.
7 Is the problem reproducible?:
7 From a problem description and investigation standpoint, the ″ideal″ problem is

7 one that is reproducible. With reproducible problems you almost always have a
7 larger set of tools or procedures available to use to help your investigation.
7 Consequently reproducible problems are usually easier to debug and solve.
2 DB2 Troubleshooting Guide

7 However, reproducible problems can have a disadvantage: if the problem is of
7 significant business impact, you don’t want it recurring. If possible, recreating the
7 problem in a test or development environment is often preferable in this case.
7 v Can the problem be recreated on a test machine?
7 v Are multiple users or applications encountering the same type of problem?
7 v Can the problem be recreated by running a single command, a set of commands,
7 or a particular application, or a standalone application?
7 v Can the problem be recreated by entering the equivalent command/query from
7 a DB2 command line?
7 Recreating a single incident problem in a test or development environment is often

7 preferable, as there is usually much more flexibility and control when
7 investigating.
7 Our example is reproducible.
7 Conclusion:
7 Describing a problem accurately and completely may be easy for some problems,
7 but very difficult for others. (The difficulty usually increases as the environmental
7 complexity increases). However, the questions that you need to ask are usually the
7 same: who, what, where, when, and how.
7 The db2support tool poses subsets of the questions described above when you
7 specify the ″-q″ option. This information is then used by DB2 Support to
7 understand the context in which you are encountering your problem. For example:
7 db2support . -q
7 Once have a good understanding of the problem situation, you can research
7 similar problems, and learn about workarounds, and fixes in the available DB2
7 troubleshooting resources.
7 Related concepts:
7 v on page 0
Identifying the version and service level of your product

The db2level command will help you determine the version and service level
(build level and FixPak number) of your DB2 instance. To determine if your DB2
instance is at the latest service level, compare your db2level output to the build
level and FixPak information on the FixPak download pages at the DB2 UDB
support web site.
A typical result of running the db2level command on a Windows system would be:
C:\>db2level
DB21085I Instance "DB2" uses "32" bits and DB2 code release "SQL08020"
with level identifier "03010106".
Informational tokens are "DB2 v8.1.7.664", "s040914", "WR21342", and FixPak "7".
Product is installed at "D:\IBM\SQLLIB".
The combination of the four informational tokens uniquely identify the precise
service level of your DB2 instance. This information is essntial when contacting
Chapter 1. Troubleshooting overview 3

IBM support for assistance. In particular, the second informational token shows the
date that the product was built. In the above example, ″s040914″ indicates a build
date of September 14th, 2004.
For JDBC or SQLJ application, if you are using the Universal Driver for SQLJ and
JDBC, you can determine the level of the driver by running the db2jcc utility:
db2jcc -version
IBM DB2 JDBC Universal Driver Architecture 2.3.63
On Windows, this command is in the sqllib\bin directory.
On UNIX/Linux, this command is in the sqllib/bin directory.
DB2 server process model

Knowledge of the DB2® process model can help you determine the nature of a
problem because it helps you to understand how the database manager and its
associated components interact.
The process model used by all DB2 servers facilitates the communication that
occurs between database servers and clients and local applications. It also ensures
that database applications are isolated from resources such as database control
blocks and critical database files.
UNIX-based environments use an architecture based on system processes. For

example, the DB2 communications listeners are created as system processes. Intel
operating systems such as Windows® use an architecture based on threads to
maximize performance. Unless explicitly noted, this discussion uses the term
″process″ to refer to both processes and threads. You can find details of the
differences between the use of Windows threads and UNIX® processes later in this
topic.
For each database being accessed, various processes are started to deal with the
various database tasks such as prefetching, communication, and logging.
Each process of a client application has a single coordinator agent that operates on
a database. A coordinator agent works on behalf of an application, and
communicates to other agents, using interprocess communication (IPC) or remote
communication protocols.
DB2 architecture provides a firewall so that applications run in a different address

space from DB2. The firewall protects the database and the database manager from
applications, stored procedures, and user-defined functions (UDFs). A firewall
maintains the integrity of the data in the databases, because it disables application
programming errors from overwriting internal buffers or files of the database
manager. The firewall also improves reliability, because application errors cannot
crash the database manager.

Remote
Client
Program
Local
Client
Program
db2fmp
F I R EW A L L
Per Instance Other

Remote db2ipccm db2sysc Threads/ Per Instance
Listeners Processes
db2agent db2agent Agent

Per Connection Per Connection
Pool
db2agntp db2agntp db2agntp db2agntp
Per Database db2pfchr db2pclnr db2loggr db2loggw db2logts db2dlock Per Database
Figure 1. Process Model for DB2 Systems
The following list provides additional details on the processes shown in the figure:
Client Programs:
Client programs run remotely or on the same machine as the database server. They
make their first contact with the database through a listener. A coordinator agent
(db2agent) is then assigned to them.
Listeners:
Client programs make initial contact with communication listeners, which are
started when DB2 is started. There is a listener for each configured communication
protocol, and an interprocess communications (IPC) listener (db2ipccm) for local
client programs. Listeners include:
v db2ipccm, for local client connections
v db2tcpcm, for TCP/IP connections
v db2snacm, for APPC connections
v db2tcpdm, for TCP/IP discovery tool requests
Agents:
All connection requests from client applications, whether they are local or remote,
are allocated a corresponding coordinator agent (db2agent). When the coordinator
agent is created, it performs all database requests on behalf of the application.

In some environments, in which the intra_parallel database manager configuration
parameter is enabled, the coordinator agent distributes the database requests to
subagents (db2agntp). These agents perform the requests for the application. Once
the coordinator agent is created, it handles all database requests on behalf of its
application by coordinating subagents (db2agntp) that perform requests on the
database.
A coordinator agent may be:

v Connected to the database with an alias. For example, ″db2agent (DATA1)″ is
connected to the database alias ″DATA1″.
v Attached to an instance. For example, ″db2agent (user1)″ is attached to the
instance ″user1″.
An additional type of agent, db2agnsc is created by DB2 to perform certain

operations in parallel. In particular, they are used in database recovery actions.
Idle agents reside in an agent pool. These agents are available for requests from
coordinator agents operating on behalf of client programs, or from subagents
operating on behalf of existing coordinator agents. The number of available agents
is dependent on the database manager configuration parameters maxagents and
num_poolagents.
db2fmp:
The fenced mode process. It is responsible for executing fenced stored procedures
and user-defined functions outside the firewall. db2fmp is always a separate
process but may be multithreaded depending on the types of routines it executes.
Database Threads/Processes:
The following list includes some of the important threads/processes used by each
database:
v db2pfchr, for buffer pool prefetchers
v db2pclnr, for buffer pool page cleaners
v db2loggr, for manipulating log files to handle transaction processing and
recovery
v db2loggw, for writing log records to the log files.
v db2logts, for collecting historical information about which logs are active when
a tablespace is modified. This information is ultimately recorded in the
DB2TSCHG.HIS file in the database directory. It is used to speed up tablespace
rollforward recovery.
v db2dlock, for deadlock detection. In a multi-partitioned database environment,
an additional process called db2glock is used to coordinate the information
gathered from the db2dlock process on each partition. db2glock runs only on the
catalog partition.
Database Server Threads and Processes:
The system controller (db2sysc) must exist in order for the database server to
function. Also, the following threads and processes may be started to carry out
various tasks:
v db2resyn, the resync agent that scans the global resync list

v db2gds, the global daemon spawner on UNIX-based systems that starts new
processes
v db2wdog, the watchdog on UNIX-based systems that handles abnormal
terminations
v db2fcmdm, the fast communications manager daemon for handling
inter-partition communication (used only in multi-partitioned databases)
v db2pdbc, the parallel system controller, which handles parallel requests from
remote nodes (used only in a partitioned database environment).
v db2cart, for archiving log files when accessing a database configured with
USEREXIT enabled
v db2fmtlg, for formatting log files, when accessing a database configured with
LOGRETAIN enabled, but with USEREXIT disabled
v db2panic, the panic agent, which handles urgent requests after agent limits have
been reached at a particular node (used only in a partitioned database
environment)
v dlasync, a monitor for the Data Links servers, if you have configured DB2 for
datalinks
Differences between Windows and UNIX:
DB2 for Windows differs from UNIX-based environments in that the database
engine is multi-threaded, not multi-processed. In Windows systems, each of the
dispatchable units on the agent side of the firewall is a thread under the process
db2sysc. This allows the database engine to let the operating system perform
task-switching at the thread level instead of the process level. For each database
being accessed, threads are started to deal with database tasks (for example,
prefetching).
Another difference is in the handling of abnormal terminations. There is no need

for a ″watchdog″ process in Windows systems, because these systems ensure that
the allocated resources are cleaned up after an abnormal termination. Thus, there is
no equivalent of the db2wdog process on the Windows systems. In addition, the
db2gds process or thread is not needed on the Windows systems, which have their
own mechanisms for starting threads.
Related concepts:
v on page 0
v “Determining active process status” on page 41
Proactive monitoring tools

Database monitoring is an important part of tuning a database server. You should
develop a database monitoring plan. This not only allows for easier detection of a
problem, but also enables you to refer back to historical data to isolate a change,
extrapolate the effects of a change, better understand user and application activity
and behaviors, and, of course, improve database performance!
Monitoring does involve overhead and this should be taken into account when
developing the monitoring plan. The frequency and volume of monitoring needs to
be controlled.
Techniques for monitoring the database:

DB2 has two main tools that allow you to access system monitor information, each
serving a particular purpose:
v Snapshot monitoring - captures a picture of the database status at a particular
point in time (the moment the snapshot is taken). Snapshot monitoring is useful
in determining the current state of the database and its applications.
v Event monitoring - provides information about specific database events which
occur over a period of time. Event monitoring is useful in identifying database
or application state changes. It is used to gather details about changes regarding
different events such as connections, tables, statements, etc.
In order to allow for effective memory usage, it is possible to configure how much
memory is used for database monitoring. The database system monitor heap size
cofiguration parameter MON_HEAP_SZ, which is a database manager
configuration parameter, specifies the amount of memory allocated on a
DB2START and is used when monitoring takes place. Setting this parameter to 0
(zero) will not allow any monitoring to take place.
Note: You must have one of SYSADM, SYSCTRL, SYSMAINT, or SYSMON

authority to perform much of the set-up related to monitoring.
Levels of monitoring:
The monitoring utilities collect information at different levels within the system:
v Database Manager - the monitor gathers details from the time the instance is
started till the time it is stopped.
v Database - this level of monitoring begins on the first connection or activation,
and will continue until deactivation or the last connection is terminated.
v Application - this level of monitoring begins when an application connects to the
database and continues until it disconnects from the database.
You can control the type of data to collect. The type of information is classified as
follows:
BUFFERPOOL ACTIVITY
LOCKS
SORTS
SQL STATEMENTS
TABLE ACTIVITY
TIMESTAMPS
UNIT OF WORK
So, for example, if you notice long lock waits, it would be logical to monitor the
LOCK group in an attempt to isolate the problem.
Snapshot and event monitors

For both snapshot and event monitors you have the option of storing monitor
information in files or SQL tables, viewing it on screen (directing it to
standard-out), or processing it with a client application. What follows are brief
overviews to get you started with snapshot monitors and event monitors. For more
information, refer to the System Monitor Guide and Reference, or the ″Monitoring″
section of the Information Center.

Setting up snapshot monitoring
The DB2® monitor maintains a running tally of valuable system information. You
can get a summary of system status at any time by issuing the GET SNAPSHOT
command. The parameters of the GET SNAPSHOT command determine the level
of detail, and the type of monitoring information returned.
Some of the statistics returned by this command provide point- in-time

information, others provide cumulative information since the last reset monitor
command was performed. An example of a cumulative counter would be
″deadlocks detected″. This holds the number of deadlocks since the last reset
monitor command. An example of an event-specific field would be ″lock list
memory in use″. This field would contain the number of bytes currently used for
the lock list and can change from one iteration of a snapshot to the next, as
applications connect and disconnect and obtain and release locks respectively.
As an example, try using the sample database to take a snapshot. From a CLP
session, issue the following commands:
DB2 CONNECT TO sample
DB2 RESET MONITOR ALL
DB2 GET SNAPSHOT FOR DATABASE ON sample
DB2 "SELECT * FROM syscat.bufferpools"
DB2 GET SNAPSHOT FOR DATABASE ON sample
Creating an event monitor

The Event Monitor facility within DB2 also allows for a great deal of flexibility. The
initial step is to first create an event monitor.
As an example, try creating an event monitor for the SAMPLE database:

DB2 CREATE EVENT MONITOR evmonex1 FOR STATEMENTS WRITE TO FILE
’c:\temp\testfile’ MAXFILES 3 MAXFILESIZE 1024 NONBLOCKED APPEND
The above command will create an event monitor called evmonex1 to capture all
statement events. A maximum of 3 files of 4MB each will be created. The file
directory (in this example ″c:\temp\testfile″) does not have to exist at CREATE
EVENT MONITOR time. However, if it does not occur by the time an event
monitor is activated, an error will occur (i.e. SQL1614N An I/O error occurred
when activating an event monitor. Reason code = ″2″. SQLSTATE=58030).
Activating an event monitor

To enable the newly created event monitor to collect information, you must
activate it. For example:
DB2 SET EVENT MONITOR evmonex1 STATE 1
To see if an event monitor is active or inactive, issue the SQL function

EVENT_MON_STATE in a query against the table,SYSCAT.EVENTMONITORS, for
example:
DB2 SELECT evmonname, EVENT_MON_STATE(evmonname) FROM syscat.eventmonitors
A returned value of 0 indicates that the specified event monitor is inactive, and 1
indicates that it is active.
For more example commands, including how to disable and drop event monitors,
see Collecting information about database system events.

Formatting an event monitor
The output from file or pipe event monitors (as opposed to table event monitors)
are written out in an unreadable format. In order to make the output readable, you
need to run a tool called db2evmon:
db2evmon -db <database alias> -evm <event monitor name>
As an exercise, try generating and then formatting the output of the evmonex1
event monitor created earlier.
db2evmon -db sample -evm evmonex1
Try running that command again after generating a statement event, for example
by issuing
’DB2 "SELECT * FROM EMPLOYEE"’
. There will now be much more information, since an event has been triggered.
Basic monitoring commands

The following commands track current monitor settings, enabling functional
groups, and ″clear out″ existing monitor data. These commands can be executed
from a DB2 CLP session:
To obtain the current state of the functional groups listed earlier, issue the
following command:
DB2 GET MONITOR SWITCHES
If you see that a particular group’s ″switch″ is ″on″, you know that this group will
be monitored.
Note: Event monitors are only affected by the time and timestamp information
switch. All other switch settings effect snapshot monitors only and have no
effect on the data collected by event monitors.
To enable or disable the monitor switches, use the command:

DB2 UPDATE MONITOR SWITCHES USING <switch name> ON | OFF
For partitioned database systems, you can set monitor switches specifically for a
certain partition, or globally for all partitions. To set a monitor switch for a specific
partition, issue the following command:
db2 update monitor switches using <switch name> ON|OFF at dbpartitionnum <partition number>
To set a monitor switch for all partitions, issue the command:

db2 update monitor switches using <switch name> on global
To reset monitor elements to zero, thereby allowing you to start with fresh system
monitor information, use the following command (which can also be qualified to
occur only on particular partitions):
DB2 RESET MONITOR ALL | FOR DATABASE <database alias>
Related reference:
v Snapshot monitor
v Snapshot monitor CLP commands
v Event monitors
v Creating an event monitor

v Formatting file or pipe event monitor output from a command line
v Event monitor sample output
Configuring health indicators for health monitors

The health monitor checks the state of your system using health indicators to
determine if an alert should be issued. Preconfigured actions can be taken in
response to alerts. The health monitor can also log alerts in the administration
notification log and send notifications by e-mail or pager. For information on how
to set up health indicators via the Health Center, see Configuring health indicators
using Health Center
The following example provides an overview of the steps necessary to configure a

health indicator and then monitor for health alerts on that indicator from the CLP.
The scenario is that you have decided that you want to be alerted when a
particular database is in a non-normal state. This will involve monitoring the
health indicator ″db.db_op_status″, the database operational state health indicator.
To determine the current settings for this health indicator on the SAMPLE
database, use the following command:
DB2 GET ALERT CFG FOR DATABASE ON SAMPLE
The following is the default output resulting from this request for configuration
information:
Alert Configuration
Indicator Name = db.db_op_status

Default = Yes
Type = State-based
Sensitivity = 0
Formula = db.db_status;
Actions = Disabled
Threshold or State checking = Enabled
...
Thus the state checking is enabled for this health indicator. If you wanted to alter
any of these fields, you could do so using the command UPDATE ALERT CFG, for
example:
DB2 UPDATE ALERT CFG FOR DATABASE ON SAMPLE using db.db_op_status SET THRESHOLDSCHECKED NO
To return to the default values, use the RESET ALERT CFG command, for example:
DB2 RESET ALERT CFG FOR DATABASE ON SAMPLE USING db.db_op_status
Note: The table space has to be in the non-normal state when the health indicator
evaluates in order to generate the alert. So, if the state doesn’t persist long
enough, then it’s possible that an alert will not be generated.
In DB2 version 8.2 FixPak 7, the refresh interval for this particular health indicator
is 5 minutes. Therefore if you want to test this alert situation, perform an action
that will put the database in a non-normal state (for example by quiescing it) and
then wait 5 minutes. Unfortunately, there is no method for you to determine the
refresh intervals. Most of them range between 5 and 60 minutes.
Once that time has passed, obtain a health monitor snapshot:

DB2 QUIESCE DATABASE IMMEDIATE
<Wait **>
DB2 GET HEALTH SNAPSHOT FOR DATABASE ON sample
Database Health Snapshot

Snapshot timestamp = 01/14/2005 15:02:03.96836
Database name = SAMPLE
Database path = D:\DB2\NODE0000\SQL00002\
Input database alias = SAMPLE
Operating system running at database server = NT
Location of the database = Local
Database highest severity alert state = Attention
Health Indicators:
Value = 2
Evaluation timestamp = 01/14/2005 15:00:27.296000
Alert state = Attention
If you were unaware that the database was quiesced, you could use the GET
RECOMMENDATIONS command in order to obtain the recommended actions for
this particular alert situation:
DB2 GET RECOMMENDATIONS FOR HEALTH INDICATOR db.db_op_status FOR DATABASE ON sample
Problem:
Value = 2
Evaluation timestamp = 01/14/2005 15:05:27
Alert state = Attention
Partition = 0
Additional information =
Recommendations:
Recommendation: Unquiesce the database.

Rank: 1
...
If you have quiesced your database to perform this test, unquiesce it now using
any one of the methods described in the GET RECOMMENDATIONS output. For
example:
DB2 UNQUIESCE DATABASE
DB2 CONNECT RESET
Related reference:
v Introduction to the health monitor
v Health monitor interfaces
v Health indicator configuration
v Health monitor sample output
Health monitor output

The two pieces of information that you will need in order to understand health
monitor output are as follows:
1. The meaning of the health indicator. See Health indicators summary for
definitions of all of the indicators.
2. The meaning of the attribute values displayed for each health indicator. Some
of the attributes (evaluation flag, thresholds, sensitivity) define when the health

db2diag - db2diag.log analysis tool
monitor will generate an alert for the health indicator. Other attributes (action
flag, actions) define what the health monitor does upon generating the alert.
For more information about these attributes, see Health indicator configuration.
Two examples of health snapshot output and their meaning are provided here:
Health monitor sample output.
Interpreting event monitor and snapshot output

To begin interpreting system monitor output, you must first know how to find the
meaning of the various monitor elements. For definitions of monitor elements that
can be referred by snapshot monitoring, refer to Snapshot monitor logical data
groups and monitor elements. For definitions of monitor elements that can be
logged by event monitors, refer to Event monitor logical data groups and monitor
elements.
Diagnosing the source of the problem
Diagnostic tools
7 db2diag - Analyzing db2diag.log files
7 In DB2 version 8, the primary log file intended for use by database and system
7 administrators is the Administration Notification log. The db2diag.log file, on the
7 other hand, is intended for use by DB2 customer support for troubleshooting
7 purposes.
7 The db2diag tool serves to filter and format the volume of information available in
7 the db2diag.log.
7 Example 1: Filtering the db2diag.log by database name:
7 If there are several databases in the instance, and you wish to only see those
7 messages which pertain to the database ″SAMPLE″, you can filter the db2diag.log
7 as follows:
7 db2diag -g db=SAMPLE
7 Thus you would only see db2diag.log records which contained ″DB: SAMPLE″,
7 such as:
7 2005-01-14-15.16.52.984000-480 E105657825H410 LEVEL: Error
7 PID : 2200 TID : 2148 PROC : db2syscs.exe
7 INSTANCE: DB2 NODE : 000 DB : SAMPLE
7 APPHDL : 0-254 APPID: *LOCAL.DB2.050114231611
7 FUNCTION: DB2 UDB, base sys utilities, sqleDatabaseUnquiesce, probe:2
7 MESSAGE : ADM7509W Database unquiesce request has completed successfully.
7 Example 2: Filtering the db2diag.log by process id:
7 The following command can be used to display all severe error messages produced
7 by processes running on partitions 0,1,2, or 3 with the process ID (PID) 2200:
7 db2diag -g level=Severe,pid=2200 -n 0,1,2,3
7 Note that this command could have been written a couple of different ways,
7 including ″db2diag -l severe -pid 2200 -n 0,1,2,3″. These commands would
7 successfully retrieve db2diag.log records which meet these requirements, such as:

db2diag - db2diag.log analysis tool
7 2005-01-14-12.18.51.515000-480 I105644317H448 LEVEL: Severe

7 PID : 2200 TID : 2108 PROC : db2syscs.exe
7 INSTANCE: DB2 NODE : 000 DB : SAMPLE
7 APPHDL : 0-47 APPID: *LOCAL.DB2.050114195055
7 FUNCTION: DB2 UDB, catalog services, sqlrldre, probe:60
7 RETCODE : ZRC=0x800D002D=-2146631635=SQLM_RC_EVACT "drop active mon"
7 DIA8053C Event monitor cannot be deactivated.
7 Example 3: Formatting the db2diag tool output.
7 The following command filters all records occurring after January 1, 2005
7 containing non-severe and severe errors logged on partitions 0,1 or 2. It outputs
7 the matched records such that the time stamp, partition number and level appear
7 on the first line, pid, tid and instance name on the second line, and the error
7 message follows thereafter:
7 db2diag -time 2005-01-01 -node "0,1,2" -level "Severe, Error"
7 Partition: %node Message Level: %{level} \nPid: %{pid} Tid: %{tid} Instance: %{instance}\nMessage: @
7 An example of the output produced is as follows:

7 Time: 2005-01-05-14.59.58.984000 Partition: 000 Message Level: Severe
7 Pid: 2616 Tid: 2452 Instance: DB2
7 Message: Backup Terminated.
7 For more information, issue the following commands:

7 v db2diag -help provides a short description of the options
7 v db2diag -h brief provides descriptions for all options without examples
7 v db2diag -h notes provides usage notes and restrictions
7 v db2diag -h examples provides a small set of examples to get started
7 v db2diag -h tutorial provides examples for all available options
7 v db2diag -h all provides the most complete list of options
7 Related reference:
7 v db2diag - db2diag.log analysis tool Command
7 v Administration logs, error logs and first failure data capture
db2greg - Displaying and altering the Global Registry (UNIX)

The Global Registry lives in /var/db2/global.reg (/var/opt/db2/global.reg on
HP-UX) and only exists on UNIX and Linux platforms. It consists of three different
record types:
v ″Service″: Service records contain information at the product level - i.e. version,
install path, etc.
v ″Instance″: Instance records contain information at the instance level - i.e.
Instance name, instance path, version, the ″start-at-boot″ flag, etc.
v ″Variable″: Variable records contain information at the variable level - i.e.
Variable name, Variable value, and Comment
One can view (and manipulate with root access) the Global Registry with the
″db2greg″ tool. This tool is located in sqllib/bin, and in the install directory under
bin as well (for use when logged in as root). You should only use this tool if
requested to do so by DB2 Customer Support.
db2look - Mimicking databases

There are many times when it is advantageous to be able to create a database that
is similar in structure to another database. For example, rather than testing out
new applications or recovery plans on a production system, it makes more sense to

db2look - DB2 Statistics and DDL Extraction Tool
create a test system that is similar in structure and data, and to then do the tests
against it instead. This way, the production system will not be affected by the
adverse performance impact of the tests or by the accidental destruction of data by
an errant application. Also, when you are investigating a problem (such as invalid
results, performance issues, and so on), it may be easier to debug the problem on a
test system that is identical to the production system.
You can use the db2look tool to extract the required DDL statements needed to
reproduce the database objects of one database in another database. The tool can
also generate the required SQL statements needed to replicate the statistics from
the one database to the other, as well as the statements needed to replicate the
database configuration, database manager configuration, and registry variables.
This is important because the new database may not contain the exact same set of
data as the original database but you may still want the same access plans chosen
for the two systems.
The db2look tool is described in detail in the DB2 Command Reference but you can
view the list of options by executing the tool without any parameters. A more
detailed usage can be displayed using the -h option.
Using db2look to mimic the tables in a database:
To extract the DDL for the tables in the database, use the -e option. For example,
create a copy of the SAMPLE database called SAMPLE2 such that all of the objects
in the first database are created in the new database:
C:\>db2 create database sample2
DB20000I The CREATE DATABASE command completed successfully.
C:\>db2look -d sample -e > sample.ddl
-- USER is:
-- Creating DDL for table(s)
-- Binding package automatically ...
-- Bind is successful
Note: If you want the DDL for the user-defined spaces, database partition groups
and buffer pools to be produced as well, add the ″-l″ flag after ″-e″ in the
command above. The default node groups, buffer pools, and table spaces
will not be extracted. This is because they already exist in every database by
default. If you wish to mimic these, you must alter them yourself manually.
Bring up the file sample.ddl in a text editor. Because you want to execute the DDL
in this file against the new database, you must change the CONNECT TO
SAMPLE statement to CONNECT TO SAMPLE2. If you used the ″-l″ option, you
may need to alter the path associated with the tablespace commands, such that
they point to appropriate paths as well. While you are at it, take a look at the rest
of the contents of the file. You should see CREATE TABLE, ALTER TABLE, and
CREATE INDEX statements for all of the user tables in the sample database:
...
------------------------------------------------
-- DDL Statements for table "DB2"."ORG"
------------------------------------------------
CREATE TABLE "DB2"."ORG" (

"DEPTNUMB" SMALLINT NOT NULL ,
"DEPTNAME" VARCHAR(14) ,
"MANAGER" SMALLINT ,

"DIVISION" VARCHAR(10) ,
"LOCATION" VARCHAR(13) )
IN "USERSPACE1" ;
...
Once you have changed the connect statement, execute the statements, as follows:
C:\>db2 -tvf sample.ddl > sample2.out
Take a look at the sample2.out output file -- everything should have been executed
successfully. If errors have occurred, the error messages should state what the
problem is. Fix those problems and execute the statements again.
As you can see in the output, DDL for all of the user tables are exported. This is
the default behavior but there are other options available to be more specific about
the tables included. For example, to only include the STAFF and ORG tables, use
the -t option:
C:\>db2look -d sample -e -t staff org > staff_org.ddl
To only include tables with the schema DB2, use the -z option:
C:\>db2look -d sample -e -z db2 > db2.ddl
Mimicking statistics for tables:
If the intent of the test database is to do performance testing or to debug a

performance problem, it is essential that access plans generated for both databases
are identical. The optimizer generates access plans based on statistics, configuration
parameters, registry variables, and environment variables. If these things are
identical between the two systems then it is very likely that the access plans will
be the same.
If both databases have the exact same data loaded into them and runstats is
performed on both, the statistics should be identical. However, if the databases
contain different data or if only a subset of data is being used in the test database
then the statistics will likely be very different. In such a case, you can use db2look
to gather the statistics from the production database and place them into the test
database. This is done by creating UPDATE statements against the SYSSTAT set of
updatable catalog tables as well as RUNSTATS commands against all of the tables.
The option for creating the statistic statements is -m. Going back to the
SAMPLE/SAMPLE2 example, gather the statistics from SAMPLE and add them
into SAMPLE2:
C:\>db2look -d sample -m > stats.dml
-- USER is:
-- Running db2look in mimic mode
As before, the output file must be edited such that the CONNECT TO SAMPLE
statement is changed to CONNECT TO SAMPLE2. Again, take a look at the rest of
the file to see what some of the runstats and update statements look like:
...
-- Mimic table ORG
RUNSTATS ON TABLE "DB2"."ORG" ;
UPDATE SYSSTAT.INDEXES
SET NLEAF=-1,
NLEVELS=-1,
FIRSTKEYCARD=-1,
FIRST2KEYCARD=-1,
FIRST3KEYCARD=-1,

FIRST4KEYCARD=-1,
FULLKEYCARD=-1,
CLUSTERFACTOR=-1,
CLUSTERRATIO=-1,
SEQUENTIAL_PAGES=-1,
PAGE_FETCH_PAIRS=’’,
DENSITY=-1,
AVERAGE_SEQUENCE_GAP=-1,
AVERAGE_SEQUENCE_FETCH_GAP=-1,
AVERAGE_SEQUENCE_PAGES=-1,
AVERAGE_SEQUENCE_FETCH_PAGES=-1,
AVERAGE_RANDOM_PAGES=-1,
AVERAGE_RANDOM_FETCH_PAGES=-1,
NUMRIDS=-1,
NUMRIDS_DELETED=-1,
NUM_EMPTY_LEAFS=-1
WHERE TABNAME = ’ORG’ AND TABSCHEMA = ’DB2 ’;
...
As with the -e option that extracts the DDL, the -t and -z options can be used to
specify a set of tables.
Extracting configuration parameters and environment variables:
The optimizer chooses plans based on statistics, configuration parameters, registry

variables, and environment variables. As with the statistics, db2look can be used to
generate the necessary configuration update and set statements. This is done using
the -f option. For example:
C:\>db2look -d sample -f > config.txt
-- USER is:
-- This CLP file was created using DB2LOOK Version 8.2
-- Timestamp: 2/16/2005 1:26:50 AM
-- Database Name: SAMPLE
-- Database Manager Version: DB2/NT Version 8.2.1
-- Database Codepage: 1252
-- Database Collating Sequence is: UNIQUE

CONNECT TO SAMPLE;
--------------------------------------------------------
-- Database and Database Manager configuration parameters
--------------------------------------------------------
UPDATE DBM CFG USING cpuspeed 3.739392e-007;

UPDATE DBM CFG USING intra_parallel NO;
UPDATE DBM CFG USING comm_bandwidth 1.000000;
UPDATE DBM CFG USING federated NO;
UPDATE DBM CFG USING fed_noauth NO;
UPDATE DB CFG FOR SAMPLE USING locklist 50;

UPDATE DB CFG FOR SAMPLE USING dft_degree 1;
UPDATE DB CFG FOR SAMPLE USING maxlocks 22;
UPDATE DB CFG FOR SAMPLE USING avg_appls 1;
UPDATE DB CFG FOR SAMPLE USING stmtheap 2048;
UPDATE DB CFG FOR SAMPLE USING dft_queryopt 5;
---------------------------------
-- Environment Variables settings
---------------------------------

COMMIT WORK;
CONNECT RESET;
TERMINATE;
Note: Only those parameters and variables that affect access plan generation will
be included.
Related reference:
v db2look - DB2 Statistics and DDL Extraction Tool Command
db2nstck and db2_call_stack - Generating EDU call stacks

The term EDU stands for ″engine dispatch unit″ and refers to a thread (Windows)
or a process (UNIX) that is doing work on behalf of DB2. The db2nstck (Windows)
and db2_call_stack (UNIX) commands are used to generate call stacks for the
EDUs that are running under an instance. A call stack, also known as a stack
traceback, shows the processing path that an EDU is currently in. The function that
the EDU is currently executing is at the top of the stack, the function that called
that one is below it, and so on.
These commands are mainly used when it appears that the DB2 engine has
become ″hung″. (In most cases, problems that appear to be with DB2 are in reality
caused by application problems. For example, a long-running transaction may be
holding locks that all other applications are waiting on. All of the applications will
appear to be hung but they are really just waiting for one or more locks to be
released. Situations like this can usually be identified using DB2’s snapshot and
event monitor tools.
Operating system commands (such as iostat and ps in UNIX) can be used to

determine if EDUs are actually doing any work. If it appears that none of the
EDUs are doing any processing, an engine hang may be a possibility. By
generating multiple call stacks for each EDU (with a sufficient amount of time in
between -- 1 to 2 minutes for example), you can compare the call stacks to see if
the EDUs are in a different processing path from one stack to the next. If all of the
EDUs are stuck in the same function that was shown in the previous call stack
then there may in fact be a problem with the DB2 engine. In that case, the call
stacks must be provided to IBM so that the hang can be analyzed.
In UNIX, the db2_call_stack command generates call stacks for both single and
multi-node instances. The call stacks are placed in files in the diagnostic directory
(sqllib/db2dump by default). Prior v8.2.2 each EDU has its own file which is called
tprocessID.nodenumber. For v8.2.2 Each EUD has its own file which is called
t<processID>.<nodenumber>.xml The files are textbased and can be viewed using
any editor.
Here is an example of a portion of a call stack on UNIX:

*** Start stack traceback ***
0xD024E2C8 select + 0xAC

0x205E0244 sqlorqueInternal__FP13sqlo_que_descP12SQLO_MSG_HDRil + 0x100
0x205E008C sqlorque2 + 0xAC
0x10004840 sqleRunSysCtlr__Fv + 0x10C
0x10002D44 sqleSysCtlr__Fv + 0x8C
0x205C6FFC sqloRunInstance + 0x75C

0x100028EC DB2main + 0x8A4

0x10003470 main + 0xC
*** End stack traceback ***
Note: Running db2_call_stack will temporarily affect the performance of the

server. It is often preferable to use a ″kill -36″ (signal 36 is for AIX, 21 for
sun, 29 for hp, 23 for linux.) command directly against the db2sysc process
on UNIX platforms. Running ps will show every db2 process as db2sysc. ps
-elf will show only one db2sysc per logical partition, which is the parent of
all db2 processes, and this is the process against which kill -36 should be
run. See Diagnostic tools for UNIX-based systems.
In Windows, the db2nstck command is used for single-node instances and

db2_call_stack is used for multi- node instances. All of the call stacks will be
placed into a single file in the diagnostic directory (sqllib\instance by default). The
db2nstck command tells you the name of the file name. For example:
C:\>db2nstck
The stack dump has been saved in the file C:\SQLLIB\DB2\P2292.000
This file is in a binary format. To convert it to readable text, you need the db2xprt
tool and the DB2 for Windows .pdb files (i.e. symbol files). See Formatting and
interpreting trap files for information about where to find this tool and how to
use it.
Related reference:
v Trap files
9 db2pd - Monitoring and troubleshooting DB2

9 The db2pd tool is a problem determination tool which contains quick and
7 immediate information from the DB2 memory sets. It is similar to the Informix
7 ″onstat″ tool, and runs very quickly as it is not acquiring any locks or latches and
7 runs outside of the DB2 engine resources.
7 The tool collects this information without acquiring any latches or using any
7 engine resources. It is therefore possible (and expected) to retrieve information that
7 is changing while db2pd is collecting information; hence the data may not be
7 completely accurate. If null pointers are encountered, a signal handler is used to
7 prevent db2pd from aborting abnormally. This can result in messages such as
7 ″Changing data structure forced command termination″ to appear in the output.
7 Nonetheless, the tool can be helpful for problem determination. Two benefits to
7 collecting information without latching include faster retrieval and no competition
7 for engine resources.
7 What follow are some examples of problem scenarios in which db2pd can be used
7 to expedite problem determination.
7 Scenario 1: Catching a lock timeout using the ″-catch″ db2pd option:
7 This option is UNDOCUMENTED, yet enabled on customer systems.
7 The purpose of db2pd -catch is to allow the user to catch any sqlcode (and reason
7 code), zrc code, or ecf code and capture the information needed to solve the error
7 code.

db2pd - Monitor and troubleshoot DB2 UDB
7 The primary reaction is to execute the db2cos (callout script) that can dynamically
7 be altered to run any db2pd command, os command, or any other command
7 needed to solve the problem. The template db2cos file is located in sqllib/cfg and
7 must be moved to sqllib in order to be called.
7 Usage:
7
7 -catch clear | status | <errorCode> [<action>] [count=<count>]
7 Sets catchFlag to catch error or warning.
7
7
7 Error Codes:
7 <sqlCode>[,<reasonCode>] / sqlcode=<sqlCode>[,<reasonCode>]
7 ZRC (hex or integer)
7 ZRC #define (such as SQLP_LTIMEOUT)
7 ECF (hex or integer)
7 "deadlock" or "locktimeout"
7
7 Actions:
7 [db2cos] (default) Run sqllib/db2cos callout script
7 [stopdb2trc] Stop db2trc
7 [dumpcomponent] Dump component flag
7 [component=<componentID>] Component ID
7 [lockname=<lockname>] Lockname for catching specific lock
7 (lockname=000200030000001F0000000052)
7 [locktype=<locktype>] Locktype for catching specific lock
7 (locktype=R or locktype=52)
7 Examples (see last example for a step-by-step locktimeout catch):

7 1. Catch sqlcode -911 reason code 68 and run db2cos
7 db2pd -catch -911,68 db2cos OR
7 db2pd -catch -911,68
7
7 Error Catch #1
7 Sqlcode: -911
7 ReasonCode: 68
7 ZRC: 0
7 ECF: 0
7 Component ID: 0
7 LockName: Not Set
7 LockType: Not Set
7 Current Count: 0
7 Max Count: 255
7 Bitmap: 0x261
7 Action: Error code catch flag enabled
7 Action: Execute sqllib/db2cos callout script
7 2. Catch sqlcode -911 (and any reason code) and run db2cos
7 db2pd -catch -911
7
7 Input ECF string ’-911’ parsed as 0xFFFFFC71 (-911).
7 Error Catch #2
7 Sqlcode: -911
7 ReasonCode: 0
7 ZRC: 0
7 ECF: 0
7 Component ID: 0
7 LockName: Not Set
7 LockType: Not Set
7 Current Count: 0
7 Max Count: 255
7 Bitmap: 0x61

7 3. Catch SQLP_LTIMEOUT and run db2cos

7 db2pd -catch SQLP_LTIMEOUT
7
7 Input ZRC string ’SQLP_LTIMEOUT’ parsed as 0x80100044 (-2146435004).
7 Error Catch #3
7 Sqlcode: 0
7 ReasonCode: 0
7 ZRC: -2146435004
7 ECF: 0
7 Component ID: 0
7 LockName: Not Set
7 LockType: Not Set
7 Current Count: 0
7 Max Count: 255
7 Bitmap: 0xA1
7 4. Catch SQLP_LTIMEOUT and run db2cos, but only fire the script one time.
7 db2pd -catch SQLP_LTIMEOUT count=1
7
7 Input ZRC string ’SQLP_LTIMEOUT’ parsed as 0x80100044 (-2146435004).
7 Error Catch #4
7 Sqlcode: 0
7 ReasonCode: 0
7 ZRC: -2146435004
7 ECF: 0
7 Component ID: 0
7 LockName: Not Set
7 LockType: Not Set
7 Current Count: 0
7 Max Count: 1
7 Bitmap: 0xA1
7 5. Only clear -911,68 setting
7 db2pd -catch clear -911,68
7 Error catch setting cleared for sqlCode -911 reasonCode 68
7 6. Clear all settings
7 db2pd -catch clear all
7 All error catch flag settings cleared.
7 7. Step-by-step instructions to catch a lock timeout.
7 v Copy the db2cos template into sqllib.
7 cp sqllib/cfg/db2cos sqllib/db2cos
7 v Set the error catch setting. You can use -911,68, -911, SQLP_LTIMEOUT, or
7 locktimeout. If fact, if you know the lock type or lock name in advance, you
7 can use use the locktype or lockname suboptions to filter out unwanted lock
7 catches.
7 db2pd -catch locktimeout count=1
7 Error Catch #1
7 Sqlcode: 0
7 ReasonCode: 0
7 ZRC: -2146435004
7 ECF: 0
7 Component ID: 0
7 LockName: Not Set
7 LockType: Not Set
7 Current Count: 0
7 Max Count: 1
7 Bitmap: 0xA1

7 v Reproduce a lock timeout.

7 db2 update staff set id = 0
7 DB21034E The command was processed as an SQL statement because it was not a
7 valid Command Line Processor command. During SQL processing it returned:
7 SQL0911N The current transaction has been rolled back because of a deadlock
7 or timeout. Reason code "68". SQLSTATE=40001
7 v Notice the db2diag.log tells us what happened.
7 2005-01-06-15.31.59.017134-300 I416442C274 LEVEL: Event
7 PID : 9875676 TID : 1 PROC : db2pd
7 INSTANCE: jmcmahon NODE : 000
7 FUNCTION: DB2 UDB, RAS/PD component, pdErrorCatch, probe:30
7 START : Error catch set for ZRC -2146435004
7
7 2005-01-06-15.36.07.084657-300 I416717C403 LEVEL: Event
7 PID : 10379460 TID : 1 PROC : db2agent (SAMPLE)
7 INSTANCE: jmcmahon NODE : 000 DB : SAMPLE
7 APPHDL : 0-71 APPID: *LOCAL.jmcmahon.050106202920
7 FUNCTION: DB2 UDB, trace services, pdInvokeCalloutScript, probe:10
7 START : Invoking sqllib/db2cos script from lock manager sqlplnfd
7
7 2005-01-06-15.36.09.274708-300 I417121C386 LEVEL: Event
7 PID : 10379460 TID : 1 PROC : db2agent (SAMPLE)
7 INSTANCE: jmcmahon NODE : 000 DB : SAMPLE
7 APPHDL : 0-71 APPID: *LOCAL.jmcmahon.050106202920
7 FUNCTION: DB2 UDB, trace services, pdInvokeCalloutScript, probe:20
7 STOP : Completed invoking sqllib/db2cos script
7 v Now look in sqllib/db2dump/db2cos.rpt which is the default output file that
7 contains the db2pd output generated when the error was caught.
7 Lock Timeout Caught
7 Thu Jan 6 15:36:07 EST 2005
7 Instance jmcmahon
7 Datbase: SAMPLE
7 Partition Number: 0
7 PID: 10379460
7 TID: 1
7 Function: sqlplnfd
7 Component: lock manager
7 Probe: 999
7 Timestamp: 2005-01-06-15.36.07.084474
7 AppID: *LOCAL.jmcmahon.050106202920
7 AppHdl:
7
7 <snip>
7 Database Partition 0 -- Database SAMPLE -- Active -- Up 0 days 00:06:53
7
7 Locks:
7 Address TranHdl Lockname Type Mode Sts Owner Dur HldCnt Att Rlse
7 0x402C6B30 3 00020003000000040000000052 Row ..X W* 3 1 0 0 0x40
7 v Just look for the ″W*″ as this is the lock that experienced the timeout. You
7 can map this to a transaction, application, agent, and even SQL statement
7 with all of the db2pd output provided. You can get strategic and narrow
7 down the output or use other commands to collect the information you need.
7 For example, you could change the db2pd options to use the ″-locks wait″
7 option that only prints locks with a wait status and their waiter. You could
7 also put in -fmtlock, -app, and -agent options if that’s what you need.
7 "LOCKTIMEOUT")
7 echo "Lock Timeout Caught" >> $HOME/sqllib/db2dump/db2cos.rpt
7 date >> $HOME/sqllib/db2dump/db2cos.rpt
7 echo "Instance " $instance >> $HOME/sqllib/db2dump/db2cos.rpt
7 echo "Datbase: " $database >> $HOME/sqllib/db2dump/db2cos.rpt
7 echo "Partition Number:" $dbpart >> $HOME/sqllib/db2dump/db2cos.rpt
7 echo "PID: " $pid >> $HOME/sqllib/db2dump/db2cos.rpt
7 echo "TID: " $tid >> $HOME/sqllib/db2dump/db2cos.rpt

7 echo "Function: " $function >> $HOME/sqllib/db2dump/db2cos.rpt

7 echo "Component: " $component >> $HOME/sqllib/db2dump/db2cos.rpt
7 echo "Probe: " $probe >> $HOME/sqllib/db2dump/db2cos.rpt
7 echo "Timestamp: " $timestamp >> $HOME/sqllib/db2dump/db2cos.rpt
7 echo "AppID: " $appid >> $HOME/sqllib/db2dump/db2cos.rpt
7 echo "AppHdl: " $apphld >> $HOME/sqllib/db2dump/db2cos.rpt
7 db2pd -db $database >> $HOME/sqllib/db2dump/db2cos.rpt
7 db2pd -db $database -locks wait -fmtlock -app -age >> $HOME/sqllib/db2dump/db2cos.rpt
7 ;;
7 It’s as simple as that!
7 Related reference:
7 v db2pd - Monitor and Troubleshoot DB2 Command
7 Scenario 2: Identifying a lock wait.
7 There are several types of locks reported with db2pd -locks that can cause an
7 application to wait. Common ones are table, pool, and row where other
7 possibilities are Internal V (dynamic sql cache), Internal P (package cache), and
7 CatCache (catalog cache). Prior to db2pd, there was no way to identify the Internal
7 V, Internal P, or CatCache lock.
7 db2pd -locks output showing Catcache, Row, Internal V, and Internal P.

7 Locks:
7 Address TranHdl Lockname Type Mode Sts Owner
7 0x00000002202EA3E0 2 0000000200000B0520A4774043 CatCache ..X G 2
7 0x00000002202E30D8 2 00000002000004070000000052 Row .NS G 2
7 0x00000002202DCC28 2 00000002000000010001860056 Internal V ..S G 2
7 0x00000002202D88F0 2 53514C4532453047668358B641 Internal P ..S G 2
7
7
7 Dur HldCnt Att Rlse
7 255 0 0 0x40
7 1 0 0 0x40
7 1 0 0 0x40
7 1 1 0 0x40
7 db2pd -catalogcache output showing the lock 0000000200000B0520A4774043 maps

7 to a SYSTABLES entry for the SYSIBM table.
7 Catalog Cache:
7 Configured Size 655360
7 Current Size 353416
7 Maximum Size 4294950912
7 High Water Mark 393216
7
7 SYSTABLES:
7 Address Schema Name Type TableID TbspaceID
7 0x0000000220A47740 SYSIBM SYSFUNCTIONS V 0 0
7
7 LastRefID CatalogCacheLoadingLock CatalogCacheUsageLock Sts
7 10615 000100000000000220A4774043 0000000200000B0520A4774043 V
7 db2pd -dynamic output showing that lock 00000002000000010001860056 maps to an

7 sql statement creating a view. Since the dynamic cache has three tiers, one must
7 identify the lock in the Variation section then map the Variation to an Environment
7 and finally map the Environment to a Statement.
7 Dynamic Cache:
7 Current Memory Used 1019927
7 Total Heap Size 1271398
7 Cache Overflow Flag 0

7 Number of References 150

7 Number of Statement Inserts 77
7 Number of Statement Deletes 1
7 Number of Variation Inserts 150
7 Number of Statements 76
7
7 Dynamic SQL Statements:
7 Address AnchID StmtUID NumEnv NumVar NumRef NumExe Text
7 0x0000000220ACF220 134 2 2 2 2 1 create view
7 syscat.librarybindfiles (libschema, libname, libversion, package_schema,
7 package_name, package_version, bindfile_path) as select
7 l.libschema, l.libname, b.libversion, b.package_schema, b.package_name, b.package_version,
7 b.bindfile_path from sysibm.syslibrarybindfiles as b,
7 sysibm.syslibraries as l where b.lib_id = l.lib_id
7
7 Dynamic SQL Environments:
7 Address AnchID StmtUID EnvID Iso QOpt Blk
7 0x0000000220ACF480 134 2 1 CS 5 U
7 0x0000000220ACFE60 134 2 2 CS 5 U
7
7 Dynamic SQL Variations:
7 Address AnchID StmtUID EnvID VarID NumRef Typ Lockname
7 0x0000000220ACF6C0 134 2 1 1 1 1 00000002000000010001860056
7 db2pd -static output showing that lock 53514C4532453047668358B641 maps to a

7 Package SQLE2E0G.
7 Static Cache:
7 Number of Package Inserts 2
7 Number of Section Inserts 0
7
7 Packages:
7 Address Schema PkgName Version UniqueID NumSec
7 0x0000000220A01780 JMCMAHON SQLE2E0G AAAAARAU 0
7
7 UseCount NumRef Iso QOpt Blk Lockname
7 0 1 CS 5 U 53514C4532453047668358B641
7 Scenario 3: Mapping an application to a dynamic SQL statement.
7 Starting with Saturn, db2pd -applications reports the Current and Last Anchor ID
7 and Statement Unique ID for dynamic SQL statements. This allows direct mapping
7 from an application to a dynamic sql statements.
7 db2pd -app -dyn
7
7 Applications:
7 Address AppHandl [nod-index] NumAgents CoorPid Status
7 0x00000002006D2120 780 [000-00780] 1 10615 UOW-Executing
7
7 C-AnchID C-StmtUID L-AnchID L-StmtUID Appid
7 163 1 110 1 *LOCAL.jmcmahon.050202200412
7
7 0x0000000220A02760 163 1 2 2 2 1 CREATE VIEW MYVIEW
7 0x0000000220A0B460 110 1 2 2 2 1 CREATE VIEW YOURVIEW
7 Scenario 4: Monitoring memory usage.

7 The db2pd -memsets and -mempools options report statistics about DB2 Memory
7 Sets and Memory Pools which can be very useful when trying to understand
7 memory usage.
7 Memory Sets:
7 Name Address Id Size Key DBP Type Ov OvSize
7 SAMPLE 0x0000000220000000 290831 46055424 0x0 0 1 Y 8978432
7 App780 0xFFFFFFFF77F00000 1487884 131072 0x0 0 3 N 0
7
7 Memory Pools:
7 Address MemSet PoolName Id Overhead LogSz LogUpBnd LogHWM
7 0x0000000220000CB0 SAMPLE utilh 5 0 304 20496384 304
7 0x0000000220000BB0 SAMPLE undefh 59 0 0 12345678 0
7 0x0000000220000AB0 SAMPLE pckcacheh 7 30592 768781 Unlimited 768781
7 0x00000002200009B0 SAMPLE catcacheh 8 0 194696 Unlimited 194696
7 0x00000002200008B0 SAMPLE bph 16 12064 4265856 Unlimited 4265856
7 0x00000002200003B0 SAMPLE lockh 4 0 492160 524288 492160
7 0x00000002200002B0 SAMPLE dbh 2 59168 3800760 8683520 3800760
7
7 PhySz PhyUpBnd PhyHWM Bnd BlkCnt CfgParm
7 16384 20496384 16384 Ovf 0 UTIL_HEAP_SZ
7 0 12353536 0 Ovf 0 n/a
7 819200 Unlimited 819200 Ovf 0 PCKCACHESZ
7 262144 Unlimited 262144 Ovf 0 CATALOGCACHE_SZ
7 4292608 Unlimited 4292608 Ovf 0 n/a
7 540672 Unlimited 540672 Ovf 0 n/a
7 524288 524288 524288 Ovf 0 LOCKLIST
7 3932160 8683520 3932160 Ovf 0 DBHEAP
7 Scnenario 5: Memory block usage using FFDC feature. (This is undocumented, yet
7 enabled on customer systems). You must set DB2MEMDBG=FFDC to enable this feature.
7 Usage
7 -memblock [dbms, fcm, fmp, appctl <id>, private (windows)] [scrape (unix)]
7 This option will report extended memory blocks from the following
7 sets/pools/heaps:
7 v DBMS
7 v FCM
7 v FMP
7 v AppCtl (shared memory set id must be supplied and database must be
7 supplied)
7 v Database
7 v Private (windows only)
7 v The ″scrape″ option on Unix does not use OSS calls to search for blocks. Instead,
7 it starts at the beginning of the memory set and walks it to the end, 4 bytes at a
7 time. This can be useful if there is suspicion that a memory block has been
7 ″orphaned″ by the OSS layer.
7 First, turn on FFDC extended memory blocks:

7 db2set DB2MEMDBG=FFDC
7 Example for DBMS set:

7 db2pd -memb OR
7 db2pd -memb dbms
7 Printing all blocks in DBMS set.
7 Address PoolID PoolName BlockID Size I P N LOC File
7 0x30944048 62 resynch 2 92 1 0 0 2314 1219138542
7 0x3092AF98 62 resynch 1 102444 1 0 0 124 4075170607
7 0x30109748 57 ostrack 6 4860032 1 0 0 2987 4279232714
7 0x300E5CC8 57 ostrack 5 140032 1 0 0 2974 4279232714
7 0x300E0048 57 ostrack 1 64 1 0 0 2910 4279232714
7 0x300E00A8 57 ostrack 2 208 1 0 0 2923 4279232714
7 0x300E0198 57 ostrack 3 80 1 0 0 2939 4279232714
7 0x300E0208 57 ostrack 4 80 1 0 0 2949 4279232714
7 0x30924048 70 apmh 5 16016 1 0 0 141 2389885231
7 0x30927EF8 70 apmh 6 124 1 0 0 2760 1219138542
7 You can then use the db2fntohash tool to map the File hash value to a filename.
7 If engn/pd/db2filenames.lst does NOT exist, you can create it. This should be
7 built with each nightly build. If engn/pd/db2filenames.lst DOES exist,
7 db2fntohash -l <hashvalue> will provide the filename.
7 db2fntohash db2filenames.in db2filenames.lst
7
7 db2fntohash -l 1219138542 db2filenames.lst
7 db2fntohash: hash value "1219138542" => file name "sqlesysc.C"
7
7 db2fntohash: hash value "4075170607" => file name "sqlerlst.C"
7
7 db2fntohash: hash value "4279232714" => file name "startdbm.C"
7 Example for FCM set:

7 db2pd -memb fcm
7 Printing all blocks in FCM set.
7 0x813C4048 79 fcmrqb 19 16400 1 0 0 2547 1272798017
7 0x813B8048 79 fcmrqb 18 47120 1 0 0 2375 1272798017
7 0x813AC048 79 fcmrqb 17 47120 1 0 0 2375 1272798017
7 0x813A0048 79 fcmrqb 16 47120 1 0 0 2375 1272798017
7 Example for FMP set:

7 db2pd -memb fmp
7 Printing all blocks in FMP set.
7 0x90009FA8 59 undefh 1 122916 1 0 0 357 4133190529
7 Example for DB set:

7 db2pd -memb -db sample
7 Printing all blocks in Database SAMPLE set.
7 0x4093C048 5 utilh 1 64 1 0 0 150 3032704123
7 0x4093C0A8 5 utilh 2 232 1 0 0 194 1779766316
7 0x40940048 7 pckcacheh 2 408 1 0 0 1082 19703507
7 0x409401F8 7 pckcacheh 3 1060 1 0 0 2119 3048315900
7 0x40940738 7 pckcacheh 14 408 1 0 0 1082 19703507
7 Example for AppCtl:

7 db2pd -memb appctl -db sample
7 Printing all blocks in Appctl set id 566624315.
7 0xA370C048 3 appctlh 1 64 1 0 0 150 3032704123
7 0xA370C0A8 3 appctlh 2 2992 1 0 0 1413 2658538710

7 0xA370CC78 3 appctlh 3 88 1 0 0 2547 1436638871

7 0xA370D1E8 3 appctlh 8 144 1 0 0 778 91986655
7 0xA370D298 3 appctlh 9 400 1 0 0 4284 2614547685
7 Example for all instance scope sets (DBMS, FCM, FMP):

7 db2pd -memb all
7 0x30980048 74 fcm 4 16400 1 0 0 1224 2927312349
7 0x3091FFB8 74 fcm 2 393232 1 0 0 1009 2927312349
7 0x30914048 74 fcm 1 4112 1 0 0 818 1272798017
7 0x30915078 74 fcm 3 4112 1 0 0 1042 2927312349
7 0x309D6F88 11 monh 80 69699 1 0 0 772 3881733182
7 0x309C2F88 11 monh 79 69699 1 0 0 772 3881733182
7 <snip>
7
7
7 0x813C4048 79 fcmrqb 19 16400 1 0 0 2547 1272798017
7 0x813B8048 79 fcmrqb 18 47120 1 0 0 2375 1272798017
7 0x813AC048 79 fcmrqb 17 47120 1 0 0 2375 1272798017
7 <snip>
7
7
7 0x90009FA8 59 undefh 1 122916 1 0 0 357 4133190529
7 Example for all instance and database scope sets (DBMS, FCM, FMP, DB, AppCtl):
7 db2pd -memb all -db sample
7 0x30980048 74 fcm 4 16400 1 0 0 1224 2927312349
7 0x3091FFB8 74 fcm 2 393232 1 0 0 1009 2927312349
7 0x30914048 74 fcm 1 4112 1 0 0 818 1272798017
7 0x30915078 74 fcm 3 4112 1 0 0 1042 2927312349
7 0x309D6F88 11 monh 80 69699 1 0 0 772 3881733182
7 0x309C2F88 11 monh 79 69699 1 0 0 772 3881733182
7 0x309BC048 11 monh 70 184 1 0 0 130 88219158
7 <snip>
7
7
7 0x813C4048 79 fcmrqb 19 16400 1 0 0 2547 1272798017
7 0x813B8048 79 fcmrqb 18 47120 1 0 0 2375 1272798017
7 0x813AC048 79 fcmrqb 17 47120 1 0 0 2375 1272798017
7 <snip>
7
7
7 0x90009FA8 59 undefh 1 122916 1 0 0 357 4133190529
7
7
7 0xA370C048 3 appctlh 1 64 1 0 0 150 3032704123
7 0xA370C0A8 3 appctlh 2 2992 1 0 0 1413 2658538710
7 <snip>
7
7
7 0x4093C048 5 utilh 1 64 1 0 0 150 3032704123

7 0x4093C0A8 5 utilh 2 232 1 0 0 194 1779766316

7 0x40940048 7 pckcacheh 2 408 1 0 0 1082 19703507
7 0x409401F8 7 pckcacheh 3 1060 1 0 0 2119 3048315900
7 Example for AppCtl sets (database required):

7 db2pd -db sample -memb appctl
7 0xA370C048 3 appctlh 1 64 1 0 0 150 3032704123
7 0xA370C0A8 3 appctlh 2 2992 1 0 0 1413 2658538710
7 0xA370CC78 3 appctlh 3 88 1 0 0 2547 1436638871
7 0xA370D1E8 3 appctlh 8 144 1 0 0 778 91986655
7 0xA370D298 3 appctlh 9 400 1 0 0 4284 2614547685
7 <snip>
7
7
7 0x4093C048 5 utilh 1 64 1 0 0 150 3032704123
7 0x4093C0A8 5 utilh 2 232 1 0 0 194 1779766316
7 0x40940048 7 pckcacheh 2 408 1 0 0 1082 19703507
7 0x409401F8 7 pckcacheh 3 1060 1 0 0 2119 3048315900
7 0x40940738 7 pckcacheh 14 408 1 0 0 1082 19703507
7 Scenario 6: Producing a shared memory dump with db2pd -bindump

7 (undocumented, yet enabled on customer systems). Reading the shared memory
7 dump (and now core files with Saturn) using db2pd -binload (internal option).
7 Usage
7 -bindump [-db <database>] [-inst]
7 v DBMS, FCM, FMP sets dumped with -bindump.
7 v DBMS, DB, AppCtl sets dumped with -db <database>.
7 v DBMS, FCM, FMP, DB, AppCtl sets dumped with -db <database> and -inst.
7 Example -bindump:
7 db2pd -bind
7 Sending -bindump output to /home/jmcmahon/db2pd.bin
7 Dumping DBMS set starting at 0x30000000 with size 52690944
7 Dumping FCM set starting at 0x80000000 with size 43270144
7 Dumping FMP set starting at 0x90000000 with size 257654784
7 Example -bindump -db sample:

7 db2pd -bind -db sample
7 Dumping DB set for database SAMPLE starting at 0x40000000 with size 65716224
7 Dumping AppCtl set starting at 0xa0000000 with size 192839680
7 Example -bindump -db sample -inst:

7 db2pd -bind -db sample -inst
7 Dumping DB set for database SAMPLE starting at 0x40000000 with size 65716224
7 Dumping AppCtl set starting at 0xa0000000 with size 192839680
7 Dumping FCM set starting at 0x80000000 with size 43270144
7 Dumping FMP set starting at 0x90000000 with size 257654784
7 Usage
7 -binload <shared memory dump> | <corefile> [any db2pd option]

7 This is very useful when you have a shared memory dump from the -bindump
7 option. You can run ANY db2pd option just like it is from a live system.
7 Note that the default name is ″db2pd.bin″ which is the default when the -bindump
7 option is used without a filename.
7 Starting with Saturn code (FP9), the -binload option will look for the string ″core″
7 and treat the file as a core dump if the filename starts with ″core.″ This
7 functionality has only had limited testing. Please report any issues to
7 db2raspdcore@ca.ibm.com.
7 Note: Run db2pd in interactive mode and then run each desired option. That will
7 prevent having to load the shared memory dump each time you run db2pd
7 with the -binload option.
7 Example -binload:
7 db2pd -binl -memp
7 Reading memory dump from file db2pd.bin.
7 Allocated segment at address 0x30000000 with shmid 524307.
7 Allocated segment at address 0xa0000000 with shmid 524306.
7
7
7 Database Partition 0 -- Active -- Up 0 days 02:55:14
7
7
7 Memory Pools:
7 Address MemSet PoolName Id Overhead LogSz LogUpBnd
7 0x30000840 DBMS fcm 74 16240 417792 450560
7 0x300007A0 DBMS monh 11 24192 142032 368640
7 0x30000700 DBMS resynch 62 12112 102504 3244032
7 0x30000660 DBMS ostrack 57 13184 5000400 5046272
7 0x300005C0 DBMS apmh 70 0 38730 2129920
7 0x30000520 DBMS kerh 52 32 31488 376832
7 0x30000480 DBMS bsuh 71 3360 3598864 17137664
7 0x300003E0 DBMS sqlch 50 0 762400 786432
7 0x30000340 DBMS pmth 80 0 72 216540
7 0x300002A0 DBMS krcbh 69 0 19228 32768
7 0x30000200 DBMS debug 87 16240 12353536 12353552
7 0x80000700 FCM fcmrqb 79 16240 835632 1548336
7 0x80000660 FCM fcmce 78 16240 225328 696368
7 0x800005C0 FCM fcma 75 16240 237616 712752
7 0x80000520 FCM fcmrqb 79 16240 835632 1548336
7 0x80000480 FCM fcmce 78 16240 225328 696368
7 0x800003E0 FCM fcma 75 16240 237616 712752
7 0x80000340 FCM fcmbp 13 404848 17011260 17980332
7 0x800002A0 FCM debug 87 16240 12353536 12517392
7 0x80000200 FCM eduah 72 5984 256024 419888
7 0x900002A0 FMP undefh 59 8032 122900 245163840
7 0x90000200 FMP debug 87 16240 12353536 12517392
7
7 LogHWM PhySz PhyUpBnd PhyHWM Bnd BlkCnt CfgParm
7 417792 458752 458752 458752 Ovf 4 n/a
7 142116 180224 376832 180224 Ovf 13 MON_HEAP_SZ
7 102504 131072 3244032 131072 Ovf 2 n/a
7 5000400 5029888 5046272 5029888 Ovf 6 n/a
7 39010 49152 2129920 49152 Ovf 21 n/a
7 32396 49152 376832 49152 Ovf 11 n/a
7 3605820 3620864 17137664 3620864 Ovf 55 n/a
7 762400 786432 786432 786432 Ovf 163 n/a
7 252 16384 229376 16384 Ovf 2 n/a
7 19228 32768 32768 32768 Ovf 3 n/a

7 12353536 12369920 12369920 12369920 Ovf 1 n/a

7 835632 917504 1556480 917504 Phy 19 n/a
7 225328 294912 704512 294912 Phy 15 n/a
7 237616 294912 720896 294912 Phy 15 n/a
7 835632 917504 1556480 917504 Phy 19 n/a
7 225328 294912 704512 294912 Phy 15 n/a
7 237616 294912 720896 294912 Phy 15 n/a
7 17011260 17448960 17989632 17448960 Phy 38 n/a
7 12353536 12369920 12533760 12369920 Phy 1 n/a
7 256024 262144 425984 262144 Phy 1 n/a
7 122900 131072 245170176 131072 Phy 1 n/a
7 12353536 12369920 12533760 12369920 Phy 1 n/a
7
7 Removed segment with shmid 524307.
7 Example -binload -db sample:

7 db2pd -binl -memp -db sample
7
7
7
7
7 Memory Pools:
7 0x400008E0 SAMPLE utilh 5 0 264 20496384
7 0x40000840 SAMPLE undefh 59 0 0 12345678
7 0x400007A0 SAMPLE pckcacheh 7 17696 205415 Unlimited
7 0x40000700 SAMPLE catcacheh 8 0 43232 Unlimited
7 0x40000660 SAMPLE bph 16 11344 4238784 Unlimited
7 0x400005C0 SAMPLE bph 16 112 533248 Unlimited
7 0x400003E0 SAMPLE bph 16 112 74496 Unlimited
7 0x40000340 SAMPLE lockh 4 32 475568 507904
7 0x400002A0 SAMPLE dbh 2 32816 3603997 10715136
7 0x40000200 SAMPLE debug 87 16240 12353536 12353552
7 0xA0000480 AppCtl appctlh 3 0 48 793018
7 0xA00003E0 AppCtl appctlh 3 112 14019 793018
7 0xA0000340 AppCtl agsh 17 400 38276 114851888
7 0xA00002A0 AppCtl mbosh 84 0 24 163864
7 0xA0000200 AppCtl debug 87 16240 12353536 12517392
7
7
7 1560 16384 20496384 16384 Ovf 2 UTIL_HEAP_SZ
7 0 0 12353536 0 Ovf 0 n/a
7 205415 278528 Unlimited 278528 Ovf 16 PCKCACHESZ
7 43232 65536 Unlimited 65536 Ovf 8 CATALOGCACHE_SZ
7 4238784 4276224 Unlimited 4276224 Ovf 2 n/a
7 533248 557056 Unlimited 557056 Ovf 2 n/a
7 271104 294912 Unlimited 294912 Ovf 2 n/a
7 140032 163840 Unlimited 163840 Ovf 2 n/a
7 74496 98304 Unlimited 98304 Ovf 2 n/a
7 475568 507904 507904 507904 Ovf 2 LOCKLIST
7 4696564 3670016 10715136 4849664 Ovf 326 DBHEAP
7 12353536 12369920 12369920 12369920 Ovf 1 n/a

7 10116 16384 802816 16384 Phy 1 APP_CTL_HEAP_SZ

7 18449 327680 802816 327680 Phy 34 APP_CTL_HEAP_SZ
7 38276 57344000 114868224 57344000 Log 26 APPGROUP_SHARE_HEAP
7 24 16384 180224 16384 Phy 1 n/a
7 12353536 12369920 12533760 12369920 Phy 1 n/a
7
7
7 Example -binload interactive mode:

7 db2pd
7 db2pd> You are running db2pd in interactive mode.
7 db2pd> If you want command line mode, rerun db2pd with valid options.
7 db2pd> Type -h or -help for help.
7 db2pd> Type q to quit.
7 db2pd> -binl
7 db2pd> -memp
7
7
7
7
7 Memory Pools:
7 0x30000840 DBMS fcm 74 16240 417792 450560
7 0x300007A0 DBMS monh 11 24192 142032 368640
7 0x30000700 DBMS resynch 62 12112 102504 3244032
7 0x30000660 DBMS ostrack 57 13184 5000400 5046272
7 0x300005C0 DBMS apmh 70 0 38730 2129920
7 0x30000520 DBMS kerh 52 32 31488 376832
7 0x30000480 DBMS bsuh 71 3360 3598864 17137664
7 0x300003E0 DBMS sqlch 50 0 762400 786432
7 0x30000340 DBMS pmth 80 0 72 216540
7 0x300002A0 DBMS krcbh 69 0 19228 32768
7 0x30000200 DBMS debug 87 16240 12353536 12353552
7 0x80000700 FCM fcmrqb 79 16240 835632 1548336
7 0x80000660 FCM fcmce 78 16240 225328 696368
7 0x800005C0 FCM fcma 75 16240 237616 712752
7 0x80000520 FCM fcmrqb 79 16240 835632 1548336
7 0x80000480 FCM fcmce 78 16240 225328 696368
7 0x800003E0 FCM fcma 75 16240 237616 712752
7 0x80000340 FCM fcmbp 13 404848 17011260 17980332
7 0x800002A0 FCM debug 87 16240 12353536 12517392
7 0x80000200 FCM eduah 72 5984 256024 419888
7 0x900002A0 FMP undefh 59 8032 122900 245163840
7 0x90000200 FMP debug 87 16240 12353536 12517392
7
7 db2pd> -age
7
7
7 417792 458752 458752 458752 Ovf 4 n/a
7 142116 180224 376832 180224 Ovf 13 MON_HEAP_SZ
7 102504 131072 3244032 131072 Ovf 2 n/a
7 5000400 5029888 5046272 5029888 Ovf 6 n/a
7 39010 49152 2129920 49152 Ovf 21 n/a

7 32396 49152 376832 49152 Ovf 11 n/a

7 3605820 3620864 17137664 3620864 Ovf 55 n/a
7 762400 786432 786432 786432 Ovf 163 n/a
7 252 16384 229376 16384 Ovf 2 n/a
7 19228 32768 32768 32768 Ovf 3 n/a
7 12353536 12369920 12369920 12369920 Ovf 1 n/a
7 835632 917504 1556480 917504 Phy 19 n/a
7 225328 294912 704512 294912 Phy 15 n/a
7 237616 294912 720896 294912 Phy 15 n/a
7 835632 917504 1556480 917504 Phy 19 n/a
7 225328 294912 704512 294912 Phy 15 n/a
7 237616 294912 720896 294912 Phy 15 n/a
7 17011260 17448960 17989632 17448960 Phy 38 n/a
7 12353536 12369920 12533760 12369920 Phy 1 n/a
7 256024 262144 425984 262144 Phy 1 n/a
7 122900 131072 245170176 131072 Phy 1 n/a
7 12353536 12369920 12533760 12369920 Phy 1 n/a
7
7
7
7 Agents:
7 Current agents: 8
7 Idle agents: 6
7 Active agents: 2
7 Coordinator agents: 1
7
7
7 Address AppHandl [nod-index] AgentPid Priority Type State ClientPid LkTmOt DBName
7 0x315898C0 0 [000-00000] 155688 0 Idle n/a 0 n/a
7 0x31588880 0 [000-00000] 1204334 0 Idle n/a 0 n/a
7 0x314DE190 0 [000-00000] 700478 0 Idle n/a 0 n/a
7 0x315890A0 0 [000-00000] 643262 0 Idle n/a 0 n/a
7 0x31588060 0 [000-00000] 1134718 0 Idle n/a 0 n/a
7 0x314DD730 0 [000-00000] 676026 0 Panic Idle n/a 0 n/a
7 0x314DF1D0 55 [000-00055] 614516 0 SubAgent Active 794820 NotSet SAMPLE
7 0x314DE9B0 55 [000-00055] 1093854 0 Coord Active 794820 NotSet SAMPLE
7 db2pd> q
7
7 Userid ClientNm Rowsread Rowswrtn
7 n/a n/a 0 0
7 n/a n/a 0 0
7 n/a n/a 0 0
7 n/a n/a 0 0
7 n/a n/a 0 0
7 n/a n/a 0 0
7 jmcmahon db2bp 40 0
7 jmcmahon db2bp 12 0
7
7 Scenario 7: Figuring out which application is using up your tablespace.
7 Using db2pd -tcbstats, the number of Inserts can be identified for a table. Here is
7 temp table TEMP1.
7 TCB Table Stats:
7 Address TbspaceID TableID TableName SchemaNm Scans ObjClass UDI
7 0x000000022094AA58 4 2 TEMP1 SESSION 0 Temp 0
7
7 DataSize IndexSize PgReorgs NoChgUpdts Reads FscrUpdates
7 1 0 0 0 0 0

7
7 Inserts Updates Deletes OvFlReads OvFlCrtes LfSize LobSize
7 21586 0 0 0 0 0 0
7 which can be mapped to tablespace 4 in db2pd -tablespaces output.

7 Tablespaces:
7 Address Id Type Content AS AR PageSize ExtentSize Auto Prefetch
7 0x0000000220942F80 4 DMS UsrTmp No No 4096 32 Yes 32
7
7 BufID BufIDDisk State TotPages UsablePgs UsedPgs PndFreePgs
7 1 1 0x00000000 10000 9952 128 0
7
7 FreePgs InitSize IncSize IIP MaxSize LastResize
7 9824 0 0 No 0 None
7
7 LRF HWM MinRecTime NQuiescers FSC NumCntrs MaxStripe Name
7 No 128 0 0 On 1 0 TEMPSPACE2
7
7 Containers:
7 Address TspId ContainNum Type TotalPages UseablePgs
7 0x0000000220377CE0 4 0 File 10000 9952
7
7 StripeSet Container
7 0 /export/home/jmcmahon/tempspace2a
7 (The UsablePgs vs. TotalPages is where they would notice the space filling)
7 Once this is known, we can identify the dynamic sql statement using a table called
7 ″TEMP1.″
7 db2pd -db sample -dyn
7
7
7 Dynamic Cache:
7 Number of Statement Inserts 32
7 Number of Statement Deletes 13
7 Number of Variation Inserts 21
7 Number of Statements 19
7
7 0x0000000220A08C40 78 1 2 2 3 2 declare global temporary
7 table temp1 (c1 char(6)) not logged
7 0x0000000220A8D960 253 1 1 1 24 24 insert into session.temp1
7 values(’TEST’)
7 And finally, map this to -app output to identify the application.

7 Applications:
7 Address AppHandl [nod-index] NumAgents CoorPid Status
7 0x0000000200661840 501 [000-00501] 1 11246 UOW-Waiting
7
7 C-AnchID C-StmtUID L-AnchID L-StmtUID Appid
7 0 0 253 1 *LOCAL.jmcmahon.050202160426
7
7 Also, the -agent output will show the number of row written as verification.
7
7 Address AppHandl [nod-index] AgentPid Priority Type DBName
7 0x0000000200698080 501 [000-00501] 11246 0 Coord SAMPLE

7
7 State ClientPid Userid ClientNm Rowsread Rowswrtn LkTmOt
7 Inst-Active 26377 jmcmahon db2bp 22 9588 NotSet
7 Scenario 8: Monitoring recovery. db2pd -recovery shows several counters to make

7 sure recovery is progressing. Current Log and Current LSN provide the log
7 position. CompletedWork counts the number of bytes completed thus far.
7 Recovery:
7 Recovery Status 0x00000401
7 Current Log S0000005.LOG
7 Current LSN 000002551BEA
7 Job Type ROLLFORWARD RECOVERY
7 Job ID 7
7 Job Start Time (1107380474) Wed Feb 2 16:41:14 2005
7 Job Description Database Rollforward Recovery
7 Invoker Type User
7 Total Phases 2
7 Current Phase 1
7
7 Progress:
7 Address PhaseNum Description StartTime CompletedWork
7 0x0000000200667160 1 Forward Wed Feb 2 16:41:14 2005 2268098 bytes
7 0x0000000200667258 2 Backward NotStarted 0 bytes
7 Scenario 9: Understanding how much resource a transaction is using. db2pd

7 -transactions provides the number of locks, first lsn, last lsn, logspace used, and
7 space reserved. This can be useful for understanding the behavior of any
7 transaction.
7 Transactions:
7 Address AppHandl [nod-index] TranHdl Locks State Tflag
7 0x000000022026D980 797 [000-00797] 2 108 WRITE 0x00000000
7 0x000000022026E600 806 [000-00806] 3 157 WRITE 0x00000000
7 0x000000022026F280 807 [000-00807] 4 90 WRITE 0x00000000
7
7 Tflag2 Firstlsn Lastlsn LogSpace SpaceReserved
7 0x00000000 0x000001072262 0x0000010B2C8C 4518 95450
7 0x00000000 0x000001057574 0x0000010B3340 6576 139670
7 0x00000000 0x00000107CF0C 0x0000010B2FDE 3762 79266
7
7 TID AxRegCnt GXID
7 0x000000000451 1 0
7 0x0000000003E0 1 0
7 0x000000000472 1 0
7 Scenario 10: Monitoring log usage. db2pd -logs is useful to monitor log usage for
7 a database. Watching the Pages Written output tells whether log usage is
7 progressing or not.
7 Logs:
7 Current Log Number 4
7 Pages Written 464
7
7 Address StartLSN State Size Pages Filename
7 0x000000022022FEB8 0x000000FA0000 0x00000000 1000 597 S0000000.LOG
7 0x000000022022FF78 0x000001388000 0x00000000 1000 5 S0000001.LOG
7 0x0000000220008E78 0x000001770000 0x00000000 1000 3 S0000002.LOG
7 0x0000000220A57F58 0x000001B58000 0x00000000 1000 1000 S0000003.LOG
7 0x0000000220A32598 0x000001F40000 0x00000000 1000 1000 S0000004.LOG
7 For Saturn, db2pd -logs has some new information:

7 Logs:
7 Current Log Number 2
7 Pages Written 846
7 Method 1 Archive Status Success

7 Method 1 Next Log to Archive 2

7 Method 1 First Failure n/a
7 Method 2 Archive Status Success
7 Method 2 Next Log to Archive 2
7 Method 2 First Failure n/a
7
7 Address StartLSN State Size Pages Filename
7 0x000000023001BF58 0x000001B58000 0x00000000 1000 1000 S0000002.LOG
7 0x000000023001BE98 0x000001F40000 0x00000000 1000 1000 S0000003.LOG
7 0x0000000230008F58 0x000002328000 0x00000000 1000 1000 S0000004.LOG
7 Two problems can be identified with this output.

7 1. If there is a problem with archiving, Archive Status will be set to Failure
7 indicating the most recent log archive failed. Or if there is an ongoing archive
7 failure preventing logs from archiving at all First Failure will be set.
7 2. If log archiving is proceeding very slowly, Next Log to Archive will be behind
7 Current Log Number. This can cause the log path to fill up, which in turn, can
7 prevent any data changes from occurring in the database when the log path is
7 completely filled.
7 Scenario 11: Sysplex list. Without db2pd -sysplex, the only way to report the
7 sysplex list is with db2trc.
7 Sysplex List:
7 Alias: LAKETCP
7 Location Name: LAKEERIE
7 Count: 1
7
7 IP Address Port Priority Connections Status PRDID
7 9.26.89.87 448 1 0 0
db2support - Collecting environment information

When it comes to collecting information for a DB2 problem, the most important
DB2 utility you need to run is db2support.
The db2support utility is designed to automatically collect all DB2 and system
diagnostic information available. It also has an optional interactive ″Question and
Answer″ session, which poses questions about the circumstances of your problem.
Using db2support avoids possible user errors, as you don’t need to manually type
commands such as ″GET DATABASE CONFIGURATION FOR <database name>″ or ″LIST
TABLESPACES SHOW DETAIL″. Also, you don’t require instructions on which
commands to run or what files to collect, therefore information-gathering for
problem determination is quicker.
Note: The db2support utility should be run by a user with SYSADM authority,
such as an instance owner, so that the utility can collect all of the necessary
information without an error. If a user without SYSADM authority runs
db2support, SQL errors (SQL1092) might result when the utility runs
commands such as ″query client″ or ″list active databases″.
If you’re using db2support to help convey information to IBM software support,

run db2support while the system is experiencing the problem that led you to
request support. That way, the tool will collect operating system performance
information. You can also run db2support after the fact for problems requiring
post-mortem analysis (for example, traps or crashes).
Executing db2support -h brings up the complete list of possible options you can
run the utility with. The following basic invocation is usually sufficient for

db2support - Problem Analysis and Environment Collection Tool
collecting most of the information required to debug a problem (note that if the -c
option is used the utility will establish a connection to the database):
db2support <output path> -d <database name> -c
The output is conveniently collected and stored in a compressed ZIP archive,

db2support.zip, so that it can be transferred and extracted easily on any system.
The db2support tool adds the output of the various commands that it runs to the
archive as files without encapsulating them in a directory (as of DB2 v.7, FixPak 7).
Previously, generated command output was stored in either db2support.html or
detailed_system_info.html. db2support no longer store all it’s output files under
one output directory. Rather, db2support creates multiple subdirectories, and
organizes each output file under an appropriate subdirectory. This avoids a mass
of files existing under a single directory, and having to search through the list of
files to find the one of interest. The subdirectory names help organize and
categorize the type of data that is captured. For example, the database manager
configuration, which has the output filename of dbm.supp_cfg, is stored under the
subdirectory named DB2CONFIG. The diagnostic files and contents that exist under
DIAGPATH (e.g., db2diag.log) is stored under the subdirectory named DB2DUMP.
The type of information that db2support captures depends on the way the
command is invoked, whether or not the database manager has been started, and
whether it’s possible to connect to the database.
The db2support utility collects the following information under all conditions:
v db2diag.log
v All trap files
v Lock list files
v Dump files
v Buffer pool and table space (SQLSPCS.1 and SQLSPCS.2) control files (with -d)
v Various system related files
v Output from various system commands
v db config (with -d)
v dbm config files
v Log File Header file (with -d)
v Recovery History File (with -d)
v db2cli.ini
Depending on the circumstances, db2support may also collect:

v Active log files
v Contents of the db2dump directory (in other words, what was not collected
above)
v Core files (-a for all core files, -r for only the most recent core file)
v Extended system information (-s)
The HTML report db2support.html will always include the following information:
v PMR number, if one exists (if -n was specified)
v Operating system and level (for example, AIX 4.5)
v DB2 release information
v Engine library header information
v 32- vs 64-bit environment
v DB2 install path information

v Contents of db2nodes.cfg (for DB2 UDB Enterprise Server Edition)
v Number of CPUs and disks and how much memory
v List of databases on this instance
v Registry information and environment, including path and libpath
v Disk freespace for current filesystem and inodes for UNIX
v JDK level
v dbm config
v Listing of the database recovery history file
v ’ls -lR’ (or Windows equivalent) of the sqllib directory
v The result of the LIST NODE DIRECTORY command
v The result of the LIST ADMIN NODE DIRECTORY command
v The result of the LIST DCS DIRECTORY command
v The result of the LIST DCS APPLICATIONS EXTENDED command
v List of all installed software
The following information appears in the db2support.html file when -s is specified:

v Detailed disk information (partition layout, type, LVM information, and so on)
v Detailed network information
v Kernel statistics
v Firmware versions
v Other platform specific commands
The db2support.html file contains the following information if DB2 has been
started:
v Client connection state
v db/dbm config (db cfg require -d option)
v CLI config
v Memory pool info (size and consumed). Complete data if -d option used
v The result of the LIST ACTIVE DATABASES command
v The result of the LIST DATALINKS MANAGERS command
v The result of the LIST DCS APPLICATIONS command
The db2support.html file contains the following information if -c has been specified
and a connection to the database can be made
v Number of user tables
v Approximate size of DB data
v Database snapshot
v Application snapshot
v Buffer pool information
v The result of the LIST APPLICATIONS command
v The result of the LIST COMMAND OPTIONS command
v The result of the LIST DATABASE DIRECTORY command
v The result of the LIST INDOUBT TRANSACTIONS command
v The result of the LIST NODEGROUPS command
v The result of the LIST NODES command

v The result of the LIST ODBC DATA SOURCES command

v The result of the LIST PACKAGES/TABLES command
v The result of the LIST TABLESPACE CONTAINERS command
v The result of the LIST TABLESPACES command
v The result of the LIST DRDA IN DOUBT TRANSACTIONS command
Related reference:
v db2support - Problem Analysis and Environment Collection Tool Command
Platform specific tools

Diagnostic tools (Windows)
The following diagnostic tools are available for Windows NT, Windows 2000 and
Windows XP operating systems:
Event log, performance monitor, and other administrative tools
The Administrative Tools folder provides a variety of diagnostic
information, including access to the event log and access to performance
information.
Task Manager
The Task Manager shows all of the processes running on the Windows
server, along with details about memory usage. Use this tool to find out
which DB2 processes are running, and to diagnose performance problems.
Using this tool, you can determine memory usage, memory limits, swapper
space used, and memory leakage for a process.
To open the Task Manager, press <Ctrl> + <Alt> + <Delete>, and click
Task Manager from the available options.
Dr. Watson
The Dr. Watson utility is invoked in the event of a General Protection Fault
(GPF). It logs data that may help in diagnosing a problem, and saves this
information to a file. You must start this utility by typing drwatson on the
command line.
DB2-supplied tools
DB2 supplies administration and development tools to help you identify
DB2 problems such as the administration notification logs.
ODBC and CLI traces
CLI traces help to identify problems in CLI and ODBC applications.
SNA server tracing
If you have the SNA server installed, it provides a tracing facility. To use it,
go to the SNA Server window.
Related concepts:
v on page 0
v “Diagnostic tools (UNIX)” on page 38
Diagnostic tools (UNIX)

This section describes some essential UNIX-based commands for troubleshooting
and performance monitoring. For details on any one of these commands, precede it
with man on the command line. Use these commands to gather and process data
that can help identify the cause of a problem you are having with your

UNIX®-based system. Once the data is collected, it can be examined by someone

who is familiar with the problem, or provided to DB2® UDB customer support if
requested.
Troubleshooting commands (AIX): The following AIX system commands are

useful for DB2 troubleshooting:
errpt
The errpt command reports system errors such as hardware errors and network
failures.
v For an overview that shows one line per error, use errpt
v For a more detailed view that shows one page for each error, use errpt -a
v For errors with an error number of ″1581762B″, use errpt -a -j 1581762B
v To find out if you ran out of paging space in the past, use errpt | grep SYSVMM
v To find out if there are token ring card or disk problems, check the errpt output
for the phrases ″disk″ and ″tr0″
lsps
The lsps -a command monitors and displays how paging space is being used.
lsattr
This command displays various operating system parameters. For example, use the
following command to find out the amount of real memory on the node:
lsattr -l sys0 -E
This example lets you see the maximum number of processes per user.
xmperf
For AIX systems using Motif, this command starts a graphical monitor that collects
and displays system-related performance data. The monitor displays 3-dimensional
diagrams for each node in a single window, and is good for high-level monitoring.
However, if activity is low, the output from this monitor is of limited value.
Troubleshooting commands (UNIX): The following UNIX-based system

commands are also useful for DB2 troubleshooting. These commands are for all
UNIX-based systems, including AIX, unless otherwise noted.
df
The df command lets you see if file systems are full.

v To see how much free space is in all file systems (including mounted ones), use
df
v To see how much free space is in all file systems with names containing ″dev″,
use df | grep dev
v To see how much free space is in your home file system, use df /home
v To see how much free space is in the file system ″tmp″, use df /tmp
v To see if there is enough free space on the machine, check the output from the
following commands: df /usr , df /var , df /tmp , and df /home

truss
Available for SVR4 UNIX-based environments such as Solaris Operating

Environment, and Silicon Graphics IRIX, this command is useful for tracing system
calls in one or more processes. The truss command is not available for AIX.
pstack
Available for Solaris 2.5.1 or later, the /usr/proc/bin/pstack command displays

stack traceback information. The /usr/proc/bin directory contains other tools for
debugging processes that appear to be suspended.
Performance Monitoring Tools:
The following tools are available for monitoring the performance of your
UNIX-based system.
vmstat
This command is ideal for monitoring paging rate, which can be found under the
page in (pi) and page out (po) columns. Other important columns are the amount
of allocated virtual storage (avm) and free virtual storage (fre).
This command is useful for determining if something is suspended or just taking a

long time.
iostat
This command is useful for monitoring I/O activities. You can use the read and
write rate to estimate the amount of time required for certain SQL operations (if
they are the only activity on the system).
This command is also useful for determining if something is suspended or just

taking a long time.
netstat
This command lets you know the network traffic on each node, and the number of
error packets encountered. It is useful for isolating network problems.
system file
Available for Solaris Operating Environment, the /etc/system file contains

definitions for kernel configuration limits such as the maximum number of users
allowed on the system at a time, the maximum number of processes per user, and
the inter-process communication (IPC) limits on size and number of resources.
These limits are important because they affect DB2 performance on a Solaris
Operating Environment machine. See the Quick Beginnings information for further
details.
Commands for DB2 in a partitioned database environment (UNIX): For

multi-partitioned systems, you can run UNIX-based commands on all database
partitions by surrounding the commands with quotation marks and preceding
them with one of the following prefix commands:
db2_all

Provides information for all logical nodes. The following example shows active
applications for all logical nodes:
db2_all ";db2 LIST APPLICATIONS"
The semicolon (;) is optional and it improves performance by issuing commands

simultaneously to all nodes.
rah
Provides information for all physical nodes. This command is useful for filtering
out multiple entries that might occur using db2_all when there are multiple logical
nodes on machines. The following example lists the first three lines of the
hardware error log for each physical node:
rah ";errpt | head -3"
The semicolon (;) is optional and it improves performance by issuing commands

simultaneously to all nodes.
The following commands are also useful for multi-partitioned systems:
spmon
If using multiple nodes on RS/6000 SP systems, you may need to check if the high
performance switch (HPS) is running on all workstations.
To view the status of all nodes, use one of the following commands from the
control workstation:
v spmon -d for ASCII output
v spmon -g for a graphical user interface
Alternatively, use the command netstat -i from a node workstation to see if the
switch is down. If the switch is down, there is an asterisk (*) beside the node
name. For example:
css0* 65520 <Link>0.0.0.0.0.0
The asterisk does not appear if the switch is up.
Related reference:
v Diagnostic tools for Windows operating systems
v Administration logs, error logs and first failure data capture
Determining active process status

There are commands that can be run either within DB2 or at the operating system
level, in order to determine the status of various DB2 processes. Several of these
are demonstrated here. Use these commands to look for discrepancies between
DB2(R) processes that are running and DB2 processes that you expect to be there.
The db2_local_ps command (Linux/UNIX):
On Linux/UNIX platforms, all of the DB2 processes running under an instance can
be displayed using the db2_local_ps command:
/home/db2v8> db2_local_ps
(/home/db2v8) $ db2_local_ps

Node 1
UID PID PPID C STIME TTY TIME CMD
db2v8 41292 40012 0 Jan 19 - 0:00 db2sysc 1
db2v8 28270 41292 0 Jan 19 - 6:34 db2fcmdm 1
db2v8 37768 41292 0 Jan 19 - 0:00 db2gds 1
db2v8 38928 41292 0 Jan 19 - 0:00 db2panic (idle) 1
db2v8 39578 41292 0 Jan 19 - 0:00 db2tcpcm 1
db2v8 40284 41292 0 Jan 19 - 0:00 db2resync 1
db2v8 40910 41292 0 Jan 19 - 0:00 db2ipccm 1
root 41030 41292 0 Jan 19 - 0:00 db2ckpwd 1
db2v8 42058 41292 0 Jan 19 - 0:00 db2pdbc 1
root 43348 41292 0 Jan 19 - 0:00 db2ckpwd 1
root 43864 41292 0 Jan 19 - 0:00 db2ckpwd 1
db2v8 44380 41292 0 Jan 19 - 0:00 db2spmrsy 1
db2v8 33514 37768 0 Jan 19 - 0:00 db2srvlst 1
db2v8 44896 37768 0 Jan 19 - 0:00 db2spmlw 1
db2v8 45532 37768 0 10:12:05 - 0:03 db2cart 1
db2v8 45738 40910 0 10:12:04 - 0:00 db2agent (instance) 1
db2v8 48518 40910 0 10:13:50 - 1:20 db2agent (idle) 1
Node 2 ...
Note that no processes will be shown if the instance is stopped. Run the db2start
command if no processes are listed.
The db2stat command (Windows):
On Windows platforms, all of the DB2 processes running under an instance can be
displayed using the db2stat command:
C:\>db2stat
Environment Strings
--> DB2INSTANCE=DB2
--> DB2TEMPDIR=D:\IBM\SQLLIB\
DB2 Processes
DB2JDS 960 x3C0
DB2LICD 1004 x3EC
DB2SEC 1020 x3FC
DB2RCMD 1036 x40C
DB2SYSTRAY 1508 x5E4
DB2SYSCS 1780 x6F4
DB2FMP 1424 x590
DB2STAT 1988 x7C4
One thing to note in the Windows case is that because DB2 is thread-based (not
process-based), you will only see one process (DB2SYSCS) for all of an instance’s
EDUs. It is obvious that the same degree of information is not returned in
Windows as is returned in UNIX, but it is still useful at times to know the process
ID for a running instance. For example, you can use the Task Manager utility to
determine the CPU and memory usage for a given process ID.
The ps command is a UNIX®-based system command that returns process status

information about active processes (to standard output). Use it to look for
discrepancies between DB2® processes that are running and DB2 processes that
you expect to be there.

Flags control the types of information displayed for each active process, and may
be applied simultaneously to give a cumulative effect. You can access information
on the command syntax and flag options by using the man ps command on a
system command prompt.
Example: To show all processes of the instance ID ″svtdbm″ use: ps -fu svtdbm
The following sample shows typical output from this command. Additional
processes would appear on multi-partitioned systems.
svtdbm 1 5112 2 27894 3 0 10:48:17 4 - 0:00 db2ipccm 5

svtdbm 6135 27894 0 10:48:17 - 0:00 db2gds
svtdbm 14329 27894 0 10:48:17 - 0:00 db2resyn
svtdbm 15356 27894 0 10:48:17 - 0:00 db2snacm 6
svtdbm 18682 27894 0 10:48:17 - 0:00 db2tcpcm 6
svtdbm 27894 31989 0 10:48:17 - 0:00 db2sysc 7
svtdbm 33275 27894 0 10:48:17 - 0:00 db2tcpim 6
svtdbm 39939 18682 7 10:48:19 - 0:00 db2agent
svtdbm 56074 34761 2 10:48:37 pts/10 0:00 db2
Legend:
1. The instance ID “(return to example)”
2. The process identifier (pid) “(return to example)”
3. The parent process identifier “(return to example)”
4. The timestamp “(return to example)”
5. The name of the process. The related links for more information.
“(return to example)”
6. The communication listeners (in this sample, APPC and TCP/IP listeners and
the TCP/IP interrupt manager) are up “(return to example)”
7. The system controller process
On UNIX-based systems other than AIX® and SCO OpenServer, the db2sysc
process is the only process shown for all server-side processes (for example,
agents, loggers, page cleaners, and prefetchers). On Solaris systems, you can see
these side processes with the command /usr/ucb/ps axw. “(return to example)”
Related reference:
v Understanding the DB2 process model
Trace files
Basic trace diagnostics
If you experience a recurring and reproducible problem with DB2, tracing
sometimes allows you to capture additional information about it. Under normal
circumstances, you should only use a trace if asked to by DB2 Customer Support.
The process of taking a trace entails setting up the trace facility, reproducing the
error and collecting the data.
The amount of information gathered by a trace grows rapidly. When you take the
trace, capture only the error situation and avoid any other activities whenever
possible. When taking a trace, use the smallest scenario possible to reproduce a
problem.

The process of performing traces often has a global effect on the behavior of a DB2
instance. The degree of performance degradation is dependent on the type of
problem and on how many resources are being used to gather the trace
information.
DB2 Customer Support should provide the following information when traces are
requested:
v Simple, step by step procedures
v An explanation of where each trace is to be taken
v An explanation of what should be traced
v An explanation of why the trace is requested
v Backout procedures (i.e. how to disable all traces)
Though you should be counseled by DB2 Customer Support as to which traces to

obtain, here are some general guidelines as to when you’d be asked to obtain
particular traces:
v If the problem occurs during installation, and the default installation logs are not
sufficient to determine the cause of the problem, installation traces are
appropriate. See Tracing installation problems
v If the problem occurs in one of the GUI (Graphical User Interface) tools, and the
same actions succeed when performed via explicit commands in the DB2
command window, then a Control Center trace is appropriate. Note that this will
only capture problems with tools that can be launched from the Control Center.
v If the problem manifests in a CLI application, and the problem cannot be
recreated outside of the application, then a CLI trace is appropriate.
v If the problem manifests in a JDBC application, and the problem cannot be
recreated outside of the application, then a JDBC trace is appropriate.
v If the problem is directly related to information that is being communicated at
the DRDA layer, a DRDA trace is appropriate.
v For all other situations where a trace is feasible, a DB2 trace is most likely to be
appropriate.
Trace information is not always helpful in diagnosing an error. For example, it may
not capture the error condition in the following situations:
v The trace buffer size you specified was not large enough to hold a complete set
of trace events, and useful information was lost when the trace stopped writing
to the file or wrapped.
v The traced scenario did not re-create the error situation.
v The error situation was re-created, but the assumption as to where the problem
occurred was incorrect. For example, the trace was collected at a client
workstation while the actual error occurred on a server.
DB2 trace files

The db2trc command controls the trace facility provided with DB2. The trace
facility records information about operations and formats this information into a
readable form.
Please keep in mind that there is added overhead when a trace is running so
enabling the trace facility may impact your system’s performance.

In general, DB2 Support and development teams use DB2 traces to debug
customer problems. You might run a trace to gain further information about a
problem that you are investigating, but its use is rather limited without knowledge
of the DB2 source code.
Nonetheless, it is important to know how to correctly turn on tracing and how to

dump trace files, just in case you need to send one to the DB2 support and
development teams.
Note: You will need one of SYSADM, SYSCTRL or SYSMAINT authority to use
db2trc
Command syntax:
To get a general idea of the options available, execute the db2trc command without
any parameters:
C:\>db2trc Usage: db2trc (chg|clr|dmp|flw|fmt|inf|off|on) options
Command parameters:
chg|change
Change the trace mask, maxSysErrors or maxRecordSize
clr|clear
Clear the trace
dmp|dump
Dump the trace to a binary trace file
flw|flow
Show control flow of the trace
fmt|format
Format the trace
inf|info|information
Get information on the trace off Turn the trace off on Turn the trace on
For more information about a specific db2trc command parameter, use the -u
option. For example, to see more information about turning the trace on, execute
the following command:
db2trc on -u
This will provide information about all of the additional options (labeled as
″facilities″) that can be specified when turning on a DB2 trace.
The most important option that you need to be aware for turning on trace is -L.
This specifies the size of the memory buffer that will be used to store the
information being traced. The buffer size can be specified in either bytes or
megabytes. (To specify megabytes append either ″M″ or ″m″ after the value). The
trace buffer size must be a power of two megabytes. If you specify a size that does
not meet this requirement, the buffer size will automatically be rounded down to
the nearest power of two.
If the buffer is too small, information might be lost. By default only the most
recent trace information is kept if the buffer becomes full. If the buffer is too large,
it might be difficult to send the file to the DB2 support team.

If tracing an operation that is relatively short (such as a database connection), a

size of approximately 8MB might be sufficient:
C:\> db2trc on -l 8M Trace is turned on
However, if you are tracing a larger operation or if a lot of work is going on at the
same time, a larger trace buffer may be required.
On most platforms, tracing can be turned on at any time and works as described
above. However, there are certain situations to be aware of:
1. On multiple database partition systems, you must run a trace for each physical
(as opposed to logical) database partition.
2. On Solaris platforms, if the trace is turned off after the instance has been
started, a very small buffer will be used regardless of the size specified. To
effectively run a trace on Solaris, turn the trace on before starting the instance
and ″clear″ it as necessary afterwards.
DB2 trace memory buffers: The most important option that you need to be aware
for turning on trace is -l. This specifies the size of the memory buffer that will be
used to store the information being traced. The buffer size can be specified in
either bytes or megabytes. (To specify megabytes append either ″M″ or ″m″ after
the value). The trace buffer size must be a power of two megabytes. If you specify
a size that does not meet this requirement, the buffer size will automatically be
rounded down to the nearest power of two.
If the buffer is too small, information might be lost. By default only the most
recent trace information is kept if the buffer becomes full. If the buffer is too large,
it might be difficult to send the file to the DB2 support team.
If tracing an operation that is relatively short (such as a database connection), a

size of approximately 8MB might be sufficient:
C:\> db2trc on -l 8M
Trace is turned on
However, if you are tracing a larger operation or if a lot of work is going on at the
same time, a larger trace buffer may be required.
On most platforms, tracing can be turned on at any time and works as described
above. However, there are certain situations to be aware of:
1. On multiple database partition systems, you must run a trace for each physical
(as opposed to logical) database partition.
2. On Solaris platforms, if the trace is turned on after the instance has been
started, a very small buffer will be used regardless of the size specified. To
effectively run a trace on Solaris, turn the trace on before starting the instance
and ″clear″ it as necessary afterwards.
Dumping a trace file: Once the trace facility has been enabled using the on
option, all subsequent work done by the instance will be traced.
While the trace is running, you can use the clr option to clear out the trace buffer.
All existing information in the trace buffer will be removed.
C:\>db2trc clr
Trace has been cleared
Once the operation being traced has finished, use the dmp option followed by a
trace file name to dump the memory buffer to disk. For example:

C:\>db2trc dmp trace.dmp

Trace has been dumped to file
The trace facility will continue to run after dumping the trace buffer to disk. To
turn tracing off, use the off option:
C:\>db2trc off
Trace is turned off
Formatting a DB2 trace file: The dump file created above is in a binary format
that is not readable.
To verify that a trace file can be read, format the binary trace file to show the flow
control and send the formatted output to a null device. The following example
shows the command to perform this task:
db2trc flw example.trc nul
The output for this command will explicitly tell you if there is a problem reading
the file, and whether or not the trace was wrapped.
At this point, the dump file could be sent to DB2 Support. They would then format
it based on your DB2 service level. However, you may sometimes be asked to
format the dump file into ASCII format before sending it. This is accomplished via
the flw and fmt options. You must provide the binary dump file along with the
name of the ASCII file that you want to create:
C:\>db2trc flw trace.dmp trace.flw
C:\Temp>db2trc flw trace.dmp trace.flw
Total number of trace records : 18854
Trace truncated : NO
Trace wrapped : NO
Number of trace records formatted : 1513 (pid: 2196 tid 2148 node: -1)
Number of trace records formatted : 100 (pid: 1568 tid 1304 node: 0)
...
C:\>db2trc fmt trace.dmp trace.fmt

C:\Temp>db2trc fmt trace.dmp trace.fmt
Trace truncated : NO
Trace wrapped : NO
Total number of trace records : 18854
Number of trace records formatted : 18854
If this output indicates ″Trace wrapped″ is ″YES″, then this means that the trace
buffer was not large enough to contain all of the information collected during the
trace period. A wrapped trace might be okay depending on the situation. If you
are interested in the most recent information (this is the default information that is
maintained, unless the -L option was used to specify otherwise), then what is in
the trace file might be sufficient. However, if you are interested in what happened
at the beginning of the trace period or if you are interested in everything that
occurred, you might want to redo the operation with a larger trace buffer.
Another thing to be aware of is that on UNIX platforms, DB2 will automatically

dump the trace buffer to disk when it shuts the instance down due to a severe
error. Thus if tracing is enabled when an instance ends abnormally, a file will be
created in the diagnostic directory and its name will be db2trdmp.###, where ### is
the node number. This does not occur on Windows platforms. You have to dump
the trace manually in those situations.
To summarize, the following is an example of a common use of db2trc:

db2trc on -l 8M
db2trc clr
<execute command>
db2trc dump db2trc.dmp
db2trc off
db2trc flw db2trc.dmp <filename>.flw
db2trc fmt db2trc.dmp <filename>.fmt
db2trc fmt -c db2trc.dmp .fmtc
DRDA trace files

Before analyzing DRDA traces, you need to understand that DRDA is an open
standard for definition of data and communication structures. i.e. DRDA comprises
a set of rules about how data should be organized for transmission and how
communication of that information should occur. These rules are defined in the
following reference manuals:
v DRDA V3 Vol. 1: Distributed Relational Database Architecture
v DRDA V3 Vol. 2: Formatted Data Object Content Architecture
v DRDA V3 Vol. 3: Distributed Data Management Architecture
PDF versions of these manuals are available on www.opengroup.com.
The db2drdat utility records the data interchanged between the DB2 Connect
server (on behalf of the database client) and the host or iSeries™ database server.
As a database administrator (or application developer), you may find it useful to

understand how this flow of data works, because this knowledge can help you
determine the origin of a particular problem. For example, if you issue a
CONNECT TO database statement for a host or iSeries database server, but the
command fails and you receive an unsuccessful return code. If you understand
exactly what information was conveyed to the host or iSeries database server
management system, you may be able to determine the cause of the failure even if
the return code information is general. Many failures are caused by simple user
errors.
Command syntax:
To get a general idea of the options available, issue the following command:
db2drdat -help
This will show the command syntax:

db2drdat on [-i] [-c] [-r] [-s] [-l=LENGTH]
db2drdat off [-t=TRACEFILE] [-p=PROCESSID]
Command parameters:
on Turns the DRDA trace on
off Turns the DRDA trace off
-i Include timestamps
-c Trace SQLCAs
-r Trace DRDA receive buffers
-s Trace DRDA send buffers
-l=LENGTH
Sets the size of the trace buffer (in bytes). The default value is one
megabyte.

-t=TRACEFILE
TRACEFILE is the file name where the captured trace will be stored. The
default file name is db2drdat.dmp.
-p=PROCESSID
PROCESSID is the ID of a process to be traced. If PROCESSID is not
specified, all process IDs for the DB2 instance are traced.
Usage:
The following is an example of a common use of db2drdat:

db2drdat on
<Execute command>
db2drdat off
<Collect the trace file db2drdat.dmp>
Output from db2drdat lists the data streams exchanged between the DB2 Connect
workstation and the host or iSeries database server management system. Data sent
to the host or iSeries database server is labeled SEND BUFFER and data received
from the host or iSeries database server is labeled RECEIVE BUFFER.
If a receive buffer contains SQLCA information, it will be followed by a formatted

interpretation of this data and labeled SQLCA. The SQLCODE field of an SQLCA
is the unmapped value as returned by the host or iSeries database server. The send
and receive buffers are arranged from the oldest to the most recent within the file.
Each buffer has:
v The process ID
v A SEND BUFFER, RECEIVE BUFFER, or SQLCA label. The first DDM command
or object in a buffer is labeled DSS TYPE.
The remaining data in send and receive buffers is divided into five columns,
consisting of:
v A byte count.
v Columns 2 and 3 represent the DRDA(R) data stream exchanged between the
two systems, in ASCII or EBCDIC.
v An ASCII representation of columns 2 and 3.
v An EBCDIC representation of columns 2 and 3.
Here is an example of a trace entry in a db2drda trace file:

2 data DB2 DRDA Communication Manager sqljcSend fnc (3.3.54.5.0.1177)
pid 102974 tid 1 cpid -1 node 0 probe 1177
bytes 464
SEND BUFFER(AR):
EXCSAT RQSDSS (ASCII) (EBCDIC)

0 1 2 3 4 5 6 7 8 9 A B C D E F 0123456789ABCDEF 0123456789ABCDEF
0000 00B4D041000100AE 1041006E115E8482 ...A.....A.n.^.. ..}........>.;db
0010 F282974040404040 4040404040404040 ...@@@@@@@@@@@@@ 2bp
0020 4040F0F0F0F1F9F2 F3C5F0F0F0000000 @@.............. 0001923E000...
0030 0000000000000000 0000000000000000 ................ ................
0040 0000000000000000 000000000060F0F0 .............`.. .............-00
0050 F0F1A2A495404040 4040404040404040 .....@@@@@@@@@@@ 01sun
0060 4040404040404040 4040404040404040 @@@@@@@@@@@@@@@@
0070 C4C4C2F2C8D4E340 0018140414030007 .......@........ DDB2HMT........
0080 2407000714740005 240F000714400007 $....t..$....@.. ...............

0090 000D1147D8C4C2F2 61F6F0F0F0000B11 ...G....a....... ....QDB2/6000...

00A0 6D99818398A485A3 000C115AE2D8D3F0 m..........Z.... _racquet...]SQL0
00B0 F8F0F2F0 .... 8020
For more information, see the DB2 for OS/390 Reference for Remote DRDA Requesters
and Servers, the Distributed Relational Database Reference, and the Distributed Data
Management Architecture Level 3: Reference.
Related reference:
v db2drdat - DRDA Trace Command
JDBC trace files

There are two different methods for capturing JDBC traces, depending on the
which type of JDBC driver your application uses.
Enabling JDBC traces: The JDBC trace is enabled by adding specific entries to the
db2cli.ini file.
Note: There are lots of keywords that can be added to the db2cli.ini file that can
affect application behavior. These keywords can resolve or be the cause of
application problems. There are also some keywords that are not covered in
the CLI documentation. Those are only available from DB2 Service and
Support. If you have keywords in your db2cli.ini file that are not
documented, it is likely that they were a recommendation from the DB2
support team.
By default, the location of the DB2™ CLI/ODBC configuration keyword file is in

the sqllib directory on Window platforms, and in the sqllib/cfg directory of the
database instance running the CLI/ODBC applications on UNIX™ platforms. If the
ODBC Driver Manager is used to configure a User Data Source on the Windows™
platform, a db2cli.ini may be created in the user’s home (profile) directory. The
environment variable DB2CLIINIPATH can also be used to override the default
and specify a different location for the file.
Step 1. Create a path for the trace files.
It is important to create a path that every user can write to. For example,
on Windows:
mkdir c:\temp\trace
On UNIX:
mkdir /tmp/trace
chmod 777 /tmp/trace
Step 2. Update the CLI configuration keywords
This can be done by either (A) manually editing the db2cli.ini file
or(B)using the UPDATE CLI CFG command.
Option A: Manually Editing the db2cli.ini file.
a. Open up the db2cli.ini file in a plain text editor.
b. Add the following section to the file (or if the COMMON section
already exists, just append the variables):
[COMMON]
JDBCTrace=1
JDBCTracePathName=<path>
JDBCTraceFlush=1
where <path> is, for example, C:\temp\trace on Windows, or

/tmp/trace on UNIX platforms.

c. Save the file with at least one blank line at the end of the file. (This
prevents some parsing errors.)
Option B: Using UPDATE CLI CFG commands to update the db2cli.ini
file. Issue the following commands:
db2 UPDATE CLI CFG FOR SECTION COMMON USING JDBCTrace 1
db2 UPDATE CLI CFG FOR SECTION COMMON USING JDBCTracePathName <path>
where <path> is, for example, C:\temp\trace on Windows, or /tmp/trace
on UNIX platforms.
db2 UPDATE CLI CFG FOR SECTION COMMON USING JDBCTraceFlush 1
Step 3. Confirm the db2cli.ini configuration
Issue the following command to verify that the correct keywords are set
and being picked up:
db2 GET CLI CFG FOR SECTION COMMON
Step 4. Restart the application
The db2cli.ini file is only read when the application starts, therefore, for
any changes to take effect, the application must be restarted.
If tracing a JDBC stored procedure, this means restarting the DB2 instance.
Step 5. Capture the error
Run the application until the error is generated, then terminate the
application. If it is possible, reduce the situation, such that the only JDBC
applications that are running at the time of trace are those related to the
problem recreation. This makes for much clearer trace files.
Step 6. Disable the JDBC trace
Set the JDBCTrace=0 keyword in the [COMMON] section of the db2cli.ini
manually, or issue:
db2 UPDATE CLI CFG FOR SECTION COMMON USING Trace 0
db2 UPDATE CLI CFG FOR SECTION COMMON USING JDBCTrace 0
Restart any applications that are running and tracing.
Step 7. Collect the trace files
The JDBC trace files will be written to the path specified in the
JDBCTracePathName keyword. The filenames generated will all end with
a .trc extension. All files generated in the trace path at the time of the
problem recreation are required.
When you use the trace facility to diagnose application issues, keep in mind that it
does have an impact on application performance and that it affects all applications,
not only your test application. This is why it is important to remember to turn it
off after the problem has been identified.
Enabling DB2 Universal JDBC traces: Option A:
If you use the DataSource interface to connect to a data source, then use the
DataSource.setTraceLevel() and DataSource.setTraceFile() method to enable tracing.
Option B:
If you use the DriverManager interface to connect to a data source, the easiest way
to enable tracing will be to set the logWriter on DriverManager before obtaining a
connection.
For example:
DriverManager.setLogWriter(new PrintWriter(new FileOutputStream("trace.txt")));

Option C:
If you are using the DriverManager interface, you can alternatively specify the
traceFile and traceLevel properties as part of the URL when you load the driver.
For example:
String databaseURL = "jdbc:db2://hal:50000/sample:traceFile=c:/temp/foobar.txt;" ;
More details on each of these options are provided in JDBC and SQLJ problem
diagnosis with the DB2 Universal JDBC Driver.
Refer to one of the following:

v Obtaining traces of applications that use CLI-based Legacy Type 2 JDBC Driver,
or
v Obtaining traces of applications that use the DB2 Universal JDBC Driver
Obtaining traces of applications that use CLI-based Legacy Type 2 JDBC

Driver: The CLI-based Legacy Type 2 JDBC Driver for Linux, UNIX™, and
Windows™ offers comprehensive tracing facilities. By default, these facilities are
disabled and use no additional computing resources. When enabled, the trace
facilities generate one or more text log files whenever an application accesses the
driver). These log files provide detailed information about:
v the order in which JDBC functions were called by the application
v the contents of input and output parameters passed to and received from JDBC
functions
v the return codes and any error or warning messages generated by JDBC
functions
The CLI trace offers very little information about the internal workings of the DB2
CLI driver.
This type of trace is applicable for situations where a problem is encountered in:
v a JDBC application which uses the CLI-based Legacy Type 2 JDBC driver
v DB2 JDBC stored procedures1
Note: Internally, the CLI-based Legacy Type 2 JDBC Driver makes use of the DB2
CLI driver for database access. For example, the Java getConnection()
method is internally mapped by the CLI-based Legacy Type 2 JDBC Driver
to the DB2 CLI SQLConnect() function. As a result, Java developers might
find a DB2 CLI trace to be a useful complement to the DB2 JDBC trace. See
CLI Traces.
Related reference:
v CLI/ODBC/JDBC trace facility
v CLI and JDBC trace files
1. Most of the DB2 CLI trace information and instructions which follow are generic and apply to both applications and stored
procedures equally. However, unlike applications which are clients of a database server (and typically execute on a machine
separate from the database server), stored procedures execute at the database server. Therefore, the following additional steps
must be taken when tracing DB2 CLI stored procedures:
– Ensure the trace keyword options are specified in the db2cli.ini file located at the DB2 server.
– Ensure all keywords are configured correctly prior to database startup time (that is, when the db2start command is issued).
Changing trace settings while the database server is running may have unpredictable results.

v db2cli.ini initialization file
Obtaining traces of applications that use the DB2 Universal JDBC Driver: To
obtain data for diagnosing SQLJ or JDBC problems with the DB2 Universal JDBC
Driver, collect trace data and run utilities that format the trace data. You should
run the trace and diagnostic utilities only under the direction of IBM(R) software
support.
Related reference:
v JDBC and SQLJ problem diagnosis with the DB2 Universal JDBC Driver
v Example of tracing under the DB2 Universal JDBC Driver
First failure data capture

First-failure data capture (FFDC) is a general term applied to the set of diagnostic
information that DB2(R) captures automatically when errors occur. This
information reduces the need to reproduce errors to get diagnostic information.
FFDC information can be found in the following files:
Administration Notification Logs:
When significant events occur, DB2 writes information to the administration

notification log. The information is intended for use by database and system
administrators. Many notification messages provide additional information to
supplement the SQLCODE that is provided. The type of event and the level of
detail of the information gathered are determined by the NOTIFYLEVEL
configuration parameter.
db2diag.log:
Diagnostic information about errors is recorded in this text log file. This
information is used for problem determination and is intended for DB2 customer
support. The level of detail of the information is determined by the DIAGLEVEL
db2dasdiag.log:
Diagnostic information about errors encountered specifically in the DAS (DB2

Administration Server) is recorded in this text file.
Dump files:
For some error conditions, extra information is logged in external binary files
named after the failing process ID. These files are intended for use by DB2
customer support.
Trap files:
The database manager generates a trap file if it cannot continue processing because
of a trap, segmentation violation, or exception.
Core files (UNIX only):

When DB2 terminates abnormally, the operating system generates a core file. The
core file is a binary file that contains information similar to the DB2 trap files. Core
files may also contain the entire memory image of the terminated process.
Related reference:
v Interpreting the administration and error logs
v Dump files
v Trap files
v Interpreting platform specific error logs
v notifylevel - Notify level configuration parameter
v diaglevel - Diagnostic error capture level configuration parameter
First failure data capture locations

The DIAGPATH variable, specified in the database manager configuration, gives
the fully qualified path to the first failure data capture (FFDC) storage directory.
The default value for DIAGPATH is a null string. It is recommended that you keep
this default value. If you choose to change the value, it is recommended that you
use a centralized location, especially if there are multiple database instances.
To change the DIAGPATH value, use the following command:

DB2 UPDATE DBM CFG USING DIAGPATH <path>
The default value of NULL means that the first failure data capture (FFDC)
information is placed in the following locations:
v For Windows® systems:
– If the DB2INSTPROF environment variable is not set: db2path\db2instance
(where db2path is the path referenced in the DB2PATH environment variable,
and db2instance is the environment variable containing the ID of the instance
owner).
– If the DB2INSTPROF environment variable is set: x:\db2instprof\db2instance,
where x is the drive referenced in the DB2PATH environment variable,
db2instprof is the instance profile directory, and db2instance is the environment
variable containing the ID of the instance owner.
Note: On Windows NT®, Windows 2000 and Windows XP systems, the DB2
administration notification log is found in the event log and can be
reviewed through the Windows Event Viewer. On other operating
systems, the administration notification log for the instance is called
instance_name.nfy. The diagnostics log (db2diag.log) is still located in the
DIAGPATH path.
v For UNIX® operating systems:
– $HOME/sqllib/db2dump, where $HOME is the home directory of the instance
owner.
We recommend that you clean out the DIAGPATH directory periodically to keep it
from becoming too large.
Related reference:
v diagpath - Diagnostic data directory path configuration parameter

Setting the error capture level for the administration notification

log file
The information that DB2 records in the administration notification log is
determined by the NOTIFYLEVEL setting. To check the current setting, issue:
DB2 GET DBM CFG
Look for the following variable:

Notify Level (NOTIFYLEVEL) = 3
The meaning of the different notification levels is defined in notifylevel - Notify

level configuration parameter.
To alter the setting, use the command:
DB2 UPDATE DBM CFG USING NOTIFYLEVEL X
where X is the desired notification level.
Interpreting administration notification log file entries

Use a text editor to view the administration notification log file on the machine
where you suspect a problem to have occurred. The most recent events recorded
are the furthest down the file. Generally, each entry contains the following parts:
v A timestamp
v The location reporting the error. Application identifiers allow you to match up
entries pertaining to an application on the logs of servers and clients.
v A diagnostic message (usually beginning with ″DIA″ or ″ADM″) explaining the
error.
v Any available supporting data, such as SQLCA data structures and pointers to
the location of any extra dump or trap files.
If the database is behaving normally, this type of information is not important and
can be ignored.
v The Administration logs grow continuously . When they get too large, back
them up and then erase the file. A new set of files is generated automatically the
next time they are required by the system.
The following example shows the header information for a sample log entry, with
all the parts of the log identified.
Note: Not every log entry will contain all of these parts.
2002-02-05-03.14.39.020766 ▌1▐ Instance:db2inst1 ▌2▐ Node:000 ▌3▐
PID:89198(db2agent (MYDB )) ▌4▐ Appid:*LOCAL.db2inst1.020205091435 ▌5▐
recovery manager ▌6▐ sqlpresr ▌7▐ Probe:1 ▌8▐ Database:MYDB ▌9▐
ADM1530E ▌10▐ Crash recovery has been initiated. ▌11▐
Legend:
1. A timestamp for the message.
2. The name of the instance generating the message.
3. For multi–partition systems, the partition generating the message. (In a
non-partitioned database, the value is ″000″.)>
4. The DB2 component that is writing the message. For messages written by user
applications using the db2AdminMsgWrite API, the component will read “User
Application”.

5. Identification of the application for which the process is working. In this

example, the process generating the message is working on behalf of an
application with the ID *LOCAL.db2inst1.020205091435.
To identify more about a particular application ID, either:
v Use the db2 list applications command on a DB2 UDB server or db2 list
dcs applications on a DB2 UDB Connect gateway to view a list of
application IDs. From this list, you can determine information about the
client experiencing the error, such as its node name and its TCP/IP address.
v Use the db2 get snapshot for application command to view a list of
application IDs.
6. The DB2 component that is writing the message. For messages written by user
applications using the db2AdminMsgWrite API, the component will read “User
Application”.
7. The name of the function that is providing the message. This function
operates within the DB2 subcomponent that is writing the message. For
messages written by user applications using the db2AdminMsgWrite API, the
function will read “User Function”.
To find out more about the type of activity performed by a function, look at
the fourth letter of its name. In this example, the letter ″p″ in the function
″sqlpresr″ indicates a data protection problem. (Logs could be damaged, for
example.)
The following list shows some of the letters used in the fourth position of the
function name, and the type of activity they identify:
b Buffer pools
c Communication between clients and servers
d Data management
e Engine processes
o Operating system calls (such as opening and closing files)
p Data protection (such as locking and logging)
r Relational database services
s Sorting
x Indexing
8. Unique internal identifier. This number allows DB2 customer support and
development to locate the point in the DB2 source code that reported the
message.
9. The database on which the error occurred.
10. When available, a message indicating the error type and number as a
hexadecimal code.
11. When available, message text explaining the logged event.
Example Log Files
Example 1
A power outage causes your DB2 server machine to reboot. While rebooting, some
of the file systems do not remount properly.
You want to check on your database’s integrity after the outage, so you start the
instance and connect to the database. The connection is successful. In the
administration notification log, you see the following entries:

2002-02-05-03.14.33.576365 Instance:db2inst1 Node:000

PID:140546(db2star2) Appid:none
base sys utilities startdbm Probe:911
ADM7513W Database manager has started.

PID:89198(db2agent (MYDB )) Appid:*LOCAL.db2inst1.020205091435
buffer pool services sqlbStartPoolsErrorHandling Probe:39
ADM6080E The tablespace "TS1" (ID "3"), was put OFFLINE and in
ROLLFORWARD_PENDING. Tablespace state is 0x"00004080". ▌1▐
*
recovery manager sqlpresr Probe:1 Database:MYDB ▌2▐
ADM1530E Crash recovery has been initiated. ▌3▐
*
recovery manager sqlpresr Probe:350 Database:MYDB
ADM1533W Database has recovered. However, one or more tablespaces are
offline. ▌4▐
*
recovery manager sqlpresr Probe:370 Database:MYDB
ADM1531E Crash recovery has completed successfully. ▌5▐
Legend:
1. Table space ″TS1″ has been put offline and is marked roll-forward pending.
This has probably happened because TS1 was located on one of the file systems
that did not get remounted during the machine’s reboot.
2. The name of the database being recovered was MYDB.
3. A notification that database crash recovery has initialized.
4. Crash recovery indicates that the database was recovered but one or more table
spaces were not recovered because they were offline. The table space that was
offline is indicated in an earlier message (1).
5. A notification that the database crash recovery was successful.
From here, the db2 list tablespace containers command can identify the file
systems associated with table space TS1. Your system administrator can put the
table space back online by remounting the file systems. You can then roll forward
the tablespace.
Example 2
While attempting to create a new table, the following SQLCODE is returned:

SQL0968C The file system is full. SQLSTATE=57011
The SQLCODE documentation suggests that you examine the administration

notification for additional details. Looking in the administration notification log,
you see the following message:
PID:88170(db2agent (FOO )) Appid:*LOCAL.db2inst1.020205182029
buffer pool services sqlbSMSDirectWrite Probe:99 Database:FOO
ADM6017E Tablespace "SYSCATSPACE" (ID "0") is full. Detected on container
"/home/db2inst1/db2inst1/NODE0000/SQL00001/SQLT0000.0" (ID "0"). Please check
the underlying file system space or user limits.

The message indicates the container on which the file system error occurred. Using
this information, ensure that there is enough file space on the container. If the file
space is sufficient, then ask your system administrator to investigate any system
resource limits. For example, on UNIX®, the ulimit command can restrict the
maximum file size that a user can access.
Related reference:
v Location of logs and files
v notifylevel - Notify level configuration parameter.
Setting the diagnostic log file error capture level

The DB2 diagnostic log is a file that contains text information logged by DB2(R).
This information is used for problem determination and is intended for DB2
customer support. See Location of logs and files for details on its location.
The information that DB2 records in the db2diag.log is determined by the

DIAGLEVEL setting. To check the current setting, issue:
DB2 GET DBM CFG
Look for the following variable:

Diagnostic error capture level (DIAGLEVEL) = 3
The meaning of the different notification levels is defined in diaglevel - Diagnostic

error capture level configuration parameter
To alter the setting, use the command:

DB2 UPDATE DBM CFG USING DIAGLEVEL X
where X is the desired notification level.
Note: If you are diagnosing a problem that can be reproduced, it is recommended

that you use DIAGLEVEL 4 while performing the problem determination.
Interpreting diagnostic log file entries

Use a text editor to view the diagnostic log file on the machine where you suspect
a problem to have occurred. The most recent events recorded are the furthest
down the file.
Note: The Administration logs grow continuously. When they get too large, back
them up and then erase the file. A new set of files is generated automatically
the next time they are required by the system.
The following example shows the header information for a sample log entry, with
all the parts of the log identified.
Note: Not every log entry will contain all of these parts. Only the first several
fields (timestamp to TID) and FUNCTION will be present in all the
db2diag.log records.
2005-01-20-00.28.11.406000-480 1 I2841H435 2 LEVEL: Error 3
PID : 740 4 TID : 2640 5 PROC : db2syscs.exe 6
INSTANCE: DB2 7 NODE : 000 8 DB : SAMPLE 9
APPHDL : 0-7 10 APPID: *LOCAL.DB2.050120082811 11
FUNCTION: 12 DB2 UDB, data protection, sqlpsize, probe:20
RETCODE : 13 ZRC=0x860F000A=-2045837302=SQLO_FNEX "File not found."
DIA8411C A file "" could not be found.

Legend:
1. A timestamp and timezone for the message.
2. The record ID field. The db2diag.log’s recordID specifies the file offset at
which the current message is being logged (e.g. ″2841″) and the message
length (e.g. ″435″) for the platform where the DB2 diagnostic log was
created.
3. The diagnostic level associated with an error message. e.g. Info, Warning,
Error, Severe, or Event
4. The process ID
5. The thread ID
6. The process name
7. The name of the instance generating the message.
8. For multi-partition systems, the partition generating the message. (In a
non-partitioned database, the value is ″000″.)
9. The database name.
10. The application handle. This value aligns with that used in db2pd output
and lock dump files. It consists of the coordinator partition number
followed by the coordinator index number, separated by a dash.
11. Identification of the application for which the process is working. In this
example, the process generating the message is working on behalf of an
application with the ID *LOCAL.db2inst1.020205091435.
A TCP/IP-generated application ID is composed of three sections
1. IP address: It is represented as a 32-bit number displayed as a
maximum of 8 hexadecimal characters.
2. Port number: It is represented as 4 hexadecimal characters.
3. A unique identifier for the instance of this application.
Note: When the hexadecimal versions of the IP address or port number

begin with 0-9, they are changed to G-P respectively. For example,
″0″ is mapped to ″G″, ″1″ is mapped to ″H″, and so on. The IP
address, AC10150C.NA04.006D07064947 is interpreted as follows:
The IP address remains AC10150C, which translates to 172.16.21.12.
The port number is NA04. The first character is ″N″, which maps to
″7″. Therefore, the hexadecimal form of the port number is 7A04,
which translates to 31236 in decimal form.
To identify more about a particular application ID, either:
v Use the db2 list applications command on a DB2 UDB server or db2
list dcs applications on a DB2 UDB Connect gateway to view a list of
application IDs. From this list, you can determine information about the
client experiencing the error, such as its node name and its TCP/IP
address.
v Use the db2 get snapshot for application command to view a list of
application IDs.
v Use the db2pd -applications -db sample command.
12. The product name (″DB2 UDB″), component name (″data protection″), and
function name (″sqlpsize ″) that is writing the message (as well as the
probe point (″20″) within the function).

13. The return code (if any) returned by a called function. This field consists of
a type (″ZRC″), the return code value, and the corresponding error
description.
Note: Timestamps in the db2diag.log contain a time zone in DB2 version 8.2. For
example: 2005-01-20-00.28.11.406000-480, where ″-480″ is the difference
between UTC (Coordinated Universal Time, formerly known as GMT) and
local time at the application server in minutes. Thus -480 represents
UTC - 8 hours, i.e. PST (Pacific Standard Time).
Now that you have seen a sample db2diag.log entry, here is a list of all of the
possible fields:
<timestamp><timezone> <recordID> LEVEL: <level> (<source>)
PID : <pid> TID : <tid> PROC : <procName>
INSTANCE: <instance> NODE : <node> DB : <database>
APPHDL : <appHandle> APPID: <appID>
FUNCTION: <prodName>, <compName>, <funcName>, probe:<probeNum>
MESSAGE : <messageID> <msgText>
CALLED : <prodName>, <compName>, <funcName> OSERR: <errorName> (<errno>)
RETCODE : <type>=<retCode> <errorDesc>
ARG #N : <typeTitle>, <typeName>, <size> bytes
... argument ...
DATA #N : <typeTitle>, <typeName>, <size> bytes
... data ...
Interpreting the db2diag.log file informational record

The first message in db2diag.log should always be an informational record.
The only exception is the case when the first message in a log file is produced by a
component using ossLog API calls (e.g. the DAS, genreg, etc) in which case there
will not be an informational record output.
An example of an informational record is as follows:

2004-08-08-19.43.54.155770-240 I1A1044 LEVEL: Event
PID : 5710078 TID : 1 PROC : gtf
INSTANCE: db2inst1 NODE : 000
FUNCTION: DB2 UDB, RAS/PD component, _pdlogInt, probe:120
START : New db2diag.log file
DATA #1 : Build Level, 144 bytes
Instance "db2inst1" uses "64" bits and DB2 code release "SQL08020"
with level identifier "03010106".
Informational tokens are "DB2 v8.1.1.64", "n040805", "U498350", FixPak "7".
DATA #2 : System Info, 192 bytes
System: AIX titanic_1 2 5 0029317A4C00
CPU: total:32 online:8
Physical Memory(MB): total:32000 free:60
Virtual Memory(MB): total:71936 free:38311
Swap Memory(MB): total:39936 free:38251
Kernel Params: msgMaxMessageSize:4194304 msgMaxQueueSize:4194304
shmMax:68719476736 shmMin:1 shmIDs:131072 shmSegments:0 semIDs:131072
semNumPerID:65535 semOps:1024 semMaxVal:32767 semAdjustOnExit:16384
Information in this record is only valid at the time when this file
was created (refer to this record’s time stamp for that information).
The Informational record is output for ″db2start″ on every logical partition. This
results in multiple informational records: one per logical partition. Since the
informational record contains memory values which are different on every
partition, this information may be useful.

Interpreting an SQL structure in the db2diag.log file

For severe errors, an SQLCA structure is dumped into the db2diag.log. For details
on the SQLCA fields, see the related topics below.
The following diagram provides an example of a diagnostics log with an SQLCA

dump.
2003-11-04-12.21.16.559205-300 I381956-688 LEVEL: Severe
PID : 47918 TID : 1 PROC : db2agent (SAMPLE)
INSTANCE: payroll NODE : 000 DB : SAMPLE
APPHDL : 0-25 APPID: *LOCAL.payroll.0E9CB4172114
DATA #1 : SQLCA, PD_DB2_TYPE_SQLCA, 136 bytes 1
sqlcaid : SQLCA sqlcabc: 136 sqlcode: -980 2 sqlerrml: 0
sqlerrmc: 3
sqlerrp : sqlrita
sqlerrd 4 : (1) 0xFFFFE101 5 (2) 0x00000000 (3) 0x00000000
(4) 0x00000000 (5) 0x00000000 (6) 0x00000000
sqlwarn : (1) (2) (3) (4) (5) (6)
(7) (8) (9) (10) (11)
sqlstate:
Legend:
1. Beginning of the SQLCA entry.
2. The SQL state (when negative, an error has occurred).
3. Any reason codes associated with the SQL error code.
4. Sometimes there are several errors leading to the final SQL error code.
These errors are shown in sequence in the sqlerrd area and may be
described by calling db2diag -rc error_code if they are ZRC or ECF errors
(i.e. start with 0x8 or 0x9).
The hexadecimal representation of an SQL error.
Related reference:
v SQLCA (SQL communications area)
v db2diag - Analyzing db2diag.log files
v diaglevel - Diagnostic error capture level configuration parameter
Dump files
Dump files are created when an error occurs for which there is additional
information that would be useful in diagnosing a problem (such as internal control
blocks). Every data item written to the dump files has a timestamp associated with
it to help with problem determination. Dump files are in binary format and are
intended for DB2® customer support representatives.
When a dump file is created or appended, an entry is made in the db2diag.log

indicating the time and the type of data written. These db2diag.log entries
resemble the following:

2004-12-05-06.26.55.664042-360 I5969607C424 LEVEL: Severe

PID : 646 TID : 1 PROC : db2agent (instance) 0
INSTANCE: db2inst1 NODE : 000 DB : SAMPLE
APPHDL : 0-9 APPID: * LOCAL.DB2.050120082811
FUNCTION: DB2 UDB, DRDA Application Server, sqljsSignalHandler, probe:10
MESSAGE : DIA0505I Execution of a component signal handling function has begun.
2004-12-05-06.26.55.696441-360 I5970032C166 LEVEL: Severe

PID:646 TID:1 NODE:000 Title: **** DRDA ASCB ****
Dump File:/home/db2inst1/sqllib/db2dump/6461.000▌1▐
Legend:
▌1▐ In this UNIX® example, SECTION STMT data is stored in a file named
56772.000 located in the /home/db2/sqllib/db2dump directory.
Notes:
v For partitioned database systems, the file extension identifies the partition
number. For example, the following entry indicates that the dump file was
created by a DB2 process running on partition 10:
Dump File: /home/db2/sqllib/db2dump/56772.010 Data : SECTION STMT
Related concepts:
v on page 0
v on page 0
v on page 0
v on page 0
v “System core files (UNIX)” on page 69
Trap files
DB2® generates a trap file if it cannot continue processing because of a trap,
segmentation violation, or exception.
All signals or exceptions received by DB2 are recorded in the trap file. The trap file
also contains the function sequence that was running when the error occurred. This
sequence is sometimes referred to as ″function call stack″ or ″stack traceback.″ The
trap file also contains additional information about the state of the process when
the signal or exception was caught.
DB2 customer support may require trap files for problem analysis. They are
located in the directory specified by the DIAGPATH database manager
On UNIX®-based systems, the first letter in their names is ″t″, followed by a

process identifier (PID). The file extension is the partition number (000 on single
partition databases).
On Windows® systems, each trap file is named Pxxxxx.yyy where xxxxx is the PID
and yyy is the database partition number (or 000 on single partition databases). If
the trap file is generated because of an exception, it will have the extension .TRP.
Examples:
v t56772.000 is a trap file for the process with pid 56772.

v t56772.010 is a trap file for the process with pid 56772. It was created by a DB2
process running on partition 10.
Trap file generation:
DB2 can generate a trap files on demand, but you should only do this if requested
by DB2 Customer Support.
Refer to db2nstck and db2_call_stack - Generating EDU call stacks if you are
required to generate trap files.
Related reference:
v db2nstck and db2_call_stack - Generating EDU call stacks
v System core files - UNIX only
v Dr Watson log and Windows event logs
The trap files need not be formatted on UNIX platform.
Formatting trap files (Windows)

A tool called db2xprt.exe is available to let you format trap files (*.TRP). It can be
found under the SQLLIB/BIN directory. This tool formats DB2 UDB’s binary trap
files into a human readable ASCII file. Trap files are located in the instance
directory (DB2INSTPROF) by default or in the diagnostic data directory path if the
DIAGPATH database manager configuration parameter is set.
Authorization
You must have access to the DIAGPATH directory.
Command syntax
>>-db2xprt--+----------+--+----+--+----+-- infile --+---------+><
+-/p-- path -+ ’-/m-’ ’-/n-’ ’- outfile -’
’-/v-------’
Command parameters
/p path
A semicolon (;) separated path that points to the location or locations
where the binary files and PDB files are located.
/v Displays version information.
/m Formats a memory dump along with the rest of the trap file.
/n Format data without regard to line number information.
infile Specifies the input file.
outfile
Specifies the output file.
For example, if a trap file called ″DB30882416.TRP″ had been produced in your
DIAGPATH, you could format it as follows:
db2xprt DB30882416.TRP DB30882416.FMT

Analyzing trap files

Before analyzing the stack trace in a trap file, it is important to understand the
signal or exception which generated the trap file.
In a trap file on a Windows platform, this can be seen in the trap file as follows:
The following information is for pid <3088> tid <2416>
Exception C0000005 Occurred
Exception Address = 6C843B8C
Other Unknown access fault
On UNIX platforms, this information will look more like this:

Signal #11 (SIGSEGV): si_addr is 0x00000078, si_code is 0x00000033 (SEGV_ACCERR:Invalid permissions f
Resource Limits
Data seg top [sbrk(0)] = 0x20736010
Cur data size (bytes) = 0x000000000EFFFE00
These exception and signal numbers can be looked up in operating system header
files. For example, ″signal #11 (sigsegv)″ is defined on AIX in
/usr/include/sys/signal.h as follows:
#define SIGSEGV 11 /* (*) segmentation violation */
Common signals and exceptions that cause trap file generation

Below are the most common traps and exceptions for which you will see trap files:
1. Invalid memory accesses. These are indicated by exception C0000005 on Intel
platforms, and Signal #11 (SIGSEGV) on Unix. These occur when memory is
not mapped (doesn’t exist) or when a process does not have permission to
perform the desired action (e.g. write to protected memory).
2. Illegal instructions. These are indicated by C000001C on Intel, or Signal #4
(SIGILL) on Unix. They are usually caused by a bad (often NULL) function
pointers, stack overwrites (overwriting ″saved″ instruction addresses) or
possibly memory corruption.
3. Stack overflow exception. This can occur on Windows platforms, and is marked
by an Exception C00000FD. It may also rarely occur on UNIX platforms. It can
occur when the stack limit is very low.
4. SIGKILL. This Signal #9 can be caused by someone manually (or through a
script) killing a DB2 process, in which case, no trap file is generated.
5. DB2 self-triggered diagnostic trap files (aka ″programming signals″). There are
cases where DB2 will force a trap file to be dumped for diagnostic purposes by
sending itself a programming signal, but the process does not terminate (the
instance is not abnormally ended). The user generated trap files (db2nstck or
db2_call_stack) will also have these signal numbers. These traps are indicated
by:
v Signal #36 on AIX
v Signal #21 on Solaris
v Signal #29 on HP/UX
v Signal #23 on LINUX
v Windows platforms will have a trap file <tid>.000 with an entry stating that
no Exception record is present.
Sometimes these diagnostic trap files are just informative. Other times they are
an indication of a more serious condition and will result in an abort or some
other ″fatal″ symptom soon after, e.g. a ″real″ trap.

6. SIGABRT. These are triggered by calling abort(). Most SIGABRT (Signal #6)
traps on UNIX platforms need not be investigated. There are scenarios where
the DB2 instance will intentionally be brought down with an abort() call (Signal
#6) if the situation is sufficiently severe. The abort() itself and the resulting trap
files and error entries are not a direct symptom of the problem and should not
be analyzed. In these cases, DB2 has not really trapped in the true sense. On
AIX, errpt -a will indicate a Signal #6, i.e. SIGABRT, which should also produce
a core file, i.e. a dump of the process image receiving the signal. In these cases,
the root problem may not be a DB2 problem. Certain types of data corruption
and other severe errors may cause us to bring ourselves down (aka ″panic″).
Note: You will often see several of these traps files when a SIGSEGV or SIGILL
has been encountered on another process. The parent of a trapping process
will call abort() to bring the instance down in it’s ″SIGCHLD″ signal
handler, which essentially means it sends itself a SIGABRT. Thus you will
typically see one process performing the illegal operation receiving a
SIGSEGV or SIGILL followed by it’s parent receiving a SIGABRT.
Analyzing the stack trace back

Once you have determined which trap files are most likely to be pertinent to the
problem investigation, it is time to look at the stack traceback (also called ″function
sequence″ or ″function call stack″) in the trap file.
Here’s an example from an Windows trap file:

++++++++++++++++++++++++++++++++++++++++++++
Failing instruction at 6D7969E5 offset: 00000615 in
<?sqler_callbDrdaInput@@YAHPAUUCintfc@@@Z>
<D:/wsdb/db2nt_v81fp3/s030728/engn/include/clientbiconv.h:2067>
++++++++++++++++++++++++++++++++++++++++++++
Stack calling chain:

--EBP -----EIP-----ARGS--
45FAD058 6D7969E5 3D6ECE30 6D790D6A 3D6ECE30 00000000 offset: 00000615 in
<?sqler_callbDrdaInput@@YAHPAUUCintfc@@@Z>
<D:/wsdb/db2nt_v81fp3/s030728/engn/include/clientbiconv.h:2067>
45FAD654 6D790D6A 3D6ECE30 45FAD670 00000000 45FAD6A4 offset: 0000041A in

<?sqlerCallDL@@YAHPAUUCintfc@@PAUUCstpInfo@@@Z>
<sqlerEngineCallbacks.C:1060>
45FAD6A8 6D1A3F5F 3D6ECE30 45FAD790 00001CA3 0000004F offset: 000002E3 in

<?sqljs_ddm_excsqlstt@@YAHPAUUCintfc@@PAUsqljsDDMObject@@@Z>
<sqljsexe.C:1096>
45FAD6CC 6D189182 17108040 45FAD790 3D6ECE30 71C00000 offset: 0000004E in

<?sqljsParseRdbAccessed@@YAHPAUsqljsDrdaAsCb@@PAUsqljsDDMObject@@PAUUCintfc@@@Z>
<sqljspar.C:348>
45FAD7C4 6D189DA8 17108040 3D6ECE30 00000000 171086C6 offset: 00000234 in

<?sqljsParse@@YAHPAUsqljsDrdaAsCb@@PAUUCintfc@@@Z>
<sqljspar.C:761>
45FAD804 6D180AE9 171086C6 000000E7 00AAAC00 00000000 offset: 00001041 in

<?sqljsDrdaAsDriver@@YAHPAUsqlcc_init_struct@@@Z>
<sqljsqam.C:1103>
45FAD838 6D18094C 00000000 00000000 00000000 00000000 offset: 00000EA4 in

<?sqljsDrdaAsDriver@@YAHPAUsqlcc_init_struct@@@Z>
<sqljsqam.C:1597>
Without access to DB2 source code, you cannot go much further in your analysis of
this information. To determine whether or not you are encountering a known
problem, however, you should identify the top two of three functions in the stack
traceback and search for existing APARs that address traps on one or all of these
functions.
For example in this case, the functions of interest would be:

v ″sqler_callbDrdaInput″
v ″sqlerCallDL″
v ″sqljs_ddm_excsqlstt″
APARs which address problems which had caused DB2 instances to trap should
contain information about the stack traceback which typifies the trap. Thus you are
able to ascertain whether or not the trap you are encountering is a known problem
by searching on those top stack functions. See How to effectively search for
known problems for more information. In this case, you would find that APAR
JR19727 is a very good match for this stack traceback.
If you discover that there are no existing APARs or work-arounds available which
match your situation, then at that point you will require assistance from DB2
Customer Support.
Related reference:
v Trap files
v System core files - UNIX only
v Dr Watson log and Windows event logs - Windows only
Platform specific error logs

There are many other files and utilities available outside of DB2 to help analyze
problems. Often they are just as important to determining root cause as the DB2
files available. Some of this information is contained in logs and traces within the
following areas:
v Operating systems
v Applications and third-party vendors
v Hardware
Based on your operating environment, there may be more places outside of what
has been described here, so be aware of all of the potential areas you ma y need to
investigate when debugging problems in your system.
Operating systems:
Every operating system has its own set of diagnostic files to keep track of activity
and failures. The most common (and usually most useful) is an error report or
event log. Here is a list of how this information can be collected:
v AIX: the /usr/bin/errpt -a command
v Solaris: /var/adm/messages* files or the /usr/bin/dmesg command
v Linux: the /var/log/messages* files or the /bin/dmesg command
v HP-UX: the /var/adm/syslog/syslog.log file or the /usr/bin/dmesg command
v Windows : the system, security, and application event log files and the
windir\drwtsn32.log file (where windir is the Windows install directory)

There are always more tracing and debug utilities for each operating system. Refer
to your operating system documentation and support material to determine what
further information is available.
Applications and third-party vendors:
Each application should have its own logging and diagnostic files. These files will
complement the DB2 set of information to provide you with a more accurate
picture of potential problem areas.
Hardware:
Hardware devices usually log informatio n into operating system error logs.
However, sometimes additional information is required. In those cases, you need to
identify what hardware diagnostic files and utilities may be available for piece of
hardware in your environment. An example of such a case is when a bad page, or
a corruption of some type is reported by DB2. Usually this is reported due to a
disk problem, in which case the hardware diagnostics would need to be
investigated. Please refer to your hardware documentation and support material to
determine what further information is available.
In summary, to completely understand and evaluate a problem, you may need to

collect all information available from DB2, your applications, the operating system
and underlying hardware. The db2support tool automates the collection of most
DB2 and operating system information that you will need, but you should still be
aware of any information outside of this that may help the investigation.
Operating system errors

A problem with a system resource may be indicated by a clear error in the
db2diag.log or in the Administration Notification log or the db2diag.log. For
example:
2004-12-14-06.16.09.234346-300 E16317C370 LEVEL: Error (OS)
PID : 1240500 TID : 1 PROC : db2test
FUNCTION: DB2 UDB, oper system services, sqloxltc_app, probe:15
CALLED : OS, -, unspecified_system_function
OSERR : EINVAL (22) "A system call received a parameter that is not valid."
In situations where the meaning of the operating system error is not as clearly
defined, or when you wish to see more information about the meaning of that
error, you must refer to the error definition file on your server.
On most UNIX systems, system errors can be found in /usr/include/sys/errno.h.

On Linux, the error numbers are located in /usr/include/asm/errno.h. Here are
some of the most common:
#define EPERM 1 /* Not super-user */
#define ENOENT 2 /* No such file or directory */
#define EIO 5 /* I/O error */
#define ENOMEM 12 /* Not enough core */
#define EACCES 13 /* Permission denied */
#define ETXTBSY 26 /* Text file busy */
#define EFBIG 27 /* File too large */
#define ENOSPC 28 /* No space left on device */
On Windows, system error header files are installed on the system with a compiler
or the Windows SDK. Here are a few constants:

#define ERROR_FILE_NOT_FOUND 2L
#define ERROR_ACCESS_DENIED 5L
#define ERROR_INVALID_ACCESS 12L
You can also invoke the net helpmsg command on Windows, which may provide
information about the error., For example,
C:\>net helpmsg 5
Access is denied.
For both UNIX and Windows platforms, besides searching any problem databases
you may have access to, you can use resources on the Web to research errors.
Searching the following locations using the error constant as a search keyword (for
example ″ENOSPC″, instead of 28), along with the operating system API being
called (if known), will often provide a few hints about the meaning or cause of the
error:
v AIX library: http://www-1.ibm.com/servers/aix/library/index.html
v HP-UX: http://docs.hp.com/
v Solaris: http://docs.sun.com/
v Windows: http://support.microsoft.com/
v And, of course, don’t forget the DB2 online support site:
http://www.ibm.com/software/data/db2/udb/support
Setting up the system error log (UNIX)

Prerequisites:
You must have root authority.
Procedure:
To route alerts to the system error log, perform the following the steps:
1. Add the following line to the /etc/syslog.conf file.
user.warn fully_qualified_file_name
where:
v user is the facility to log. This includes DB2 and any other applications
logging messages with the same facility name.
v warn is the priority over which messages are logged. Available choices are:
alert Only alert messages are logged
err Error and alert messages are logged
warn Warning, error, and alert messages are logged
info Information, warning, error, and alert messages are logged
v fully_qualified_file_name is the file (along with its fully qualified path) where
the messages will be logged and the SQLCA will be dumped. This file will
not be created by the system; you must create it in the specified path.
2. Restart the syslog daemon so that it will know about the new configuration in
the syslog.conf file:
a. (AIX systems only) Use the refresh -s syslogd command.
b. Use the kill -1 pid_of_syslogd where pid_of_syslogd is the process ID of the
syslogd process. You can obtain this process ID by issuing the ps -fu syslogd
command.

3. Check to see if information is being logged into the syslog file by issuing:
ps -fu db2sysc
kill -36 db2sysc.process.id
4. Check the file at fully_qualified_file_name (as defined in the /etc/syslog.conf
file). If there is information in the file, then the system error log has been
enabled to capture the information.
The log file may grow quickly, and you will have to reduce its size periodically.
You must use kill -1 pid_of_syslog after you issue the following commands:
mv logfile logfile.old
touch logfile
On AIX, you can include the following line in the crontab that you run as part of
your regular system maintenance instead of using the kill command:
refresh -s syslogd
Related concepts:
v on page 0
System core files (UNIX)

If a program terminates abnormally, a core file is created by the system to store a
memory image of the terminated process. Errors such as memory address
violations, illegal instructions, bus errors, and user-generated quit signals cause
core files to be dumped.
The core file is named ″core″, and is placed in the directory where the application
was running. Note that system core files are distinct from DB2® core files.
Related concepts:
v “Determining active process status” on page 41
Accessing system core file information (UNIX)

The dbx system command helps you determine which function caused a system
core file to be created. This is a simple check that will help you identify whether
the database manager or DB2 Connect is in error, or whether an operating system
or application error is responsible for the problem.
Prerequisites:
You must have the dbx command installed.
Procedure:
To determine which function caused the core file dump to occur:

1. Enter the following command from a UNIX-based command prompt:
dbx program_name core_filename
where program_name is the name of the program that terminated abnormally,

and core_filename is the name of the file containing the core file dump.
The core_filename parameter is optional. If you do not specify it, the default
name ″core″ is used.
2. To obtain symbolic information, compile the application using the ″-g″ option.

3. To end the dbx command, type quit at the dbx prompt.

4. For the HP-UX operating system, use the xdb command for similar function.
5. On Version 4.1 of AIX, ensure that the full core option has been enabled using
the chdev command or smitty.
6. The dbx command provides much more function than is described in this
section. To find out more, enter man dbx from a UNIX-based command prompt.
Example of the dbx Command: The following example shows how to use the
dbx command to read the core file for a program called ″main″.
1. At a command prompt, enter:
dbx main
2. Output similar to the following appears on your display:
dbx version 3.1 for AIX.
Type ’help’ for help.
reading symbolic information ...
[using memory image in core]
segmentation.violation in freeSegments at line 136
136 (void) shmdt((void *) pcAdress[i]);
3. The name of the function that caused the core dump is ″freeSegments″. If the
function name begins with ″db2″, ″sql″, or ″ddcs″, it may indicate an error in
the database manager or DB2 Connect products. Enter where at the dbx
prompt to display the program path to the point of failure.
(dbx) where
freeSegments(numSegs = 2, iSetId = 0x2ff7f730, pcAddress = 0x2ff7f758, line
136
in "main.c"
main (0x1, 2ff7f7d4), line 96 in "main.c"
In this example, the error occurred at line 136 of freeSegments, which was
called from line 96 in main.c.
4. To end the dbx command, type quit at the dbx prompt.
Accessing event logs (Windows)

Windows event logs can also provide useful information. While the system event
log tends to be the most useful in the case of DB2 crashes or other mysterious
errors related to system resources, it is worthwhile obtaining all three types of
event logs:
v System
v Application
v Security
The method used to launch the Windows Event Viewer will differ, depending on
whether you are using Windows XP, Windows 2003, Windows 2000 or Windows
NT.
For example, to open the Event Viewer on Windows XP, click Start, click Control
Panel, click Performance and Maintenance, click Administrative Tools, and then
double-click Event Viewer.
Exporting event logs (Windows)

From the event viewer, you can export event logs in two formats
v .evt format, which can be loaded back into an event viewer (for example on
another machine) or

v in text format.
Event viewer format is easy to work with since you can use the GUI to switch the
chronology order, filter for certain events, and advance forwards or backwards.
Text files provide one significant advantage - you can trust the timestamps! When
you export event logs in .evt format, the timestamps are in Coordinated Universal
Time and get converted to the local time of the machine in the viewer. If you are
not careful, you can miss key events because of time zone differences. Text files are
also easier to search, but once you load an event log from another machine into the
event viewer, it is easy enough to export it again in text format.
Accessing the Dr. Watson log file (Windows)

The Dr. Watson log, drwtsn32.log, is a chronology of all the exceptions that have
occurred on the system. The DB2 trap files are more useful than the Dr. Watson
log, though it can be helpful in assessing overall system stability and as a
document of the history of DB2 traps. The default path is
<install_drive>:\Documents and Settings \All Users\Documents\DrWatson
7 Combining DB2 and OS Diagnostics

7 Introduction to system configuration and user environment information:
7 Diagnosing some problems related to memory, swap files, CPU, disk storage, and
7 other resources requires a thorough understanding of how a given operating
7 system manages these resources. At a minimum, defining resource-related
7 problems requires knowing how much of that resource exists, and what resource
7 limits may exist per user. (The relevant limits are typically for the user ID of the
7 DB2 instance owner.)
7 Here is some of the important configuration information that you need to obtain:
7 v Operating system patch level, installed software, and upgrade history
7 v Number of CPUs
7 v Amount of RAM
7 v Swap and file cache settings
7 v User data and file resource limits and per user process limit
7 v IPC resource limits (message queues, shared memory segments, semaphores)
7 v Type of disk storage (for example EMC, Shark, Network Access Storage solution)
7 v What else is the machine used for? Does DB2 compete for resources?
7 v Where does authentication occur?
7 Most platforms have straightforward commands for retrieving resource

7 information. However, you will rarely need to obtain that information manually,
7 since the db2support utility collects this data and much more. The
7 detailed_system_info.html file produced by db2support (when the options ″-s″ and
7 ″-m″ are specified) contains the syntax for many of the operating system commands
7 used to collect this information.
7 The following exercises are intended to help you discover system configuration
7 and user environment information in various DB2 diagnostic files. The first tutorial
7 familiarizes you with the steps involved in running the db2support utility.
7 Subsequent exercises cover trap files, which provide more DB2-generated data that
7 can be useful in understanding the user environment and resource limits.

7 Exercise 1: Running the db2support command

7 1. On each platform that you have available, start DB2 with the db2start
7 command.
7 2. Assuming you already have the SAMPLE database available, create a directory
7 for storing the output from db2support.
7 3. Change to that directory and issue:
7 DB2SUPPORT <directory> -D sample -S -M
7 4. Review the console output, especially the types of information that are
7 collected.
7 You should see output like this (when run on Windows):
7 ...
7 Collecting "System files"
7 "db2cache.prf"
7 "db2dbamr.prf"
7 "db2diag.bak"
7 "db2eventlog.000"
7 "db2imdbd.dmp"
7 "db2misc.prf"
7 "db2nodes.cfg"
7 ...
7 Collecting "Detailed operating system and hardware information"
7 Collecting "System resource info (disk, CPU, memory)"
7 Collecting "Operating system and level"
7 Collecting "JDK Level"
7 Collecting "DB2 Release Info"
7 Collecting "DB2 install path info"
7 Collecting "Registry info"
7 ...
7 Creating final output archive
7 "db2support.html"
7 "db2_sqllib_directory.txt"
7 "detailed_system_info.html"
7 "userResponse.xml"
7 "db2supp_system.zip"
7 "dbm_detailed.supp_cfg"
7 "SAMPLE_node0_detailed_db.supp_cfg"
7 "dbmsnap.supp_out"
7 "db2diag.log"
7 db2support is now complete.
7 An archive file has been produced: "db2support.zip"
7 5. Now use a Web browser to view the detailed_system_info.html file. On each of
7 your systems, identify the following information:
7 v Number of CPUs
7 v Amount of RAM
7 v Operating system level
7 v User environment
7 v User resource limits (UNIX ulimit command)
7 Exercise 2: Locating environment information in a DB2 trap file

7 1. On a UNIX system, ensure DB2 is started, then issue
7 db2_call_stack
7 The call stacks are placed in files in the diagnostic directory (sqllib/db2dump
7 by default). Each EDU has its own file which is called tprocessID.node number.
7 2. Locate the following in one of the trap files:
7 v DB2 code level

7 v Data seg top (this is the maximum private address space that has been
7 required)
7 v Cur data size (this is the maximum private address space limit)
7 v Cur core size (this is the maximum core file limit)
7 v Signal Handlers (this information may not appear in all trap files)
7 v Environment variables (this information may not appear in all trap files)
7 v map output (shows loaded libraries)
7 Example trap file from AIX (truncated):

7 db2 build information: DB2 v8.1.0.48 s040212 SQL08015
7 timestamp: 2005-01-21-19.54.15.306950
7 instance name: db2isnt1.001
7 EDU name : db2sysc 1
7 Signal #36
7 uname: S:AIX R:2 V:5 M:000786694C00 N:steel
7 thread id : 1 (0x1)
7 ...
7 Signal #36: si_pid is 0, si_uid is 0, si_value is 00000000
7 ...
7 Resource Limits
7 Data seg top [sbrk(0)] = 0x20116000
7 Cur data size (bytes) = 0x000000000EFFFE00
7 Cur stack size (bytes) = 0x0000000010000000
7 Cur core size (bytes) = 0x7FFFFFFFFFFFFFFF
7 ...
7 Correlating DB2 and system events or errors:
7 System messages and error logs are too often ignored. You can save hours, days,
7 and even weeks on the time it takes to solve a problem if you take the time to
7 perform one simple task at the initial stage of problem definition and investigation.
7 That task is to compare entries in different logs and take note of any that appear to
7 be related both in time and in terms of what resource the entries are referring to.
7 While not always relevant to problem diagnosis, in many cases the best clue is
7 readily available in the system logs. If we can correlate a reported system problem
7 with DB2 errors we will have often identified what is directly causing the DB2
7 symptom. Obvious examples are disk errors, network errors, and hardware errors.
7 Not-so obvious are problems reported on different machines, for example domain
7 controllers or NIS servers, which may affect connection time or authentication.
7 System logs can be investigated in order to assess stability, especially when

7 problems are reported on brand new systems. Intermittent traps occurring in
7 common applications may be a sign that there is an underlying hardware problem.
7 Here is some other information provided by system logs.

7 v Significant events such as when the system was rebooted
7 v Chronology of DB2 traps on the system (and errors/traps/exceptions from other
7 software that is failing)
7 v Kernel panics, out-of-filesystem-space, and out-of-swap-space errors (which may
7 prevent the system from creating/forking a new process)
7 System logs can help to rule out crash entries in the db2diag.log as causes for
7 concern. One consistent DB2 crash recovery investigation was resolved by
7 discovering that the system was rebooting - it turned out that the cleaning staff
7 was unplugging the computer every morning at the same time!

7 If you see crash recovery in DB2 Administration Notification or DB2 diagnostic

7 logs with no preceding errors, the DB2 crash recovery is likely a result of a system
7 shutdown.
7 This principle of correlating information extends to logs from any source and to
7 any identifiable user symptoms. For example, it can be very useful to identify and
7 document correlating entries from another application’s log even if you can’t fully
7 interpret them.
7 The summation of this information is a very complete understanding of your

7 server and of all of the varied events which are occurring at the time of the
7 problem.
7 Related concepts:
7 v on page 0

Chapter 2. Specific troubleshooting techniques
Installation issues
This section will help you identify the source of common installation problems.
Note: The first thing you need to do is confirm whether the installation
prerequisites are met
Before installing DB2, you should check to make sure that your environment meets
the minimum hardware and software requirements as described in the Quick
Beginnings documentation. Links to such documentation is included in the Related
reference below.
If your system does not meet the minimum requirements, then the DB2 installer
could fail. For example, a failure could occur for a simple reason such as not
having enough disk space, not having the prerequisite software installed, or kernel
parameters not being set according to minimum requirements.
Once you have eliminated the environment or non-DB2 factors, you can focus on
the techniques described in subsequent sections.
Related reference:
v Disk and memory requirements for non-partitioned DB2 UDB Enterprise Server
Edition (Windows and UNIX)
v Disk and memory requirements for partitioned DB2 UDB Enterprise Server
Edition (Windows and UNIX)
v Disk and memory requirements for DB2 UDB clients (Windows and UNIX)
v Installation requirements for non-partitioned DB2 servers (AIX)
v Installation requirements for partitioned DB2 servers (Windows)
v DB2 Information Integrator installation worksheet
v Installation requirements for Query Patroller server (UNIX)
v DB2 PDF and printed documentation
Location of installation error logs

The type of installation error logs produced will depend on the method of
installation used.
db2setup wizard:
This DB2 installation program captures all installation information, including

errors, in a number of files:
v db2setup.log or db2wi.log file captures the current (or most recent) DB2
installation information including errors.
v db2setup.his or db2.log contains a history of all DB2 installations on your
machine
v db2setup.err file captures any error output that is returned by Java (for example,
exceptions and trap information)

In some cases you can specify where these files are created, but by default they are
located in the following location:
v On Windows operating systems: the ″My Documents″\DB2LOG\ directory (the
location of the ″My Documents″ directory will depend on the settings on your
computer).
v On UNIX systems: the /tmp directory
db2_install script:
This UNIX installation method generates an error log file in /tmp called
db2_install_log.<pid>, where <pid> is the process ID which performed the
installation.
Response file installation:
If you are installing via a response file on UNIX, you will be using the ″db2setup″
command, and the same installation error logs will be produced as described
above for the db2setup wizard.
If you are installing via a response file on Windows, you can specify the name and
location of the error log using the ″/L″ flag on the ″setup.exe″ command. If you do
not specify the log file’s name, DB2 names it db2.log. The db2.log file is located in
the My Documents\db2log folder.
Native installation tools:
Installation error logs may be produced, but the type and location of the file will
depend on the tool used. For example, if smit is used, the error log file can usually
be found in /smit.log.
Related reference:
v Installation methods for DB2 UDB (Windows and UNIX)
v Interpreting installation error logs
Installation methods for DB2 UDB (Windows and UNIX)

This topic provides information about DB2® UDB installation methods. The
following table shows the installation methods that are available by operating
system.
Table 1. Installation method by operating system.
Installation method Windows® UNIX®
DB2 Setup wizard Yes Yes
db2_install script No Yes
Response file installation Yes Yes
Native installation tools No Yes
The following list describes installation methods for DB2 UDB.

DB2 Setup wizard
The DB2 Setup wizard is a GUI installer available on both UNIX and
Windows operating systems. The DB2 Setup wizard provides an

easy-to-use interface for installing DB2 UDB and for performing initial
setup and configuration tasks. The DB2 Setup wizard can also be used to
create instances and response files.
On UNIX systems, the DB2 Setup wizard replaces the text-based installer
interface (db2setup).
db2_install script
The db2_install script uses the operating systems’s native installation
utility to install DB2 UDB. The db2_install script prompts for a DB2
product keyword. This script installs all components for the DB2 product
you specify, in English only. You cannot select or deselect components or
interface language support. The db2_install script does not perform user
and group creation, instance creation, or configuration. It installs the DB2
components to your system. This method of installation might be preferred
in cases where greater control over the installation setup process than the
GUI installer provides is required.
Response file installation
A response file is an ASCII file that contains setup and configuration
values. The file is passed to the DB2 setup program and the installation is
performed according to the values that have been specified. There are a
number of ways to create a response file:
v Using the response file generator (Windows)
v Using the DB2 Setup wizard (UNIX and Windows)
v By customizing sample response files that are provided for each DB2
product (UNIX and Windows)
Using the response file generator, you can create a response file that
replicates an existing installation. For example, you might install a DB2
client, fully configure the client, then generate a response file to replicate
the installation and configuration of the client to other computers. The
response file generator is available only on Windows.
The DB2 Setup wizard can create a response file for both UNIX and
Windows installations. The selections you make as you proceed through
the DB2 Setup wizard are recorded in a response file that you can save to a
location on your system.
For your convenience, you can create a response file without performing
an installation. This feature can be useful in an environment where a DBA
does not have the authority required to perform an installation. The DBA
can create a response file for the installation and provide it to the system
administrator who installs the product on the DBA’s behalf.
An alternative to using the response file generator or the DB2 Setup
wizard to create a response file is to manually modify a sample response
file. Sample response files are provided on the DB2 product CD-ROM.
Native installation tools
Installing DB2 using your operating system’s native installation provides
the greatest control over the installation process, but it is also more difficult
than the other installation methods. When installing a particular DB2
product, you have to ensure that the required components are installed
and that component dependencies are maintained. Advanced knowledge of
both DB2 and your operating environment is required. User and group
creation, instance creation, and configuration must be performed manually.
Chapter 2. Specific troubleshooting techniques 77

Interpreting Windows installation error logs
db2.log and db2.wi log errors:
Sometimes referred to as an ″installation history log file″, db2.log contains a history

of all DB2 installations that have occurred.
Note: Your installation history log file may not be named ″db2.log″ if you
specified a different file name and location using the -l parameter on the
setup.exe command.
The db2wi.log file contains the log of the current (or most recent) installation.
Information is written to it as the installation events occur, unlike the history log
file where the log is not actually written until the end of the installation. At the
end of the installation, the only difference between what is written to the
db2wi.log and what is appended to the history log file is that the former also
contains a list of Windows Installer properties with current values. This list of
properties is very long and is typically only looked at by DB2 Support.
Depending on the problem you are trying to solve, all of the information in the log
files may be useful, despite the complexity and amount of information contained
in the files. The difficult part is narrowing down the problem and finding the parts
of the log files that are most relevant to the problem.
Note: When you run the installation with tracing turned on, the logging is done in
verbose mode, thereby creating a log with much more information. Do not
confuse the trace file with the log file. They are different files. However, by
turning trace on, you increase the logging level of the installation. If you are
familiar with Windows Installer you will recognize that the log created
when enabling trace is the verbose log that you would typically get by
adding the /l*v parameter to windows installer based installations.
Ensuring that the log you are reading is the correct log:
In many cases you may find yourself debugging an installation that you did not
initiate. If you are unsure whether the log you are viewing is the correct log, there
are some hints that you can use to verify that it is the correct log. The best way to
determine this is to look at the time of the installation, as well as information
about the product being installed. The product shown can be the product code or
the path to the installation database used to install the product. If it is the path to
the installation database, notice that the file name contains the name of the
product. However, if it lists a globally unique identifier (GUID) for example,
{D8F53726-C7AD-11D4-9155-00203586D551}, then it is a bit more difficult. Here is a
list of the product codes for the DB2 UDB products that you can use to
cross-reference information found in the product field.
Product ID code
DB2 UDB Enterprise Server {D8F53726-C7AD-11D4-9155-
Edition 00203586D551}
DB2 UDB Workgroup Server {7A28F948-4945-4BD1-ACC2-
Edition ADC081C24830}
DB2 UDB Personal Edition {C0AA883A-72AE-495F-9601-
49F2EB154E93}

Product ID code
DB2 Warehouse Manager {84AF5B86-19F9-4396-8D99-
11CD91E81724}
DB2 Datalinks Manager {1D16CA65-F7D9-47E5-BB26-
C623A44832A3}
DB2 Information Integrator {273F8AB8-C84B-4EE6-85E7-
Relational Wrappers D7C5270A6D08}
DB2 Connect Enterprise {9C8DFB63-66DE-4299-AC6B-
Edition 37D799A728A2}
DB2 Connect Personal {F1912044-6E08-431E-9B6D-
Edition 90ED10C0B739}
DB2 Administration Client {ABD23811-AA8F-416B-9EF6-
E54D62F21A49}
DB2 Application {68A40485-7F7F-4A91-9AB6-
Development Client D67836E15CF2}
DB2 Run-time Client {63F6DCD6-0D5C-4A07-
B27C-3AE3E809D6E0}
DB2 Spatial Extender {F6846BF9-F4B5-4BB2-946D-
3926795D5749}
DB2 Information Integrator {DD30AEB3-4323-40D7-
Non-Relational Wrappers AB39-735A0523DEF3}
DB2 Warehouse Manager {5FEA5040-22E2-4760-A88C-
Connectors 73DE82BE4B6E}
DB2 Query Patroller {7A8BE511-8DF3-4F22-B61A-
AF0B8755E354}
DB2 Cube Views {C8FEDF8F-84E8-442F-A084-
0A0F6A772B52}
DB2 UDB Express Edition {58169F10-CA30-4F40-8C6D-
C6DA8CE47E16}
DB2 Run-time Client Lite {07C9CEE7-4529-4E60-95D3-
6B6EF6AC0E81}
For example, consider the following excerpt of a log file. The date on which the
installation took place is February 2, 2005. The GUID is {D8F53726-C7AD-11D4-
9155-00203586D551} which matches the DB2 UDB Enterprise Server Edition:
=== Verbose logging started: 2/14/2005 15:40:03 Build type: SHIP UNICODE 2.00.2600.1183
Calling process: C:\WINNT\system32\msiexec.exe ===
MSI (c) (8C:88): Resetting cached policy values
MSI (c) (8C:88): Machine policy value ’Debug’ is 0
MSI (c) (8C:88): ******* RunEngine:
******* Product: {D8F53726-C7AD-11D4-9155-00203586D551}
******* Action:
******* CommandLine: **********
Determining installation success or failure:
In some cases all you may be interested in is whether the installation was
successful. To determine the success or failure of an installation look at the end of
the db2wi.log file for a line that looks like the following:
MSI (s) (98:8C): Product: DB2 Enterprise Server Edition
-- Installation operation completed successfully.

Likewise, if the installation failed, look for a line at the end of the db2wi.log file
that looks like the following:
MSI (s) (40:10): Product: DB2 Enterprise Server Edition -- Installation operation failed.
Diagnosing major errors:
If you received a failure the next step is to determine the cause of the failure. A
general tip that allows you to find the error quickly is to search for Return value 3
in the log file. Once you find this in the log file you will usually see further text
detailing what the problem is.
For example, if you had logged into the Windows server using a user ID that does
not belong to the Administrators group, and then tried to install DB2, the
db2wi.log file would contain an error as follows:
...
MSI (s) (40:10): Doing action: VerifyPrereqsCA
Action start 15:47:17: VerifyPrereqsCA.
MSI (s) (40:10): Creating MSIHANDLE (20) of type 790542 for thread 572
1: In order to install the product "DB2 Enterprise Server Edition", the user running the
installation must have at minimum the authority of a member of the user group "Administrators".
Action ended 15:47:59: VerifyPrereqsCA. Return value 3.
MSI (c) (BC:3C): Doing action: SetupCompleteError
Action start 15:47:59: SetupCompleteError.
...
In some cases the error comes directly from Windows Installer. In these cases the
error can be difficult to understand, but there are methods to get more information
about the error. In some cases you may only be presented with an error number
along with some strings separated by commas. You can look up these error types
directly from the Microsoft Web site at http://www.msdn.microsoft.com
Diagnosing minor errors:
In some cases the installation may complete successfully but with the occurrence of
a minor configuration error. When these types of errors occur it means that the
installation completed, however an error occurred during the configuration stage
of the installation. When this occurs the installation exit code will be set to 1. The
most common place to look for these types of errors is during the execution of the
action that performs most of the up and running operations. The output from this
task can be found by searching for a line that looks like the following:
when installation traces are enabled:

Action 10:27:03: DeferredCallURE_CA.
MSI (s) (9C:DC): Executing op:
CustomActionSchedule(Action=DeferredCallURE_CA,ActionType=3137,Source=BinaryData,Target=CallURECA,)
when installation traces are not enabled:
Action 10:27:03: DeferredCallURE_CA.
The up and running portion of the installation which runs during the Custom
Action DeferredCallURE_CA is organized into tasks. The success status of each is
reported to the log which can be used to determine if the task was successful, or if
a problem occurred. A successful task will output a line in the log file that looks
like the following:
1: Creating/migrating DB2 instances:.......Success
Alternatively, if the task failed you would see a line that looks like the following:

1: Creating/migrating DB2 instances:.......Failure
If you need to find more information about a particular task that failed, or for
further details about what the task did, you can look at the lines immediately
before the overall result of the task. The following example shows that an instance
″DB2″ was created successfully, and that some DBM Config variables were set. For
example:
1: The instance ″DB2″ has been created successfully.
1: The value ″SVCENAME=db2c_DB2″ was set in the DBM CFG file for the ″DB2″
instance.
1: The value ″DB2COMM=TCPIP″ was set in the Profile Registry for the ″DB2″
instance.
1: Creating/migrating DB2 instances:.......Success
Response file errors:
If you have installed DB2 using a response file, there are some common problems
which you may encounter. The most common response file problem is that the
installation cannot find the response file that was specified by the -u option of
setup.exe because the location was specified incorrectly. In a case like this you
would see lines in the log file that look like the following:
Action start 0:23:55: DetectAndSetInstallPathCA.
Action ended 0:23:55: DetectAndSetInstallPathCA. Return value 1.
Action start 0:23:55: InitSilentInstallCA.
1: Failed to access the response file: "c:\db2ese.rsp".
Action ended 0:23:55: InitSilentInstallCA. Return value 3.
Action ended 0:23:55: INSTALL.
Return value 3.
Right before the Return Value 3, which shows that the action InitSilentInstallCA
failed, you can see some information about the error. In this case the response file
c:\db2ese.rsp cannot be accessed because the path does not exist. This problem
can be corrected by simply correcting the response file path given in the -u
parameter to setup.exe. Another possible cause of response file installation failure
can be that the user running the installation does not have permission to access the
file.
Keyword errors:
One type of error that can occur is caused by an invalid keyword in the response
file installation. The response file that is passed into the installation is validated for
two main types of problems before the installation begins. The first type of
validation that occurs is for syntax. Certain keywords have length limits, or accept
only certain values. If an invalid value is specified for a keyword, or if a keyword
is entered that is not recognized, the installation will exit. The second type of
validation is semantic. At this stage values are checked to make sure they are
compatible with the system and with each other. This stage is only performed if
the syntax checking does not find any errors.
Keyword syntax validation:
Response file validation is done during the InitSilentInstallCA custom action. To

find this you can either search for this method directly (if you know that the

problem was with a keyword), or you can use the method described above by
searching for Return Value 3. The following examples illustrate some of the error
types that might be displayed as a result of the response file syntax validation:
MSI (s) (40:54): Doing action: InitSilentInstallCA

MSI (s) (40:54): Creating MSIHANDLE (5120) of type 790542 for thread 1876
1: ERROR:A Response file error occurred.
The value "INVALID_VALUE" is not valid for the keyword "KILL_PROCESSES".
1: ERROR:A Response file error occurred. Unknown keyword "THIS_IS_AN_INVALID_KEYWORD" at line "116".
1: ERROR:One or more errors occurred while checking the syntax of the response file "c:\db2ese.rsp".
Please correct the errors and run the install again.
1: ERROR:Unable to set the response file "c:\db2ese.rsp" in the up and running engine.
1: Failed to initialize silent install. Action ended 0:28:05: InitSilentInstallCA.
Return value 3.
In the above example, the keyword for KILL_PROCESSES is given a value of

INVALID_VALUE; however, the only values accepted are YES or NO. The keyword
THIS_IS_AN_INVALID_KEYWORD was also in the response file, however this is not
recognized by the DB2 Setup wizard.
Response file semantic validation:
Semantic response file validation is also done during the InitSilentInstallCA

custom action. The errors that occur during the semantic validation are displayed
in the log file. After reading the errors that are in the log file it should be apparent
what needs to be corrected. For example:
1: ERROR:The TCP/IP entries are invalid.
Please specify a valid service name and port number.
1: ERROR:The response file specified "c:\db2ese.rsp" is not valid.
1: Failed to initialize silent install.
Action ended 1:51:03: InitSilentInstallCA. Return value 3.
In this example, the log file indicates that the TCP/IP entries specified in the
response file are not valid. The combination of service name and port number
conflicted with values that were already on the system.
Related reference:
v Location of installation error logs
v Tracing installation problems
Interpreting UNIX installation error logs

db2setup.log, db2setup.his and db2setup.err:
The db2setup.log file contains the log of the current (or most recent) installation.
The information from the db2setup.log is appended to the db2setup.his (sometimes
referred to as an ″installation history log file″) once the installation has ended.
The db2setup.err file captures any error output that is returned by Java (for
example, exceptions and trap information).
Depending on the problem you are trying to solve, all of the information in the log
files may be useful, despite the complexity and amount of information contained

in the files. The difficult part is narrowing down the problem and finding the parts
of the log files that are most relevant to the problem.
An example of what will appear in a db2setup.log is as follows:

DB2 Setup log file started at: Fri Feb 18 15:42:22 EST 2005 EST
============================================================
Operating system information: AIX 5.2

Product to install: DB2 Universal Database Enterprise Server Edition
Installation type: Typical
Previously Installed Components:
Java(TM) Runtime Environment (JRE)
Java(TM) Development Kit (JDK)
Selected Components:
Product Messages - en_US.iso88591
Base Client Support
DB2 LDAP Support
...
Enabling Asynchronous I/O:.......Success
Checking license agreement acceptance:.......Success
Command to be run:
"/usr/sbin/installp -acgqX -d
’/net2/d_v8_inst30/v8_inst30/db2aix5L64_v82/.s040812/ese.sbcsaix1/db2/aix’ ’db2_08_01.msg.en_US.is
...
The appearance of the subsequent section of the db2setup.log will depend on the
platform, since DB2 will use the operating system’s native installation utility. Once
all of the filesets have been processed, the following messages in the db2setup.log
will indicate whether the installation has completed successfully:
Installing DB2 file sets:.......Success
Registering DB2 licenses:.......Success
Setting default global profile registry variables:.......Success
Creating the DB2 Administration Server:.......Success
Initializing instance list:.......Success
Creating DB2 instances:.......Success
Updating existing DB2 instances:.......Success
Configuring the DB2 Administration Server:.......Success
Updating global profile registry:.......Success
DB2 Setup log file finished at: Fri Feb 18 16:36:52 2005 EST
============================================================
db2_install_log.<pid>:
The db2_install_log.<pid> will be produced when you perform the installation

using the db2_install script. The log will contain a list of all of the filesets
processed and whether or not they were applied successfully. The format of the file
will vary, since db2_install uses the operating system’s native installation utility.
For example, on AIX the entries for each fileset would appear as follows:
Pre-installation Verification...
+---------------------------------------------------------------------+
Verifying selections...done
Verifying requisites...done
Results...
SUCCESSES
---------
Filesets listed in this section passed pre-installation
verification and will be installed.
Selected Filesets
-----------------
db2_08_01.db2.rte 8.1.1.64 # Run-time Environment

<< End of Success Section >>
...
+----------------------------------------------------------------------+
Installing Software...
+----------------------------------------------------------------------+
installp: APPLYING software for:

db2_08_01.db2.rte 8.1.1.64
. . .
Finished processing all filesets. (Total time: 19 secs).
+----------------------------------------------------------------------+
Summaries:
+----------------------------------------------------------------------+
Installation Summary
--------------------
Name Level Part Event Result
------------------------------------------------------------------------
db2_08_01.db2.rte 8.1.1.64 USR APPLY SUCCESS
+----------------------------------------------------------------------+
To determine the success or failure of the installation, look at the end of the file for
a line that looks like the following:
The installation logfile can be found in /tmp/db2_install_log.91190.
db2_install program completed successfully.
Response file errors:
If you have installed DB2 using a response file, there are some common problems
which you may encounter. The most common response file problem is that the
installation cannot find the response file that was specified by the -r option of
db2setup because the location was specified incorrectly. In a case like this, the
installation would fail and you would see lines in the db2setup.log file that look
like the following:
DB2 Setup log file started at: Fri Feb 18 17:36:56 2005 EST
============================================================

ERROR:DB2 Setup is unable to open the response file "/home/db2inst1/db2ese.rsp".
============================================================
In this case the response file c:\db2ese.rsp cannot be accessed because an

incorrect response file path was specified in the db2setup command. Simply
verifying the correct path and re-running the db2setup command with the correct
information can correct this problem. Another possible cause of response file
installation failure can be that the user running the installation does not have
permission to access the file.
Keyword errors:
One type of error that can occur is caused by an invalid keyword in the response
file installation. The response file that is passed into the installation is validated for

two main types of problems before the installation begins. The first type of
validation that occurs is for syntax. Certain keywords have length limits, or accept
only certain values. If an invalid value is specified for a keyword, or if a keyword
is entered that is not recognized, the installation will exit. The second type of
validation is semantic. At this stage values are checked to make sure they are
compatible with the system and with each other. This stage is only performed if
the syntax checking does not find any errors.
Keyword syntax validation:
The following example illustrates one of the error types that might be displayed is
an error is encountered during the response file syntax validation:
The response file specified "/home/lcawley/db2ese.rsp" is not valid.
For more information please see the DB2 installation log at "/tmp/db2setup.log".
The db2setup.log shows the following:

============================================================

ERROR:A Response file error occurred. The value "MAYBE" is not valid for the
keyword "LIC_AGREEMENT".
ERROR:One or more errors occurred while checking the syntax of the response
file "/home/lcawley/db2ese.rsp". Please correct the errors and run the install
again.
============================================================
In the above example the keyword for LIC_AGREEMENT is given a value of

MAYBE, however the only values accepted are ACCEPT or DECLINE. This
problem can be resolved by correcting the value in the response file and running
db2setup again.
Response file semantic validation:
The following example illustrates one of the error types that might be displayed is
an error is encountered during the response file syntax validation:
DBI1702E The specified service name or port number conflicts
with existing values in the TCP/IP services file.
...
The response file specified "/home/lcawley/db2ese.rsp" is not valid.
For more information please see the DB2 installation log at "/tmp/db2setup.log".
The db2setup.log shows the following:

============================================================

ERROR:DBI1702E The specified service name or port number conflicts
with existing values in the TCP/IP services file.
Explanation:
The service name or port number entered by the user conflicts with
existing values in the TCP/IP services file. The service name may
already be used with a different port number, or the port number
may already be used with a different service name.

User Response:
Specify a service name and port number that does not conflict
with existing entries in the services file.
============================================================
After reading the errors that are in the log file it should be apparent what needs to
be corrected. In this example, the log file indicates that the TCP/IP entries
specified in the response file are not valid. The combination of service name and
port number conflicted with values that were already defined on the system.
Related reference:
v Tracing installation problems
Tracing installation problems

v setup - Install DB2 Command
v db2setup - Install DB2 Command
Troubleshooting a DB2 Information Integrator installation
Instance creation and update issues
Determining the cause of instance creation problems

If the problem with instance creation occurs during DB2 installation, then you need
to review the error message and the installation error logs . Forexample, if you
neglect to drop the instances prior to uninstalling DB2 on UNIX-based systems,
db2setup may return the following error when you attempt to reinstall DB2 and
recreate the instance:
The user you selected already has an sqllib directory in their home directory. This user cannot be u
In this situation, you need to verify whether the sqllib directory (e.g.
/home/db2inst1/sqllib does indeed exist. It is possible in this situation that
removing the sqllib directory will not be sufficient to resolve the problem, since if
you did not drop the instances prior to uninstalling DB2, they may still appear in
the DB2 UDB Instance Profile Registry. Such a situation would result in errors in
the installation error log and a failure to create the instance. An example of such
errors from db2setup.log on AIX is as follows:
Installing DB2 file sets:.......Success
Registering DB2 licenses:.......Success
Setting default global profile registry variables:.......Success
Creating the DB2 Administration Server:.......Success
ERROR:Could not switch current DB2INSTANCE to "db2inst1". The return code is
"-2029059916".
Initializing instance list:.......Success

ERROR:DBI1122E Instance db2inst1 cannot be updated.
Explanation:

An attempt was made to update an instance. This instance cannot be updated
because:
v This ″db2iupdt″ command cannot be used to update this instance.
v The instance is still active.
User Response:
Ensure that you are using the correct version of the ″db2iupdt″ command. Also
ensure that there are no db2 processes running at the instance. Retry the command.
DBI1079I Output is saved in the log file /tmp/db2iupdt.log.113908.
In this case you may need to manually remove the old instance from the instance
profile registry and then re-attempt the installation. In this example, that would
mean removing ″db2inst1″ from the list of instances in the
/var/db2/V8.1/profiles.reg file. This problem can be avoided of course if you
ensure that you follow the instructions when uninstalling DB2 and drop the
instances prior to uninstalling the DB2 product.
If your instance creation problem is not occurring during installation, or if

everything other than the instance creation succeeded during the installation, you
should try to debug the problem using either the db2isetup tool (on UNIX
platforms only) or the db2icrt script.
Additional trace information can be obtained when using db2icrt and db2isetup
on UNIX-based systems.
Example 1:
DB2DIR/instance/db2icrt -d -u db2fenc1 db2inst1
where DB2DIR is the DB2 installation directory, will produce additional trace points
in the /tmp/db2icrt.log file.
Example 2:
DB2DIR/instance/db2isetup -t /home/db2inst1/isetup.trc
where DB2DIR is the DB2 installation directory, will produce a trace file as well as a
default /tmp/db2icrt.log file.
Related reference:
v Creating an instance using db2icrt
v db2isetup - Start Instance Creation Interface Command
v Setting environment variables on UNIX systems
Identifying instance update issues (UNIX)

Since most instance creation and update problems happen in the UNIX
environment, this section focuses only on the UNIX platform. Note however that
the command does exist in DB2 Enterprise Server Edition environments on
Windows.
To identify an instance update issue, you should first check the diagnostic log file
/tmp/db2iupdt.log.pid. For example, if you attempt to update an instance called
db2inst1 as follows:

db2iupdt -u db2fenc1 db2inst1
But it fails with ″DBI1069E Unexpected error. Function = terminate_instance,

Return code = 13.″, you should check the /tmp/db2iupdt.log.<pid>. The error
log’s full name will be indicated at the time of the error. For example:
Explanation:
All processed and failed operations have been saved into this log file.
User Response:
Do not modify this file in any way. This file is for IBM Technical Support reference.
Here is an example of the db2iupdt.log:

Program name = db2istop
Instance name = db2inst1
Home dir = /home/db2inst1
Version = 81
Product dir = /usr/opt/db2_08_01
db2stop return code = SQL6033W
exit code = 13
DBI1069E Unexpected error. Function = terminate_instance, Return code = 13.
Explanation:
An unexpected error occurred during the execution of this program.

...
DBI1250E Applications are still using instance db2inst1.
Explanation:
There are applications that are still running that are using the specified instance.
All applications using this instance must be terminated before the command can be
completed successfully. You can get a list of the applications that are currently
using the instance by issuing the command:
db2 list applications
User Response:
You can either wait for the applications to end by themselves, or you can explicitly
force the applications to end. You can logon as the instance owner and run the
command
...
DBI1122E Instance db2inst1 cannot be updated.
...
Explanation:
All processed and failed operations have been saved into this log file.

Based on the above error log, the causeof the instance update problem is that there
are applications still using the instance at the time that the ″db2iupdt″ was
attempted.The actions db2iupdt.log is very clear in this case.
If the actions necessary are not this clear, you can debug this further by running
db2iupdt with the -d debug option. In order to redirect the debug information to a
file, you may use shell redirection.
For example:
db2iupdt -d db2inst1 2>&1 | tee db2iupdt.debug.out
The output file in this example is called db2iupdt.debug.out. Here is an example

of the output you would see in db2iupdt.debug.out file for the same error as was
described previously:
...
+ [ 0 -eq 0 ]
+ /usr/bin/echo ## exit function terminate_instance
+ /usr/bin/tee -a /tmp/db2iupdt.log.47370
## exit function terminate_instance
+ return 13
status=13
+ [ 13 -ne 0 ]
...
+ /usr/bin/ps -elf
+ /usr/bin/grep db2inst1
200001 A root 4948 47370 1 60 20 1b83 228 30e5cdd0 20:46:15 pts/2 0:00 grep db2inst1
40001 A db2inst1 9820 41098 0 60 20 b855 552 31e311cc 20:46:06 - 0:00 db2loggw (LISE) 0
40001 A db2inst1 18806 41098 0 60 20 c959 648 c0041500 20:46:08 - 0:00 db2pfchr 0
40001 A db2inst1 26426 41098 0 60 20 2de7 608 31e31a0c 20:46:05 - 0:00 db2loggr (LISE) 0
40001 A db2inst1 31840 54756 0 60 20 5a3 1144 20:45:20 - 0:00 db2tcpcm 0
...
40001 A db2inst1 54756 30608 0 60 20 89f0 1416 20:45:19 - 0:00 db2sysc 0
+ /usr/bin/ipcs -a
+ /usr/bin/grep db2inst1
q 8126477 0xffffffff --rw------- db2inst1 usr db2inst1 usr 0 0
4194304 0 0 no-entry no-
...
+ display_msg /usr/opt/db2_08_01/msg/en_US.iso88591/db2install.cat 250
DBI1250E Applications are still using instance %s.\n db2inst1
+
From the above output, the terminate_instance function returnsa value of 13. If
you look at the scripts in the DB2DIR/instance directory, where DB2DIR represents
/usr/opt/db2_08_01 on AIX, and /opt/IBM/db2/V8.1 on all other UNIX-based
systems, you will see that terminate_instance is defined in the db2iutil script.
Note: If you have a FixPak or modification level installed in an alternate path, the
DB2DIR directory is usr/opt/db2_08_FPn on AIX and opt/IBM/db2/V8.FPn on
all other UNIX-based systems, where n represents the number of the FixPak
or modification level.
Within db2iutil you can see that terminate_instance stops all applications, CLP
backend processes and the database manager for a specified instance.
The return code (13) is obtained from the db2istop command here:

# Stop all applications and the database manager for the instance.
status=${FALSE?}
/usr/bin/su ${sysadm?} -c "${PROGDIR?}/db2istop ${argforce?} \
${nodenum?} ${instname?} ${insthome?} ${db2version?} \
${db2proddir?} ${TMPFILE3?}"
status=$?
The db2istop script shows the meaning of the value 13:

EXIT_DB2STOP_ERR=13
Other db2iupdt errors on UNIX can be investigated using the log file and debug
traces in the same manner.
Related reference:
v Updating instance configuration on UNIX
v Updating instance configuration on Windows
Avoiding fixpak upgrade problems

Since most FixPak upgrade problems can be avoided by simply following the steps
as outlined in the FixPak readme, you should review it prior to doing the FixPak
upgrade.
As a general rule, you need to do the following things before installing a FixPak:
v Verify that there is enough disk space on your system.
v Verify that the required prerequisites are installed (as listed in the FixPak
readme).
v Stop all database processes (i.e. stop all instances as well as the DB2
Administration Server, the Fault Monitor and the License Daemon).
v For AIX only, unload the DB2 libraries with /usr/sbin/slibclean as root.
If you are downloading the FixPak image from the official IBM FTP site, make sure
that the size of the downloaded image is the same as the one on the FTP site.
Note: The installFixpak script does not provide a trace option. installAltFixPak
simply checks for the necessary pre-requisites and then calls the operating
system’s native install tool with certain customized options
Related reference:
v Applying the latest FixPak (Windows and UNIX)
License issues
Before you start debugging the license problem, you should have a better
understanding of it, so you should ask yourself these questions:
1. What is the SQL error code? Check the Message Reference guide for further
explanation and take the corrective action.
2. Is the license for a new installation? If this is a new installation, then you will
need to verify that the DB2 license key matches with the version and product
of DB2.
3. If this is not a new installation, has something else changed in the system? For
example, perhaps there has been a change to the operating system (OS) or to
file permissions, or perhaps a DB2 FixPak upgrade was installed.

4. Did you install the permanent DB2 license key? This is important because your
DB2 server could be running in ″Try-and-Buy″ mode if the permanent license
key has not been installed, and you cannot start DB2 because the ″Try-and-Buy″
mode has expired.
For example, let’s say you receive the following SQL code
SQL8007W There are "90" day(s) left in the evaluation period for the product "DB2
Enterprise Server Edition". For evaluation license terms and conditions, refer to the
IBM Evaluation Agreement in the EVALUATE.AGR file, located in the following
directory: "C:\PROGRA~1\SQLLIB".
According to the Message Reference, SQL8007W indicates, ″A valid license key

has not been installed for this product. The evaluation period will expire
after the specified number of days.″
In this example, you would first make sure that you have installed the license key
for ″DB2 Enterprise Server Edition.″ You could use the license management tool to
do so. (The license management tool is discussed next.) If the license key is not
installed, then you need to install it. Second, find out if there is any change made
to the operating system.
Note: Finding out what has changed will give you a better idea as to what could
have caused the DB2 licensing error.
Troubleshooting licensing problems

In order to verify that you have a valid DB2 license, use the license management
tool called db2licm. To find out how to use it, run db2licm with the -h option to
display help information.
Run ″db2licm -l″ on your test system to list all the DB2 products with license
information. If this command does not return the appropriate license information,
then this could mean either a valid DB2 license key is not installed or there is a
potential permission problem with the nodelock file.
To confirm whether a valid DB2 license is installed, compare the license key inside
the nodelock file with the license key file (for example, db2ese.lic is the license key
file for DB2 UDB Enterprise Service Edition). Here are the different locations for
the nodelock file based on the platforms:
v AIX - /var/ifor
v HP-UX, Linux and Solaris - /var/lum
v Windows - \Program Files\SQLLIB\license
Inside the nodelock file, comments line start with ’#’, and the license key is usually
proceeded and followed by a comment line. For example,
#
<You will see the actual license key here.>
#[admin_comment] "IBM Toronto Lab" "DB2 Enterprise Edition" ...
If the license key is not yet installed, use db2licm to install it. If the license key
does match the one on the product CD, check whether there is a permission
problem with the nodelock file.
In UNIX, you should verify the file permission for the nodelock is set to rw-r--r--
and owned by root.

Related reference:
v License management
Migration issues
This section will help you identify the source of common migration problems
encountered when migrating DB2 Universal Database (UDB) Version 7 to DB2
UDB Version 8.
If you are migrating to DB2 UDB Version 8 from another database product, you
should visit the DB2 UDB Migrate Now! Web site:
www.ibm.com/software/data/db2/migration/
Related reference:
v Installation issues
v Migrating to DB2 UDB Version 8.2
v Overview of migrating to DB2 Information Integrator
Migration recommendations
Consider the following recommendations when planning your database migration:
Back up log files before migration when DB2® UDB uses replication
7 If you use replication for your DataJoiner® and DB2 UDB data, you must
7 archive all of the DB2 log files before migration.
7 For complete information on migrating your DB2 replication environment,
7 see the IBM® DB2 Information Integrator Migration Guide: Migrating to
7 SQL Replication Version 8 at
7 http://www.ibm.com/software/data/integration/db2ii/support.html.
DataJoiner instance migration
If you want to migrate an instance of DataJoiner or DB2 UDB on which
you are running the Capture or Apply programs for DB2 replication, you
must prepare to migrate your replication environment before you
migrating the instance.
7 For complete information on migrating your DB2 replication environment,
7 see the IBM DB2 Information Integrator Migration Guide: Migrating to
7 SQL Replication Version 8 at
7 http://www.ibm.com/software/data/integration/db2ii/support.html.
Perform hardware and operating system upgrades separately from DB2 UDB
migration
Performing hardware and operating system upgrades separately from DB2
migration simplifies problem determination if you encounter migration
difficulties. If you upgrade your software or hardware prior to migrating to
DB2, ensure that your system is operating acceptably before attempting
DB2 migration.
7 Dropping the detailed deadlocks event monitor
7 At the same time a database is created, a detailed deadlocks event monitor
7 is also created. As with any monitor, there is some overhead associated
7 with this event monitor. If you do not want the detailed deadlocks event
7 monitor, then the event monitor can be dropped using the command:
7 DROP EVENT MONITOR db2detaildeadlock

7 To limit the amount of disk space that this event monitor consumes, the
7 event monitor deactivates, and a message is written to the administration
7 notification log, once it has reached its maximum number of output files.
7 Removing output files that are no longer needed allows the event monitor
7 to activate again on the next database activation.
Back-level server support
As you move your environment from DB2 Version 7 to DB2 Version 8, if
you migrate your DB2 clients to Version 8 before you migrate all of your
DB2 servers to Version 8, there are several restrictions and limitations. To
avoid the known restrictions and limitations, migrate all of your DB2
servers to Version 8 before you migrate any of your DB2 clients to Version
8. These restrictions and limitations are not associated with DB2 Connect™;
nor with zSeries®, OS/390®, or iSeries™ database servers.
For example:
v bring your DB2 UDB Version 7 clients up to the latest FixPak
v migrate your DB2 Servers to DB2 UDB Version 8 32-bit and then update
to FixPak 8
v migrate your DB2 Connect server to Version 8 and then update to
FixPak 8
v migrate your DB2 clients to DB2 UDB Version 8 and then update to
FixPak 8
v migrate your DB2 UDB Version 8 servers to 64-bit
Benchmark DB2 performance
Run a number of test queries before migrating DB2. Record the exact
environment conditions when queries are run. Also, keep a record of the
db2expln command output for each test query. Compare the results before
and after migration. This practice can help to identify and correct any
performance degradation.
Devise a migration planning document
You should create a migration planning document. Consider using this
sample document as a template:
http://d02dbs88.southbury.ibm.com/support/swg/dmgtech.nsf/f2bc0e4fed485c5f48256bb
Devise a plan to back out of a migration
There is no utility to reverse a migration. If you must back out of a
migration, you might have to remove DB2 Version 8 code from your
system, reinstall the previous version of DB2 to recreate back-level
instances, and restore your database backups. If you have to back out of a
migration, current database backups and detailed records of database and
database configuration settings are essential.
Migrating instances with DB2 DataPropagator™ replication
Before migrating an instance of DataJoiner or DB2 UDB on which you are
running the Capture or Apply programs for DB2 DataPropagator, read the
migration documentation for DB2 DataPropagator Version 8. You must
prepare to migrate your replication environment before you migrate the
DB2 or DataJoiner instance. You must also perform specific tasks
immediately after the migration of your DB2 or DataJoiner instance.
Migration documentation for DB2 DataPropagator Version 8 can be found
at the http://www.ibm.com/software/data/dpropr/library.html Web site.
Related concepts:
v Benchmark testing

v Explain tools
Related tasks:
v Migrating DB2 UDB (Windows)
v Migrating DB2 UDB (UNIX
Related reference:
v DB2 Universal Database planned incompatibilities
v Version 8 incompatibilities with previous releases
v Version 7 incompatibilities with previous releases
Migration restrictions
You should be aware of the following restrictions before you migrate to DB2 UDB
Version 8:
v Migration is supported only from:
– DB2 UDB Version 6.x or Version 7.x. (All platforms supported in Version 6.x
and Version 7.x; Linux must be at Version 6 FixPak 2.)
– DB2 DataJoiner V2.1.1 32-bit (AIX, Windows NT, and Solaris Operating
Environment).
v Issuing the migrate database command from a DB2 UDB Version 8 client to
migrate a database to a DB2 Version 8 server is supported; however, issuing the
migration command from an DB2 UDB Version 6 or Version 7 client to migrate a
database to a DB2 UDB Version 8 server is not supported.
v When migrating from DB2 DataJoiner V2.1.1, DB2 Information Integrator is
required to support non-IBM data sources.
v Migration between platforms is not supported. For example, you cannot migrate
a database from a DB2 server on Windows to a DB2 server on UNIX.
v Migrating a partitioned database system that has multiple computers requires
that database migration be performed after DB2 UDB Version 8 is installed on all
participating computers. Any DB2 migration commands need to be run on each
of the participating computers.
v Windows allows only one version of DB2 UDB to be installed on a computer.
For example, if you have DB2 UDB Version 7 installed and install DB2 UDB
Version 8, DB2 UDB Version 7 is removed during the installation. All instances
are migrated during DB2 installation on Windows operating systems.
v User objects within your database cannot have DB2 UDB Version 8 reserved
schema names as object qualifiers. These reserved schema names include:
SYSCAT, SYSSTAT, and SYSFUN.
v User-defined distinct types using the names BIGINT, REAL, DATALINK, or
REFERENCE must be renamed before migrating the database.
v You cannot migrate a database that is in one of the following states:
– Backup pending
– Roll-forward pending
7 – One or more table spaces in an abnormal state
– Transaction inconsistent
v Restoration of back-level (DB2 Version 6.x or Version 7.x) database backups is
supported, but rolling forward of back-level logs is not supported.
v Database transactions executed between database backup time and the time DB2
UDB Version 8 migration is completed are not recoverable.

7 Reverse FixPak upgrade restrictions
7 Beginning with DB2 UDB Version 8 FixPak 1, the maximum number of tablespaces
7 supported in a database increases from 4096 to 32768.
7 For the most part, you should be able to move your database from DB2 UDB
7 Version 8 to DB2 UDB Version 8 FixPak 1 without noticing a change or having to
7 do anything special to use the new tablespace limit.
7 Note the following restrictions related to moving a database from DB2 UDB
7 Version 8 FixPak 1 (or a later DB2 UDB Version 8 FixPak) back to the DB2 Version
7 8 level:
7 v If you want to move from DB2 Version 8.2 back to DB2 Version 8.1, you will
7 need to run the db2demigdbd command before going back to DB2 Version 8.1.
7 The db2demigdbd is a reverse database directory files tool that restores your
7 database directory to its Version 8.1 format.
7 v Moving a database that contains a tablespace ID higher than 4096 to DB2
7 Version 8 from DB2 Version 8 FixPak 1 or later is unsupported. Attempting to do
7 so will result in anomalous behavior and improper operation.
7 v Restoring a database image that contains a tablespace ID higher than 4096 on
7 DB2 Version 8 is unsupported. Attempting to do so results in anomalous
7 behavior and improper operation.
7 v When moving from DB2 UDB Version 8 FixPak 1 (or a later DB2 UDB Version 8
7 FixPak) back to DB2 UDB Version 8, the log-skipping functionality is disabled
7 until such time as the DB2TSCHG.HIS file is removed.
Migrating databases
Prerequisites:
You require SYSADM authority.
Restrictions:
Migration is supported only from:

v DB2 Version 6.x or Version 7.x. (all platforms supported in Version 6.x and
Version 7.x)
v DB2 DataJoiner Version 2.1.1 (AIX, Windows NT, and Solaris Operating
Environment).
7 No database migration is required if the database has been migrated to a DB2

7 Version 8 FixPak level.
Procedure:
To migrate a DB2 database:

1. Migrate the database using the db2 migrate database command.
DB2 MIGRATE DATABASE command
►► MIGRATE DATABASE database-alias ►

DB

► ►◄
USER username
USING password
where:
DATABASE database-alias
Specifies the alias of the database to be migrated to the currently
installed version of the database manager.
USER username
Identifies the user name under which the database is to be migrated.
USING password
The password used to authenticate the user name. If the password is
omitted, but a user name was specified, the user is prompted to enter
it.
7 2. Optional: Update statistics for local tables within the database. When database
migration is complete, old statistics that are used to optimize query
performance are retained in the catalogs. However, DB2 Version 8 has statistics
that are modified or do not exist in DB2 Version 6 or DB2 Version 7. To take
advantage of these statistics, you may want to execute the runstats command
on tables, particularly those tables that are critical to the performance of your
SQL queries.
3. Optional: Rebind packages. During database migration, all existing packages
are invalidated. After the migration process, each package is rebuilt when it is
used for the first time by the DB2 Version 8 database manager. You can run the
db2rbind command to rebuild all packages stored in the database.
4. Optional: Revoke EXECUTE privileges on external stored procedures that
contain SQL data access from PUBLIC. During database migration, EXECUTE
privileges are granted to PUBLIC for all existing functions, methods, and
external stored procedures. This will cause a security exposure for external
stored procedures that contain SQL data access which allow users to access
SQL objects for which they would not otherwise have privileges. Revoke the
privileges by entering the db2undgp - r command.
5. Optional: Migrate DB2 Explain tables.
6. Optional: If you recorded configuration settings before migration, you might
want to compare pre-migration configuration settings to current configuration
settings to verify successful migration. Verify:
v database configuration parameter settings
v database manger configuration parameter settings
v tablespaces records
v packages records
Note: During migration, the database configuration parameter maxappls is set

to automatic. If you want it set to a different value, you should update it
manually.
Application Development issues

Applications or stored procedures are written in a variety of languages using a set
of standard APIs to manipulate database data.
It is important to know the layers involved in DB2 applications in order to

diagnose issues. The DB2 Call Level Interface (CLI), which is equivalent in

functionality to ODBC, is often used underneath other drivers, so it may be
necessary to debug different layers to determine problems. The figure below shows
the typical application flow:
Figure 2.
By following the DB2 Application Development manuals as well as the specific

documentation related to your programming interface, you can avoid most errors
involved in creating and executing DB2 applications. DB2 product manuals for
application development can be found in both HTML and PDF formats at
http://www-306.ibm.com/software/data/db2/udb/support/manualsv8.html .
Related references:
v DB2 Supported Programming Interfaces
v DB2 supported development software
Debugging and Optimizing an Application

You can debug and optimize your application while you develop it.
Procedure:
To debug and optimize your application:

v Prototype your SQL statements. You can use the command line processor, the
Explain facility, analyze the system catalog views for information about the
tables and databases that your program is manipulating, and update certain
system catalog statistics to simulate production conditions.
v Use the flagger facility to check the syntax of SQL statements in applications
being developed for DB2 Universal Database for z/OS and OS/390, or for
conformance to the SQL92 Entry Level standard. This facility is invoked during
precompilation.
v Make full use of the error-handling APIs. For example, you can use
error-handling APIs to print all messages during the testing phase.
v Use the database system monitor to capture certain optimizing information for
analysis.
Proper application error handling

Often applications are difficult to troubleshoot because not enough error
information is provided by the application.
The method used to retrieve additional error information is different depending on

the type of application, Instructions for how to devise appropriate error handling
can be found in the Application Development documentation, in particular in the
sections listed in the Related Reference below.
Example 1: Adding error handling calls to a simple CLI application
Sample1.c is a simple CLI application that has a problem upon execution.
When the application is run, it uses the current user ID and password to connect
with a database named SAMPLE. It allocates the appropriate environment and
database handles, then connects to the database.

Thereafter it performs a fetch as follows:
sqlrc = SQLAllocHandle( SQL_HANDLE_STMT, hdbc, &hstmt ) ;
if ( sqlrc != SQL_SUCCESS )
{
printf("\n Error allocating statement handle.\n");
return( 1 ) ;
}
sqlrc = SQLFetch( hstmt ) ;

{
printf("\n Error Fetching.\n");
return( 1 ) ;
}
sqlrc = SQLFreeHandle( SQL_HANDLE_STMT, hstmt );

{
printf("\n Error freeing statement handle.\n");
return( 1 ) ;
}
When the program is executed, the following error occurs:

Error Fetching.
The first step in diagnosing this problem is to examine the code, in order to find
where this error is generated. In this case, you would find that the SQLFetch call
must have returned an SQL_ERROR. Since the application didn’t request the error
details, it is difficult to determine the cause. You could obtain a CLI trace (see
″Enabling CLI traces″)and look for the line which includes the SQLFetch call. It
would be similar to the following:
SQLFetch( hStmt=1:1 )
---> Time elapsed - +2.200000E-004 seconds
SQLFetch( )
<--- SQL_ERROR Time elapsed - +1.238000E-003
seconds
This shows the error being returned, but no error details. In many cases, detailed
diagnostic information is returned only when an application requests it. Taking a
CLI trace may reveal more information than the application returns.
If you were to rewrite this sample program to gather proper error information, you
would need to add calls to the function SQLGetDiagRec. For example, you could
create a simple procedure called GetCLIErrorInfo which calls SQLGetDiagRec and
which is in turn called after the fetch.
...
void GetCLIErrorInfo( SQLSMALLINT htype, /* handle type identifier */
SQLHANDLE hndl /* handle */ )
{
...
while ( SQLGetDiagRec( htype,
hndl,
i,
sqlstate,
&sqlcode,
message,
SQL_MAX_MESSAGE_LENGTH + 1,
&length
) == SQL_SUCCESS ) {
printf( "\n SQLSTATE = %s\n", sqlstate ) ;
printf( " Native Error Code = %ld\n", sqlcode ) ;
printf( "%s\n", message ) ;

i++ ;}
...}
int main()
...
sqlrc = SQLFetch( hstmt ) ;
{
printf("/n Error Fetching./n");
GetCLIErrorInfo( SQL_HANDLE_STMT, hstmt );
}
...
Once such error handling has been added, the sample program would return
output as follows:
Error Fetching.
SQLSTATE = HY010
Native Error Code = -99999
[IBM][CLI Driver] CLI0125E Function sequence error.SQLSTATE=HY010
-------------------------
Application Complete.
By checking the ODBC Standard or DB2 CLI documentation, you will learn that a
function sequence error is returned from an SQLFetch call if you do not call
SQLExecute first. The statement needs to be executed before you can fetch any data
it may return. This is a simple example of a situation that can be resolved only
after appropriate error handling has been added to the application.
Related Reference:
v Diagnostics in CLI applications overview
v Error-Handling Considerations in Partitioned Database Environments
v Error Message Retrieval in an Application
v Handling an SQLException under the DB2 JDBC Type 2 Driver
v Handling an SQLException under the DB2 Universal JDBC Driver
v Handling SQL errors in an SQLJ application
v Returning error messages from SQL procedures
Displaying the contents of a bind file using the db2bfd tool

There are times when it is beneficial to examine the contents of a DB2 bind file.
For example, a vendor might supply an executable and bind file and you might
want to see what kinds of SQL statements the application might execute. Also,
consider the case where an errant application is deleting data from a table that is
supposed to remain read-only but you aren’t sure which application is responsible.
By examining the bind files for all of the applications that run against the database,
you might be able to find a DELETE statement against the table in question and
from there dig further into the application’s logic to determine what is going
wrong.
The db2bfd tool can be used to display these kinds of thing as well as some other
information.
Execute the tool without any parameters to see its usage.

C:\>db2bfd
Usage: db2bfd [ [-b] [-h] [-s] [-v] ] <filespec>
Where: <filespec> is a V7 or V8 bind file

Options: -b = display bind file header
-h = display this information
-s = display SQL statements
-v = display host variable declarations
Example 1: Use db2bfd -s to see the SQL statements contained in the

db2sampl.bnd file
On Windows® operating systems:

db2bfd -s %DB2PATH%\bnd\db2sampl.bnd
(where %DB2PATH% is a variable that determines where DB2® is installed)
On UNIX®:
db2bfd -s $HOME/sqllib/bnd/db2sampl.bnd
(where $HOME is the home directory of the instance owner)
Next, use the -v option to see the corresponding host variables that would have
been defined in the source code for db2sampl:
On Windows® operating systems:

db2bfd -v %DB2PATH%\bnd\db2sampl.bnd
(where %DB2PATH% is a variable that determines where DB2® is installed)
On UNIX®:
db2bfd -v $HOME/sqllib/bnd/db2sampl.bnd
(where $HOME is the home directory of the instance owner)
Related reference:
v db2bfd - Bind File Description Tool Command
v Package Creation Using the BIND Command
Resolving compilation errors

Errors that occur during compilation are often due to programming errors or
environments that are not correctly set up for application creation.
Here are some basic steps to ensure that your environment is setup correctly:
v Ensure makefiles contain the correct compiler information
v Set all compiler-related variables for include files and libraries
v Ensure any command line arguments point to the correct DB2 directories
Here are some of the common mistakes that can lead to linker errors:
v Improper API/function calls
v Multiple versions of linked libraries in the LIBPATH
v API/function definitions not matching the code contained in the library files
Refer to the samples included with DB2, as they include scripts and instructions to
compile and link the sample applications.

Example 1: Investigating a compilation error in a simple CLI application written in
C/C++
When attempting to compile an application (named test.c), the following command

was used:
cl test.c -I "%DB2PATH%\include" -link"%DB2PATH%\lib\db2cli.lib"
The following error was encountered:

sample.c
sample.c(19) : warning C4047: ’function’ : ’long ’ differs in levels of indirect
ion from ’long *’
sample.c(19) : warning C4024: ’SQLAllocHandle’ : different types for formal and
actual parameter 2
sample.c(19) : error C2198: ’SQLAllocHandle’ : too few actual parameters
The first step in problem determination is to check the line of code that the
compiler objects to. In this case, it is as follows:
sqlrc = SQLAllocHandle( SQL_HANDLE_ENV, &henv ) ;
{
printf("/n Error allocating environment handle./n");
return -1;
}
Since the error indicated that there were too few parameters, you would then
check the CLI documentation for SQLAllocHandle. You would see that the call is
indeed incorrect. It is missing one input handle. You can find the SQLAllocHandle
call documented at
http://publib.boulder.ibm.com/infocenter/db2help/topic/com.ibm.db2.udb.doc/ad/r0000556.htm
This is a simple error common to all types of programming. Errors such as this,
received at compile time, can usually be resolved by checking the related
documentation and correcting the code as necessary.
Related reference:
v DB2 supported development software
Avoiding linker errors

Compiler and linker errors for DB2 applications are often the same as those you
encounter when creating any type of application. The linking step sometimes
results in difficulty creating the executable.
Ensure your environment is correctly installed and you are using the proper DB2
static and shared libraries (e.g. those contained in sqllib/lib32 or sqllib/lib). This is
the best way to avoid compilation/linking problems. You should check the DB2
Application Development guides relating to your programming interface and your
compiler specific documentation. They will guide you to set up the correct
environment.
Here are some of the common items that relate to having your environment setup
correctly:
v Ensure makefiles contain the correct compiler information
v Set all compiler-related variables for include files and libraries
v Ensure any command line arguments point to the correct DB2 directories
Here are some of the common mistakes that can lead to linker errors:

v Improper API/function calls
v Multiple versions of linked libraries in the LIBPATH
v API/function definitions not matching the code contained in the library files
You can look at the samples included with DB2 as they also include scripts and
instructions to compile and link them.
Troubleshooting suspended or looping applications

It is possible that, after you start a query or application, you suspect that it is
suspended (it does not show any activity) or that it is looping (it shows activity,
but no results are returned to the application). Ensure that you have turned lock
timeouts on. In some situations, however, no error is returned. In these situations,
you may find the diagnostic tools provided with DB2® and the database system
monitor snapshot helpful.
One of the functions of the database system monitor that is useful for debugging
applications is to display the status of all active agents. To obtain the greatest use
from a snapshot, ensure that statement collection is being done before you run the
application (preferably immediately after you run DB2START) as follows:
db2_all "db2 UPDATE MONITOR SWITCHES USING STATEMENT ON"
When you suspect that your application or query is either stalled or looping, issue
the following command:
db2_all "db2 GET SNAPSHOT FOR AGENTS ON database
Error-Handling Considerations in partitioned database

environments
In a partitioned database environment, DB2® breaks up SQL statements into
subsections, each of which is processed on the partition that contains the relevant
data. As a result, an error may occur on a partition that does not have access to the
application. This condition does not occur in a nonpartitioned database
environment.
You should consider the following:

v Non-CURSOR (EXECUTE) non-severe errors
v CURSOR non-severe errors
v Severe errors
v Merged multiple SQLCA structures
v How to identify the partition that returned the error
If an application ends abnormally because of a severe error, indoubt transactions

may be left in the database. (An indoubt transaction pertains to global transactions
when one phase completes successfully, but the system fails before the subsequent
phase can complete, leaving the database in an inconsistent state.)
DB2 Development Center issues

Choose one of the following:
v Enabling the Next push button in a wizard
v Correcting a user ID or password
v Resolving routines that fail to build on another connection
v Supporting null input values to Java stored procedures

v Testing null input values to SQL user-defined functions
v Diagnosing SQL routines that fail to build
v Correcting color schemes that make text unreadable
v Diagnosing SQL routine build failures with the DSNTPSMP build utility
v Diagnosing Java routine build failures with the DSNTJSPP build utility
v Diagnosing routine build failures on DB2 UDB for OS/390 or DB2 UDB for
z/OS servers, error code -471
v Setting display for the pop-up information for fields and controls
v Diagnosing system connection exceptions when trying to establish multiple
JDBC connections (AIX)
v Maintaining correct formatting for stored procedure source code (DB2 for
iSeries)
Enabling the Next push button in a wizard

Symptom:
When using a wizard, the Next push button is not available after you type text in
the text fields.
Possible cause:
The text that you enter in each fieldmust conform to the rules for that field. The
Next pushbutton is not available until the text that you enter conforms to the
rules.If you did not receive notification that your text entries do not follow
therules, turn on the constraint checking option.
Action:
v Close the wizard.
v Open the Environment Settings notebook.
v On the User-assistance features page, select all check boxes in the
Constraintchecking group to turn on the constraint checking options.
Correcting a user ID or password

Symptom:
The user ID or password that you typed to connect to the database server is
incorrect.
Possible causes:
The authentication configuration of the database server is set to SERVER, and your
user ID and password are not valid on the server. This is the default configuration
for DB2 Universal Database.
The authentication configuration of the database server is set to CLIENT, and your
user ID and password are not valid on the client workstation.
Action:
Check the authentication configuration of the database to which you are trying to
connect. If the configuration is set to SERVER, your user ID and password must be
valid values on the server. If the configuration is set to CLIENT, your user ID and
password must be valid values on the client workstation.
Resolving routines that fail to build on another connection

Symptom:
When you copy a routine to another connection, the build fails.
Possible causes:
Building the routine to another database could have failed because of the following
reasons:
v The routine contained references to objects (such as schemas and tables) that do
not exist on the target database.
v The SPECIFIC NAME of the object being pasted already exists in the target
database.
Actions:
The following actions may resolve the symptom:

v Check the source code of the routine and remove references to objects that do
not exist on the target database, or create the objects on the target database.
v Change the specific name by editing the SPECIFIC clause of the SQL procedure
or UDF, or editing the Specific Name property of a Java stored procedure.
Supporting null input values to Java stored procedures

Symptom:
A null value is passed to a Java stored procedure input parameter that was
declared with a Java primitive type (for example, an int). DB2 generates an SQL
error message without calling the stored procedure.
Possible cause:
On a workstation server, passing null values to input parameters is always

allowed. On DB2 for OS/390 or z/OS, passing null values to input parameters is
allowed if you specify SIMPLE WITH NULLS for the linkage. The Insert Java
Stored Procedure wizard declared an input parameter using a Java primitive type.
DB2 generates an SQL error message because the input parameter that was
declared with the primitive type was passed a null value.
Action:
Change the primitive type to the corresponding Java class in the Development
Center source code editor to allow the stored procedure to support null input
values. For example, the Java primitive type int becomes Integer.
Testing null input values to SQL user-defined functions

Symptom:

A null value passed to a SQL user-defined function input parameter generates an
SQL error message without calling the UDF.
Possible cause:
There is a known DB2 restriction that prevents Java applications, including

Development Center, from sending a null value to input parameters of a
user-defined function.
Action:
To test SQL user-defined functions that accept NULL values, call them from a DB2
Command Processor session: db2 ==> values(<udfname>(<parameter
values,including NULL,separated by commas>))
Diagnosing SQL routines that fail to build

Symptom:
Building an SQL routine results in the error message ″Build not successful: -1.″ No
other error messages are shown.
On Windows, the user may get an error messages such as: ’nmake’ is not
recognized as an internal or external command, operable program or batch
file.
Possible causes:
With DB2 versions prior to 8.2, a C compiler is required to build SQL stored
procedures, so possible causes include:
v A C compiler is not installed on the database server.
v A C compiler is installed on the database server but is not properly configured.
For example, environment variables required by the C compiler may be
incorrect.
Actions:
For database servers on Windows NT, 2000, or XP:

v Ensure that Visual C++ Version 6 is installed.
v Ensure that the server’s DB2_SQLROUTINE_COMPILER_PATH is set to the
location of the file vcvars32.bat.
v Optional: Move the Visual C++ environment settings for PATH, INCLUDE, and
LIB from User settings to System settings.
Additional configuration information is available in the IBM DB2 UDB Application

Development Guide in the Building and Running Applications and Setting Up the
Application Development Environment topics.
Correcting color schemes that make text unreadable

Symptom:
You cannot read the text in the editor because the text and background are the
same or nearly the same color.

Possible cause:
Using the Environment Settings notebook, you selected a foreground and

background color that are the same or almost the same.
Actions:
v If you did not close the Environment Settings notebook or apply the color
changes, you can apply the previous color selection by clicking Reset on the
Editor page of the Environment Settings notebook.
v If you closed the Environment Settings notebook or applied the color changes,
on the Editor page of the Environment Settings notebook, choose a different
color scheme by clicking Change. The Change Colors window opens.
Diagnosing SQL routine build failures with the DSNTPSMP

build utility
Symptom:
When you try to build an SQL object to an OS/390 database server, the build fails
and returns the following error message:
DSNTPSMP - returned nnn
Rolling back...successful
procedure-name - Build failed.
Where procedure-name is the name of the routine. nnn is a numeric error code. See
the Message Reference for your operating system for an explanation of error codes.
Possible cause:
There might be a problem with the SQL object source code.
Actions:
1. Review the diagnostic information to determine why the build failed. The
DSNTPSMP build utility might return a result set of messages that describe
where and why the SQL object build failed.
2. Correct the problem that is described by the diagnostic information.
3. Rebuild the SQL object.
4. If additional information is needed for diagnosing the problem look at the
SYSLOG for the WLM procedure where DSNTPSMP executes. Often, a dataset
may fill up and be the cause for the build failure. This can be determined by
looking directly at the WLM procedure output
Diagnosing Java routine build failures with the DSNTJSPP

build utility
Symptom:
When you try to build an SQL object to an OS/390 database server, the build fails
and returns the following error message:
DSNTJSPP - returned nnn
Rolling back...sucessful
procedure-name - Build failed

Where nnn is a numeric error code. See the DB2 Universal Database for OS/390
and z/OS Messages and Codes book (GC26-9940) for an explanation of error
codes.
Possible cause:
There might be a problem with the SQL object source code.
Actions:
1. Review the diagnostic information to determine why the build failed. Obtain
available diagnostic information by selecting the Verbose build option when
using the wizard, or from the Stored Procedure Properties notebook. Selecting
this option causes DSNTJSPP to return a result set of messages to Development
Center that describe where and why the SQL object build failed.
2. Correct the problem that is described by the diagnostic information.
3. Rebuild the Java object.
4. If additional information is needed for diagnosing the problem, look at the
SYSLOG for the WLM procedure where DSNTJSPP and DSNTBIND execute.
Diagnosing routine build failures on DB2 UDB for OS/390 or

DB2 UDB for z/OS servers, error code -471
Symptom:
When you try to build an SQL or Java object to an OS/390 or z/OS database
server, the build fails and you receive the following error message:
Create object returns -471
[IBM][CLI Driver][DB2] SQL0969N
There is no message text corresponding to SQL error "-471" in the message file on this workstation
The error was returned from module "DSNX9WCA" with original tokens "DSNTPSMP 00E79002 ".
SQLSTATE=55023
Possible causes:
The Work Load Manager (WLM) was not started.
Actions:
1. Ensure that the SQL objects build utility DSNTPSMP is properly installed on
your DB2 for OS/390 or z/OS server.
2. Ensure that the Java objects build utility DSNTJSPP and DSNTBIND are
properly installed on your DB2 for OS/390 or z/OS server.
Contact your system administrator for information about obtaining the build
utility.
Setting display for the pop-up information for fields and

controls
Symptom:
In the Development Center wizards and notebooks, the pop-up information that
explains the fields and controls does not display.
Possible causes:

v The pop-up information windows are turned off.
v You are not holding the mouse pointer over a field or control long enough for
the pop-up information to appear.
v You are not pressing the F1 key when a field or control is in focus.
Actions:
v Ensure that the pop-up information windows are not turned off:
1. Open the Environment Settings notebook.
2. On the User-assistance features page, ensure that the Show pop-up
information for interface controls check box is selected.
3. In the Delay pop-up information in milliseconds field, type the time delay in
milliseconds. The value milliseconds is an integer. For example, entering 1000
as the value results in a 1000 millisecond (1 second) delay before pop-up
information appears.
v Press the F1 key on your keyboard when the field or control is in highlighted. To
move the highlight from one field or control to another, press the Tab key. If no
field or control is highlighted, the browser help displays.
Diagnosing system connection exceptions when trying to

establish multiple JDBC connections (AIX)
Symptom:
You have trouble viewing objects in the Server View and experience system
connection exceptions such as:
COM.ibm.db2.jdbc.DB2Exception: [IBM][CLI Driver] SQL1224N
A database agent could not be started to service a request, or was terminated
as a result of a database system shutdown or a force command.
SQLSTATE=55032
Possible cause:
The number of JDBC connections is limited by the number of shared memory

segments to which a single process can be attached. By default, AIX does not
permit 32-bit applications to attach to more than 11 shared memory segments per
process, of which a maximum of 10 can be used for local DB2 connections.
Development center requires multiple JDBC connections. The EXTSHM
environment variable must be configured for multiple JDBC connections.
Action:
To resolve this issue if you are using AIX version 4.3.1 or newer, set set the
environment variable EXTSHM to ON to increase the number of shared memory
segments to which a single process can be attached. EXTSHM must be exported
both in the shell where the client application is started and also in the shell where
db2start is run. To configure the EXTSHM environment variable for multiple JDBC
connections:
1. In client sessions, before starting the client application, type the following
command: export EXTSHM=ON
2. Before starting the DB2 server, type the following commands:
export EXTSHM=ON
db2set DB2ENVLIST=EXTSHM
db2set -all

3. Add the following lines to the sql/db2profile:
EXTSHM=ON
export EXTSHM
Maintaining correct formatting for stored procedure source

code (DB2 for iSeries)
Symptom:
On iSeries, the source code of a stored procedure body loses all its formatting and
appears on a single line.
Possible cause:
You created the stored procedure using:

v STRSQL
v RUNSQLSTM command
v CREATE PROCEDURE statement in embedded SQL
Stored procedures created using one of these methods will lose the formatting of
their bodies in Development Center.
Action:
If you want to make the stored procedure appear formatted correctly, you must
recreate the procedure using:
v Development Center
v the Run SQL Scripts interface in the iSeries Navigator
v the New External Procedure or New SQL Procedure interface in the iSeries
Navigator
Backup and recovery issues

Taking frequent database backups is essential for the recovery of databases in the
event of a disaster. Understanding the recovery options and high level processing
of recovery can reduce the anxiety that is commonly associated with recovery.
Database recovery is one of the crucial tasks that is performed by the database
administrator. Refer to Developing a backup and recovery strategy.
The main backup and restore processes are:

db2agent
co-ordinator of all child processes and handles all communications with
the application.
db2bm
db2 buffer manipulator - uses prefetchers to process I/O requests. Places
data into or reads data from the backup restore buffers. Passes data to the
db2med process.
db2med
db2 media controllers - performs I/O with backup destination . Data is
written from or read into the backup / restore buffers. Buffers passed back
to db2bm.

Verifying backup images using the db2ckbkp command
DB2 UDB provides a utility called db2ckbkp that can be used to validate a backup
image. For details about the options, run db2ckbkp with no arguments.
Below is an example scenario in which the db2ckbkp utility may be used.

SQL1013N The database alias name or database name "ITB20" could not be
found.
PID:1081544(db2agent (ITB20) 0) TID:1 Appid:*LOCAL.DB2INST1.021012172526
database_utilities sqludMRWarn Probe:40 Database:ITB20
DiagData
.
16f6 ffff 433a 5c44 4232 4261 636b 7570 . C:\DB2Backup
5c49 5442 3230 2e30 5c44 4232 5c4e 4f44 \ITB20.0\DB2\NOD
4530 3030 305c 4341 544e 3030 3030 5c32 E0000\CATN0000\2
3030 3230 3931 335c 3039 3237 3339 2e30 0020913\092739.0
3031 00 01.
The above error indicates that the backup image cannot be found. It is possible
that the image exists in that specified location but may be corrupted. To verify that
the backup image is good, run db2ckbkp with the -a option to dump as much
information as possible.
If there is no problem with the backup image the following message will be
returned:
Image Verification Complete - successful.
Related reference:
v db2ckbkp - Check Backup
History file analysis

Every DB2 UDB database has a history file that records administrative operations.
A recovery history file is created with each database and is automatically updated.
During a database migration, the history file is migrated as well. The history file
can be accessed by issuing the following command:
db2 list history all for <dbname>
The database history file is invaluable in a recovery scenario. The history file is
individually restorable from any backup image. If the current database is unusable
or not available, and the associated recovery history file is damaged or deleted, an
option on the RESTORE command allows only the recovery history file to be
restored. The recovery history file can then be reviewed to provide information on
which backup to use to restore the database. For example, you can restore the
history file for the sample database with this command:
db2 restore database sample history file
The recovery history file provides enough information to recover a database or

table space using a backup image.
Extract Of A History File
Op Obj Timestamp+Sequence Type Dev Earliest Log Current Log Backup ID
-- --- ------------------ ---- --- ------------ ------------ --------------
B D 20050314154131001 F D
----------------------------------------------------------------------------
Contains 2 tablespace(s):
00001 SYSCATSPACE
00002 USERSPACE1
----------------------------------------------------------------------------

Comment: DB2 BACKUP SAMPLE ONLINE
Start Time: 20050314154131
End Time:
----------------------------------------------------------------------------
EID: 4 Location: C:\SAMPLE.0\DB2\NODE0000\CATN0000\20050314
Some of the important information in the history file is

Op This is the operation that was performed. In the exampled above it is a ″B″
which stands for backup. A list of the possible values for Op can be found
in the Command Reference under the LIST HISTORY command.
Obj This is the granularity of the backup. ″D″ for database backup . ″T″ for
tablespace backup.
Earliest Log
In the case of an online backup, this is the first log required for the
rollforward operation.
Current Log
The last log that was written to when the backup completed. In the case of
an online backup, this is the minimum log required for the backup to
complete.
Related reference:
v Understanding the recovery history file
DB2 Data Links Manager issues
1 Setting up a backup strategy for DB2 Data Links Manager

1 The following system setup and backup procedures are recommended for easier
1 recovery.
1 1. Place the Data Links File Manager (DLFM) database (usually called DLFM_DB),
1 any file systems under control of the Data Links Filesystem Filter (DLFF), the
1 server archive area, and the DLFM home directory on different file systems.
1 Ensure they do not share disks.
1 2. Back up any file systems under the control of the DLFF and the DLFM_DB on
1 a regular basis. In the event of disk failure, this backup will provide added
1 recoverability.
1 Many commercial backup-and-restore programs are available for all types of
1 file systems. These programs enable you to back up your file systems onto any
1 media that you choose, and typically provide easy-to-use restore operations. An
1 example of such a program is IBM’s Tivoli Storage Manager. Tivoli Storage
1 Manager runs in all environments that the DB2 database product supports.
1 3. Back up the archive area or archive server on a regular basis, if possible.
1 The archive area itself is used as a repository for DB2 Data Links
1 Manager-initiated backups of DLFM data, and possibly linked files. Keeping
1 these backups might result in some data redundancy, but having them will
1 provide added recoverability in the event of disk failure.
1 4. Attempt to have only one DB2 database associated with one or more DB2 Data
1 Links Managers. If possible, avoid having multiple databases associated with
1 one DB2 Data Links Manager, because certain recovery scenarios will become
1 more complex than is necessary. Perform full database backups of the DB2
1 databases and the DLFM database (DLFM_DB) on the Data Links server at
1 regular intervals. Coordinate the timing of the database backups so that they
1 occur together, and ensure that their data is transactionally consistent.

1 An alternative approach to backing up a Journaled File System
1 on AIX systems
1 This section describes a high availability backup method for an AIX Journaled File
1 System environment. Using this approach enables you to perform a file system
1 backup without stopping your DB2 Data Links Manager.
1 This backup method uses the following programs:

1 v online.sh
1 v quiesce.c
1 These programs are provided with your DB2 Data Links Manager installation
1 software and are located under the /sqllib/samples/dlfm directory. The online.sh
1 script, which calls the quiesce.c program, performs the following actions:
1 v Temporarily deactivates all the tables in databases that are registered with the
1 DB2 Data Links Manager. This stops any new Data Links Manager activity.
1 v Unmounts and remounts the file system as a read-only file system.
1 v Performs a file system backup.
1 v Unmounts and remounts the file system as a read-write file system.
1 v Resets and reactivates the database tables.
1 Prerequisites:
1 To use the online.sh script, you must have a catalog entry on the DB2 Data Links
1 Manager node for each database that is registered with the DB2 Data Links
1 Manager. You must also have the complete entry for the Data Links File System
1 (DLFS) on the /etc/filesystems file.
1 Procedure:
1 To execute the backup, perform the following steps:

1 1. Copy the quiesce.c CLI source file and the online.sh shell script to directory
1 of your choice on the Data Links server where the DLFS is located.
1 You will modify and work with these copies in the rest of this procedure.
1 2. Compile quiesce.c using the following command:
1 xlC -o quiesce -ldb2 -L$HOME/sqllib/lib -I$HOME/sqllib/include quiesce.c
1 3. Modify the online.sh script to suit your environment:
1 a. Select a backup command for the do_backup function of the online.sh
1 script. At the top of the do_backup function, there are several commented
1 lines of backup options. Remove the comment characters from the lines you
1 want to use, and insert any necessary backup commands and parameters.
1 b. At the top of the script, replace the specified default DLFM_INST
1 environment variable with your Data Links File Manager instance name.
1 c. On the next line, replace the default PATH_OF_EXEC environment variable
1 with the path where your quiesce.c executable resides.
1 4. From the Data Links server where the DLFS is located, run the script as
1 follows:
1 online.sh <filesystem_name>
1 File system backup and restore utilities

To ensure your IT environment’s data integrity and recoverability, it is imperative
that you institute a regular backup scheme.

Many commercial backup-and-restore programs are available for all types of file
systems. These programs enable you to back up your file systems onto any media
that you choose, and typically provide easy-to-use restore operations. An example
of such a program is IBM’s Tivoli Storage Manager. Tivoli Storage Manager runs in
all environments that the DB2 database product supports.
Your operating system will also provide backup-and-restore utilities. Windows NT

and Windows 2000 provide the Backup utility. AIX provides the System Storage
Management utilities. On Solaris operating environments, you can use the
ufsdump and ufsrestore commands. See your operating system documentation for
more information about the backup utilities.
Important: To reduce recovery time after a crash, use an incremental backup

strategy where level 0 refers to a full backup, and levels 1 through 9 refer to
incremental backups. A level n backup backs up only those files that have changed
since a level (n-1) backup. After a level n backup, the next backup to be taken will
be a level (n+1) backup.
1 Obtaining Data Links Manager environment information using

1 the dump utility (dlfm_dump)
1 When you report problems to IBM Service, you might be asked to run the Data
1 Links File Manager (DLFM) dump utility, dlfm_dump, to provide the service
1 personnel with details of your DB2 Data Links Manager environment.
1 The dlfm_dump utility enables you to obtain a ″snapshot″ of significant data that
1 is stored in the Data Links File Manager’s own database (called DLFM_DB by default
1 at installation time). With this data, you can examine various DLFM system
1 configuration details, plus other DLFM-related data on a Data Links server. By
1 default, the output is stored in a file in the directory where the dlfm_dump utility
1 is invoked. The data that gets placed into the output file includes:
1 v The version of the current DLFM_DB.
1 v The keys currently in use for access-token generation.
1 v Security control information.
1 v Registered databases and prefixes.
1 v Data Links File System (DLFS) directory tree structures.
1 v DATALINK columns that reference this Data Links server.
1 v DB2 backups involving linked files on this Data Links server.
1 v Lists of all linked and unlinked files on this Data Links server.
1 Important:
1 v If the DLFM is managing a large number of files, both those that are currently
1 linked as well as those that were previously linked, the output file could be very
1 large. Ensure that you have sufficient space in the file system where the output
1 file is to be written. You can use the amount of space that the DLFM_DB occupies,
1 or that a backup of the DLFM_DB occupies, to estimate the space that is required
1 for the dlfm_dump output file.
1 v Because the output of the dlfm_dump utility contains sensitive security
1 information, ensure that you place the output file in a secure directory.
1 You use the following syntax to invoke the dlfm_dump utility.
1 Requirement: You must execute this command from the Data Links server using
1 the Data Links Manager Administrator user ID.

1 ►► dlfm_dump -all ►◄
-o output_filename
1
1 -all A required parameter, which specifies that all significant DLFM definitions
1 and data are to be placed into an output file.
1 -o output_filename
1 An optional parameter, which specifies the name of the output file where
1 the data will be saved. The default output file name is dlfm_dump.log.
1 The dlfm_dump utility output does not include temporary data that the DLFM_DB
1 might be maintaining, such as data that is stored during transaction processing.
1 Note: The dlfm_dump utility output is minimally formatted, and is primarily

1 intended to assist in diagnosing problems, not for general reporting
1 purposes. To produce more readable output of Data Links server data, use
1 any or all of the following commands:
1 v dlfm retrieve
1 v dlfm list registered databases
1 v dlfm list registered directories
1 v dlfm list registered prefixes
1 v dlfm list registered replication access control
1 v dlfm list registered users
1 v dlfm list upd_in_progress files for db
1 v dlfm list upd_in_progress files for prefix
1 DB2 Data Links logging manager

1 When you report problems to IBM® Service, you might be asked to perform a
1 DB2® trace to provide the service personnel with details of your DB2 Data Links
1 Manager environment. DB2 traces can be especially useful for analyzing a
1 recurring and reproducible problem.
1 The Data Links server contains its own DB2 database, which is used as a logging
1 manager to track all linked files. Therefore, you can take a DB2 trace on both the
1 DB2 host and Data Links server machines where a problem was encountered, if
1 necessary. In certain situations, you might need to run a concurrent DB2 trace on
1 several machines. For example, you might encounter a communication problem
1 between a DB2 host and a Data Links server machine.
1 During the process of performing a DB2 trace, all actions and any relevant
1 parameter values get logged. Traces should be taken when there is minimum
1 activity on the machine to prevent the capture of unnecessary information.
1 The process of performing a trace has a global effect on the behavior of a DB2
1 instance. The degree of performance degradation is dependent on the type of
1 problem and on how many resources are being used to gather the trace
1 information.
DB2 Data Links server failure

In the case of a machine failure on a Data Links server, DB2 applications
interacting with the Data Links File Manager can hang. Use the db2 force
application command to force DB2 applications off the system.

Also, the following components might require recovery after a Data Links server
machine fails:
v The DB2 database containing a table which has a DATALINK column
v The Data Links Manager database (DLFM_DB)
v Data Links Filesystem Filter (DLFF)
v File systems under the control of the DLFF that are registered to the Data Links
Manager
v The Data Links Manager archive area
DB2 Data Links failure and recovery overview

If a disk that contains files referenced by a DATALINK column fails, all the user
files, along with the directory hierarchy of the file system, might be destroyed. To
recover from such a scenario, the administrator should make periodic backups of
the file system containing the user data and directory hierarchy, so that it can be
restored. The restored file system must preserve directory and file ownerships, and
time stamps.
After restoring the file system, the directory structure must be brought up to the
point-in-time of the crash by applying the directory changes that occurred after the
file system backup was taken. After this step, the DB2® RECONCILE command
must be run on all tables referencing files on the damaged disk. The db2_recon_aid
utility is provided to simplify this task.
Following a crash, there are three possible file states:

1. Files that are in linked state, and are referenced in a DATALINK column with
the RECOVERY NO attribute, are treated as follows:
v If the file is not found on the file system, the corresponding DATALINK column
value will be set to NULL.
v If the file is found, and the DATALINK column also has the READ PERMISSION
FS and WRITE PERMISSION FS attributes defined, no additional checks will be
made to validate the correctness of the file.
v If the file is found, and the referencing DATALINK column also has the
WRITE PERMISSION BLOCKED attribute defined, the file’s modification time and
size will be checked. If there is a mismatch in the values, the DATALINK
column value will be set to NULL.
2. Files that are in a linked state, when the corresponding DATALINK columns
have the RECOVERY YES attribute, will be restored from the archive server if the
file modification time is different than the file modification time at link time, or
if the file is not found.
If the modification time of the version on the file system is different, it is
renamed with extension .MOD so that the more recent changes are not lost. The
archived version is still retrieved, and the renamed version is reported in the
exception report.
If a renamed version of the file with .MOD extension already exists, the file will
not be retrieved, the DATALINK column value will be changed to NULL, and it
will be reported in the exception report.
3. Files that are in the unlinked state on the file server are not restored or checked
for correctness.

| Determining the file system directories that are needed to
| restore to the current point in time
| After a file system restore operation, you must manually bring the file system
| directory hierarchy to the current point in time by recreating directories as
| necessary.
| Use the Data Links Manager fsysadm.log file to help you determine directories
| that you need to recreate. Data always gets appended to the fsysadm.log.
| On AIX® and Solaris™ Operating Environments, the directory changes are logged
| in the INSTHOME/sqllib/fsysadm.log file, where INSTHOME is the home directory
| of the Data Links Manager Administrator. There is one entry for each event.
| Setting the attributes of a file is also logged. The format of the entries for the
| fsysadm.log file is as follows.
| Time = <timestamp> EUID = <integer> UID = <integer> GID = <integer> Mode = <octal>
| Action = <CREATE/REMOVE/SETATTR/RENAME> Object type = <DIR/FILE> Path = <fully qualified
| source name, destination name>
| where:
| v Time is the time of the activity in local time
| v EUID is the effective user ID of the user performing the action
| v UID is the user ID attribute of the file or directory that was created, or whose
| attributes were modified
| v GID is the group ID attribute of the file or directory that was created, or whose
| attributes were modified
| v Mode is the octal representation of the mode of the file or directory
| where Action can be:

| v CREATE indicates a directory was created
| v REMOVE indicates the file or directory was removed
| v SETATTR indicates the mode of the file or directory was modified by the user
| v RENAME indicates the file was renamed
| where Object type can be:

| v DIR the directory
| v FILE the file
| and where Path is the fully qualified path of the file or directory. If the action was
| RENAME, the destination name is displayed after the path information.
| On Windows® systems, the directory changes are logged in the

| x:\sqllib\dlfm\fsysadm.log file, where x: represents the drive where you installed
| DB2® Data Links Manager. A single event can have multiple entries, depending
| upon how many users or groups of users have Access Control Lists for the given
| file or directory. The format of the first entry for the fsysadm.log file is as follows.
| Time = <timestamp> User = <string> Action = <CREATE/REMOVE/SETATTR/RENAME>
| Object type = <DIR/FILE> Path = <fully qualified source name, destination name>
| The format of any additional entries associated with the first entry is as follows:
| ACE User = <string> Access = <Hex integer> ACE Type = <Hex integer>
| ACE Flags =<Hex integer>

| where:
| v Time is the time of the activity in local time
| v User is the name of the user performing the action
| v Owner is the name of the owner of the file or directory
| v Path is the fully qualified path of the file or directory
| v ACE User is the name of a user who has an ACL entry for this file or directory
| v Access* is the set of flags indicating the types of access the user has
| v ACE Type* is the type of ACE (for example, allow/deny)
| v ACE Flags* is a set of ACE type-specific control flags
| where Action can be:

| v CREATE indicates a directory was created
| v REMOVE indicates the file or directory was removed
| v SETATTR indicates the mode of the file or directory was modified by the user
| v RENAME indicates the file was renamed
| where Object type can be:

| v DIR the directory
| v FILE the file
| and where Path is the fully qualified path of the file or directory. If the action was
| RENAME, the destination name is displayed after the path information.
| ( * ) For the definitions of these hexadecimal values, refer to the Access Control
| Entry structures in the Microsoft® SDK documentation for Windows NT® and
| Windows 2000.
Reconciling database tables with DB2 Data Links file date

using the db2_recon_aid utility
The db2_recon_aid utility provides a mechanism for checking and running
RECONCILE on tables of a database that are potentially inconsistent with the
DATALINK file data on the file server.
Like the RECONCILE utility, the db2_recon_aid utility must be run on a DB2
server containing tables with DATALINK columns to be reconciled.
On AIX systems or Solaris Operating Environments, the db2_recon_aid utility is

located in the INSTHOME/sqllib/adm directory, where INSTHOME is the home
directory of the instance owner.
On Windows systems, the db2_recon_aid utility is located in x:\sqllib\bin

directory, where x: is the drive where you installed DB2 Data Links Manager.
To run the db2_recon_aid utility, use the following syntax:
►► db2_recon_aid ►

► ▼ -db database_name ►◄
-check
-reportdir directory_name
-selective -server data_links_server_name -prefixes prefix_list
database_name
A required value, which specifies the name of the database containing the
tables with DATALINK columns that need to be reconciled.
-check A parameter that instructs the utility to list the tables that might need
reconciliation. If you use this parameter, no reconcile operations will be
performed. This parameter is required when the -reportdir parameter is
not specified.
-reportdir directory_name
Required when the -check parameter is not specified. Specifies the
directory where the utility is to place a report for each of the reconcile
operations. For each table on which reconcile was performed, files of the
format <tbschema>.<tbname>.<ext> will be created where:
v <tbschema> is the schema of the table.
v <tbname> is the table name.
v <ext> is .ulk or .exp. The .ulk file contains a list of files that were
unlinked on the Data Links server, and the .exp file contains a list of
files that were in exception on the Data Links server.
-selective
An optional parameter that instructs the utility to process only those tables
with DATALINK columns containing file references that match the
specified -server and -prefixes criteria.
v If you use this parameter, you must also both the -server and -prefixes
parameters.
v If you do not use this parameter, then all Data Links servers and their
prefixes that are registered with the specified DB2 database will either be
reconciled, or will be flagged as needing reconciliation.
-server data_links_server_name
Required when the -selective parameter is used. Specifies the name of the
Data Links server for which the reconcile operation is to be performed. The
name value must be an IP hostname that is identical to the Data Links
server hostname registered with the specified DB2 database.
If this parameter is not used, all Data Links servers that are registered with
the specified DB2 database will be reconciled.
-prefixes prefix_list
Required when the -selective parameter is used. Specifies the name of one
or more Data Links File System (DLFS) prefixes. Prefix values must start
with a slash, and must be registered with the specified Data Links file
server. Separate multiple prefix names with a colon ( : ), but do not include
any embedded spaces. For example: /dlfsdir1/smith/:/dlfsdir2/smith/.
The path in a DATALINK column value is considered to match the
prefix_list if any of the prefixes in the list are a left-most substring of the
path.
If this parameter is not used, all prefixes for all Data Links servers that are
registered with the specified DB2 database will be reconciled.

Examples:db2_recon_aid -db STAFF -check
db2_recon_aid -db STAFF -reportdir /home/smith
db2_recon_aid -db STAFF -check -selective -server
dlmserver.services.com -prefixes /dlfsdir1/smith/
db2_recon_aid -db STAFF -reportdir /home/smith -selective -server
dlmserver.services.com -prefixes /dlfsdir1/smith/:/dlfsdir2/smith/
DB2 Data Links Manager recovery scenarios

This topic presents some sample DB2® Data Links Manager failure scenarios and
the steps required to recover from them.
The scenarios use the following terms:

DLFS file system
Registered prefix (example for AIX®: /dlink)
DLFM backup directory
Directory where files are backed up (example: /dlfm/dlfm_backup)
DLFM instance directory
Instance directory of the Data Links Manager Administrator ID (example:
/home/dlfm)
DLFM DB2 database
DB2 database that contains all metadata (DLFM_DB)
DB2 database
Registered database that contains DATALINK data type (example:
CROWN)
The example DB2 database is referred to as ″CROWN″ throughout all scenarios.
Important: Some of the following scenarios might require operations to be

performed on both the DB2 node and the DB2 File Manager node.
Scenario Recovery Steps

DB2 database is lost or 1. On the DB2 host, enter the following commands. As a
was accidentally dropped, result, all of the affected tables will be put into the
but DB2 backup and log Datalink_Reconcile_Not_Possible (DRNP) state.
files are available. db2 "restore database CROWN"
db2 "rollforward database CROWN to end of logs and stop"
Important: When a db2 "connect to CROWN"
database gets dropped, it

2. Place all tables with DATALINK columns into
happens within the period
Datalink_Reconcile_Pending (DRP) state with the following
of time as specified by the
commands:
DB2 database
db2 set integrity for <table> to datalink reconcile pending
configuration parameter db2 set integrity for <table> datalink reconcile pending immediate unchecked
dl_time_drop. db2 reconcile <table> dlreport <filename>
In this example, the DB2

database has not been
dropped from the Data
Links server.

DB2 database was Requirement: You must take a backup of your DLFM_DB
explicitly dropped, but database before updating it as described in the following
DB2 backup and log files procedure.
are available.
Recommendation: Work with IBM® Service in updating your
Important: When a DLFM_DB database. This database is an essential part of your
database gets dropped, it Data Links Manager configuration.
happens within the period 1. Ensure that the drop database operation is complete, and
of time as specified by the that all files associated with that database have been
DB2 database unlinked.
configuration parameter
2. On the Data Links server enter the following commands. It
dl_time_drop.
is critical that you set the dbid value in the db2 update
statement to be exactly as found in the db2 select
statement.
db2 "connect to dlfm_db"
db2 "select dbid, dbname, dbinst, hostname from dfm_dbid"
db2 "update dfm_dbid set action=5 where dbid=x’35B3D7BE0006BF7B’"
3. On the DB2 host, enter the following commands. As a
result, all of the affected tables will be put into the
Datalink_Reconcile_Not_Possible (DRNP) state.
db2 "restore database CROWN"
db2 "connect to CROWN"
4. For each table placed in DRNP state in step 3, enter the

following commands:
db2 set integrity for <table> datalink reconcile pending immediate unchecked
db2 reconcile <table> dlreport <filename>
The DLFM_DB database is 1. On the Data Links server enter the following commands:
lost, but the backup and db2 "restore database dlfm_db"
all log files for the db2 "rollforward database dlfm_db to end of logs and stop"
DLFM_DB database are 2. On the DB2 host, enter the following command to run the
available. db2_recon_aid utility. This utility automatically runs
RECONCILE for each table with URL file references to the
affected Data Links server:
db2_recon_aid -db CROWN -reportdir <dirpath>
-selective -server <dlm_hostname> -prefixes <dlfs_prefix>
v dlm_hostname is the registered IP hostname of the
affected Data Links Manager
v dlfs_prefix is the registered prefix corresponding to the
affected Data Links File System (DLFS)
The DLFM_DB database is 1. On the Data Links server enter the following commands:
lost, a backup of the db2 "restore database dlfm_db"
DLFM_DB database is db2 "rollforward database dlfm_db to end of logs and stop"
available, but not all of 2. On the DB2 host, enter the following commands. As a
the log files are available. result, all of the affected tables will be put into the
Datalink_Reconcile_Pending (DRP) state.
3. Place all tables with data link values into DRP state by
entering the following commands:

The Data Links File 1. Restore the DLFS from your storage manager.
System (DLFS) is lost.
2. On the DB2 host, enter the following command to run the
db2_recon_aid utility. Notice that by using the -selective
options, you can perform reconciliation for just the DLFS
that was lost.
The DLFM backup Restore the DLFM backup directory from your storage
directory is lost. manager.
The DLFS file system and 1. Restore the DLFM backup directory from your storage
the DLFM backup manager.
directory are lost.
2. Restore the DLFS from your storage manager.
that was lost.
The DLFM_DB database, 1. On the Data Links server enter the following commands:
the DLFM backup db2 "restore database dlfm_db"
directory, and the DLFS db2 "rollforward database dlfm_db to end of logs and stop"
file system are lost, but 2. Restore the DLFM backup directory from your storage
the backup and all log manager.
files for the DLFM_DB 3. Restore the DLFS from your storage manager.
database are available.
that was lost.

The DLFM_DB, the DLFS 1. On the Data Links server enter the following commands:
file system, and the DLFM db2 "restore database dlfm_db"
backup directory are lost. db2 "rollforward database dlfm_db to end of logs and stop"
The backup of the 2. Restore the DLFM backup directory from your storage
DLFM_DB database is manager.
available, but not all log
files are available.
that was lost.
The DB2 database, the 1. On the Data Links server enter the following commands:
DLFM_DB database, the db2 "restore database dlfm_db"
DLFS file system and db2 "rollforward database dlfm_db to end of logs and stop"
DLFM backup directory 2. Restore the DLFM backup directory from your storage
are lost, but backup, and manager.
all log files for the
DLFM_DB database are
available. 4. On the DB2 host, enter the following commands. As a
result, all of the affected tables will be put into the
Datalink_Reconcile_Not_Possible (DRNP) state.
5. For each table placed in DRNP state in step 4, enter the
following commands to place them in DRP state:
Note: Rollforward to a point-in-time might not put tables that

have all DATALINK columns defined with RECOVERY NO
into Datalink_Reconcile_Pending (DRP) state. For all such
tables, run the RECONCILE utility. You can also use the
db2_recon_aid utility to automatically identify and run
RECONCILE against those tables.
Business intelligence issues
DB2 Data Warehouse Center logging

The log holds records until a designated count limit is reached. When the count
limit is reached, the Data Warehouse Center automatically deletes the oldest logs.
Recommendation: Set the log record count to a size that holds 3 to 4 days worth
of records.
You cannot turn off the basic logging function.

Viewing DB2 Data Warehouse Center runtime errors
Use the basic logging function to view run-time (step processing) errors.
Procedure:
To view run-time errors:

1. Open the Data Warehouse Center desktop.
2. Click Data Warehouse Center -> Work in Progress.
The Work in Progress window opens.
3. Select the step for which you want to view errors.
4. To open the Log Viewer window and display the run-time errors for the
selected step, click Log.
Viewing DB2 Data Warehouse Center build-time errors

Use the basic logging function in the Data Warehouse Center to view build-time
errors.
Procedure:
To view build-time (table import, object creation, and step promotion) errors:
1. Open the Work in Progress window.
2. Click Work in Progress -> Show Log to open the Log Viewer window and
display the build-time errors for the Data Warehouse Center.
Viewing log entries in the Data Warehouse Center

If a step or process does not run successfully, you can use the Log Viewer to find
the cause of the failure.
Procedure:
To view log entries:

1. Select the step or process.
2. Click Log to open the Log Viewer window.
DB2 Data Warehouse Center component trace data
Running warehouse agents as a user process (Windows)

If you run a warehouse agent as a system process instead of a user process, it can
cause steps to fail. When the warehouse agent runs as a system process, it is not
authorized to connect to network drivers or products because the process does not
have a user ID. If the warehouse agent runs as a user process, the warehouse agent
has the characteristics of a user, including the ability to access network drives or
programs to which the user is authorized.
Procedure:
To run the warehouse agent as a user process:

1. Change the warehouse server, warehouse logger, and warehouse agent daemon
services to run as user processes by performing the following steps:
a. Double-click the Services icon in the Windows Control Panel folder.
b. Stop the services.

c. Select the service, and click Startup.
d. Click This Account.
e. Click the ... button after the This Account field to select a user ID.
The user ID must have administration authority in Windows NT, Windows
2000, or Windows XP and authorization to any required network drive.
f. Type the password for the user ID in the appropriate fields.
g. Click OK.
h. Restart the services.
2. If you are using a supplied OLAP server program, verify that the DB2 OLAP or
Essbase client is installed on a drive that is local to the agent that is running
the program.
3. If you are using a supplied OLAP server program, verify that the ARBORPATH
variable (set on the Essbase client or administrator) specifies a drive that is
local to the agent that is running the program and is specified as a system
variable.
Running a Data Warehouse Center component trace

Follow this procedure to run a Data Warehouse Center component trace.
Procedure:
To run a Data Warehouse Center component trace:

1. Right-click the warehouse object, and click Properties.
2. Specify the trace level for the warehouse control database, ODBC connection,
server, agent, or logger as directed by IBM Software Support.
3. Click OK.
4. Restart the services as requested.
5. Perform the failing operation.
6. Repeat steps 1 through 4 to set the trace level back to 0.
After completing this procedure, turn the trace level back to 0 to prevent
performance degradation.
You can run an agent trace independently for individual steps by setting the trace
level in the step’s Properties notebook on the Processing options page.
Error logging for warehouse programs and transformers

The supplied warehouse programs and transformers write errors to log files.
Warehouse programs
Supplied warehouse programs write data to the directory that is specified
in the VWS_LOGGING environment variable. Clear the directory of the log
files after sending the log files to IBM® Software Support.
Transformers
Transformer error messages start with DWC14. Transformer error
messages, warning messages, and returned SQL codes are stored as
secondary codes. For example, if a message starts with DWC14, a
transformer (stored procedure) caused the error. If the secondary code
includes an SQLCODE, an SQL statement in the transformer caused the
error. To enable logging, specify a log table name on the Processing

Options page of the Properties notebook for the step, and add a suffix of :n
to the log table name. The value of n indicates the logging level:
0 No logging
1 Log errors only
2 Log errors and warnings (this is the default logging level)
3 Log errors, warnings, and informational messages (for example,
starting and stopping a transformer)
For example, to indicate a log table named MyLogTable that contains log
entries at log level 3 or less, specify MyLogTable:3. In the output log tables,
the message type is one of the following values:
E Error
W Warning
Q SQL code
You can include a table space name after the log table name by appending
the log level to the table space name.
For example, to indicate a log table named MyLogTable that is located in
the MyTableSpace table space and contains entries at log level 3 or less,
specify MyLogTable,MyTableSpace:3.
The output log table in the warehouse control database contains detailed error
messages, warning messages, and SQL codes. In the output log tables, the message
type is one of the following values:
E Error
W Warning
Q SQL code
Tracing errors created by the Apply program

You can trace errors that are created when you use the Apply program.
Procedure:
To enable tracing for the Apply program, set the Agent Trace value = 4 in the
Warehouse Properties page. The Agent turns on full tracing for Apply when Agent
Trace = 4.
If you do not see any data in the CD table, then the Capture program is not
started, or you did not create changed data by updating the source table.
Start error trace files

The Data Warehouse Center creates three log files automatically when the logger is
not running. The log file names are IWH2LOGC.LOG, IWH2LOG.LOG, and
IWH2SERV.LOG. The Data Warehouse Center stores the files in the directory that
is specified by the VWS_LOGGING environment variable.
The log files are:

IWH2LOGC.LOG
When the logger is not running, processes write messages to this file. The

Data Warehouse Center server and the OLE server write to this file. The
file is created only if the logger stops. The file contains the complete
content of all messages that could not be sent.
IWH2LOG.LOG
The logger creates this file when it cannot start itself or when a trace is
activated. Key diagnostic information is written to this file when the logger
cannot start itself, and cannot write to the Data Warehouse Center log. If
you hear five beeps, or receive an application error, when the logger stops,
look in this file. The server cannot start if the logger cannot start.
IWH2SERV.LOG
The server log contains the startup message and grows when the server
trace is on.
Information Catalog Center issues
Finding help for information catalog center issues

The Information Catalog Center gives you some resources to help you solve
problems. These resources are:
v Online information and messages.
The Information Catalog Center provides extensive online information and
messages to help you solve problems. When you or your users receive a
message, use the online help first to resolve the problem.
You can find help for Information Catalog Center messages and explanations in
the Message Reference.
You can also look up message help from a DB2 command line by typing db2
Information Catalog Center ICMnnnnwhere nnnn represents the id number of
the message.
v Information Catalog Center trace file.
Recovering Information Catalog Center components and data

If you experience a hardware or software failure, you can lose your information
catalog database, your descriptive data, and parts of the component. If you backed
up the necessary components and data, you can restore your system, the
Information Catalog Center, and data.
Steps:
If a system failure occurs, perform the following steps after the database server’s
hard disk is restored and before your users access the information catalog:
1. Recover your database management system and reinstall the Information
Catalog Center, as necessary.
2. Restore the information catalog databases by using your backup files.
DB2 Cube Views issues
Database connection behavior using the OLAP Center

If you cannot connect to a database using the OLAP Center, check that the version
of DB2 Cube Views matches the version of the metadata tables in the DB2 catalog.
The following table shows how the OLAP Center behaves when the versions of
DB2 Cube Views and the metadata tables in the DB2 catalog are mismatched.

Version of the metadata Behavior of the OLAP
Version of DB2 Cube Views tables in the DB2 catalog Center
Not installed Not applicable Connection fails and the
OLAP Center displays an
error message
Version 8.1 None Connection fails and the
error message
Version 8.1 Version 8.1 Connection fails and the
error message
Version 8.2 None The OLAP Center can
configure the database for
use with DB2 Cube Views,
Version 8.2
Version 8.2 Version 8.1 The OLAP Center can
migrate the database for use
with DB2 Cube Views,
Version 8.2
Version 8.2 Version 8.2 Connection is successful.
DB2 Cube Views backward compatibility with the OLAP Center

and metadata API
DB2 Cube Views Version 8.2 has limited backward compatibility support for OLAP
Center and the metadata API.
DB2 Cube Views API supports the following:

v Version 8.2 API supports Describe requests from a Version 8.1 client. Describe is
the only Version 8.1 operation that is supported from the previous release.
DB2 Cube Views OLAP Center supports the following:

v Version 8.2 OLAP Center does not support Version 8.1 API or Version 8.1
metadata tables.
v Version 8.1 OLAP Center does not support Version 8.2 API or Version 8.2
metadata tables.
v Version 8.1 OLAP Center cannot connect to a Version 8.2 DB2 database.
v OLAP Center supports the import of Version 8.1 XML. When you import a
Version 8.1 XML file, OLAP Center migrates the XML to Version 8.2, using the
Translate operation, before importing the metadata.
v OLAP Center can export both Version 8.2 and Version 8.1 XML.
DB2 Cube Views db2mdapiclient supports the following:

v Version 8.1 db2mdapiclient supports Version 8.2 API if you use Version 8.2 XML.
v Version 8.2 db2mdapiclient supports Version 8.1 API if you use Version 8.1 XML.
Improving queries issued to federated data sources

If your queries issued to a remote data source do not improve as you expected
after you optimize, ensure that the federated system is set up correctly for DB2
Cube Views.

You must complete the following steps before you can expect queries issued to
remote data sources to improve:
v Enable your federated system for DB2 Cube Views.
v Create a complete cube model that satisfies the base rules, cube model
completeness rules, and optimization rules described in Metadata object rules.
v Optimize a cube model.
v If query performance does not improve, check the steps described in
Troubleshooting summary tables.
If, after completing the steps above, the performance of your queries still do not
improve, consider the following issues:
v Make sure that all applicable constraints are defined.
v Make sure that the DB2_MAXIMAL_PUSHDOWN setting is set to yes as
described in Defining remote data sources.
v Consider collocating the dimension tables that are involved in the queries on
your federated server. Collocating the dimensions, so that a replicated copy of
the dimension table exists on the federated server, might improve performance.
Connectivity issues
Before beginning to analyze a connectivity problem, it is imperative that you
understand your environment. You should know three pieces of information (as a
minimum):
1. Operating systems
2. DB2 products, versions and FixPak levels
3. Communication protocols used between each tier
If you are working in a multi- tier model (i.e. the client tier is supplemented by
one or connectivity servers) then it often helps to draw a diagram of your
environment. For an example of a connectivity server diagram, see DB2 Connect
Enterprise Edition as a connectivity server.
Two terms which you will come across when investigating connectivity issues are:
DRDA Application Requestor (AR) and DB2 UDB DRDA Application Server (AS).
These terms come from DRDA (Distributed Relational Database Architecture),
which is a set of protocols used by DB2 to coordinate communication. The
application requester is the code that handles the application side of a distributed
connection; it is the machine that you are connecting ″from″. The application server
is the code that handles the database side of the connection; it is the machine that
you are connecting ″to″.
If you are connecting to a host system, the term ″database″ may be unclear, since
that term is used differently on some mainframes. Refer to Host databases for
clarification of the term.
Click on this link for more information on Supported and non-supported client
configurations.
Click on this link for more information on Client-to-Server communication

scenarios.

Problem determination process for connectivity
Connectivity problems often involve multiple software, hardware and
communications products. Problem determination is best approached by a process
of elimination and refinement of the available data to arrive at a conclusion (the
location of the error).
Problem determination includes narrowing the scope of the problem and

investigating the possible causes. The proper starting point is to gather the relevant
information and determine what you know, what data has not been gathered, and
what paths you can eliminate. At a minimum answer the following questions.
v Are the communication paths operational?
v Has the initial connection been successful?
v Is the problem intermittent or persistent?
v Have there been any communication network changes that would make
previous directory entries invalid?
v Has the database been started?
v Is the communication breakdown between client and DB2 Connect workstation,
DB2 Connect workstation and host or iSeries(TM) database server, all clients or
one client?
v Is the problem only encountered within a specific application?
v What can you determine by the content of the message and the tokens returned
in the message?
v Will using diagnostic tools provide any assistance at this time?
v Are other machines performing similar tasks working correctly?
v If this is a remote task, is it successful if performed locally?
Some of these questions are particularly important, and will be expanded on.
Determining if communication paths are operational

It is important to isolate whether the problem you are encountering is limited to
DB2. For example, if you are encountering network errors, is it only via DB2
connections that you are encountering network errors? This becomes especially
important when the error is intermittent, as there is a greater likelihood that the
problem is outside of DB2 in those situations.
There are tools specific to each protocol that can be used to verify that
communication between the participating machines is possible external to DB2.
Here are some examples:
TCP/IP:
ping <hostname or ip address>
Try it out! Your successful result should look something like this:
Pinging myserver [1.23.45.678] with 32 bytes of data:
Reply from 1.23.45.678: bytes=32 time=921ms TTL=253
Ping statistics for 1.23.45.678:

Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milliseconds:

Minimum = 80ms, Maximum = 942ms, Average = 666ms

APPC:
Each different APPC communication product will have its own method of
verifying that a successful LU to LU session can be established. For example, on
AIX, you could start the SNA subsystem, start the node, start the link station, and
then test a session, all via the smit utility. Please refer to your specific APPC
communication product’s documentation for more details.
Named Pipes:
Try issuing a net use command from the client to a shared folder on the server.
The PCT tool
To assist in debugging the communication setup between the client and the server,
DB2 provides a tool called the Protocol Communications Tester. The tool is
launched by means of the pct command, which is found in the sqllib/bin directory.
It is a standalone tool designed to verify that a given LAN protocol is functional,
and that two endpoints can communicate with each other via that LAN protocol.
The tool can help solve protocol configuration-type problems because it effectively
removes DB2 from the picture until communication has been proven successful. If
you cannot pass these communication tests that are external to DB2, this indicates
that either your error is in the network layer or else you have not used correct
protocol-specific values in the tool. DB2 will not be able to connect across this
network until the network problems are corrected and the proper protocolspecific
values are identified.
For usage information related to this tool, please refer to the readme.pct file
located in the same directory as the executable file. The PCT tool can test the
following protocols:
v NetBIOS
v TCP/IP
Verifying the initial connection

If it is a first-time setup, there is a strong likelihood that this is a configuration
issue. First-time connection problems caused by configuration errors are most often
also persistent. To resolve the issue, verify that the setup has been properly
performed. If the connection worked previously, then the questions you should ask
yourself are:
v How long ago did it work?
v What has changed?
These types of questions will help you to pinpoint what network changes or fix
pack level upgrades, application changes or system outages may have taken place
between the point in time when it last worked, and now. Identifying what changed
and where (i.e. on the client, gateway, or server) helps you to determine where you
should focus your energy during investigation.
Determining intermittent and persistent problems

If the problem is persistent, i.e. it can be consistently recreated with a set of simple
steps, and then it is most likely a configuration issue. Your focus should be on
confirming the setup on the client and server. Refer to Confirming connection
configuration.

If the problem is intermittent, then it will be a bit more difficult to debug. It is not
the error message itself that is different -- usually you will get an SQL30081 error
message just like you do for persistent problems. The difficulty lies in the fact that
you cannot recreate it at will, and thus have fewer chances of capturing and
diagnosing the error. To pursue intermittent error messages, you will need to
investigate the circumstances of the error, for example:
v What is the network state at the time of the problem?
v What is the frequency of the occurrence?
v Is there a patterns as to what events are occurring at the time of the problem?
Depending on how frequently the errors are occurring, traces may be hard to
obtain. It is understandable that you may not want to have traces running for
extended periods of time just to capture the error -- particularly if the problem is
occurring on a production system. Thus you frequently need to rely on the
information in the db2diag.log file for these types of problems. Intermittent errors,
idle threads, connection pooling and connection concentrator describes additional
items to take into consideration when investigating intermittent connectivity errors.
Determining the location of the communication breakdown

A simple way to begin to answer this question is to determine whether or not you
can connect directly on the server. This is a very key question, and should be
something that you ask yourself every time that you encounter a communication
error. If you cannot connect locally at the time of the error, then you can eliminate
the client and the communication layer entirely from the problem scenario and can
focus instead on recreating and diagnosing the error directly on the server.
If a local connection is indeed successful, then the next step is to simulate a remote
connection directly on the server. This is done to reduce the layers of complexity
by recreating the entire problem on one machine. This is often referred to as a
loopback. The loopback is just a method of cataloging the local database as if it
were on a remote machine, as described here. Example: Creating a loopback
connection
If the connection works via this loopback connection test, but fails from a remote
client, then you might want to investigate these areas:
1. Is there a firewall between the client and the server?
2. Are the node and database catalogs on the client similar to those used in the
loopback node. Note that the service name (if used) could be different in the
client’s node catalog, but that the port number that it maps to on each machine
should nonetheless match.
If the connection fails via the local loopback test, then you can ignore the client
entirely from this point on. If the error is persistent, it is likely a configuration
issue, and the server is not correctly set up to listen for incoming requests. You
would thus need to focus on confirming the server configuration values. Refer to
Confirming connection configuration.
If you can only recreate the problem from remote clients, then you should
determine whether all clients or one client experiences the problem.
v If the problem occurs only for a subset of the clients connecting to this server,
then you should identify the commonalties and differences between these
clients, and compare their configurations (DB2 and network).
v If the problem is occurring across all of your clients, then a simple comparison is
not an option for you.

v If the problem is intermittent but is happening across all of the servers at the
same time, then this points to a greater likelihood that the error originates on the
server or network, rather than the client.
If you have only encountered the problem within a single application, it is

beneficial to try connecting from the DB2 command prompt. Whether or not this
will help you narrow the scope of the problem will depend on whether or not the
error is easily reproducible in the application. If you can consistently recreate the
problem from your application, but can connect successfully from the DB2 prompt,
then it would be wise to examine your application.
As you determine the answers to these questions, you should find that you are
successfully narrowing the problem description and thus increase the chances of
determining where the root cause of the problem lies.
Testing connectivity using the PCT tool - example

This example illustrates the use of the TCP/IP version of the PCT tool. To test
connections made over NETBIOS protocols, the PCT instructions are very similar.
The main difference is that you use the command pctn instead of pctt.
All of the pct files can be found in INSTHOME/sqllib/bin, where INSTHOME is the
home directory of the instance owner. On Windows systems, the files are located in
the sqllib\bin subdirectory.
To perform the connectivity test, you must:

1. Set up PCT on the server
2. Start the PCT listener on the server
3. Set up PCT on the client
4. Launch a PCT communication test on the client.
Configuring PCT on the server

On your server, do the following:
1. Verify your server’s hostname and IP address via the following command:
pctt -r
D:\sqllib\bin>pctt -r
TCPIP Resources...
+--------------------------------------------+
| HostName : myserver |
| HostAddress: myserver.ibm.com |
| IPAddress : 123.456.7.890 |
+--------------------------------------------+
2. Generate a pct.ini file: pctt /gt The resultant pct.ini file will appear as follows:
#--------------------------------------------------------
# PCT.INI
# Protocol Communication Tester INI File
# This file contains the Protocol Specific Configuration
# Parameters & values to be used in conjunction with
# the Protocol Communication Tester (PCTx).
#
# Defaults will be used if the parameter is not listed.
#--------------------------------------------------------
[TCPIP]
ServerHostName =

ServerHostAddress =
ServerIPAddress =
ServiceName =
ServerPort = 49433
Ensure that you run this command in a directory to which you have write
access, or else the pct.ini file will not be created successfully.
3. Edit the pct.ini file to reflect the protocol-specific parameter values for this
server and for the particular instance that your client will be trying to connect
to. When you use this tool to debug problems in DB2, it is important to
configure the pct.ini file with the exact values that your DB2 instance will use
to listen for incoming client requests, if possible. For example, you should set
the ServiceName or ServerPort value in pct.ini to the same value as is found in
the SVCENAME field in the output from the command.
"db2 get database manager configuration"
Note:
If you have set up PCT to use the same values as the DB2 instance uses, you
must stop the instance while testing PCT, since PCT will return an error in
situations where there is contention for ports that are already in use.
You only need to enter a value for either the ServerHostName or the
ServerIPAddress. The same holds true for the ServiceName, ServerPort pair.
Either one alone will suffice.
Starting the PCT listener on the server

The PCT tool must be running on the server prior to launching it on the client,
otherwise the PCT tool on the client will fail, indicating that the connection was
refused.
To make the PCT tool start listening for incoming requests, issue the command:
pctt s
When PCT starts successfully on the server, you see output like this:
- Reading configuration parameters
==> pct.ini file was found... using file specified protocol values!
Protocol Tester values...
Client/Server : Server
Connections : 1
Buffer Size : 500
Transaction Iterations : 1
Trace : OFF
Service : Send/Recv/Verify
Keep Connections : NO
Delayed Send (secs) : 0
Local TCPIP specific values...
Hostname : myserver
HostAddress : myserver.ibm.com
IPAddress : 123.456.7.890
- Initializing the Protocol Date: 2-9-2005 Time: 22:54:52:97

| retcode = < 0> ----[ TCPIP.socket ]-----[ SUCCESS ]------------
| retcode = < 0> ----[ TCPIP.setsockopt ]-----[ SUCCESS ]------------
| retcode = < 0> ----[ TCPIP.bind ]-----[ SUCCESS ]------------
| retcode = < 0> ----[ TCPIP.listen ]-----[ SUCCESS ]------------
- Listening for Remote Clients Date: 2-9-2005 Time: 22:54:52:97
At this point the server side of the PCT tool is waiting to receive a communication
from the client.

Configuring PCT on the client
On the client, do the following:
1. Generate a pct.ini file: pctt /gt. The file will be blank, and appear exactly as it
did when performing this step on the server.
2. Edit this pct.ini file to reflect the protocol-specific parameter values for the
server to which you want to connect. In order to use this tool to debug
problems in DB2, it is important to fill out the pct.ini file with the exact values
that you used when catalogging the remote DB2 server from the client. For
example, if you used the server’s hostname (as opposed to the IP address)
when you ran the CATALOG TCPIP NODE command on the client, then do the
same in the pct.ini file. Likewise if you used a servicename (as opposed to a
port number), then do the same in pct.ini. Here is an example of an updated
pct.ini file:
#--------------------------------------------------------
# PCT.INI
# Protocol Communication Tester INI File
# This file contains the Protocol Specific Configuration
# Parameters & values to be used in conjunction with
# the Protocol Communication Tester (PCTx).
#
# Defaults will be used if the parameter is not listed.
#--------------------------------------------------------
[TCPIP]
ServerHostName = myserver.ibm.com
ServerHostAddress =
ServerIPAddress =
ServiceName =
ServerPort = 50000
Starting a PCT communication test on the client

Issue the client connection attempt using the following command: pctt c
By default, PCT will perform an endpoint--to--endpoint connect operation,

followed by send and receive operations with data verification, and finally it will
disconnect.
Successful results appear as follows:

Client/Server : Client
Connections : 1
Buffer Size : 500
Trace : OFF
Hostname : myclient
HostAddress : myclient.ibm.com
IPAddress : 99.99.99.99
[(The results of a failed connection attempt are shown later.) The information in the
above output shows the PCT tool’s default settings as well as the protocol-specific
information about your client machine.]

Server TCPIP specific Values...
Server Hostname : myserver.ibm.com
Server HostAddress : myserver.ibm.com
Server IPAddress : 123.456.7.890
Service Name :
Server Port : 50000
[The information above was picked up from your client’s pct.ini file and shows the
values that will be used to identify the server.]
- Initializing the Protocol Date: 02/09/05 Time: 19:59:17: 78
- Connecting to Remote System Date: 02/09/05 Time: 19:59:17: 93
| retcode = < 0> ----[ TCPIP.connect ]-----[ SUCCESS ]------------
- Connection established! 1 Date: 02/09/05 Time: 19:59:19: 15
At this point, the connection step of the test has been completed.
- Sending Service Request Date: 02/09/05 Time: 19:59:19: 15
| retcode = < 0> ----[ TCPIP.send ]---[ Bytes= 316 ]------------
- Receiving Date: 02/09/05 Time: 19:59:19: 15
| retcode = < 0> ----[ TCPIP.receive ]---[ Bytes= 4 ]------------
- Server info:
Computer Name :
Operating System:
Version :
Hostname :
- Send/Receive/Verify Data Loop Date: 02/09/05 Time: 19:59:19: 78
- Sending Data Date: 02/09/05 Time: 19:59:19: 78
| retcode = < 0> ----[ TCPIP.send ]---[ Bytes= 500 ]------------
- Receiving Date: 02/09/05 Time: 19:59:19: 78
[By default, the PCT tool passes some data from the server to the client, to verify
the connection. Then the connection is closed (see below).]
- Terminating the Connection Date: 02/09/05 Time: 19:59:19:156
| retcode = < 0> ----[ TCPIP.sock_close ]-----[ SUCCESS ]------------
+------------------------------------------------+
| Protocol Connection Test Completed - SUCCESS |
+------------------------------------------------+
An example of an error message which can occur if PCT fails to complete the test
successfully is as follows:
Client/Server : Client
Connections : 1
Buffer Size : 500
Trace : OFF
Hostname : myclient
HostAddress : myclient.ibm.com
IPAddress : 99.999.9.999
Server TCPIP specific Values...
Server Hostname : myserver

Server HostAddress : myserver.ibm.com
Server IPAddress : 1.23.45.678
Service Name :
Server Port : 44000
[Note the incorrect port number used in this test.]

- Initializing the Protocol Date: 02/09/05 Time: 20:07:23:406
- Connecting to Remote System Date: 02/09/05 Time: 20:07:23:421
| retcode = <10061> +----[ TCPIP.connect ]-----[ ERROR ]------------
+===== ERROR ==========================================================
| retcode = <10061> -> Connection refused
+======================================================================
The return code given here is a TCP/IP return code 10061 which maps to
CONNECTION REFUSED. The cause of the error in this example was an incorrect
ServerPort value in the client’s pct.ini file. It did not match the ServerPort value
defined in the server’s pct.ini file.
If you receive errors such as this via the PCT tool, do not go any further with
problem determination within DB2 until you have rectified these network
problems, since these problems are occurring external to DB2.
Creating a loopback connection - example

When you catalog a connection to a local database as if it were a remote database,
this is sometimes referred to as a ″loop back″. You may wish to create such a
connection temporarily if you are investigating a connectivity problem. It may also
be necessary if you have a multi-threaded application which connects to a local
database on AIX and it is running out of shared memory segments (resulting in
intermittent SQL1224N error messages).
What follows is an example illustrating how to create a loopback connection to the

SAMPLE database.
1. Examine your database directory.
Find the SAMPLE database entry in the output from the command ″db2 list
db directory″.
...
Database X entry:
Database alias = SAMPLE
Database drive = C:\DB2
Database release level = a.00
Comment =
Directory entry type = Indirect
Catalog node number = 0
Alternate server hostname =
Alternate server port number =
...
When you have finished, there will be another catalog listed here as remote,
but ultimately pointing to the same local database. There are methods of
performing loopback catalogs such that the loopback alias name matches what
the original database name was, however those methods will not be used in
this example. Those steps are often followed when an application will be using
the loopback catalog permanently, and you do not want to have to change the
way the application or script was referring to the database. We are keeping it
simple for the sake of this example.

2. Determine which TCP/IP port number your instance is listening on for
incoming requests. To verify the port number or service name that the instance
is listening on for incoming requests, issue the command ″db2 get dbm cfg″.
...
TCP/IP Service name (SVCENAME) = db2c_DB2
...
This information, along with the hostname or IP address of this server, is all
that you need in order to catalog the node.
Note: If you are creating a loopback connection in order to investigate a client

connectivity issue, then if the client has cataloged the server using an IP
address rather than a hostname, then you should do the same in your
loopback catalog for consistency.
3. Configure the loopback connection
a. Catalog the loopback node using the following command:
db2 catalog tcpip node <new name> remote <hostname/ip address> server <service name/port nu
The port number used in this command must be the same port number as
the instance is listening on for incoming requests(as determined in step 2).
If a service name is used, it must map to the same port number value as the
instance is using.
For example:
db2 catalog tcpip node loopnode remote myhostname server db2c_DB2
db2 terminate
db2 list node directory
...
Node X entry:
Node name = LOOPNODE
Comment =
Directory entry type = LOCAL
Protocol = TCPIP
Hostname = myhostname
Service name = db2cDB2
...
b. Catalog the loopback database, using the following command:
db2 catalog db <database name> as <loop back db alias> at node <loop back node name, per ab
For example:
db2 catalog db sample as loopsamp at node loopnode
db2 terminate
db2 list db directory
Database X entry:
Database alias = LOOPSAMP

Node name = LOOPNODE
Comment =
Directory entry type = Remote
Catalog database partition number = -1
Database Y entry:

Database drive = D:\DB2

Comment =
Catalog database partition number = 0
4. Test the loop-back connection:
To connect to the loopback database, enter:
db2 connect to <loop back db alias> user <userid>
You will be prompted for a password.
For example:
C:>db2 connect to loopsamp user tester
Enter current password for tester:
Database Connection Information
Database server = DB2/NT 8.2.0

SQL authorization ID = TESTER
Local database alias = LOOPSAMP
5. If you wish to remove this node and database catalog after you have completed
your tests, you can do so as follows:
db2 uncatalog node loopnode
db2 uncatalog db loopsamp
db2 terminate
Confirming TCP/IP configuration for a DB2 instance

Configuration errors are commonly the cause if your communication problem has
the following characteristics:
v It is not an intermittent error;
v The connection has either never worked, or has changed recently;
v You can connect successfully locally;
v You tried creating a loopback connection on the server and it failed. (In this
situation, you should continue to debug the problem within the loopback
connection on the server. Once you have resolved the issue there, you will know
what changes need to be made on the client (if any)).
Verifying the TCP/IP setup on the server

The steps to confirm the setup on the server are as follows:
1. Verify the existence of the database by issuing one of the following commands:
v db2 list db directory
v db2 list db directory show detail
If this is a multi-tier environment, you may see the database cataloged as a
remote database.
Database X entry:

Comment =
2. Verify that the instance is configured to start listening on the appropriate
protocol, via the command db2set -all. Look for the parameter DB2COMM.

C:\>db2set -all
[e] DB2PATH=D:\IBM\SQLLIB
[i] DB2INSTOWNER=DB2INST1
[i] DB2PORTRANGE=60000:60003
[i] DB2INSTPROF=D:\IBM\SQLLIB
[i] DB2COMM=TCPIP
[g] DB2_EXTSECURITY=YES
[g] DB2SYSTEM=LISAC02
[g] DB2PATH=D:\IBM\SQLLIB
[g] DB2INSTDEF=DB2
[g] DB2ADMINSERVER=DB2DAS00
3. Verify that the appropriate protocol-specific parameters are set in the database
manager configuration settings, so that DB2 knows what values to listen on.
For example, issue db2 get dbm cfg and look in the following section:
...
NetBIOS Workstation name (NNAME) =
TCP/IP Service name (SVCENAME) = db2cDB2
...
4. If the previous step resulted in a service name instead of a port number for the
SVCENAME field in your environment, then confirm that the value listed there
is mapped to a unique port number in the operating system’s /etc/services file.
For example:
DB2_DB2 60000/tcp #These ports are reserved for the DB2 Fast Communications Manager
DB2_DB2_1 60001/tcp
DB2_DB2_2 60002/tcp
DB2_DB2_END 60003/tcp
db2c_DB2 50000/tcp #This is the connection port for instance DB2
Starting DB2 communication listeners on the server

1. Create a backup copy of the db2diag.log, then clear the contents of the original
db2diag.log file. Raise the diagnostic level to 4, then stop and restart this DB2
instance, via the following commands:
db2 update dbm cfg using diaglevel 4
Note: You must refer to the db2diag.log instead of the notification log in this
situation, because the notification log does not contain this level of detail
even when using notification level 4.
2. Verify that the db2diag.log shows that the listeners for the specific protocol
which we are interested have been started successfully.
2005-02-11-15.59.58.828000-480 I2036155H322 LEVEL: Info
PID : 2536 TID : 2632 PROC : db2syscs.exe
INSTANCE: DB2 NODE : 000
FUNCTION: DB2 UDB, common communication, sqlcctcp_start_listen, probe:80
MESSAGE : DIA3000I "TCPIP" protocol support was successfully started.
3. If you had not performed one of the necessary setup steps, you would receive
the following error when you performed the db2start command:
SQL5043N Support for one or more communications protocols
failed to start successfully. However, core database manager
functionality started successfully.
4. This would be accompanied by specific error messages in the db2diag.log file if
DIAGLEVEL was set to 4 at the time. For example, if you I had not set the
SVCENAME parameter in the database manager configuration settings, you
would receive the following error in the db2diag.log:
2005-02-11-16.14.05.843000-480 E2206H457 LEVEL: Error
FUNCTION: DB2 UDB, common communication, sqlcctcpconnmgr, probe:50

MESSAGE : ADM7006E The SVCENAME DBM configuration parameter was not
configured. Update the SVCENAME configuration parameter using the
service name defined in the TCP/IP services file.
5. Once you have confirmed that the DB2 protocol-specific listeners are started
successfully for the instance, don’t forget to downgrade the diagnostic level, so
that you do not suffer performance problems. Returning it to the default
diagnostic level of 3 is done as follows:
Verifying the TCP/IP setup on the client

The steps to confirm the setup on the client are as follows:
1. Confirm the values that you used in your cataloging of the remote server’s
instance by issuing one of the following commands:
v db2 list node directory
v db2 list node directory show detail
C:\>db2 list node directory show detail
...
Node 1 entry:
Node name = MYNODE
Comment =
Protocol = TCPIP
Hostname = myserver
Service name = 50000
Remote instance name =
System =
Operating system type = None
...
The ″Service name″ value in this output should match the port number defined
in the server instance’s database manager configuration SVCENAME setting. If
a service name is used instead of a port number, it is important that the port
number to which the service name on the client maps (in the client’s
etc/services file) should match the port number used on the server.
2. Verify that you can ping the Hostname value exactly as it appears in the results
from step (1).
3. Another test that you can perform is to use the telnet command to check
whether there is something (in this case, hopefully a DB2 instance) listening on
a particular port number on a particular server. For example: telnet
myserver.ibm.com 50000. Telnet does not have to be enabled on the server. If
there is indeed something listening on that port then you should see that the
telnet window will open, but then hang. This means that you have indeed
reached the server, and that something is indeed listening on that port. If, on
the other hand, you receive an immediate error, then either:
a. You have used the wrong hostname or IP address values. If you used a
hostname here, try using the IP address, in case it is a hostname resolution
problem;
b. You have used the wrong service name or port number. If you used a
service name, try using the port number instead, in case the service name
was not mapped correctly in the /etc/services file on this client;
c. There is a firewall between the client and server, and it is not allowing
communication on this port. Verify with your network administrator
whether or not this is the case;
d. DB2 on the server is not listening on this port. Refer back ″Verification of
the setup on the server″ in order to confirm the setup of the server.

4. If the telnet and ping tests worked, confirm that the proper database values
appear in the database and DCS catalogs, as appropriate. Errors in these
catalogs do not usually result in communication errors, however.
At this point you have confirmed that the configuration of both the client and the
server are correct from the DB2 point of view. If you continue to receive a
communication error, these are situations where you should pursue obtaining
diagnostics like db2 traces.
Related reference:
v Configuring TCP/IP communications for a DB2 instance
v Verifying port range availability on participating computers (Windows)
v Enabling communications between database partition servers (UNIX)
Confirming NetBIOS configuration for a DB2 instance

Configuration errors are commonly the cause if your communication problem has
the following characteristics:
v It is not an intermittent error;
v The connection has either never worked, or has changed recently;
v You can connect successfully locally;
v You tried creating a loopback connection on the server and it failed. (In this
situation, you should continue to debug the problem within the loopback
connection on the server. Once you have resolved the issue there, you will know
what changes need to be made on the client (if any)).
Verifying the NetBIOS setup on the server

The steps to confirm the setup on the server are as follows:
1. Verify the existence of the database by issuing one of the following commands:
v db2 list db directory
v db2 list db directory show detail
If this is a multi-tier environment, you may see the database cataloged as a
remote database.
Database X entry:
Comment =
2. Verify that the instance is configured to start listening on the appropriate
protocol, via the command db2set -all. Look for the parameter DB2COMM.
C:\>db2set -all
[e] DB2PATH=D:\IBM\SQLLIB
[i] DB2INSTOWNER=DB2INST1
[i] DB2PORTRANGE=60000:60003
[i] DB2INSTPROF=D:\IBM\SQLLIB
[i] DB2COMM=NETBIOS
[g] DB2_EXTSECURITY=YES
[g] DB2SYSTEM=LISAC02
[g] DB2PATH=D:\IBM\SQLLIB
[g] DB2INSTDEF=DB2
[g] DB2ADMINSERVER=DB2DAS00

3. Verify that the appropriate protocol-specific parameters are set in the database
manager configuration settings, so that DB2 knows what values to listen on.
For example, issue db2 get dbm cfg and look in the following section:
...
NetBIOS Workstation name (NNAME) = server1
...
Note: The NNAME must be unique among all NetBIOS nodes in the network. The
maximum length of the nname is 8 characters.
Starting DB2 communication listeners on the server:

1. Create a backup copy of the db2diag.log, then clear the contents of the original
db2diag.log file. Raise the diagnostic level to 4via the following commands:
Note: You must refer to the db2diag.log instead of the notification log in this
situation, because the notification log does not contain this level of detail
even when using notification level 4.
2. Verify that the db2diag.log shows that the listeners for the specific protocol
which we are interested have been started successfully.
2005-02-11-16.51.44.640000-480 I2987H328 LEVEL: Info
FUNCTION: DB2 UDB, common communication, sqlccnb_start_listen, probe:62
MESSAGE : DIA3001E "NETBIOS" protocol support was successfully started.
3. If you had not performed one of the necessary setup steps, you would receive
the following error when you performed the db2start command:
SQL5043N Support for one or more communications protocols
failed to start successfully. However, core database manager
functionality started successfully.
4. This would be accompanied by specific error messages in the db2diag.log file if
DIAGLEVEL was set to 4 at the time. For example, if you had not set the SVCENAME
parameter in the database manager configuration settings, you would receive
the following error in the db2diag.log:
2005-02-11-17.04.18.750000-480 I1906H457 LEVEL: Error
FUNCTION: DB2 UDB, common communication, sqlccnbconnmgr_child, probe:34
MESSAGE : DIA3426C "NETBIOS" protocol support: Workstation name (nname) for
this server is NOT valid. Enter a valid Workstation name (nname) in
the Database Manager Configuration.
5. Once you have confirmed that the DB2 protocol-specific listeners are started
successfully for the instance, don’t forget to downgrade the diagnostic level, so
that you do not suffer performance problems. Returning it to the default
diagnostic level of 3 is done as follows:
Verifying the NetBIOS setup on the client

The steps to confirm the setup on the client are as follows:
1. Verify that an appropriate NetBIOS Workstation name (NNAME) value is set in
the database manager configuration settings on the client. For example, issue
db2 get dbm cfg and look in the following section:
...
NetBIOS Workstation name (NNAME) = client1
...

Note: The NNAME must be unique among all NetBIOS nodes in the network.
The maximum length of the nname is 8 characters.
2. Verify the local adapter that will be used for the connection. Refer to
Determining the logical adapter number of the client for the NetBIOS
connection (Windows).
3. Confirm the values that you used when cataloging the remote server’s instance
by issuing one of the following commands:
v db2 list node directory
v db2 list node directory show detai
C:\>db2 list node directory
...
Node 1 entry:
Node 3 entry:
Node name = NETNODE

Comment =
Protocol = NETBIOS
Adapter number = 0
Server NNAME = SERVER1
...
The ″Adapter number″ entry should indicate the client’s logical adapter number (as
determined in step 2). The ″Server NNAME″ entry should indicate the server’s
Workstation name (nname, as specified in the server’s database manager
configuration file).
At this point you have confirmed that the configuration of both the client and the
server are correct from the DB2 point of view. If you continue to receive a
communication error, these are situations where you should pursue obtaining
diagnostics like db2 traces.
Related reference:
v Configuring NetBIOS communications on the client using the CLP
v NetBIOS parameter values worksheet
v nname - NetBIOS workstation name configuration parameter
Testing a database connection using the Configuration

Assistant
Click on this link for more information.
Common connectivity problems

Initial connection mainframe or midrange server is not
successful
Click on this link for more information on ″Initial connection mainframe or
midrange server is not successful″.
Problems encountered after an initial connection

Click on this link for more information on ″Problems encountered after an initial
connection″.

Common DB2 Connect problems
Database crashes
In order to investigate these issues, you need to understand and distinguish the
difference between a database crash and an application crash.
Determine if the database instance has crashed:
To determine if the database instance has crashed the administration notification

log at various probe points must be examined. In this file, you may see the
following:
PID:1310914(db2agent (BIDB) 0) TID:1 Appid:*N0.db2inst2.050128190001
oper system services sqloEDUCodeTrapHandler Probe:10 Database:SAMPLE
ADM0503C An unexpected internal processing error has occurred. ALL DB2

PROCESSES ASSOCIATED WITH THIS INSTANCE HAVE BEEN SHUTDOWN. Diagnostic information has been recorde
The db2diag.log would show additional information, as follows:

2005-01-20-11.15.04.507340-360 E5806002A512 LEVEL: Severe
PID : 282632 TID : 1 PROC : db2agntp 0
FUNCTION: DB2 UDB, oper system services, sqloEDUCodeTrapHandler,
probe:10
MESSAGE : ADM0503C An unexpected internal processing error has
occurred. ALL DB2 PROCESSES ASSOCIATED WITH THIS INSTANCE HAVE BEEN
SHUTDOWN. Diagnostic information has been recorded. Contact IBM Support for further assistance.
2005-01-20-11.15.04.508711-360 E5806515A645 LEVEL: Severe

PID : 282632 TID : 1 PROC : db2agntp 0
FUNCTION: DB2 UDB, oper system services, sqloEDUCodeTrapHandler,
probe:20
DATA #1 : Signal Number Recieved, 4 bytes
11
DATA #2 : Siginfo, 64 bytes
0x0FFFFFFFFFFFE050 : 0000 000B 0000 0000 0000 0032 0000 0000
...........2....
0x0FFFFFFFFFFFE060 : 0000 0000 0000 0000 0000 0001 1170 A370
.............p.p
0x0FFFFFFFFFFFE070 : 0000 0000 0000 0000 0000 0000 0000 0000
................
0x0FFFFFFFFFFFE080 : 0000 0000 0000 0000 0000 0000 0000 0000
................
In this example, second db2diag.log entry shows that the function

sqloEDUCodeTrapHandler has returned a Signal number of 11. This means that the
DB2 signal handler has caught a signal #11. On the UNIX platform the header file
called signal.h is usually located in /usr/include/sys. In this example you will
determine that a signal #11 is a segmentation violation (SIGSEGV):
Extract of the signal.h header file

...
#define SIGBUS 10 /* (*) bus error (specification exception) */
#define SIGSEGV 11 /* (*) segmentation violation */
#define SIGSYS 12 /* (*) bad argument to system call */
...
This is the first indication that the database has indeed crashed due to a
segmentation violation and the database signal handler has caught the signal. The

next step is to determine the process ID (PID) that has crashed. We return to the
db2diag.log file, to find the ″abnormally terminated process″:
2005-01-20-11.15.04.558786-360 I5807161A433 LEVEL: Severe
PID : 2027586 TID : 1 PROC : db2gds 0
FUNCTION: DB2 UDB, oper system services, sqloEDUSIGCHLDHandler, probe:50
DATA #1 : String, 160 bytes
Detected the death of an EDU with process id 282632
The signal number that terminated this process was 11
Look for trap files (t282632.*) in the dump directory
The function sqloEDUSIGCHLDHandler at probe 50 has provided the process id of the

problematic EDU, and the name of the trap file to reference. For this example you
will get a file called t282632.000 in the DIAGPATH directory. On some platforms such
as AIX, a CORE file may be generated as well.
The trap file contains a stack traceback of all the functions on the stack for the
process that crashed. For more information about how to interpret the stack trace,
and how to search for appropriate APARs, refer to Analyzing the stack trace back
Related reference:
v Trap files
v Common signals and exceptions that cause trap file generation
db2dart and INSPECT–Database inspection tools

The db2dart and inspect tools can be used in database crash situations to ensure
that the databases have not been corrupted. An overview of each of these tools
follows.
Overview of the db2dart tool:
db2dart is a command which can be used to verify the architectural correctness of

databases and the objects within them. It can also be used to display the contents
of database control files in order to extract data from tables that might otherwise
be inaccessible.
To display all of the possible options, simply execute the db2dart utility without
any parameters. Some options that require parameters, such as the table space ID,
are prompted for if they are not explicitly specified on the command line.
By default, db2dart will create a report file with the name databaseName.RPT. For
single-partition database environments, the file is created in the current directory.
For multiple-partition database environments, the file is created under a
subdirectory in the diagnostic directory. The subdirectory is called DART####, where
#### is the partition number.
db2dart accesses the data and metadata in a database by reading them directly
from disk. Because of that, db2dart should never be run against a database that
still has active connections. If there are connections, db2dart will not know about
pages in the buffer pool, control structures in memory, etc. and may report false
errors as a result. Similarly, if you run db2dart against a database that requires
crash recovery or that has not completed roll-forward recovery, similar
inconsistencies might result due to the inconsistent nature of the data on disk.
Inspecting databases, table spaces, and tables via db2dart:

The default behaviour for db2dart is to inspect the entire database. Only the
database name must be provided in this case. For example:
C:\>db2dart sample
The requested DB2DART processing has completed successfully!

Complete DB2DART report found in:
C:\IBM\SQLLIB\DB2\DART0000\SAMPLE.RPT
Complete DB2DART report found in: SAMPLE.RPT
As the output states, the full db2dart report can be found in the file SAMPLE.RPT.
You will also notice that in this case db2dart did not find any problems with the
database.
If a database is very large and you are only interested in one table space, you can
use the /TS option. When using this option, you must either provide the table
space ID on the command line (by specifying the /TSI parameter) or you can let
db2dart prompt you for it. If you do not know the table space ID, you can obtain
it via the command ″DB2 LIST TABLESPACES″ command. For example, to inspect the
USERSPACE1 table space (which has a table space ID of 2 in the sample database),
either of these commands will work:
db2dart sample /ts /tsi 2
db2dart sample /ts <= When prompted for the table space ID, enter "2".
Similarly, a single table and its associated objects (LOBs, indexes, etc.) can be
inspected using the /T option. When using this option, you must provide either the
table name or object ID and the ID of the table space in which the table resides. To
determine the object ID and table space ID for a table, you can query the FID and
TID columns of the SYSIBM.SYSTABLES catalog table. For example, determine the
object ID and table space ID for the EMP_PHOTO table in the sample database by
executing the following query:
C:\>db2 connect to sample
Database server = DB2/NT 8.2.0
SQL authorization ID = LISAC
Local database alias = SAMPLE
C:\>db2 "select creator,name,tid,fid from sysibm.systables where name =

’EMP_PHOTO’"
CREATOR NAME TID FID
------------------ --------------------- ------ ------
DB2 EMP_PHOTO 2 8
1 record(s) selected.
C:\>db2 connect reset

DB20000I The SQL command completed successfully.
To inspect this table, execute either of the following db2dart commands:

db2dart sample /t /tsi 2 /oi 8
db2dart sample /t <= When prompted for the table ID and table space ID, enter "8 2".
As mentioned above, the table name can be specified instead of the object ID:
db2dart sample /t /tsi 2 /tn EMP_PHOTO
db2dart sample /t <= When prompted for the table name and table space ID,enter "EMP_PHOTO 2".
Dumping formatted table data via db2dart:
If a table space or table becomes corrupt for any reason (for example due to a bad
disk or disk controller), attempts to access the table through SQL may not work.
(The SQL statement may fail with an error or the database may be marked bad

and all connections will be dropped.) In such a case, entries will likely be written
to the db2diag.log file, indicating that a bad page was encountered.
2004-10-12-16.49.20.119228+120 I3292G436 LEVEL: Error
PID : 14974 TID : 605992128 PROC : db2bm.14206.5
APPHDL : 0-777 APPID: *LOCAL.db2inst1.000
FUNCTION: DB2 UDB, buffer pool services, sqlbrdpg, probe:1143
RETCODE : ZRC=0x86020001=-2046689279=SQLB_BADP "page is bad"
DIA8400C A bad page was encountered.
If you see such entries, you should run db2dart against the database (or table
space) to determine the extent of the damage.
If this happens, it may be necessary to extract all of the data possible so that the
table space and table can be rebuilt. In such a situation, the /DDEL option of
db2dart can be used to extract the table data and place it into a delimited ASCII
file. Note that due to the nature of ASCII files, some columns (such as LOB
columns) cannot be extracted from the table. db2dart will tell you if this is the
case.
When using the /DDEL option, you must provide a table space ID, object ID,
starting page number, and number of pages. To extract all of the pages, use 0 for
the starting page number and some very large number for the number of pages.
(Specifying more pages than actually exist will not cause any problems.)
The ORG table in the sample database resides in table space 2 and has an object ID
of 2. To extract all of the data from this table, execute this command:
db2dart sample /ddel
When prompted, enter either of the following lines of input:

2 2 0 1000
ORG 2 0 1000
You will then be presented with the column definitions for the table and will be
asked to specify an output file name:
Table object data formatting start.
Please enter
Table ID or name, tablespace ID, first page, num of pages:
(suffic page number with ’p’ for pool relative)
2 2 0 1000
5 of 5 columns in the table will be dumped.
Column numbers and datatypes of the columns dumped:
0 SMALLINT
1 VARCHAR() -VARIABLE LENGTH CHARACTER STRING
2 SMALLINT
Default filename for output data file is TS2T2.DEL,
do you wish to change filename used? y/n
You can choose the default or specify a new one. The output file will be created in
the current directory be default.
When the extraction is complete, you will see output as follows:

Filename used for output data file is TS2T2.DEL. If existing file, data will
be appended to it.
Formatted data being dumped ...

Dumping Page 0 ....
Table object data formatting end.
The requested DB2DART processing has completed successfully!
Overview of the INSPECT command:
An online inspection command called INSPECT was introduced in DB2 Version 8.

This command allows you to perform similar database, table space, and table
checking as is done by db2dart. There are many benefits to using the INSPECT
command including the ability to run it while there are active connections against
the database. Also, it is built into the engine which means that significant
performance gains are achieved through the use of buffer pools and prefetchers.
Feature comparison of DB2DART and INSPECT:

Table 2. Feature Comparison of DB2DART vs. INSPECT
Test Performed DB2dart V8 Inspect V8
SMS Offline Online Offline Online
Check U
Tablespace Files
DMS
Extent maps U
pointed at by
more than 1
object
Extent Maps U
Agree with
Space Maps
For Space Map U
check every
page for
consistency bit
errors
For Space Map U
check buffer
pool fields in
page header
Extent Map U
consistency bit
errors
Extent Map U
consistency bit
errors
Data Object
Consistency bits U U
error
Buffer pool U U
fields in page
header
Data Manager U U
Fields in page
header

Table 2. Feature Comparison of DB2DART vs. INSPECT (continued)
SMS Offline Online Offline Online
Each Row record U U
type and length
Rows not U U
overlapping
Special control U
rows contents
Variable length U
columns length
and position
LF/LOB U
descriptors in
table rows
Summary Total U
Pages, Used
Pages, Free
Space Percentage

Index Object -
each page Offline Online Offline Online
Consistency bits U U
error
Buffer pool U U
fields in page
header
Index Manager U U
fields in page
header
Index key U U
location, length,
overlapping
Index ordering U
of keys
Uniqueness of U
unique keys
Summary Total U
Pages, Used
Pages
Block Map Object
Check U U
consistency bit
errors
Buffer pool U U
pages in Page
Header

Index Object -
each page Offline Online Offline Online
Block Manager U U
fields in Page
Header
Summary Total U
Pages, Used
Pages
Long Field Object
Allocation U U
structures
LOB Object
Allocation U U
Structures
Summary Total U
Pages, Pages
Used
Miscellaneous
Format & Dump U
data pages
Format & dump U
index pages
Format data U
rows to
delimited ASCII
Mark an index U
invalid
Initialize a data U
page to empty
Block Map Offline Online Offline Online
Object
Check U
consistency bits
error
Check buffer U
pool fields in
page header
Check block U
manager in page
header
Related reference:
v db2dart - Database Analysis and Reporting Tool Command
v INSPECT command
v First failure data capture locations

Bringing an OFFLINE table space ONLINE
Sometimes a problem will cause a table space to go OFFLINE. After fixing the
problem, perform the following steps to return the state to ONLINE:
Prerequisites:
You will need the ability to disconnect all applications from the database and to
issue the ALTER TABLESPACE command.
Procedure:
1. Disconnect all applications and reconnect to the database. You will see that the
table space can be taken out of the OFFLINE state.
2. Use the ALTER TABLESPACE ... SWITCH ONLINE statement to bring the table
space up while the rest of the database is still up and deployed.
v If the table space can be brought up successfully after issuing the command,
or if the table space was not in the OFFLINE state to begin with, DB2 will
return an SQLCODE of 0.
v If the table space cannot be brought up successfully because there are still
problems with one or more of the containers, DB2 will return an SQLCODE
of —293. You can force the database to restart by using the RESTART ...
DROP PENDING TABLESPACE, but will have to drop any faulty table
spaces afterwards.
Related reference:
v “ALTER TABLESPACE statement” in the SQL Reference, Volume 2
v “RESTART DATABASE Command” in the Command Reference
Data movement issues

This section will help you identify the source of common load, import or export
problems.
Loading data
LOAD is a utility used to bulk-load data from a file or pipe into a database. LOAD
consists of up to three phases. The load phase loads the data, the build phase
builds the indexes if there are any, and the delete phase deletes duplicates if there
are unique indexes.
The main load processes are as follows: (X is a number identifying one of many)
v db2agent
v db2lmr Load Media Reader. This process reads the input.
v db2lbmX Load Buffer Manipulator. Writes loaded data to the database.
v db2lfrmX Load Formatter. Formats the input data into internal form.
v db2lrid Ridder. Organises data to be written to disk. Performs the index sort.
v db2lmwX Load Media Writers. Write the load copy.
LOAD failure:
What do you do when a LOAD fails and the table space is not accessible? The
following options are available.
v Restart the LOAD

v Terminate the LOAD
v Restore from the most recent backup and rollforward to a point-in-time before
the failed LOAD
v Drop and rebuild the table space.
A typical message for a failed load in the db2diag.log is as follows:

PID:51945658(db2lload 5) TID:1 Appid:*N17.db2p001.0E00B7172332
database utilities call_sqluvload Probe:40 Database:SAMPLE
Load Error: Load failed for table DB2INST1.ORG in tablespace 1.
In addition, you can issue the following command to examine the state of the table
space whose table is being loaded into:
db2 list tablespaces show detail
Typical output is as follows:

Tablespace ID = 2
Name = USERPACE1
Type = System managed space
Contents = Any data
State = 0x0004
Detailed explanation:
Quiesced: EXCLUSIVE
...
Number of quiescers = 1
Quiescer 1:
Tablespace ID = 2
Object ID = 3
Use the db2tbst tool that is provided with DB2 to determine that state of the table
space.
For the above example, you would issue the following command
db2tbst 0x0004
This returns:
State = Quiesced Exclusive
In this state, the table space is inaccessible. To remove the table space from this
state, issue:
db2 quiesce tablespaces for table staff reset
where STAFF is the name of the table being loaded into.
Prior to Version 8, load required exclusive access to table spaces that contained
objects belonging to the table being loaded. In Version 8, load operates at the table
level and no longer requires exclusive access to the table space. The load utility
does not quiesce the table spaces involved in the load operation, and uses table
space states only for load operations with the COPY NO option specified. Refer to
Load overview.
Bringing a table back online:
Load also uses the table space state LOAD PENDING while in the load or build phase,
or DELETE PENDING during the delete phase. These states persist after the load

is interrupted or the load fails. They can only be released by performing a LOAD
RESTART, a LOAD TERMINATE, or a LOAD REPLACE. See Pending states after a load
operation.
:
v Load overview
v Pending states after a load operation
v db2tbst - Get Tablespace State Command
v Load temporary files
v Load utility log records
Message files from IMPORT, EXPORT, and LOAD

Click on this link for more information.
Extender issues
DB2 Spatial and Geodetic Extenders

How to interpret DB2 Spatial Extender messages
You can work with DB2® Spatial Extender through four different interfaces:
v DB2 Spatial Extender stored procedures
v DB2 Spatial Extender functions
v DB2 Spatial Extender Command Line Processor (CLP)
v DB2 Control Center
All interfaces return DB2 Spatial Extender messages to help you determine
whether the spatial operation that you requested completed successfully or
resulted in an error.
The following table explains each part of this sample DB2 Spatial Extender
message text:
GSE0000I: The operation was completed successfully.
Table 3. The parts of the DB2 Spatial Extender message text
Message text part Description
GSE The message identifier. All DB2 Spatial Extender messages
begin with the three-letter prefix GSE.
0000 The message number. A four digit number that ranges from
0000 through 9999.
I The message type. A single letter that indicates the severity of
message:
C Critical error messages
N Non-critical error messages
W Warning messages
I Informational messages
The operation was The message explanation.
completed successfully.

The explanation that appears in the message text is the brief explanation. You can
retrieve additional information about the message that includes the detailed
explanation and suggestions to avoid or correct the problem. To display this
additional information:
1. Open an operating system command prompt.
2. Enter the DB2 help command with the message identifier and message number
to display additional information about the message. For example:
DB2 "? GSEnnnn"
where nnnn is the message number.
You can type the GSE message identifier and letter indicating the message type in
uppercase or lowercase. Typing DB2 ″? GSE0000I″ will yield the same result as
typing db2 ″? gse0000i″.
You can omit the letter after the message number when you type the command.
For example, typing DB2 ″? GSE0000″ will yield the same result as typing DB2 ″?
GSE0000I″.
Suppose the message code is GSE4107N. When you type DB2 ″? GSE4107N″ at the
command prompt, the following information is displayed:
GSE4107N Grid size value "<grid-size>" is not valid where it is used.
Explanation: The specified grid size "<grid-size>" is not valid.
One of the following invalid specifications was made when the grid index
was created with the CREATE INDEX statement:
- A number less than 0 (zero) was specified as the grid size for the
first, second, or third grid level.
- 0 (zero) was specified as the grid size for the first grid level.
- The grid size specified for the second grid level is less than the grid
size of the first grid level but it is not 0 (zero).
- The grid size specified for the third grid level is less than the grid
size of the second grid level but it is not 0 (zero).
- The grid size specified for the third grid level is greater than 0 (zero)
but the grid size specified for the second grid level is 0 (zero).
User Response: Specify a valid value for the grid size.
msgcode: -4107
sqlstate: 38SC7
If the information is too long to display on a single screen and your operating
system supports the more executable program and pipes, type this command:
db2 "? GSEnnnn" | more
Using the more program will force the display to pause after each screen of data
so that you can read the information.
DB2 Spatial Extender stored procedure output parameters

DB2® Spatial Extender stored procedures are invoked implicitly when you enable
and use Spatial Extender from the DB2 Control Center or when you use the DB2

Spatial Extender CLP (db2se). You can invoke stored procedures explicitly in an
application program or from the DB2 command line.
This topic describes how to diagnose problems when stored procedures are
invoked explicitly in application programs or from the DB2 command line. To
diagnose stored procedures invoked implicitly, you use the messages returned by
the DB2 Spatial Extender CLP or the messages returned by the DB2 Control Center.
These messages are discussed in separate topics.
DB2 Spatial Extender stored procedures have two output parameters: the message
code (msg_code) and the message text (msg_text). The parameter values indicate
the success or failure of a stored procedure.
msg_code
The msg_code parameter is an integer, which can be positive, negative, or
zero (0). Positive numbers are used for warnings, negative numbers are
used for errors (both critical and non-critical), and zero (0) is used for
informational messages.
The absolute value of the msg_code is included in the msg_text as the
message number. For example
v If the msg_code is 0, the message number is 0000.
v If the msg_code is -219 , the message number is 0219. The negative
msg_code indicates that the message is a critical or non-critical error.
v If the msg_code is +1036, the message number is 1036. The positive
msg_code number indicates that the message is a warning.
The msg_code numbers for Spatial Extender stored procedures are divided
into the three categories shown in the following table:
Table 4. Stored procedure message codes
Codes Category
0000 - 0999 Common messages
1000 - 1999 Administrative messages
2000 - 2999 Import and export messages
msg_text
The msg_text parameter is comprised of the message identifier, the
message number, the message type, and the explanation. An example of a
stored procedure msg_text value is:
GSE0219N An EXECUTE IMMEDIATE statement
failed. SQLERROR = "<sql-error>".
The explanation that appears in the msg_text parameter is the brief

explanation. You can retrieve additional information about the message
that includes the detailed explanation and suggestions to avoid or correct
the problem.
For a detailed explanation of the parts of the msg_text parameter, and

information on how to retrieve additional information about the message,
see the topic: How to interpret DB2 Spatial Extender messages.
Working with stored procedures in applications:

When you call a DB2 Spatial Extender stored procedure from an application, you
will receive the msg_code and msg_text as output parameters. You can:
v Program your application to return the output parameter values to the
application user.
v Perform some action based on the type of msg_code value returned.
Working with stored procedures from the DB2 command line:
When you invoke a DB2 Spatial Extender stored procedure from the DB2
command line, you receive the msg_code and the msg_text output parameters.
These output parameters indicate the success or failure of the stored procedure.
Suppose you connect to a database and want to invoke the ST_disable_db stored
procedure. The example below uses a DB2 CALL command to disable the database
for spatial operations and shows the output value results. A force parameter value
of 0 is used, along with two question marks at the end of the CALL command to
represent the msg_code and msg_text output parameters. The values for these
output parameters are displayed after the stored procedure runs.
call db2gse.st_disable_db(0, ?, ?)
Value of output parameters

--------------------------
Parameter Name : MSGCODE
Parameter Value : 0
Parameter Name : MSGTEXT

Parameter Value : GSE0000I The operation was completed successfully.
Return Status = 0
Suppose the msg_text returned is GSE2110N. Use the DB2 help command to
display more information about the message. For example:
"? GSE2110"
The following information is displayed:

GSE2110N The spatial reference system for the
geomentry in row "<row-number>" is invalid.
The spatial reference system’s
numerical identifier is "<srs-id>".
Explanation: In row row-number, the geometry that is

to be exported uses an invalid spatial reference system.
The geometry cannot be exported.
User Response: Correct the indicated geometry or

exclude the row from the export operation by
modifying the SELECT statement accordingly.
msg_code: -2110
sqlstate: 38S9A
DB2 Spatial Extender function messages

The messages returned by DB2® Spatial Extender functions are typically embedded
in an SQL message. The SQLCODE returned in the message indicates if an error
occurred with the function or that a warning is associated with the function. For
example:

v The SQLCODE -443 (message number SQL0443) indicates that an error occurred
with the function.
v The SQLCODE +462 (message number SQL0462) indicates that a warning is
associated with the function.
The following table explains the significant parts of this sample message:
DB21034E The command was processed as an SQL statement because it was
not a valid Command Line Processor command. During SQL processing it
returned: SQL0443N Routine "DB2GSE.GSEGEOMFROMWKT"
(specific name "GSEGEOMWKT1") has returned an error
SQLSTATE with diagnostic text "GSE3421N Polygon is not closed.".
SQLSTATE=38SSL
Table 5. The significant parts of DB2 Spatial Extender function messages
Message part Description
SQL0443N The SQLCODE indicates the type of problem.
GSE3421N The DB2 Spatial Extender message number and message type.
The message numbers for functions range from GSE3000 to

GSE3999. Additionally, common messages can be returned
when you work with DB2 Spatial Extender functions. The
message numbers for common messages range from GSE0001
to GSE0999.
Polygon is not closed The DB2 Spatial Extender message explanation.
SQLSTATE=38SSL An SQLSTATE code that further identifies the error. An
SQLSTATE code is returned for each statement or row.
v The SQLSTATE codes for Spatial Extender function errors
are 38Sxx, where each x is a character letter or number.
v The SQLSTATE codes for Spatial Extender function
warnings are 01HSx, where the x is a character letter or
number.
An example of an SQL0443 error message:

Suppose that you attempt to insert the values for a polygon into the table
POLYGON_TABLE, as shown below:
INSERT INTO polygon_table ( geometry )
VALUES ( ST_Polygon ( ’polygon (( 0 0, 0 2, 2 2, 1 2)) ’) )
This results in an error message because you did not provide the end value to
close the polygon. The error message returned is:
DB21034E The command was processed as an SQL statement because it was
not a valid Command Line Processor command. During SQL processing it
returned: SQL0443N Routine "DB2GSE.GSEGEOMFROMWKT"
(specific name "GSEGEOMWKT1") has returned an error
SQLSTATE with diagnostic text "GSE3421N Polygon is not closed.".
SQLSTATE=38SSL
The SQL message number SQL0443N indicates that an error occurred and the
message includes the Spatial Extender message text GSE3421N Polygon is not
closed.
When you receive this type of message:

1. Locate the GSE message number within the DB2 or SQL error message.
2. Use the DB2 help command (DB2 ?) to see the Spatial Extender message
explanation and user response. Using the above example, type the following
command in an operating system command line prompt:
DB2 "? GSE3421"
The message is repeated, along with a detailed explanation and recommended user
response.
DB2 Spatial Extender CLP messages

The DB2® Spatial Extender CLP (db2se) returns messages for:
v Stored procedures, if invoked implicitly.
v Shape information, if you have invoked the shape_info subcommand program
from the DB2 Spatial Extender CLP. These are informational messages.
v Migration operations.
v Import and export shape operations to and from the client.
Examples of stored procedure messages returned by the DB2 Spatial Extender

CLP:
Most of the messages returned through the DB2 Spatial Extender CLP are for DB2
Spatial Extender stored procedures. When you invoke a stored procedure from the
DB2 Spatial Extender CLP, you will receive message text that indicates the success
or failure of the stored procedure.
The message text is comprised of the message identifier, the message number, the
message type, and the explanation. For example, if you enable a database using the
command db2se enable_db testdb, the message text returned by the Spatial
Extender CLP is:
Enabling database. Please wait ...
GSE1036W The operation was successful. But

values of certain database manager and
database configuration parameters
should be increased.
Likewise, if you disable a database using the command db2se disable_db testdb
the message text returned by the Spatial Extender CLP is:
GSE0000I The operation was completed successfully.
The explanation that appears in the message text is the brief explanation. You can
retrieve additional information about the message that includes the detailed
explanation and suggestions to avoid or correct the problem. The steps to retrieve
this information, and a detailed explanation of how to interpret the parts of the
message text, are discussed in a separate topic.
If you are invoking stored procedures through an application program or from the
DB2 command line, there is a separate topic that discusses diagnosing the output
parameters.
Example of shape information messages returned by the Spatial Extender CLP:
Suppose you decide to display information for a shape file named office. Through
the Spatial Extender CLP (db2se) you would issue this command:
db2se shape_info -fileName /tmp/offices
This is an example of the information that displays:
Shape file information

----------------------

File code = 9994
File length (16-bit words) = 484
Shape file version = 1000
Shape type = 1 (ST_POINT)
Number of records = 31
Minimum X coordinate = -87.053834

Maximum X coordinate = -83.408752
Minimum Y coordinate = 36.939628
Maximum Y coordinate = 39.016477
Shapes do not have Z coordinates.
Shapes do not have M coordinates.
Shape index file (extension .shx) is present.
Attribute file information

--------------------------
dBase file code = 3
Date of last update = 1901-08-15
Number of records = 31
Number of bytes in header = 129
Number of bytes in each record = 39
Number of columns = 3
Column Number Column Name Data Type Length Decimal
1 NAME C ( Character) 16 0
2 EMPLOYEES N ( Numeric™) 11 0
3 ID N ( Numeric) 11 0
Coordinate system definition: "GEOGCS["GCS_North_American_1983",

DATUM["D_North_American_1983",SPHEROID["GRS_1980",6378137,298.257222101]],
PRIMEM["Greenwich",0],UNIT["Degree",0.017453292519943295]]"
Examples of migration messages returned by the Spatial Extender CLP:
When you invoke commands that perform migration operations, messages are
returned that indicate the success or failure of that operation.
Suppose you invoke the migration of the database mydb using the command db2se
migrate mydb -messagesFile /tmp/migrate.msg. The message text returned by the
Spatial Extender CLP is:
Migrating database. Please wait ...
GSE0000I The operation was completed successfully.
DB2 Control Center messages

When you work with DB2® Spatial Extender through the DB2 Control Center,
messages will appear in the DB2 Message window. Most of the messages that you
will encounter will be DB2 Spatial Extender messages. Occasionally, you will
receive an SQL message. The SQL messages are returned when an error involves
licensing, locking, or when a DAS service is not available. The following sections
provide examples of how DB2 Spatial Extender messages and SQL messages will
appear in the DB2 Control Center.
DB2 Spatial Extender messages:
When you receive a DB2 Spatial Extender message through the Control Center, the
entire message text appears in the text area of DB2 Message window, for example:
GSE0219N An EXECUTE IMMEDIATE statement
failed. SQLERROR = "<sql-error>".

SQL messages:
When you receive an SQL message through the Control Center that pertains to
DB2 Spatial Extender:
v The message identifier, message number, and message type appear on the left
side of the DB2 Message window, for example: SQL0612N.
v The message text appears in the text area of the DB2 Message window.
The message text that appears in the DB2 Message window might contain the SQL
message text and the SQLSTATE, or it might contain the message text and the
detailed explanation and user response.
An example of an SQL message that contains the SQL message text and the
SQLSTATE is:
[IBM][CLI Driver][DB2/NT] SQL0612N "<name>" is a duplicate name. SQLSTATE=42711
An example of an SQL message that contains the message text and the detailed
explanation and user response is:
SQL8008N
The product "DB2 Spatial Extender" does not have a valid
license key installed and the evaluation period has expired.
Explanation:
A valid license key could not be found and the evaluation

period has expired.
User Response:
Install a license key for the fully entitled version of the

product. You can obtain a license key for the product by
contacting your IBM® representative or authorized dealer.
Tracing DB2 Spatial Extender problems with the db2trc

command
When you have a recurring and reproducible DB2 Spatial Extender problem, you
can use the DB2 trace facility to capture information about the problem. The DB2
trace facility is activated by the db2trc system command. The DB2 trace facility
can:
v Trace events
v Dump the trace data to a file
v Format trace data into a readable format
Restrictions:
Activate this facility only when directed by a DB2 technical support representative.
On UNIX operating systems, you must have SYSADM, SYSCTRL, or SYSMAINT

authorization to trace a DB2 instance.
On Windows operating systems, no special authorization is required.
Procedure:
To trace the DB2 Spatial Extender events to memory, follow these basic steps:
1. Shut down all other applications.

2. Turn the trace on. The DB2 Support technical support representative will
provide you with the specific parameters for this step. The basic command is:
db2trc on
Restriction: The db2trc command must be entered at a operating–system
command prompt or in a shell script. It cannot be used in the DB2 Spatial
Extender command–line interface (db2se) or in the DB2 CLP.
You can trace to memory or to a file. The preferred method for tracing is to
trace to memory. If the problem being recreated suspends the workstation and
prevents you from dumping the trace, trace to a file.
3. Reproduce the problem.
4. Dump the trace to a file. For example:
db2trc dump january23trace.dmp
This command creates a file (january23trace.dmp) in the current directory with

the name that you specify, and dumps the trace information in that file.
You can specify a different directory by including the file path. For example, to
place the dump file in the /tmp/spatial/errors directory, the syntax is:
db2trc dump /tmp/spatial/errors/january23trace.dmp
Dump the trace immediately after the problem occurs.
5. Turn the trace off. For example:
db2trc off
6. Format the data as an ASCII file. You can sort the data two ways:
v Use the flw option to sort the data by process or thread. For example:
db2trc flw january23trace.dmp january23trace.flw
v Use the fmt option to list every event chronologically. For example:
db2trc fmt january23trace.dmp january23trace.fmt
DB2 XML Extender

Starting the trace for XML Extender
Purpose:
Records the XML Extender server activity. To start the trace, apply the on option to
dxxtrc, along with the user profile and the name of an existing directory to contain
the trace file. When the trace is turned on, the file, dxxINSTANCE.trc, is placed in
the specified directory. INSTANCE is the value of DB2INSTANCE. Each DB2 UDB
instance has its own log file. INSTANCE is the numeric UID value assigned to the
User Profile for which the trace was started. The trace file is not limited in size.
Syntax:
Starting the trace from the Qshell:
►► dxxtrc on user_profile trace_directory ►◄
Starting the trace from the iSeries Navigator:

call schema.QZXMTRC(’on’, ’user_profile’, ’trace_directory’);
Starting the trace from the OS command line:

call QDBXM/QZXMTRC PARM(on user_profile ’trace_directory’)

Parameters:
Table 6. Trace parameters
Parameter Description
user_profile The name of the user profile associated with
the job within which the XML Extender is
running.
trace_directory Name of an existing path and directory
where the dxxINSTANCE.trc is placed.
Required, no default.
Examples:
The following examples show starting the trace, with file, dxxdb2inst1.trc, in the
/u/user1/dxx/trace directory.
From the Qshell:
dxxtrc on user1 /u/user1/trace
From the iSeries Navigator:
call myschema.QZXMTRC(’on’, ’user1’, ’/u/user1/trace’);
From the OS command line:
call QDBXM/QZXMTRC PARM(on user1 ’/u/user1/trace’)
Stopping the trace

Purpose:
Turns the trace off. Trace information is no longer logged.
Recommendation: Because running the trace log file size is not limited and can
impact performance, turn trace off in a production environment.
Syntax:
Stopping the trace from the Qshell:
►► dxxtrc off user_profile ►◄
Stopping the trace from the iSeries Navigator:

call schema.QZXMTRC(’off’, ’user_profile’);
Stopping the trace from the OS command line:

call QDBXM/QZXMTRC PARM(off user_profile)
Parameters:
Table 7. Trace parameters
Parameter Description
user_profile The name of the user profile associated with
the job within which the XML Extender is
running.
Examples:
The following examples demonstrate stopping the trace.

From the Qshell:
dxxtrc off user1
From the iSeries Navigator:
call myschema.QZXMTRC(’off’, ’user1’);
From the OS command line:
call QDBXM/QZXMTRC PARM(off user1)
XML Extender UDF return codes

Embedded SQL statements return codes in the SQLCODE, SQLWARN, and
SQLSTATE fields of the SQLCA structure. This structure is defined in an SQLCA
INCLUDE file. (For more information about the SQLCA structure and SQLCA
INCLUDE file, see the DB2 Application Development Guide.)
DB2 CLI calls return SQLCODE and SQLSTATE values that you can retrieve using
the SQLError function. (For more information about retrieving error information
with the SQLError function, see the CLI Guide and Reference.)
An SQLCODE value of 0 means that the statement ran successfully (with possible
warning conditions). A positive SQLCODE value means that the statement ran
successfully but with a warning. (Embedded SQL statements return information
about the warning that is associated with 0 or positive SQLCODE values in the
SQLWARN field.) A negative SQLCODE value means that an error occurred.
DB2 associates a message with each SQLCODE value. If an XML Extender UDF
encounters a warning or error condition, it passes associated information to DB2
UDB for inclusion in the SQLCODE message.
Embedded SQL statements and DB2 UDB CLI calls that invoke the DB2 XML
Extender UDFs might return SQLCODE messages and SQLSTATE values that are
unique to these UDFs, but DB2 UDB returns these values in the same way that it
does for other embedded SQL statements or other DB2 UDB CLI calls. Thus, the
way that you access these values is the same as for embedded SQL statements or
DB2 UDB CLI calls that do not start the DB2 UDB XML Extender UDFs.
XML Extender stored procedure return codes

The XML Extender provides return codes to help resolve problems with stored
procedures. When you receive a return code from a stored procedure, check the
following file, which matches the return code with an XML Extender error message
number and the symbolic constant.
dxx_install/include/dxxrc.h
| SQLSTATE codes and associated message numbers for XML

| Extender
| Table 8. SQLSTATE codes and associated message numbers
| SQLSTATE Message Number Description
| 00000 DXXnnnnI No error occurred.
| 01HX0 DXXD003W The element or attribute specified in
| the path expression is missing from
| the XML document.
| 38X00 DXXC000E The XML Extender is unable to
| open the specified file.

| Table 8. SQLSTATE codes and associated message numbers (continued)
| 38X01 DXXA072E XML Extender tried to
| automatically bind the database
| before enabling it, but could not
| find the bind files.
| DXXC001E The XML Extender could not find
| the file specified.
| 38X02 DXXC002E The XML Extender is unable to read
| data from the specified file.
| 38X03 DXXC003E The XML Extender is unable to
| write data to the file.
| DXXC011E The XML Extender is unable to
| write data to the trace control file.
| 38X04 DXXC004E The XML Extender was unable to
| operate the specified locator.
| 38X05 DXXC005E The file size is greater than the
| XMLVarchar size, and the XML
| Extender is unable to import all the
| data from the file.
| 38X06 DXXC006E The file size is greater than the size
| of the XMLCLOB, and the XML
| Extender is unable to import all the
| data from the file.
| 38X07 DXXC007E The number of bytes in the LOB
| locator does not equal the file size.
| 38X08 DXXD001E A scalar extraction function used a
| location path that occurs multiple
| times. A scalar function can use
| only a location path that does not
| have multiple occurrence.
| 38X09 DXXD002E The path expression is syntactically
| incorrect.
| 38X10 DXXG002E The XML Extender was unable to
| allocate memory from the operating
| system.
| 38X11 DXXA009E This stored procedure is for an XML
| column only.
| 38X12 DXXA010E While attempting to enable the
| column, the XML Extender could
| not find the DTD ID, which is the
| identifier specified for the DTD in
| the document access definition
| (DAD) file.
| DXXQ060E The XML Extender could not find
| the SCHEMA ID while attempting
| to enable the column. The SCHEMA
| ID corresponds to the value of the
| location attribute of the
| nonamespacelocation tag which is
| inside the schemabindings tag in
| the DAD file.

| 38X14 DXXD000E There was an attempt to store an
| invalid document into a table.
| Validation failed.
| 38X15 DXXA056E The validation element in the
| document access definition (DAD)
| file is wrong or missing.
| DXXA057E The name attribute of a side table in
| (DAD) file is wrong or missing.
| DXXA058E The name attribute of a column in
| DXXA059E The type attribute of a column in
| DXXA060E The path attribute of a column in
| DXXA061E The multi_occurrence attribute of a
| column in the document access
| definition (DAD) file is wrong or
| missing.
| DXXQ000E A mandatory element is missing
| from the document access definition
| (DAD) file.
| DXXQ056E The specified element/attribute
| cannot be mapped to a column that
| is specified as part of a foreign key.
| Data values for foreign keys are
| determined by that of the primary
| keys; a mapping of the specified
| element/attribute in the XML
| document to a table and column is
| not necessary.
| DXXQ057E The schemabindings and DTD ID
| tags cannot exist together in the
| DAD file.
| DXXQ058E The nonamespacelocation tag inside
| the schemabindings tag is missing
| in the DAD file.
| DXXQ059E The doctype tag cannot be located
| inside the XCollection tag in the
| DAD for schema validation.
| DXXQ062E This error condition is usually
| caused by a missing
| multi_occurrence = YES
| specification on the parent
| element_node of the given element
| or attribute.

| DXXQ063E The value of the multi_occurrence
| attribute on the specified
| element_node in the document
| access definition (DAD) file is
| wrong or missing. The value must
| be ’yes’ or ’no’, case insensitive.
| DXXQ064E A key column specified in the join
| condition was not mapped to any
| element or attribute node.
| 38X16 DXXG004E A null value for a required
| parameter was passed to an XML
| stored procedure.
| 38X17 DXXQ001E The SQL statement in the document
| access definition (DAD) file or the
| statement that overrides it is not
| valid. A SELECT statement is
| required for generating XML
| documents.
| 38X18 DXXG001E XML Extender encountered an
| internal error.
| DXXG006E XML Extender encountered an
| internal error while using CLI.
| 38X19 DXXQ002E The system is running out of space
| in memory or disk. There is no
| space to contain the resulting XML
| documents.
| 38X20 DXXQ003W The user-defined SQL query
| generates more XML documents
| than the specified maximum. Only
| the specified number of documents
| are returned.
| 38X21 DXXQ004E The specified column is not one of
| the columns in the result of the SQL
| query.
| 38X22 DXXQ005E The mapping of the SQL query to
| XML is incorrect.
| 38X23 DXXQ006E An attribute_node element in the
| file does not have a name attribute.
| 38X24 DXXQ007E The attribute_node element in the
| does not have a column element or
| RDB_node.
| 38X25 DXXQ008E A text_node element in the
| file does not have a column
| element.
| 38X26 DXXQ009E The specified result table could not
| be found in the system catalog.

| 38X27 DXXQ010E The RDB_node of the
| DXXQ040E attribute_node or text_node must
| have a table.
| DXXQ011E The RDB_node of the
| attribute_node or text_node must
| have a column.
| DXXQ017E An XML document generated by
| the XML Extender is too large to fit
| into the column of the result table.
| DXXQ040E The specified element name in
| file is wrong.
| 38X28 DXXQ012E XML Extender could not find the
| expected element while processing
| the DAD.
| DXXQ016E All tables must be defined in the
| RDB_node of the top element in the
| file. Sub-element tables must match
| the tables defined in the top
| element. The table name in this
| RDB_node is not in the top element.
| 38X29 DXXQ013E The element table or column must
| have a name in the document access
| definition (DAD) file.
| DXXQ015E The condition in the condition
| element in the document access
| definition (DAD) file has an invalid
| format.
| DXXQ061E The format of the string
| representation is invalid. If the
| string is a date, time, or timestamp
| value, the syntax does not conform
| to its data type.
| 38X30 DXXQ014E An element_node element in the
| file does not have a name attribute.
| DXXQ018E The ORDER BY clause is missing
| from the SQL statement in a
| file that maps SQL to XML.
| 38X31 DXXQ019E The objids element does not have a
| column element in the document
| access definition (DAD) file that
| maps SQL to XML.
| 38x33 DXXG005E This parameter is not supported in
| this release. It will be supported in
| the future release.
| 38x34 DXXG000E An invalid file name was specified.
| 38X36 DXXA073E The database was not bound when
| you tried to enable it.

| 38X37 DXXG007E The server operating system locale
| is inconsistent with the DB2 UDB
| code page.
| 38X38 DXXG008E The server operating system locale
| can not be found in the code page
| table.
| 38X41 DXXQ048E The stylesheet processor returned an
| internal error. The XML document
| or the stylesheet might not be vaild.
| 38X42 DXXQ049E The specified output file already
| exists in this directory.
| 38X43 DXXQ050E The UDF was unable to create a
| unique file name for the output
| document in the specified directory
| because it does not have access. All
| file names that can be generated are
| in use or directory might not exist.
| 38X44 DXXQ051E One or more input or output
| parameters have no valid value.
| 38X45 DXXQ055E ICU error encountered during
| conversion operation.
|
| XML Extender messages
2. Ensure the database exists and is accessible.

DXXA000I Enabling column <column_name>. Please
Wait. 3. Determine if the database is corrupted. If it is, ask
your database administrator to recover it from a
Explanation: This is an informational message. backup.
User Response: No action required.
DXXA004E Cannot enable database <database>.
DXXA001S An unexpected error occurred in build Explanation: The database might already be enabled
<build_ID>, file <file_name>, and line or might be corrupted.
<line_number>.
User Response:
Explanation: An unexpected error occurred.
1. Determine if the database is enabled.
User Response: If this error persists, contact your 2. Determine if the database is corrupted. If it is, ask
Software Service Provider. When reporting the error, be your database administrator to recover it from a
sure to include all the message text, the trace file, and backup.
an explanation of how to reproduce the problem.
DXXA005I Enabling database <database>. Please

DXXA002I Connecting to database <database>. wait.
Explanation: This is an informational message. Explanation: This is an informational message.
User Response: No action required. User Response: No action required.
DXXA003E Cannot connect to database <database>. DXXA006I The database <database> was enabled
Explanation: The database specified might not exist or successfully.
could be corrupted. Explanation: This is an informational message.
User Response: User Response: No action required.
1. Ensure the database is specified correctly.

DXXA007E Cannot disable database <database>. DXXA013E Attempt to alter table <table_name>
failed.
Explanation: The database cannot be disabled by XML
Extender if it contains any XML columns or collections. Explanation: While attempting to enable the column,
the XML Extender could not alter the specified table.
User Response: Backup any important data, disable
any XML columns or collections, and update or drop User Response: Check the privileges required to alter
any tables until there are no XML data types left in the the table.
database.
DXXA014E The specified root ID column: <root_id>
DXXA008I Disabling column <column_name>. is not a single primary key of table
Please Wait. <table_name>.
Explanation: This is an information message. Explanation: The root ID specified is either not a key,
or it is not a single key of table table_name.
User Response: Ensure the specified root ID is the
single primary key of the table.
DXXA009E Xcolumn tag is not specified in the
DAD file.
DXXA015E The column DXXROOT_ID already
Explanation: This stored procedure is for XML
exists in table <table_name>.
Column only.
Explanation: The column DXXROOT_ID exists, but
User Response: Ensure the Xcolumn tag is specified
was not created by XML Extender.
correctly in the DAD file.
User Response: Specify a primary column for the root
ID option when enabling a column, using a different
DXXA010E Attempt to find DTD ID <dtdid> failed.
different column name.
Explanation: While attempting to enable the column,
the XML Extender could not find the DTD ID, which is
DXXA016E The input table <table_name> does not
the identifier specified for the DTD in the document
exist.
access definition (DAD) file.
Explanation: The XML Extender was unable to find
User Response: Ensure the correct value for the DTD
the specified table in the system catalog.
ID is specified in the DAD file.
User Response: Ensure that the table exists in the
database, and is specified correctly.
DXXA011E Inserting a record into
DB2XML.XML_USAGE table failed.
DXXA017E The input column <column_name> does
not exist in the specified table
the XML Extender could not insert a record into the
<table_name>.
DB2XML.XML_USAGE table.
Explanation: The XML Extender was unable to find
User Response: Ensure the DB2XML.XML_USAGE
the column in the system catalog.
table exists and that a record by the same name does
not already exist in the table. User Response: Ensure the column exists in a user
table.
DXXA012E Attempt to update DB2XML.DTD_REF
table failed. DXXA018E The specified column is not enabled for
XML data.
the XML Extender could not update the Explanation: While attempting to disable the column,
DB2XML.DTD_REF table. XML Extender could not find the column in the
DB2XML.XML_USAGE table, indicating that the
User Response: Ensure the DB2XML.DTD_REF table
column is not enabled. If the column is not
exists. Determine whether the table is corrupted or if
XML-enabled, you do not need to disable it.
the administration user ID has the correct authority to
update the table. User Response: No action required.

DXXA019E A input parameter required to enable DXXA026E Unable to drop the side table
the column is null. <side_table>.
Explanation: A required input parameter for the Explanation: While attempting to disable a column,
enable_column() stored procedure is null. the XML Extender was unable to drop the specified
table.
User Response: Check all the input parameters for the
enable_column() stored procedure. User Response: Ensure that the administrator user ID
for XML Extender has the privileges necessary to drop
the table.
DXXA020E Columns cannot be found in the table
<table_name>.
DXXA027E Could not disable the column.
Explanation: While attempting to create the default
view, the XML Extender could not find columns in the Explanation: XML Extender could not disable a
specified table. column because an internal trigger failed. Possible
causes:
User Response: Ensure the column and table name
are specified correctly. v The system is out of memory.
v A trigger with this name does not exist.
DXXA021E Cannot create the default view User Response: Use the trace facility to create a trace
<default_view>. file and try to correct the problem. If the problem
persists, contact your Software Service Provider and
Explanation: While attempting to enable a column, the
provide the trace file.
XML Extender could not create the specified view.
User Response: Ensure that the default view name is
unique. If a view with the name already exists, specify
a unique name for the default view. Explanation: XML Extender could not disable a
column because an internal trigger failed. Possible
causes:
DXXA022I Column <column_name> enabled.
v The system is out of memory.
Explanation: This is an informational message.
User Response: No response required.
User Response: Use the trace facility to create a trace
file and try to correct the problem. If the problem
DXXA023E Cannot find the DAD file. persists, contact your Software Service Provider and
Explanation: While attempting to disable a column,
the XML Extender was unable to find the document
access definition (DAD) file. DXXA029E Could not disable the column.
User Response: Ensure you specified the correct Explanation: XML Extender could not disable a
database name, table name, or column name. column because an internal trigger failed. Possible
causes:
DXXA024E The XML Extender encountered an v The system is out of memory.
internal error while accessing the system v A trigger with this name does not exist.
catalog tables.
Explanation: The XML Extender was unable to access file and try to correct the problem. If the problem
system catalog table. persists, contact your Software Service Provider and
User Response: Ensure the database is in a stable
state.
DXXA025E Cannot drop the default view Explanation: XML Extender could not disable a
<default_view>. column because an internal trigger failed. Possible
causes:
the XML Extender could not drop the default view. v The system is out of memory.
User Response: Ensure the administration user ID for
XML Extender has the privileges necessary to drop the User Response: Use the trace facility to create a trace
default view. file and try to correct the problem. If the problem

DXXA037E The specified table space name is longer
than 18 characters.
Explanation: The table space name cannot be longer
DXXA031E Unable to reset the DXXROOT_ID
than 18 alphanumeric characters.
column value in the application table to
NULL. User Response: Specify a name less than 18
characters.
the XML Extender was unable to set the value of
DXXROOT_ID in the application table to NULL. DXXA038E The specified default view name is
longer than 18 characters.
User Response: Ensure that the administrator user ID
for XML Extender has the privileges necessary to alter Explanation: The default view name cannot be longer
the application table. than 18 alphanumeric characters.
User Response: Specify a name less than 18
DXXA032E Decrement of USAGE_COUNT in characters.
Explanation: While attempting to disable the column, DXXA039E The specified ROOT_ID name is longer
the XML Extender was unable to reduce the value of than 18 characters.
the USAGE_COUNT column by one.
Explanation: The ROOT_ID name cannot be longer
User Response: Ensure that the than 18 alphanumeric characters.
DB2XML.XML_USAGE table exists and that the
User Response: Specify a name less than 18
administrator user ID for XML Extender has the
characters.
necessary privileges to update the table.
DXXA046E Unable to create the side table

DXXA033E Attempt to delete a row from the
<side_table>.
Explanation: While attempting to enable a column, the
XML Extender was unable to create the specified side
the XML Extender was unable to delete its associate
table.
row in the DB2XML.XML_USAGE table.
User Response: Ensure that the administrator user ID
User Response: Ensure that the
for XML Extender has the privileges necessary to create
DB2XML.XML_USAGE table exists and that the
the side table.
administration user ID for XML Extender has the
privileges necessary to update this table.
DXXA047E Could not enable the column.
DXXA034I XML Extender has successfully disabled Explanation: XML Extender could not enable a
column <column_name>. column because an internal trigger failed. Possible
causes:
Explanation: This is an informational message
v The DAD file has incorrect syntax.
v Another trigger exists with the same name.
DXXA035I XML Extender is disabling database
<database>. Please wait. User Response: Use the trace facility to create a trace
Explanation: This is an informational message. persists, contact your Software Service Provider and
User Response: No action is required. provide the trace file.
DXXA036I XML Extender has successfully disabled DXXA048E Could not enable the column.
database <database>. Explanation: XML Extender could not enable a
Explanation: This is an informational message. column because an internal trigger failed. Possible
causes:
User Response: No action is required.

provide the trace file. Explanation: XML Extender could not enable a
causes:
Explanation: XML Extender could not enable a
causes: v Another trigger exists with the same name.
v The DAD file has incorrect syntax. User Response: Use the trace facility to create a trace
v The system is out of memory. file and try to correct the problem. If the problem
provide the trace file. Explanation: XML Extender could not enable a
causes:
Explanation: XML Extender could not enable a
causes: v Another trigger exists with the same name.
v The DAD file has incorrect syntax. User Response: Use the trace facility to create a trace
v The system is out of memory. file and try to correct the problem. If the problem
DXXA056E The validation value <validation_value>
in the DAD file is invalid.
Explanation: The validation element in document
access definition (DAD) file is wrong or missing.
User Response: Ensure that the validation element is
Explanation: XML Extender could not disable a
specified correctly in the DAD file.
causes:
v The system is out of memory. DXXA057E A side table name <side_table_name> in
DAD is invalid.
Explanation: The name attribute of a side table in the
document access definition (DAD) file is wrong or
missing.
provide the trace file. User Response: Ensure that the name attribute of a
side table is specified correctly in the DAD file.
DXXA058E A column name <column_name> in the
Explanation: XML Extender could not disable a
DAD file is invalid.
causes: Explanation: The name attribute of a column in the
v The DAD file has incorrect syntax. document access definition (DAD) file is wrong or
missing.
v Another trigger exists with the same name. User Response: Ensure that the name attribute of a
column is specified correctly in the DAD file.

DXXA059E The type <column_type> of column DXXA065E Calling stored procedure
<column_name> in the DAD file is <procedure_name> failed.
invalid.
Explanation: Check the shared library db2xml and see
Explanation: The type attribute of a column in the if the permission is correct.
User Response: Make sure the client has permission
missing.
to run the stored procedure.
User Response: Ensure that the type attribute of a
DXXA066I XML Extender has successfully disabled
collection <collection_name>.
DXXA060E The path attribute <location_path> of
<column_name> in the DAD file is
invalid. User Response: No response required.
Explanation: The path attribute of a column in the
document access definition (DAD) file is wrong or DXXA067I XML Extender has successfully enabled
missing. collection <collection_name>.
User Response: Ensure that the path attribute of a Explanation: This is an informational message.
User Response: No response required.
DXXA061E The multi_occurrence attribute

<multi_occurrence> of <column_name> in DXXA068I XML Extender has successfully turned
the DAD file is invalid. the trace on.
Explanation: The multi_occurrence attribute of a Explanation: This is an informational message.

column in the document access definition (DAD) file is User Response: No response required.
wrong or missing.
User Response: Ensure that the multi_occurrence DXXA069I XML Extender has successfully turned
attribute of a column is specified correctly in the DAD the trace off.
file.
DXXA062E Unable to retrieve the column number User Response: No response required.
for <column_name> in table <table_name>.
Explanation: XML Extender could not retrieve the DXXA070W The database has already been enabled.
column number for column_name in table table_name Explanation: The enable database command was
from the system catalog. executed on the enabled database
User Response: Make sure the application table is User Response: No action is required.
well defined.
DXXA071W The database has already been disabled.

DXXA063I Enabling collection <collection_name>.
Please Wait. Explanation: The disable database command was
executed on the disabled database
Explanation: This is an information message.
DXXA072E XML Extender couldn’t find the bind

DXXA064I Disabling collection <collection_name>. files. Bind the database before enabling
Please Wait. it.
Explanation: This is an information message. Explanation: XML Extender tried to automatically
User Response: No action required. bind the database before enabling it, but could not find
the bind files
User Response: Bind the database before enabling it.

DXXA073E The database is not bound. Please bind DXXC002E Unable to read file.
the database before enabling it.
Explanation: The XML Extender is unable to read data
Explanation: The database was not bound when user from the specified file.
tried to enable it.
User Response: Ensure that the application user ID
User Response: Bind the database before enabling it. has read permission for the file.
DXXA074E Wrong parameter type. The stored DXXC003E Unable to write to the specified file.
procedure expects a STRING parameter.
Explanation: The XML Extender is unable to write
Explanation: The stored procedure expects a STRING data to the file.
parameter.
User Response: Declare the input parameter to be has write permission for the file or that the file system
STRING type. has sufficient space.
DXXA075E Wrong parameter type. The input DXXC004E Unable to operate the LOB Locator:
parameter should be a LONG type. rc=<locator_rc>.
Explanation: The stored procedure expects the input Explanation: The XML Extender was unable to
parameter to be a LONG type. operate the specified locator.
User Response: Declare the input parameter to be a User Response: Ensure the LOB Locator is set
LONG type. correctly.
DXXA076E XML Extender trace instance ID invalid. DXXC005E Input file size is greater than
XMLVarchar size.
Explanation: Cannot start trace with the instance ID
provided. Explanation: The file size is greater than the
XMLVarchar size and the XML Extender is unable to
User Response: Ensure that the instance ID is a valid
import all the data from the file.
iSeries user ID.
User Response: Use the XMLCLOB column type.
DXXA077E The license key is not valid. See the
server error log for more detail. DXXC006E The input file exceeds the DB2 UDB
LOB limit.
Explanation: The software license has expired or does
not exist. Explanation: The file size is greater than the size of
the XMLCLOB and the XML Extender is unable to
User Response: Contact your service provider to
import all the data from the file.
obtain a new software license.
User Response: Decompose the file into smaller
objects or use an XML collection.
DXXC000E Unable to open the specified file.
Explanation: The XML Extender is unable to open the
DXXC007E Unable to retrieve data from the file to
specified file.
the LOB Locator.
Explanation: The number of bytes in the LOB Locator
has read and write permission for the file.
does not equal the file size.
User Response: Ensure the LOB Locator is set
DXXC001E The specified file is not found.
correctly.
Explanation: The XML Extender could not find the file
specified.
DXXC008E Can not remove the file <file_name>.
User Response: Ensure that the file exists and the
Explanation: The file has a sharing access violation or
path is specified correctly.
is still open.
User Response: Close the file or stop any processes
that are holding the file. You might have to stop and
restart DB2.

DXXC009E Unable to create file to <directory> DXXD001E <location_path> occurs multiple times.
directory.
Explanation: A scalar extraction function used a
Explanation: The XML Extender is unable to create a location path that occurs multiple times. A scalar
file in directory directory. function can only use a location path that does not
have multiple occurrences.
User Response: Ensure that the directory exists, that
the application user ID has write permission for the User Response: Use a table function (add an ’s’ to the
directory, and that the file system has sufficient space end of the scalar function name).
for the file.
DXXD002E A syntax error occurred near position
DXXC010E Error while writing to file <file_name>. <position> in the search path.
Explanation: There was an error while writing to the Explanation: The path expression is syntactically
file file_name. incorrect.
User Response: Ensure that the file system has User Response: Correct the search path argument of
sufficient space for the file. the query. Refer to the documentation for the syntax of
path expressions.
DXXC011E Unable to write to the trace control file.
DXXD003W Path not found. Null is returned.
Explanation: The XML Extender is unable to write
data to the trace control file. Explanation: The element or attribute specified in the
path expression is missing from the XML document.
has write permission for the file or that the file system User Response: Verify that the specified path is
has sufficient space. correct.
DXXC012E Cannot create temporary file. DXXG000E The file name <file_name> is invalid.
Explanation: Cannot create file in system temp Explanation: An invalid file name was specified.
directory.
User Response: Specify a correct file name and try
User Response: Ensure that the application user ID again.
has write permission for the file system temp directory
or that the file system has sufficient space for the file.
DXXG001E An internal error occurred in build
<build_ID>, file <file_name>, and line
DXXC013E The results of the extract UDF exceed <line_number>.
the size limit for the UDF return type.
Explanation: XML Extender encountered an internal
Explanation: The data returned by an extract UDF error.
must fit into the size limit of the return type of the
User Response: Contact your Software Service
UDF, as defined in the DB2 UDB XML Extender
Provider. When reporting the error, be sure to include
Administration and Programming guide. For example,
all the messages, the trace file and how to reproduce
the results of extractVarchar must be no more than 4000
the error.
bytes (including the terminating NULL).
User Response: Use an extract UDF that has a larger
DXXG002E The system is out of memory.
size limit for the return type: 254 bytes for
extractChar(), 4 KB for extractVarchar(), and 2 GB for Explanation: The XML Extender was unable to
extractClob(). allocate memory from the operating system.
User Response: Close some applications and try
DXXD000E An invalid XML document is rejected. again. If the problem persists, refer to your operating
system documentation for assistance. Some operating
Explanation: There was an attempt to store an invalid
systems might require that you reboot the system to
document into a table. Validation has failed.
correct the problem.
User Response: Check the document with its DTD
using an editor that can view invisible invalid
characters. To suppress this error, turn off validation in
the DAD file.

DXXG004E Invalid null parameter. DXXM001W A DB2 UDB error occurred.
Explanation: A null value for a required parameter Explanation: DB2 encountered the specified error.
was passed to an XML stored procedure.
User Response: See any accompanying messages for
User Response: Check all required parameters in the futher explanation and refer to DB2 UDB messages and
argument list for the stored procedure call. codes documentation for your operating system.
DXXG005E Parameter not supported. DXXQ000E <Element> is missing from the DAD file.
Explanation: This parameter is not supported in this Explanation: A mandatory element is missing from
release, will be supported in the future release. the document access definition (DAD) file.
User Response: Set this parameter to NULL. User Response: Add the missing element to the DAD
file.
DXXG006E Internal Error CLISTATE=<clistate>,
RC=<cli_rc>, build <build_ID>, file DXXQ001E Invalid SQL statement for XML
<file_name>, line <line_number> generation.
CLIMSG=<CLI_msg>.
Explanation: The SQL statement in the document
Explanation: XML Extender encountered an internal access definition (DAD) or the one that overrides it is
error while using CLI. not valid. A SELECT statement is required for
generating XML documents.
User Response: Contact your Software Service
Provider. Potentially this error can be caused by User Response: Correct the SQL statement.
incorrect user input. When reporting the error, be sure
to include all output messages, trace log, and how to
DXXQ002E Cannot generate storage space to hold
reproduce the problem. Where possible, send any
XML documents.
DADs, XML documents, and table definitions which
apply. Explanation: The system is running out of space in
memory or disk. There is no space to contain the
resulting XML documents.
DXXG007E Locale <locale> is inconsistent with DB2
UDB code page <code_page>. User Response: Limit the number of documents to be
generated. Reduce the size of each documents by
Explanation: The server operating system locale is
removing some unnecessary element and attribute
inconsistent with DB2 UDB code page.
nodes from the document access definition (DAD) file.
User Response: Correct the server operating system
locale and restart DB2.
DXXQ003W Result exceeds maximum.
Explanation: The user-defined SQL query generates
DXXG008E Locale <locale> is not supported.
more XML documents than the specified maximum.
Explanation: The server operating system locale can Only the specified number of documents are returned.
not be found in the code page table.
User Response: No action is required. If all
User Response: Correct the server operating system documents are needed, specify zero as the maximum
locale and restart DB2. number of documents.
DXXG017E The limit for XML_Extender_constant has DXXQ004E The column <column_name> is not in the
been exceeded in build build_ID, file result of the query.
file_name, and line line_number.
Explanation: The specified column is not one of the
Explanation: Check the XML Extender Administration columns in the result of the SQL query.
and Programming Guide to see whether your
User Response: Change the specified column name in
application has exceeded a value in the limits table. If
the document access definition (DAD) file to make it
no limit has been exceeded, contact your Software
one of the columns in the result of the SQL query.
Service Provider. When reporting the error, include all
Alternatively, change the SQL query so that it has the
output messages, trace files, and information on how to
specified column in its result.
reproduce the problem such as input DADs, XML
documents, and table definitions.
User Response: Correct the server operating system
locale and restart DB2.

DXXQ005E Wrong relational mapping. The element DXXQ011E RDB_node element of <node_name> does
<element_name> is at a lower level than not have a column in the DAD file.
its child column <column_name>.
Explanation: The RDB_node of the attribute_node or
Explanation: The mapping of the SQL query to XML text_node must have a column.
is incorrect.
User Response: Specify the column of RDB_node for
User Response: Make sure that the columns in the attribute_node or text_node in the document access
result of the SQL query are in a top-down order of the definition (DAD) file.
relational hierarchy. Also make sure that there is a
single-column candidate key to begin each level. If such
DXXQ012E Errors occurred in DAD.
a key is not available in a table, the query should
generate one for that table using a table expression and Explanation: XML Extender could not find the
the DB2 UDB built-in function generate_unique(). expected element while processing the DAD.
User Response: Check that the DAD is a valid XML
DXXQ006E An attribute_node element has no name. document and contains all the elements required by the
DAD DTD. Consult the XML Extender publication for
Explanation: An attribute_node element in the
the DAD DTD.
document access definition (DAD) file does not have a
name attribute.
DXXQ013E The table or column element does not
User Response: Ensure that every attribute_node has
have a name in the DAD file.
a name in the DAD file.
Explanation: The element table or column must have
a name in the document access definition (DAD) file.
DXXQ007E The attribute_node <attribute_name> has
no column element or RDB_node. User Response: Specify the name of table or column
element in the DAD.
Explanation: The attribute_node element in the
document access definition (DAD) does not have a
column element or RDB_node. DXXQ014E An element_node element has no name.
User Response: Ensure that every attribute_node has Explanation: An element_node element in the
a column element or RDB_node in the DAD. document access definition (DAD) file does not have a
name attribute.
DXXQ008E A text_node element has no column User Response: Ensure that every element_node
element. element has a name in the DAD file.
Explanation: A text_node element in the document
access definition (DAD) file does not have a column DXXQ015E The condition format is invalid.
element.
Explanation: The condition in the condition element in
User Response: Ensure that every text_node has a the document access definition (DAD) has an invalid
column element in the DAD. format.
User Response: Ensure that the format of the
DXXQ009E Result table <table_name> does not exist. condition is valid.
Explanation: The specified result table could not be
found in the system catalog. DXXQ016E The table name in this RDB_node is not
defined in the top element of the DAD
User Response: Create the result table before calling
file.
the stored procedure.
Explanation: All tables must be defined in the
RDB_node of the top element in the document access
DXXQ010E RDB_node of <node_name> does not
definition (DAD) file. Sub-element tables must match
have a table in the DAD file.
the tables defined in the top element. The table name in
Explanation: The RDB_node of the attribute_node or this RDB_node is not in the top element.
text_node must have a table.
User Response: Ensure that the table of the RDB node
User Response: Specify the table of RDB_node for is defined in the top element of the DAD file.
attribute_node or text_node in the document access
definition (DAD) file.

DXXQ017E The column in the result table DXXQ024E Can not create table <table_name>.
<table_name> is too small.
Explanation: The specified table can not be created.
Explanation: An XML document generated by the
User Response: Ensure that the user ID creating the
XML Extender is too large to fit into the column of the
table has the necessary authority to create a table in the
result table.
database.
User Response: Drop the result table. Create another
result table with a bigger column. Rerun the stored
DXXQ025I XML decomposed successfully.
procedure.
Explanation: An XML document has been
decomposed and stored in a collection successfully.
DXXQ018E The ORDER BY clause is missing from
the SQL statement. User Response: No action is required.
Explanation: The ORDER BY clause is missing from
the SQL statement in a document access definition DXXQ026E XML data <xml_name> is too large to fit
(DAD) file that maps SQL to XML. in column <column_name>.
User Response: Edit the DAD file. Add an ORDER BY Explanation: The specified piece of data from an XML
clause that contains the entity-identifying columns. document is too large to fit into the specified column.
User Response: Increase the length of the column
DXXQ019E The element objids has no column using the ALTER TABLE statement or reduce the size
element in the DAD file. of the data by editing the XML document.
Explanation: The objids element does not have a
column element in the document access definition DXXQ028E Cannot find the collection
(DAD) file that maps SQL to XML. <collection_name> in the XML_USAGE
table.
User Response: Edit the DAD file. Add the key
columns as sub-elements of the element objids. Explanation: A record for the collection cannot be
found in the XML_USAGE table.
DXXQ020I XML successfully generated. User Response: Verify that you have enabled the
collection.
Explanation: The requested XML documents have
been successfully generated from the database.
DXXQ029E Cannot find the DAD in XML_USAGE
table for the collection <collection_name>.
Explanation: A DAD record for the collection cannot
DXXQ021E Table <table_name> does not have
be found in the XML_USAGE table.
column <column_name>.
User Response: Ensure that you have enabled the
Explanation: The table does not have the specified
collection correctly.
column in the database.
User Response: Specify another column name in DAD
DXXQ030E Wrong XML override syntax.
or add the specified column into the table database.
Explanation: The XML_override value is specified
incorrectly in the stored procedure.
DXXQ022E Column <column_name> of <table_name>
should have type <type_name>. User Response: Ensure that the syntax of
XML_override is correct.
Explanation: The type of the column is wrong.
User Response: Correct the type of the column in the
DXXQ031E Table name cannot be longer than
document access definition (DAD).
maximum length allowed by DB2.
Explanation: The table name specified by the
DXXQ023E Column <column_name> of <table_name>
condition element in the DAD is too long.
cannot be longer than <length>.
User Response: Correct the length of the table name
Explanation: The length defined for the column in the
in document access definition (DAD).
DAD is too long.
User Response: Correct the column length in the
document access definition (DAD).

DXXQ032E Column name cannot be longer than DXXQ038E The SQL statement is too long:
maximum length allowed by DB2. SQL_statement
Explanation: The column name specified by the Explanation: The SQL statement specified in the
condition element in the DAD is too long. <SQL_stmt> element of DAD exceeds the allowed
number of bytes.
User Response: Correct the length of the column
name in the document access definition (DAD). User Response: Reduce the length of the SQL
statement to less than or equal to 32765 bytes for
Windows and UNIX, or 16380 bytes for OS/390 and
DXXQ033E Invalid identifier starting at <identifier>
iSeries.
Explanation: The string is not a valid DB2 UDB SQL
identifier.
DXXQ039E Too many columns specified for a table
User Response: Correct the string in the DAD to in the DAD file.
conform to the rules for DB2 UDB SQL identifiers.
Explanation: A DAD file used for decomposition or
RDB composition can have a maximum of 100
DXXQ034E Invalid condition element in top text_node and attribute_node elements that specify
RDB_node of DAD: <condition> unique columns within the same table.
Explanation: The condition element must be a valid User Response: Reduce the total number of text_node
WHERE clause consisting of join conditions connected and attribute_node elements that refer to unique
by the conjunction AND. columns within the same table 100 or less.
User Response: See the XML Extender documentation

for the correct syntax of the join condition in a DAD. DXXQ040E The element name <element_name> in the
DAD file is invalid.
DXXQ035E Invalid join condition in top RDB_node Explanation: The specified element name in the
of DAD: <condition> document access definition (DAD) file is wrong.
Explanation: Column names in the condition element User Response: Ensure that the element name is
of the top RDB_node must be qualified with the table typed correctly in the DAD file. See the DTD for the
name if the DAD specifies multiple tables. DAD file.
User Response: See the XML Extender documentation

for the correct syntax of the join condition in a DAD. DXXQ041W XML document successfully generated.
One or more override paths specified is
invalid and ignored.
DXXQ036E A Schema name specified under a DAD
condition tag is longer than allowed. Explanation: Specify only one override path.
Explanation: An error was detected while parsing text User Response: Ensure that the element name is
under a condition tag within the DAD. The condition typed correctly in the DAD file. See the DTD for the
text contains an id qualified by a schema name that is DAD file.
too long.
User Response: Correct the text of the condition tags DXXQ043E Attribute <attr_name> not found under
in document access definition (DAD). element <elem_name>.
Explanation: The attribute <attr_name> was not
DXXQ037E Cannot generate <element> with multiple present in element <elem_name> or one of its child
occurrences. elements.
Explanation: The element node and its descendents User Response: Ensure the attribute appears in the
have no mapping to database, but its multi_occurrence XML document everywhere that the DAD requires it.
equals YES.
User Response: Correct the DAD by either setting the DXXQ044E Element <elem_name> does not have an
multi_occurrence to NO or create a RDB_node in one ancestor element<ancestor>.
of its descendents.
Explanation: According to the DAD, <ancestor> is an
ancestor element of <elem_name> . In the XML
document, one or more element <elem_name> does not
have such an ancestor.
User Response: Ensure that the nesting of elements in

the XML document conforms to what is specified in the specified directory, change to a directory with available
corresponding DAD. file names.
DXXQ045E Subtree under element <elem_name> DXXQ051E No input or output data.

contains multiple attributes
Explanation: One or more input or output parameters
named<attrib_name>.
have no valid value.
Explanation: A subtree under <elem_name> in the
User Response: Check the statement to see if required
XML document contains multiple instances of
parameters are missing.
attribute<attrib_name> , which according to the DAD, is
to be decomposed into the same row. Elements or
attributes that are to be decomposed must have unique DXXQ052E An error occurred while accessing the
names. DB2XML.XML_USAGE table.
User Response: Ensure that the element or attribute in Explanation: Either the database has not been enabled
the subtree has a unique name. or the table DB2XML.XML_USAGE has been dropped.
User Response: Ensure that the database has been
DXXQ046W The DTD ID was not found in the enabled and the table DB2XML.XML_USAGE is
DAD. accessible.
Explanation: In the DAD, VALIDATION is set to YES,
but the DTDID element is not specified. No validation DXXQ053E An SQL statement failed : msg
check is performed.
Explanation: An SQL statement generated during
User Response: No action is required. If validation is XML Extender processing failed to execute.
needed, specify the DTDID element in the DAD file.
User Response: Examine the trace for more details. If
you cannot correct the error condition, contact your
DXXQ047E Parser error on line linenumber column softwaresService provider. When reporting the error, be
colnumber: msg sure to include all the messages, the trace file and how
to reproduce the error.
Explanation: The parser could not parse the document
because of the reported error.
DXXQ054E Invalid input parameter: param
User Response: Correct the error in the document,
consulting the XML specifications if necessary. Explanation: The specified input parameter to a stored
procedure or UDF is invalid.
DXXQ048E Internal error - see trace file. User Response: Check the signature of the relevant
stored procedure or UDF, and ensure the actual input
Explanation: The stylesheet processor returned an
parameter is correct.
internal error. The XML document or the stylesheet
might not vaild.
DXXQ055E ICU error: uerror
User Response: Ensure the XML document and the
stylesheet are valid. Explanation: ICU error encountered during conversion
operation.
DXXQ049E The output file already exists. User Response: Report the error to your software
service provider. Include trace file, error message, and
Explanation: The specified output file already exists in
instructions to reproduce the error.
this directory.
User Response: Change the output path or file name
DXXQ056E Element/attribute xmlname cannot be
for the output document to a unique name or delete
mapped to the column designated as
the existing file.
part of the foreign key (column column
in table table).
DXXQ050E Unable to create a unique file name.
Explanation: The specified element/attribute cannot
Explanation: The UDF was unable to create a unique be mapped to a column that is specified as part of a
file name for the output document in the specified foreign key. Data values for foreign keys are
directory because it does not have access, all file names determined by that of the primary keys; a mapping of
that can be generated are in use or directory might not the specified element/attribute in the xml document to
exist. a table and column is not necessary.
User Response: Ensure that the UDF has access to the User Response: Remove the RDB_node mapping to

the specified column and table in the DAD. reflects the multiplicity of child element_nodes.
DXXQ057E The schemabindings and dtdid tags DXXQ063E The multi_occurrence attribute value on
cannot exist together in the DAD file. elementname in the DAD file is invalid.
Explanation: The schemabindings and dtdid tags Explanation: The value of the multi_occurrence
cannot exist together in the DAD file. attribute on the specified element_node in the
User Response: Check that either the schemabindings
missing. The value must be ’yes’ or ’no’, case
tag or the dtdid tag exists in the DAD file, but not
insensitive.
both.
User Response: Ensure that the multi_occurrence
attribute is specified correctly in the DAD file.
DXXQ058E The nonamespacelocation tag inside the
schemabindings tag is missing in the
DAD file. DXXQ064E Column column not found in foreign
table table.
Explanation: The nonamespacelocation tag inside the
schemabindings tag is missing in the DAD file. Explanation: A key column specified in the join
condition was not mapped to any element or attribute
User Response: Add the nonamespacelocation tag to
node.
the schemabindings tag.
User Response: Check to make sure the join condition
specified in the DAD file is correct, and all key
DXXQ059E The doctype tag cannot be located
columns are mapped to element or attribute nodes.
inside the XCollection tag in the DAD
for schema validation.
DXXQ065I All triggers relating to XML enabled
Explanation: The doctype tag cannot be located inside
columns have been successfully
the XCollection tag in the DAD for schema validation.
regenerated.
User Response: Remove the doctype tag inside the
Explanation: This is an informational message only.
Xcollection tag for schema validation.
DXXQ060E Attempt to find SCHEMA ID schemaid
failed. DXXQ066E The primary key for table tablename does
not exist.
Explanation: The XML Extender could not find the
SCHEMA ID while attempting to enable the column. Explanation: XML Extender could not determine the
The SCHEMA ID corresponds to the value of the primary key for table tablename. Check that the primary
location attribute of the nonamespacelocation tag which key for the table was not dropped after the column was
is inside the schemabindings tag in the DAD file. enabled for XML.
User Response: Check that the correct value for the User Response: Alter the table to add the primary key
SCHEMA ID is specified in the DAD file. specified as the ROOT ID when the column was
enabled for XML.
DXXQ061E The format of the string is invalid.
DXXQ067E Attempt to action failed.
Explanation: The format of the string representation is
invalid. If the string is a date, time, or timestamp value, Explanation: While attempting to action, a SQL error
the syntax does not conform to its data type. occurred.
User Response: Check that the format of the date, User Response: Contact your Software Service
time, or timestamp value conforms to the format for its Provider. When reporting the error, be sure to include
data type. the XML Extender trace file.
DXXQ062E No rows of result set for table are left to DXXQ068E Cannot set current SQLID to [userid].
produce a XML value for element. SQLCODE = [sqlcode].
Explanation: This error condition is usually caused by Explanation: While attempting to set current sqlid to a
a missing multi_occurrence = YES specification on the secondary authorization id, a SQL error occurred.
parent element_node of the given element or attribute.
User Response: Check that you are specifying a valid
User Response: Check the DAD that the value of secondary authorization id and that you have
multi_occurrence on the parent element_node correctly authorization for the id.

Locking issues
This section will provide an overview of when locking issues may arise and how
to investigate them. The main locking issues that you will encounter will be: lock
waits, timeouts, escalations and deadlocks.
Lock wait and lock timeout:
An application that makes a request for a lock that is not compatible with the
existing locks on the object, or a lock request not already satisfied will be placed
into a lock request pending queue. The lock request will continue to be held for
the waiting application until either timeout period is exceeded or a deadlock is the
cause of the result.
The LOCKTIMEOUT database parameter, measured in seconds, is used to configure

how long a period an application will wait for a lock to be made available. If the
timeout period is exceeded, the waiting application receives an SQL0911 error
message with a return code 68 and the application’s unit of work is automatically
rolled back by the database manager. The default value for LOCKTIMEOUT is -1,
which causes lock timeouts to be disabled. That is, an application that is waiting
for a lock will continue to wait indefinitely or until the lock is released. For
transaction environments, the recommended starting value is 30 seconds, however,
tuning may be necessary in order to find a more appropriate value for this field.
Lock escalation:
Lock escalation can occur in two different scenarios:

1. A single application requests a lock that will exceed its allowable number of
locks.
2. An application triggers lock escalation because the maximum number of
database locks on the system has been exceeded.
In both cases, the database manager will attempt to free up memory allocated to
locking by obtaining table locks and releasing existing row locks. The desired effect
is to make more lock memory available for additional applications.
The following two database parameters have a direct affect on lock escalation:
v locklist- the total number of 4k pages allocated for lock storage.
v maxlocks- the allowable percentage of locklist that can be used by a single
application
Tuning and monitoring may be necessary to find a balance for these values.
Workload, and query behavior dictate locking patterns and how applications will
use lock memory.
Deadlocks:
Deadlocks occur when applications cannot complete a unit of work due to

conflicting lock requests that cannot be resolved until the unit of work is
completed. The DB2 deadlock detector is responsible for handling deadlocks.
When a deadlock is detected, the deadlock detector will choose a victim that will
be automatically rolled back and issued an SQL0911 R.C. 2. By rolling back the
victim, the lock conflict is removed, and the other application can continue
processing.

The frequency at which the DB2 deadlock detector checks for deadlocks can be
controlled via DLCHKTIME(measured in milliseconds, from 1000 to 600000).
Setting this value high will cause the deadlock check time to be increased but the
overhead of deadlock detection will be removed. This could potentially result in
applications stuck in deadlock for prolonged periods of time. Setting the deadlock
check time to a smaller value allows for deadlocks to be detected sooner, however
it also introduces additional overhead for the checking.
Related reference:
v Factors that affect locking
v Locks and performance
v Guidelines for locking
v Deadlocks between applications
v Specifying a lock wait mode strategy
v Correcting lock escalation problems
v Deadlock Prevention for Multiple Contexts
Tracking deadlocks
Snapshots provide valuable insight into the number of deadlocks that have
occurred. If you find that the number of deadlocks is abnormally large, you should
consider enabling an event monitor to track the deadlock events more closely. By
default, all databases have an event monitor named DB2DETAILDEADLOCK defined,
which keeps track of DEADLOCKS WITH DETAILS. The DB2DETAILDEADLOCK event
monitor starts automatically when the database starts.
The following example demonstrates how a deadlock can be analyzed using an

event monitor: Event monitor sample output.
Alternatively, you can analyze the deadlocks using the db2pd tool. Refer to db2pd
sample output.
These examples shows how to identify both a deadlock event and the applications
involved. You could potentially correct this issue by adjusting the lock mode or
isolation level of the application. The DB2 Administration Guide is an excellent
reference for different isolation levels and their effect on concurrency.
Related reference:
v Locks and concurrency control
v Deadlocks between applications [link to
http://publib.boulder.ibm.com/infocenter/db2help/topic/com.ibm.db2.udb.doc/admin/c00054
Tracking lock waits and timeouts

A common user symptom of a locking problem is an application hang. A hang
usually appears as ″Lock Wait″ within the database engine. To confirm that an
application is in ″Lock Wait″, the database administrator should take a lock
snapshot and an application snapshot.
When there is reason to believe that a locking conditions is being encountered the
following steps are suggested:
Step 1. Take a lock snapshot
"db2 get snapshot for locks on sample"
Step 2. Take an application snapshot

"db2 get snapshot for applications on sample"
Step 3. Get a list of applications connected to the database
"db2 list applications all"
The list of monitor elements that are of interest in such cases can be found here:
Lock wait information monitor elements
Here is a truncated example of snapshot output:

Locks held by application = 46
Lock waits since connect = 12
Time application waited on locks (ms) = 96443
Deadlocks detected = 0
Lock escalations = 0
Exclusive lock escalations = 0
Number of Lock Timeouts since connected = 5
Total time UOW waited on locks (ms) = Not Collected
In the example, the application experienced 5 lock timeouts during its connection
to the database. Also, we see that the application waited 96443 milliseconds for
locks to be released. The average wait time can be calculated as follows:
(Time application waited for Locks / Lock waits since connect)
In the above case we see that each lock wait had an average time of 8036
milliseconds. In some environments (for example decision support systems) this
might be considered normal, in others (for example online transaction processing)
it would be considered excessive. Of course, user needs and business requirements
should be used as measures.
In cases where you find lock timeouts becoming excessive when compared to
normal operating levels, you may have an application that is holding locks for an
extremely long period of time. It is also possible that if the number of lock
timeouts is extremely low but the response time for queries is sluggish, the lock
timeout parameter may be set too high.
Example 1: Identifying the cause of a lock timeout and lock wait
Please note, to simplify and not introduce unnecessary workload onto your
environment, we’re going to model a long running query by disabling autocommit.
1. Open two DB2 command windows, and connect to sample from both windows
2. In the first window, issue the following query:
a. db2 +c ″insert into staff values (27,’Jones’,2,’Mgr’,13,35000,0)″
3. From the second window issue the following query:
a. db2 ″select * from staff″
4. Obtain a snapshot
a. db2 get snapshot for applications on sample
5. View the snapshot
6. Issue a commit or a rollaback in both windows:
a. db2 commit
Look for the ″Application Status″ field. You should see one of the applications in
″UOW waiting″ state, and the other in ″Lock-wait″ state. In the above case, you can
see that application issuing the SELECT had made a lock request, but the
application performing the insert was holding the requested lock. In the example,

you had forced the application to never release locks until a commit, but the
principle can be modeled to be the same in cases where locks are released after
long periods of time.
This simple example demonstrates how to use a snapshot to find an application in

Lock Wait. The next course of action may be to focus on what locks the application
was waiting for and why those locks were not being released. If you re-do the
example, but this time enable the monitor switches for locks (i.e. using ″db2 update
monitor switches using lock on″) prior to performing the test, you will retrieve
additional monitor elements such as ″ID of agent holding lock″, ″Application ID
holding lock″, ″Lock name″, and ″Lock mode requested″.
Again, in this simplified case we actually forced the lock wait to happen, however
when pursuing such issues, you may want to investigate the frequency of commits
for the application, the locking semantics (isolation level) used or the size of an
application’s unit of work.
Related reference:
v Snapshot monitor sample output
Analyzing a lock snapshot

What follows is an example of how you might analyze snapshot output if you
were investigating a lock wait scenario. To begin, open a lock snapshot that you
have captured and search for an application status of ″Lockwait″. If you do not
have one, you can create sample snapshots using the scenario described in
Tracking lock waits and timeouts.
Extract Of A Snapshot For Application Handle 1 (with lock monitor switch

enabled):
Application handle = 1
Application status = Lock-wait
...
Application country/region code = 1
DUOW correlation token = *LOCAL.DB2.050312015701
Application name = db2bp.exe
Application ID = *LOCAL.DB2.050312015701
...
Take note of the application handle which is 1 in this example.
Determine the application that is holding the lock(s) that application handle 1 is
waiting on:
.........................
ID of agent holding lock = 0
Application ID holding lock = *LOCAL.DB2.050312015656
........................
It can now be determined that the application with handle 1 is waiting on an

application with handle 0. Determine the state of application handle 0.
...................
Application handle = 0
Application status = UOW Waiting

...
Application name = db2bp.exe
Application ID = *LOCAL.DB2.050312015656
...
Note that the application is in ″UOW Waiting″ state. This means that application is
either doing some other processing or has an open UOW (unit of work). If the
application is idle, i.e. there is no more work to do then the status will read ″Idle″.
An idle application does not hold any database locks.
Examine the type of lock which application 1 is requesting. For example:

...............................
Lock object type = Row
Lock mode = Exclusive Lock (X)
Lock mode requested = Next Key Share (NS)
Name of tablespace holding lock = USERSPACE1
Schema of table holding lock = DB2
Name of table holding lock = STAFF
...................................
The lock snapshot shows that application handle 1 is requesting a Next Key Share
(NS) lock but is waiting on application handle 0 that has an Exclusive Lock (X).
Analyze the application snapshot to get more details about application handle 0.
An (X) lock is not compatible with an (NS) lock, hence the Lock-Wait. The DB2
UDB Administration Guide outlines lock compatibilities.
Extract Of An Application Snapshot For Application Handle 0 (with uow and

timestamp monitor switches enabled):
.................
Elapsed time of last completed uow (sec.ms)= 0.000000
UOW start timestamp = 03/14/2005 12:16:07.872350
UOW stop timestamp =
UOW completion status =
.................
If you have the statement monitor switch enabled at the time of the snapshot, you
will also see the following:
.................
Rows read = 2
Rows written = 1
Rows deleted = 0
Rows updated = 0
Rows inserted = 0
Rows fetched = 0
...
Blocking cursor = NO
Dynamic SQL statement text:
insert into staff values (27,’Jones’,2,’Mgr’,13,35000,0)
The application snapshot confirms that the UOW has begun but has not yet been
completed. This indicates that the UOW is still open. The snapshot also tells us
that the application handle is executing a dynamic SQL statement, an insert into
the table STAFF in this case.

.............................
Number of SQL requests since last commit = 2
Commit statements = 0
Rollback statements = 0
Dynamic SQL statements attempted = 2
.............................
...
Application handle 0 has not yet committed and is holding 4 locks. Details about
these locks can be seen in the lock snapshot, for example:
................................
List Of Locks
Lock Name = 0x02000300270000000000000052
Lock Attributes = 0x00000008
Release Flags = 0x40000000
Lock Count = 1
Hold Count = 0
Lock Object Name = 39
Object Type = Row
Tablespace Name = USERSPACE1
Table Schema = DB2
Table Name = STAFF
Mode = X
..................................
One of the 4 locks that application handle 0 is holding is an X lock. This lock name
is the same as the one which the lock snapshot shows application 1 to be waiting
for.
From the above analysis it can be determined that the cause of this locking
condition is that an application is executing an INSERT into a table which requires
exclusive access and has been granted all the required locks but has not
committed. Any application with a weaker lock will have to wait for the UOW to
commit. (This is the default behavior for the Cursor Stability (CS) isolation level.)
Application handle 0 should commit more frequently.
Identifying the owner of a lock via db2pd

Click on this link for more information
Performance issues
Performance problems cover a wide range of scenarios:
v Identifiable query performing slower than expected
v Workload or batch job not completing as soon as expected, reduction in
transaction rate or throughput

v Overall system slowdown
v Suspected bottleneck in some type of system resource such as CPU, I/O,
memory
v Query or workload consuming more resource than expected or available
v Comparison is being made between one system and another
v Query, application, DB2, or system hangs
There are some subtleties in the scenarios depicted above. For problem diagnosis
purposes, it is important to clarify whether something is not meeting expectations
or is exceeding resource capacity. Sometimes it is both.
For problem determination purposes, hangs can be lumped together with

performance problems because many investigative strategies apply to both. In
addition, it may be not be possible at first to define the problem as a hang versus a
performance problem. To a user waiting for a response, a long-running job can
look like a hang even if in fact much activity can be taking place on behalf of the
application on the database server. There can also be a significant buildup of
activity during a severe system slowdown such that all or most commands appear
to hang on a system.
In addition to characterizing the problem correctly in terms of where the symptom

is observed (query/application/system resource) and what is wrong with it
(slowness or too much resource used), you require many other pieces of
information to put the problem in context. The following questions serve to
quickly determine the best place to start looking for a potential cause.
When did the problem first occur?
If the problem has been occurring for some time, and if a database monitor
schedule has been implemented, you can use historical data to find differences.
This will allow you to focus on changes in system behavior and then focus on why
these changes were introduced. Refer to Proactive monitoring tools. It is also
important to consider whether any recent changes occurred, such as hardware or
software upgrades, a new application rollout, additional users, etc.
Is the performance issue constant or intermittent?
If the poor performance is continual, check to see if the system has started to
handle more load or if a shared database resource has become a bottleneck. Other
potential causes of performance degradation include increased user activity,
multiple large applications, or removal of hardware devices. If performance is poor
only for brief periods begin by looking for common applications or utilities
running at these times. If users report that a group of applications are experiencing
performance issues, then you can begin your analysis by focusing on these
applications.
Does the problem appear to be system-wide or isolated to DB2 and its applications?
System-wide performance problems suggest an issue outside of DB2. It’s possible

that something at the operating system level needs to be addressed.
If isolated to one application, is there a particular query that appears to be problematic?

If one application is problematic, then you can further examine if users are
reporting a query or set of queries that are experiencing a slowdown. You might be
able to isolate the issue to one application and a potential group of queries.
Is there any commonality to the poor performance or does it appear to be random?
You should determine if any common database tables, table space, indexes, etc. are
involved. If so, this suggests that these objects are a point of contention. Other
areas to potentially focus on are referential integrity constraints, foreign key
cascades, locking issues etc.
Looking at DB2:
It’s possible that after your initial review, you identify that there’s a strong
possibility that DB2 is the source of the performance issues. In this case, ask the
following questions:
Are users reporting abnormally long response time for all applications?
1. Have you detected a large amount of I/O? Here are some possible causes:
a. Overloaded devices.
b. Heavy SORTS or use of temporary table spaces.
c. Table scan of a large table.
d. Insufficient buffer pool.
2. Have you detected a high CPU usage? Possible causes include:
a. Heavy SORTS.
b. Insufficient buffer pool sizes.
3. Both (1) and (2)
a. Increased number of users?
b. Large spilled SORTS.
c. Application design?
4. Neither (1) and (2)?
a. Are applications waiting on locks? Refer to Locking issues.
b. Are applications waiting on data?
Are users reporting abnormally long response time for one application or query? Possible
causes include:
v SORTS.
v Concurrency issues.
v Stale statistics.
These are general questions to ask yourself when faced with a performance issue.
This list is by no means exhaustive given the size and complexity of modern
systems, performance issues can surface from a multitude of possibilities.
Regardless, these questions will start you on a logical path to determining where
the problem came from.
Related reference:
v Elements of performance [Link to ]
v Developing a performance improvement process [Link to ]
v Performance tuning guidelines [link to ]

Data To Collect For Optimizer Problems
Analyzing access plans

Before analyzing an access plan, there are three basic things to note:
1. Configuration parameters: The optimizer considers many items when
generating the most efficient access plan. Refer to Configuration parameters
that affect query optimization to ensure that you are aware of the configuration
choices that you have made which will affect the optimizer output. Effectively
designing and tuning a system is key to improved access plans - and faster
query run times.
2. Optimization classes: When an SQL query is compiled by DB2, a number of
optimization techniques are used. Using more optimization techniques increases
compile times and potentially system resource usage. Depending on the nature
of the environment, it may be necessary to limit the number of techniques
applied to optimizing the query. The methods by which you can set the
optimization class can be found here: Setting the optimization class. If you
know that the statements being executed within an application are relatively
simple, for example that there are few joins, you may not want to use too many
resources to optimize for higher speed. That is, you may find that the run-time
differences between using a lower optimization level versus a higher one are
minimal. If you have a case where extremely complex queries are being
executed, a higher optimization level may be desirable, as more potentially cost
effective access paths will be considered.
3. Database statistics: The DB2 optimizer bases its decisions based on database
statistics that are stored in the system catalog. As such, it is good practice to
ensure the statistics are current for the database. RUNSTATS is a DB2 utility
that allows you to update table statistics. See RUNSTATS Command.
Viewing the access plan:
There are three methods of viewing an access plan. Refer to Explain tools. Note
that prior to using these tools, you must create the EXPLAIN tables for the
database. Refer to Explain tables.
To better understand access plans, you need to first understand these basic
components:
Cardinality
<PLAN OPERATOR>
Cost
I/O Cost
The card or cardinality represents the number of rows that DB2 had estimated to
be the output of the operator. The cost represents the CPU cost of this and
previous operations, and IO cost represents the cost of the operator on the system’s
IO subsystems. For example, a snippet from db2exfmt output may appear as
follows:
383.332 <- Cardinality
MSJOIN <- Plan Operator
( 5)
448.972 <- Cost
14.2765 <- I/O Cost
/---+---\

For an overview of the output from explain tools, refer to Description of db2expln
and dynexpln output.
Tracking cardinality - how to measure its accuracy:
You may often find yourself reviewing an access plan that suggests that only a few
rows will be returned. You need to examine and evaluate the cardinality estimates
to determine how accurate these really are. For example:
100
FETCH
/-------------+-----------\
100 perfdb
ISCAN Table1
| 1000
Index: perfdb
Index1
1000
The above plan fragment indicates that the FETCH operation will return an
expected 100 rows. If you suspect that more than 100 rows should be returned (or
less. the principle is the same), you can determine and compare the ″true
cardinality″ with the estimated. To do so, you need to first isolate the predicates
applied at the ISCAN. You can find these in the detailed operator section of the
access plan (this has been reduced to show only the important portions):
3) IXSCAN: (Index Scan)
Cumulative Total Cost: 30.9743
Cumulative CPU Cost: 149356
Cumulative I/O Cost: 1
Cumulative Re-Total Cost: 4.15582
Cumulative Re-CPU Cost: 103895
Cumulative Re-I/O Cost: 0
Cumulative First Row Cost: 30.9054
Estimated Bufferpool Buffers: 2
Predicates:
----------
2) Start Key Predicate
Relational Operator: Equal (=)
Subquery Input Required: No
Filter Factor: 0.1
Predicate Text:
--------------
(Q2.column1 = 5)
Taking the predicate applied here, you can formulate an SQL statement to test the
validity of the cardinality estimate:
db2 select count(*) from table1 where column1=5
If you find that the output of the above query generates a value significantly
different than the estimated value, the statistics may have become stale and a
RUNSTATS may be necessary.
Tracking sorts that spilled to disk:
Another useful technique of isolating query bottlenecks is to determine if a sort

spilled. This is an extension of the techniques developed for verifying cardinalities.
Examine the change in operator costing. Here is an example:
3.65665e+07
TBSCAN
( 15)

6.87408e+06
1.45951e+06
|
3.65665e+07
SORT
( 16)
6.14826e+06
TBSCAN
( 17)
2.00653e+06
1.14206e+06
You see that from the SORT(16) operator to the TBSCAN(15) operator, the cost
went from 6.14e6 to 6.87e6. This jump in COST for the operation directly following
a sort indicates that the SORT has spilled to disk, and has created temporary tables
in order to handle overflow rows. To determine the estimated number of pages
expected to spill, jump to the operator details section of the plan:
15) TBSCAN: (Table Scan)
.
.
.
Estimated Bufferpool Buffers: 163976
The estimated buffer pool buffers indicate that 163976 pages were expected to spill.
Possible corrections include increasing sortheap or defining an index.
Related reference:
v Description of db2expln and dynexpln output
v Explain tools
v Examples of db2expln and dynexpln output
v Guidelines for analyzing explain information
v The SQL compiler process
v Guidelines for using explain information
v Optimization class guidelines
Tuning practices
Sort overflows
DB2 Connect performance troubleshooting

Problem determination particular to multi-partition environments

In a multi-partition environment, the way to identify the partitions that are
involved in a problem varies from problem to problem:
v Unexpected messages or SQL codes. In a partitioned database environment, the
database partition where an application is submitted is the coordinator partition.
Whenever a DB2 operation fails, the error message is returned to the coordinator
partition. Then, based on the nature of the operation and diagnostic data
collected from the coordinator partition, you can identify the rest of the nodes
involved in this failed operation so that you can gather further diagnostic data
from them.

v Abends (Abnormal ending). In a partitioned database environment, a DB2
instance could abend on all partitions or on some of the partitions. If you are not
sure which partitions were shut down abnormally, check the DIAGPATH
directories to see whether any trap files were generated at the specific time of
the problem and where the trap files reveal the node numbers of partitions
where the abend occurred. Or you may look through the db2diag.log files to
search for the related EDUCodeTrapHandler entries, which are written into
db2diag.log whenever there is an instance crash incident. The db2diag.log entry
shows the node number information as well.
v Database or data corruption. Normally the first time you notice a database
corruption problem is when you fail to connect to a certain database partition or
fail to access a certain database object. You can immediately conclude that the
partition (or partitions) where the database object resides probably has the
corruption. If you are able to schedule a long enough window, you may further
use the db2dart utility to inspect all or part of the partitions and accurately
identify the corrupted partitions.
v Loops and hangs. When a specific operation or set of operations hangs, you
might take a few application snapshots one minute apart on each partition, in
order to determine what the status of the application is and whether any work is
really being done. If the application status is ″executing″ but no counters are
increasing, then that’s the partition you need to focus on.
v Slow performance. For an operation with slow performance, you might again
take a few application snapshots one minute apart on each partition, in order to
determine what the status of the application is and how the work is being done.
If the status is ″executing″ and counters like rows-read or written are increasing,
you might be able to evaluate the performance based on how fast the counters
increase. Some DB2 utilities provide query commands, such as load query and
rollforward query, which you can use to identify the partitions where
performance is slower than you expect.
Related reference:
v Data partitioning
v Recovering from transaction failures in a partitioned database environment
Common problems in multi-partition environments

Partitioned instance set-up
When you are using DB2 Extended Enterprise Server Edition, there are some
additional steps that you must take if you wish to use multiple database partitions.
Refer to Enabling data partitioning in a database for an overview of the
considerations.
If you find that you cannot start the instance successfully after enabling data
partitioning, some of the common causes are:
v The .rhosts file has not been configured properly, on UNIX and Linux platforms
v The fast communication manager (FCM) ports have not been reserved
Examples of these problem scenarios are provided below.
Problem 1: .rhosts file has not been configured
On UNIX and Linux platforms, when you have set up a partitioned instance and
are trying to start the instance for the first time, you may encounter an SQL6048
error. For example:

%db2start
03-17-2005 18:13:32 0 0 SQL6048N A communication error occurred during
START or STOP DATABASE MANAGER processing.
03-17-2005 18:13:34 1 0 SQL6048N A communication error occurred during
START or STOP DATABASE MANAGER processing.
SQL1032N No start database manager command was issued. SQLSTATE=57019
If the .rhosts file has not been properly configured, you will see an indication of
the source of the problem in the db2diag.log file:
PID:59060(db2start) TID:1 Appid:none
oper system services sqloPdbExecuteRemoteCmd Probe:50
0xF01FD223 : ... myhost.ibm.com

PID:59060(db2start) TID:1 Appid:none
oper system services sqloPdbExecuteRemoteCmd Probe:50
0x2FF1D75C : 7273 6864 3A20 3038 3236 2D38 3133 2050 rshd: 0826-813 P
0x2FF1D76C : 6572 6D69 7373 696F 6E20 6973 2064 656E ermission is den
0x2FF1D77C : 6965 642E 0A ied..
...
Since DB2 ESE uses the rsh command to execute some commands such as db2start
on all of the partitions, each partition must have the authority to perform remote
commands on the other partitions. This can be done by updating the .rhosts file in
the instance home directory. Refer to Enabling the execution of remote commands
(UNIX).
After you edit the .rhosts file properly, run db2start again. It should now be
successful on all partitions.
Problem 2: Too few (or no) FCM ports have been reserved
This situation can occur on all platforms. It is usually typified by a SQL6031 error
on db2start. For example:
SQL6031N Error in the db2nodes.cfg file at line number "3". Reason code "12".
To find out what reason code 12 refers to, execute the db2 ? command:
% db2 ? SQL6031
...
(12) The port value at line "<line>" of the db2nodes.cfg file in
the sqllib directory is not in the valid port range defined for
your DB2 instance id in the services file (/etc/services on
UNIX-based systems).
...
In a partitioned database environment, most communication between database

partitions is handled by the Fast Communications Manager (FCM). To enable the
FCM at a database partition and allow communication with other database
partitions, you must create a service entry in the /etc/services file. The FCM uses
the specified port to communicate. If you have defined multiple partitions on the
same host, you must define a range of ports. Refer to the appropriate link for
further information:
v Verifying port range availability on participating computers (UNIX)
v Verifying port range availability on participating computers (Windows)
Related reference:

v Creating a node configuration file
v Determining problems with rah on UNIX-based platforms
Log full situation during redistribution: After adding a database partition and
updating the nodegroup definition, you need to use the REDISTRIBUTE NODEGROUP
utility to move the data to the appropriate database partitions. Partition
redistribution is an insert-delete procedure. DB2 logs each insert and delete SQL on
the related partitions. If the log space size is not big enough, it’s possible for the
partition redistribution operation to fail with an SQL0964 (″The transaction log for
the database is full″) error when a large number of records need to be redistributed
across partitions. For example:
% db2 "redistribute nodegroup ibmdefaultgroup uniform"
SQL6064N SQL error "-964" occurred during data redistribution.
% db2 ? sql0964
SQL0964C The transaction log for the database is full.
You can confirm the error message in the db2diag.log file:

2005-03-18-13.02.47.734287-300 E13748A504 LEVEL: Error
PID : 7176390 TID : 1 PROC : db2agntp (SAMPLE) 0
APPHDL : 0-52 APPID: *N0.lcawley.050318180122
FUNCTION: DB2 UDB, data protection, sqlpgResSpace, probe:2860
MESSAGE : ADM1823E The active log is full and is held by application handle "52". Terminate this
APPLICATION.
2005-03-18-13.02.47.737550-300 I14253A469 LEVEL: Error

PID : 7176390 TID : 1 PROC : db2agntp (SAMPLE) 0
FUNCTION: DB2 UDB, data protection, sqlpWriteLR, probe:6680
RETCODE : ZRC=0x85100009=-2062548983=SQLP_NOSPACE
"Log File has reached its saturation point"
DIA8309C Log file was full.
...
2005-03-18-13.02.47.827614-300 I15116A685 LEVEL: Error
PID : 5865510 TID : 1 PROC : db2agent (SAMPLE) 0
INSTANCE: lcawley NODE : 000 DB : SAMPLE
DATA #1 : SQLCA, PD_DB2_TYPE_SQLCA, 136 bytes
sqlcaid : SQLCA sqlcabc: 136 sqlcode: -964 sqlerrml: 0
sqlerrmc:
sqlerrp : sqlridel
sqlerrd : (1) 0x85100009 (2) 0x00000009 (3) 0x00000000
(4) 0x00000000 (5) 0x00000000 (6) 0x00000000
sqlwarn : (1) (2) (3) (4) (5) (6)
(7) (8) (9) (10) (11)
sqlstate:
From the db2diag.log file, you can tell that partition (NODE) 0 hit the log full
error. You now need to increase the log space size on that partition. You can either
increase the log file size or the number of the log files. The following example
chooses to increase the log file size:
Switch to partition 0:
% db2 terminate
% export DB2NODE=0
Increase the log space size by increasing the LOGFILSIZ db cfg parameter:
% db2 update db cfg for sample using LOGFILSIZ 1000

Then you can retry the redistribute operation. If the data redistribution operation
fails, some tables may be redistributed while others are not. This occurs because
data redistribution is performed one table at a time. So you cannot recover from
the failing redistribution operation by simply reissuing that command. Refer to
Redistribution error recovery for more details.
Related reference:
v Log space requirements for data redistribution
v Redistributing data across partitions
v Redistribution error recovery
Dropping a database partition: Click on this link for more information.
SQL1035 (database in use) when taking offline backups in parallel: The version
recovery method requires loading a backup copy of the database. The database
will be restored to exactly the same state that it was in when it was backed up.
Using the version recovery method, you must schedule and perform full backups
of the database on a regular basis.
You may have the following experience: All database partitions are inactive at the
moment. However, offline backup of all partitions in parallel fails with SQL1035N
(The database is currently in use.) on most of the partitions.
You try to take offline backup of all partitions in parallel as follows:

% db2 force applications all
% db2_all ";db2 backup database sample to /database/backup"
host1: SQL1035N The database is currently in use. SQLSTATE=57019
host1: db2 backup database ... completed rc=4
host2: SQL1035N The database is currently in use. SQLSTATE=57019
host2: db2 backup database ... completed rc=4
host3: Backup successful. The timestamp for this backup image is :
20050318135804
host3: db2 backup database ... completed ok
You can confirm the error in the db2diag.log file:

2005-03-18-13.58.04.542575-300 I45063A686 LEVEL: Error
PID : 8085644 TID : 1 PROC : db2agent (SAMPLE) 1
APPHDL : 1-58 APPID: *N1.db2inst1.050318185804
DATA #1 : SQLCA, PD_DB2_TYPE_SQLCA, 136 bytes
sqlcaid : SQLCA sqlcabc: 136 sqlcode: -1035 sqlerrml: 0
sqlerrmc:
sqlerrp : SQLESUBC
sqlerrd : (1) 0x00000000 (2) 0x00000000 (3) 0x00000000
(4) 0x00000000 (5) 0x00000000 (6) 0x00000000
sqlwarn : (1) (2) (3) (4) (5) (6)
(7) (8) (9) (10) (11)
sqlstate:
This occurs because the database backup utility requires exclusive access to the
catalog partition.
To properly backup the database in parallel, you may use the following
commands:
% db2_all "<<+0<; db2 backup database sample to /database/backup"
% db2_all "<<-0<; db2 backup database sample to /database/backup"

This assumes that CATALOG PARTITION is partition 0. It is also a good
illustration of why the CATALOG PARTITION should contain catalog data ONLY.
Related reference:
v BACKUP DATABASE command
v Version recovery

Chapter 3. Searching knowledge bases
DB2 troubleshooting resources
A wide variety of troubleshooting and problem determination information is
available to assist you in using DB2® products.
DB2 documentation:
Troubleshooting information can be found throughout the DB2 Information Center,

as well as throughout the PDF books that make up the DB2 library. You can refer
to the ″Support and troubleshooting″ branch of the DB2 Information Center
navigation tree (in the left pane of your browser window) to see a complete listing
of the DB2 troubleshooting documentation.
DB2 Technical Support Web site:
Refer to the DB2 Technical Support Web site if you are experiencing problems and
want help finding possible causes and solutions. The Technical Support site has
links to the latest DB2 publications, TechNotes, Authorized Program Analysis
Reports (APARs), FixPaks and the latest listing of internal DB2 error codes, and
other resources. You can search through this knowledge base to find possible
solutions to your problems.
Access the DB2 Technical Support Web site at

http://www.ibm.com/software/data/db2/udb/support

Chapter 4. Getting fixes
Release Notes
Information Integrator Release Notes
DB2 Release Notes
How to effectively search for known problems

There are many resources available that describe known problems, including DB2
APARs, whitepapers, redbooks, technotes and manuals. It is important to be able
to effectively search these (and other) resources in order to quickly determine
whether a solution already exists for the problem you are experiencing.
Before searching, it is important to have a clear understanding of your problem

situation. If you do not, refer to the Introduction to problem determination.
Once you have a clear understanding of what the problem situation is, you need to
compile a list of search keywords to increase your chances of finding the existing
solutions. Here are some tips:
1. Use multiple words in your search. The more pertinent search terms you use,
the better your search results will be.
2. Start with specific results, and then go to broader results if necessary. i.e. If too
few results are returned, then remove some of the less pertinent search terms
and try it again. Alternatively, if you are uncertain which keywords to use, you
can perform a broad search with a few keywords, look at the type of results
that you receive, and be able to make a more informed choice of additional
keywords.
3. Sometimes it is more effective to search for a specific phrase. For example, if
you enter: ″administration notification file″ (with the quotes) you will get only
those documents that contain the exact phrase in the exact order in which you
type it. (As opposed to all documents that contain any combination of those
three words).
4. Use wildcards. If you are encountering a specific SQL error, search for
″SQL5005<wildcard>″, where wildcard is the appropriate wildcard for the
resource you’re searching. This is likely to return more results than if you had
merely searched for ″SQL5005″ or ″SQL5005c″.
5. If you are encountering a situation where your instance ends abnormally and
produces trap files, search for known problems using the first two or three
functions in the trap or core file’s stack traceback. If too many results are
returned, try adding keywords ″trap″, ″abend″ or ″crash″.
6. If you are searching for keywords that are operating-system-specific (such as
signal numbers or errno values), try searching on the constant name, not the
value. i.e. Search for ″EFBIG″ instead of the error number 27. For information
about matching operating system error numbers to their constant names, see
Platform specific error logs and Formatting and interpreting trap files.
In general, search terms that are successful often involve:

v Words that describe the command run
v Words that describe the symptoms

v Tokens from the diagnostics
Related reference:
v Introduction to problem determination
v Platform specific error logs
v Formatting and interpreting trap files
Authorized Program Analysis Reports (APARs) and fix packs

A DB2 FixPak contains updates and fixes for problems (Authorized Program
Analysis Reports, or ″APARs″). These APARs address problems found during
testing at IBM, as well as fixes for problems reported by customers.
APARs have unique identifiers. For example, ″IY53671″ and ″JR19727″. Each DB2
APAR is specific to a particular version of DB2 (e.g. DB2 version 8 or DB2 version
7).
Every FixPak is accompanied by a document, called APARLIST.TXT, that lists the

APARs it contains.
FixPaks are cumulative. This means that the latest FixPak for any given version of
DB2 contains all of the updates from previous FixPaks for the same version of
DB2. It is recommended that you keep your DB2 environment running at the latest
FixPak level to ensure problem-free operation.
Viewing APAR content:
To get a better understanding of APARs, go to the DB2 UDB Support website for
Linux, Windows and UNIX platforms:
http://www.ibm.com/software/data/db2/udb/support/
In the search bar, enter one of the APAR identifiers mentioned above, i.e. ″IY53671″
or ″JR19727″.
Here is what you should see for IY61009, for example:
IY61009: Cannot directly change a tablespace privilege from yes to grant.
A fix is available
DB2 Universal Database Version 8 FixPak 8
APAR status
Closed as program error.
Error description
Users who have privileges to access a tablespace without the
grant option cannot be granted permission with the grant option
without having their privilege on the tablespace dropped first.
Users who are granted access with the grant option up front do
not have a problem.
Local fix
Drop the user’s access to the tablespace and then reissue the
grant using the ’with grant option’.

Problem summary
Users Affected: All
Problem Description:
Users who may think they have access to a tablespace with the
grant option will not be able to grant other users access to the
tablespace.
Catalog table SYSCAT.TBSPACEAUTH.USEAUTH will be ’Y’ instead of

’G’ for any user who was granted access to the tablespace
without the grant option and then granted access to the
tablespace with the grant option.
Problem Summary:
Users are unable to grant access to a tablespace when they have
been given explicit grant access with the grant option.
Problem conclusion
Problem was first fixed in Version 8.2 FixPak 8 (s041221)
Temporary fix
Comments
APAR information
APAR number IY61009
Reported component name DB2 UDB ESE AIX
Reported component ID 5765F4100
Reported release 810
Status CLOSED PER
PE NoPE
HIPER NoHIPER
Submitted date 2004-08-24
Closed date 2005-02-04
Last modified date 2005-02-09
APAR is sysrouted FROM one or more of the following:
APAR is sysrouted TO one or more of the following:
Modules/Macros
DEFSRXXX
Fix information
Applicable component levels
R810 PSY UP
Look at various fields in the APAR.:

v The abstract is the title. It is a brief description of the defect.
v The error description describes in detail the symptoms experienced. Depending on
the complexity of the problem, this information will allow you to confirm if you
are encountering the defect described.
v Both the local fix and temporary fix describe circumvention (if available), and
corrective steps to work around the defect (if available).
v The problem summary may go into some detail about what causes the problem.
v The problem conclusion typically indicates which FixPak will contain the APAR.
This information is also specified under the heading ″Fixes are available″.
Chapter 4. Getting fixes 203

Open APARs:
Go back to the main page on the DB2 Support website. Click on the ″APARs″ link.
It should take you to the following page:
http://www.ibm.com/software/data/db2/udb/support/apars.html
Here you will see lists of (among other things) ″Open APARs″. Open APARs are
those which are currently being worked upon or are waiting to be included in the
next available FixPak. You can view lists of all of these outstanding APARs on this
website.
Open APARs are likely to be resolved in the next available FixPak for the version
that they are opened against. There are two exceptions to this:
v an APAR might be identified too late in a FixPak’s development and testing
cycle to be included, or
v an APAR might be closed, to be fixed in a future release (closed as ″FIN″). The
explanation of a FIN APAR is:
If a defect is determined to be of lower impact which does not require an

immediate, permanent fix, [IBM] may defer the fix for a future release. APARs will
reflect deferred fixes with a closing code of ″FIN ″(Fixed If there is a Next release)
to designate plans for inclusion in a future release. -- IBM Software Support
Handbook http://techsupport.services.ibm.com/guides/handbook.html
Normal vs. HIPER APARs:
On this web page

(http://www.ibm.com/software/data/db2/udb/support/apars.html), you can also
see information about ″HIPER APARs″.
High-Impact PERvasive (HIPER) APARs are critical bugs of which all DB2
customers should be aware. Bugs that might result in data loss, system outages,
function loss, poor performance, or are pervasive in the field are typically
categorized as HIPER APARs.
A HIPER APAR has the same format as all other APARs. If you click on one of
these HIPER APARs, though, you will see that there is a flag at the bottom of the
APAR information:
Yes HIPER Yes HIPER
What to do about HIPER APARs:
You should review each HIPER APAR that is not resolved in the DB2 version and
FixPak that you are running. This helps you assess the risk exposure of staying at
a particular FixPak level.
Reading the abstract of the less severe open APARs also helps you assess the risk.
Related reference:
v Applying the latest FixPak (Windows and UNIX)

Notices
IBM may not offer the products, services, or features discussed in this document in
all countries. Consult your local IBM representative for information on the
products and services currently available in your area. Any reference to an IBM
product, program, or service is not intended to state or imply that only that IBM
product, program, or service may be used. Any functionally equivalent product,
program, or service that does not infringe any IBM intellectual property right may
be used instead. However, it is the user’s responsibility to evaluate and verify the
operation of any non-IBM product, program, or service.
IBM may have patents or pending patent applications covering subject matter
described in this document. The furnishing of this document does not give you
any license to these patents. You can send license inquiries, in writing, to:
IBM Director of Licensing
IBM Corporation
North Castle Drive
Armonk, NY 10504-1785
U.S.A.
For license inquiries regarding double-byte (DBCS) information, contact the IBM
Intellectual Property Department in your country/region or send inquiries, in
writing, to:
IBM World Trade Asia Corporation
Licensing
2-31 Roppongi 3-chome, Minato-ku
Tokyo 106, Japan
The following paragraph does not apply to the United Kingdom or any other
country/region where such provisions are inconsistent with local law:
INTERNATIONAL BUSINESS MACHINES CORPORATION PROVIDES THIS
PUBLICATION “AS IS” WITHOUT WARRANTY OF ANY KIND, EITHER
EXPRESS OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
WARRANTIES OF NON-INFRINGEMENT, MERCHANTABILITY, OR FITNESS
FOR A PARTICULAR PURPOSE. Some states do not allow disclaimer of express or
implied warranties in certain transactions; therefore, this statement may not apply
to you.
This information could include technical inaccuracies or typographical errors.

Changes are periodically made to the information herein; these changes will be
incorporated in new editions of the publication. IBM may make improvements
and/or changes in the product(s) and/or the program(s) described in this
publication at any time without notice.
Any references in this information to non-IBM Web sites are provided for
convenience only and do not in any manner serve as an endorsement of those Web
sites. The materials at those Web sites are not part of the materials for this IBM
product, and use of those Web sites is at your own risk.
IBM may use or distribute any of the information you supply in any way it
believes appropriate without incurring any obligation to you.

Licensees of this program who wish to have information about it for the purpose
of enabling: (i) the exchange of information between independently created
programs and other programs (including this one) and (ii) the mutual use of the
information that has been exchanged, should contact:
IBM Canada Limited
Office of the Lab Director
8200 Warden Avenue
Markham, Ontario
L6G 1C7
CANADA
Such information may be available, subject to appropriate terms and conditions,

including in some cases payment of a fee.
The licensed program described in this document and all licensed material
available for it are provided by IBM under terms of the IBM Customer Agreement,
IBM International Program License Agreement, or any equivalent agreement
between us.
Any performance data contained herein was determined in a controlled

environment. Therefore, the results obtained in other operating environments may
vary significantly. Some measurements may have been made on development-level
systems, and there is no guarantee that these measurements will be the same on
generally available systems. Furthermore, some measurements may have been
estimated through extrapolation. Actual results may vary. Users of this document
should verify the applicable data for their specific environment.
Information concerning non-IBM products was obtained from the suppliers of

those products, their published announcements, or other publicly available sources.
IBM has not tested those products and cannot confirm the accuracy of
performance, compatibility, or any other claims related to non-IBM products.
Questions on the capabilities of non-IBM products should be addressed to the
suppliers of those products.
All statements regarding IBM’s future direction or intent are subject to change or
withdrawal without notice, and represent goals and objectives only.
This information may contain examples of data and reports used in daily business
operations. To illustrate them as completely as possible, the examples include the
names of individuals, companies, brands, and products. All of these names are
fictitious, and any similarity to the names and addresses used by an actual
business enterprise is entirely coincidental.
COPYRIGHT LICENSE:
This information may contain sample application programs, in source language,

which illustrate programming techniques on various operating platforms. You may
copy, modify, and distribute these sample programs in any form without payment
to IBM for the purposes of developing, using, marketing, or distributing
application programs conforming to the application programming interface for the
operating platform for which the sample programs are written. These examples
have not been thoroughly tested under all conditions. IBM, therefore, cannot
guarantee or imply reliability, serviceability, or function of these programs.
Each copy or any portion of these sample programs or any derivative work must
include a copyright notice as follows:

© (your company name) (year). Portions of this code are derived from IBM Corp.
Sample Programs. © Copyright IBM Corp. _enter the year or years_. All rights
reserved.
Trademarks
The following terms are trademarks of International Business Machines
Corporation in the United States, other countries, or both, and have been used in
at least one of the documents in the DB2 UDB documentation library.
ACF/VTAM iSeries
AISPO LAN Distance
AIX MVS
AIXwindows MVS/ESA
AnyNet MVS/XA
APPN Net.Data
AS/400 NetView
BookManager OS/390
C Set++ OS/400
C/370 PowerPC
CICS pSeries
Database 2 QBIC
DataHub QMF
DataJoiner RACF
DataPropagator RISC System/6000
DataRefresher RS/6000
DB2 S/370
DB2 Connect SP
DB2 Extenders SQL/400
DB2 OLAP Server SQL/DS
DB2 Information Integrator System/370
DB2 Query Patroller System/390
DB2 Universal Database SystemView
Distributed Relational Tivoli
Database Architecture VisualAge
DRDA VM/ESA
eServer VSE/ESA
Extended Services VTAM
FFST WebExplorer
First Failure Support Technology WebSphere
IBM WIN-OS/2
IMS z/OS
IMS/ESA zSeries
The following terms are trademarks or registered trademarks of other companies

and have been used in at least one of the documents in the DB2 UDB
documentation library:
Microsoft, Windows, Windows NT, and the Windows logo are trademarks of
Microsoft Corporation in the United States, other countries, or both.
Intel and Pentium are trademarks of Intel Corporation in the United States, other
countries, or both.
Java and all Java-based trademarks are trademarks of Sun Microsystems, Inc. in the
United States, other countries, or both.
Notices 207
UNIX is a registered trademark of The Open Group in the United States and other
countries.
Linux is a trademark of Linus Torvalds in the United States, other countries, or

both.
Other company, product, or service names may be trademarks or service marks of

others.

Index
A T
APAR 1 table spaces
Authorized Program Analysis Reports 1 bringing ONLINE 151
threads
in DB2 4
C tools
diagnostic
coordinator agent
Windows 38
description 4
troubleshooting
core files
introduction to problem determination
identification 69
tutorial 1
UNIX systems 69
operating system diagnostics
tutorial 71
tutorials
D introduction to problem
databases determination 1
processes 4 operating system problem
diagnostic tools determination 71
UNIX 69
Windows 38
dump files
error reports 61
H
high impact pervasive APARs 1
HIPER APARs 1
I
IBM Software Support Handbook 1
O
OFFLINE table space
bringing ONLINE 151
P
problem determination tutorials
introduction to problem
determination 1
operating system diagnostics 71
process model
for DB2 4
process status
viewing
UNIX 41
S
Software Support Handbook 1
system command (UNIX)
dbx 69
system core files, UNIX 69
system processes 4

Contacting IBM
In the United States, call one of the following numbers to contact IBM:
v 1-800-IBM-SERV (1-800-426-7378) for customer service
v 1-888-426-4343 to learn about available service options
v 1-800-IBM-4YOU (426-4968) for DB2 marketing and sales
In Canada, call one of the following numbers to contact IBM:

v 1-800-IBM-SERV (1-800-426-7378) for customer service
v 1-800-465-9600 to learn about available service options
v 1-800-IBM-4YOU (1-800-426-4968) for DB2 marketing and sales
To locate an IBM office in your country or region, check IBM’s Directory of

Worldwide Contacts on the web at http://www.ibm.com/planetwide
Product information
Information regarding DB2 Universal Database products is available by telephone
or by the World Wide Web at http://www.ibm.com/software/data/db2/udb
This site contains the latest information on the technical library, ordering books,
product downloads, newsgroups, FixPaks, news, and links to web resources.
If you live in the U.S.A., then you can call one of the following numbers:
v 1-800-IBM-CALL (1-800-426-2255) to order products or to obtain general
information.
v 1-800-879-2755 to order publications.
For information on how to contact IBM outside of the United States, go to the IBM
Worldwide page at www.ibm.com/planetwide

IBMR
Printed in USA
GC09-2850-01
Spine information:
IBM ® ®
IBM DB2 Universal Database DB2 Troubleshooting Guide

Troubleshooting Guide: IBM DB2 Universal Database

Uploaded by

Copyright:

Available Formats

Troubleshooting Guide: IBM DB2 Universal Database

Uploaded by

Document Information

Original Title

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

Troubleshooting Guide: IBM DB2 Universal Database

Uploaded by

Copyright:

Available Formats

What are some important steps in resolving a problem?

What are some important steps in resolving a problem?

What types of questions help understand the environment a problem occurs in?

What types of questions help understand the environment a problem occurs in?

® ®

IBM DB2 Universal Database IBM

© Copyright IBM Corp. 1993, 2000 iii

iv DB2 Troubleshooting Guide

7 What is the problem?:

7 What is the environment and configuration?:

© Copyright IBM Corp. 1993, 2000 1

7 Part of identifying where a problem is occurring is understanding the environment

7 When does the problem happen?:

7 Developing a detailed time line of events leading up to a failure is another

7 Under which conditions does the problem happen?:

7 Is the problem reproducible?:

7 From a problem description and investigation standpoint, the ″ideal″ problem is

2 DB2 Troubleshooting Guide

7 Recreating a single incident problem in a test or development environment is often

7 Our example is reproducible.

Identifying the version and service level of your product

Chapter 1. Troubleshooting overview 3

IBM DB2 JDBC Universal Driver Architecture 2.3.63

On Windows, this command is in the sqllib\bin directory.

On UNIX/Linux, this command is in the sqllib/bin directory.

DB2 server process model

UNIX-based environments use an architecture based on system processes. For

DB2 architecture provides a firewall so that applications run in a different address

4 DB2 Troubleshooting Guide

Per Instance Other

db2agent db2agent Agent

Figure 1. Process Model for DB2 Systems

Chapter 1. Troubleshooting overview 5

A coordinator agent may be:

An additional type of agent, db2agnsc is created by DB2 to perform certain

Database Server Threads and Processes:

6 DB2 Troubleshooting Guide

Differences between Windows and UNIX:

Another difference is in the handling of abnormal terminations. There is no need

Proactive monitoring tools

Techniques for monitoring the database:

Chapter 1. Troubleshooting overview 7

Note: You must have one of SYSADM, SYSCTRL, SYSMAINT, or SYSMON

Snapshot and event monitors

8 DB2 Troubleshooting Guide

Some of the statistics returned by this command provide point- in-time

Creating an event monitor

As an example, try creating an event monitor for the SAMPLE database:

Activating an event monitor

To see if an event monitor is active or inactive, issue the SQL function

Chapter 1. Troubleshooting overview 9

Basic monitoring commands

To enable or disable the monitor switches, use the command:

To set a monitor switch for all partitions, issue the command:

10 DB2 Troubleshooting Guide

Configuring health indicators for health monitors

The following example provides an overview of the steps necessary to configure a

Indicator Name = db.db_op_status

Once that time has passed, obtain a health monitor snapshot:

* End stack traceback *