SG 246983

Front cover
ABCs of z/OS System

Programming
Volume 3
DFSMS, Data set basics, SMS
Storage management software

and hardware
Catalogs, VSAM, DFSMStvs
Paul Rogers
Redelf Janssen
Andre Otto
Rita Pleus
Alvaro Salla
Valeria Sokal
ibm.com/redbooks
International Technical Support Organization
ABCs of z/OS System Programming Volume 3
March 2010
SG24-6983-03
Note: Before using this information and the product it supports, read the information in Notices on
page ix.
Fourth Edition (March 2010)
This edition applies to Version 1 Release 11 of z/OS (5694-A01) and to subsequent releases and
modifications until otherwise indicated in new editions.
Copyright International Business Machines Corporation 2010. All rights reserved.

Note to U.S. Government Users Restricted Rights -- Use, duplication or disclosure restricted by GSA ADP Schedule
Contract with IBM Corp.
Contents
Notices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
Trademarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .x
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi
The team who wrote this book . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi
Now you can become a published author, too! . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xii
Comments welcome. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xii
Stay connected to IBM Redbooks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii
Chapter 1. DFSMS introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.1 Introduction to DFSMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Data facility storage management subsystem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 DFSMSdfp component . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.4 DFSMSdss component . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.5 DFSMSrmm component . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.6 DFSMShsm component . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.7 DFSMStvs component . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
Chapter 2. Data set basics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

2.1 Data sets on storage devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.2 Data set name rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.3 DFSMSdfp data set types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.4 Types of VSAM data sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.5 Non-VSAM data sets. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.6 Extended-format data sets and objects. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.7 Data set striping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.8 Data set striping with z/OS V1R11 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.9 Large format data sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.10 Large format data sets and TSO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
2.11 IGDSMSxx parmlib member support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2.12 z/OS UNIX files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.13 Data set specifications for non-VSAM data sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
2.14 Locating an existing data set. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
2.15 Uncataloged and cataloged data sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
2.16 Volume table of contents (VTOC) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
2.17 VTOC and DSCBs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
2.18 VTOC index structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
2.19 Initializing a volume using ICKDSF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
Chapter 3. Extended access volumes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

3.1 Traditional DASD capacity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
3.2 Large volumes before z/OS V1R10 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
3.3 zArchitecture data scalability. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
3.4 WLM controlling PAVs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
3.5 Parallel Access Volumes (PAVs) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
3.6 HyperPAV feature for DS8000 series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
3.7 HyperPAV implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
3.8 Device type 3390 and 3390 Model A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
3.9 Extended access volumes (EAV) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
Copyright IBM Corp. 2010. All rights reserved. iii

3.10 Data sets eligible for EAV volumes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
3.11 EAV volumes and multicylinder units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
3.12 Dynamic volume expansion (DVE) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
3.13 Using dynamic volume expansion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
3.14 Command-line interface (DSCLI) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
3.15 Using Web browser GUI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
3.16 Select volume to increase capacity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
3.17 Increase capacity of volumes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
3.18 Select capacity increase for volume . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
3.19 Final capacity increase for volume . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
3.20 VTOC index with EAV volumes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
3.21 Device Support FACILITY (ICKDSF) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
3.22 Update VTOC after volume expansion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
3.23 Automatic VTOC index rebuild - z/OS V1R11. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
3.24 Automatic VTOC rebuild with DEVMAN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
3.25 EAV and IGDSMSxx parmlib member . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
3.26 IGDSMSxx member BreakPointValue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
3.27 New EATTR attribute in z/OS V1R11 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
3.28 EATTR parameter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
3.29 EATTR JCL DD statement example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
3.30 Migration assistance tracker . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
3.31 Migration tracker commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
Chapter 4. Storage management software . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

4.1 Overview of DFSMSdfp utilities. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
4.2 IEBCOMPR utility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
4.3 IEBCOPY utility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
4.4 IEBCOPY: Copy operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
4.5 IEBCOPY: Compress operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
4.6 IEBGENER utility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
4.7 IEBGENER: Adding members to a PDS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
4.8 IEBGENER: Copying data to tape . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
4.9 IEHLIST utility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
4.10 IEHLIST LISTVTOC output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
4.11 IEHINITT utility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
4.12 IEFBR14 utility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
4.13 DFSMSdfp access methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
4.14 Access method services (IDCAMS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
4.15 IDCAMS functional commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
4.16 AMS modal commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
4.17 DFSMS Data Collection Facility (DCOLLECT) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
4.18 Generation data group (GDG). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
4.19 Defining a generation data group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
4.20 Absolute generation and version numbers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
4.21 Relative generation numbers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
4.22 Partitioned organized (PO) data sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
4.23 PDS data set organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
4.24 Partitioned data set extended (PDSE) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
4.25 PDSE enhancements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
4.26 PDSE: Conversion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
4.27 Program objects in a PDSE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
4.28 Sequential access methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
4.29 z/OS V1R9 QSAM - BSAM enhancements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
iv ABCs of z/OS System Programming Volume 3

4.30 Virtual storage access method (VSAM) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
4.31 VSAM terminology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
4.32 VSAM: Control interval (CI) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
4.33 VSAM data set components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
4.34 VSAM key sequenced cluster (KSDS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
4.35 VSAM: Processing a KSDS cluster . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
4.36 VSAM entry sequenced data set (ESDS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
4.37 VSAM: Typical ESDS processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
4.38 VSAM relative record data set (RRDS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
4.39 VSAM: Typical RRDS processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
4.40 VSAM linear data set (LDS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
4.41 VSAM: Data-in-virtual (DIV) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
4.42 VSAM: Mapping a linear data set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
4.43 VSAM resource pool . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
4.44 VSAM: Buffering modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
4.45 VSAM: System-managed buffering (SMB) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
4.46 VSAM buffering enhancements with z/OS V1R9 . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
4.47 VSAM SMB enhancement with z/OS V1R11 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
4.48 VSAM enhancements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
4.49 Data set separation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
4.50 Data set separation syntax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
4.51 Data facility sort (DFSORT) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190
4.52 z/OS Network File System (z/OS NFS). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
4.53 DFSMS optimizer (DFSMSopt) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
4.54 Data Set Services (DFSMSdss) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
4.55 DFSMSdss: Physical and logical processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
4.56 DFSMSdss: Logical processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
4.57 DFSMSdss: Physical processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
4.58 DFSMSdss stand-alone services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
4.59 Hierarchical Storage Manager (DFSMShsm) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204
4.60 DFSMShsm: Availability management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
4.61 DFSMShsm: Space management. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
4.62 DFSMShsm: Storage device hierarchy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
4.63 ML1 enhancements with z/OS V1R11 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
4.64 DFSMShsm z/OS V1R11 enhancements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
4.65 ML1 and ML2 volumes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
4.66 Data set allocation format and volume pool determination . . . . . . . . . . . . . . . . . . . . 217
4.67 DFSMShsm volume types. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
4.68 DFSMShsm: Automatic space management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
4.69 DFSMShsm data set attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
4.70 RETAINDAYS keyword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
4.71 RETAINDAYS keyword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
4.72 DFSMShsm: Recall processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
4.73 Removable media manager (DFSMSrmm) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
4.74 Libraries and locations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
4.75 What DFSMSrmm can manage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
4.76 Managing libraries and storage locations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
Chapter 5. System-managed storage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239

5.1 Storage management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240
5.2 DFSMS and DFSMS environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
5.3 Goals and benefits of system-managed storage. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242
5.4 Service level objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
Contents v
5.5 Implementing SMS policies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
5.6 Monitoring SMS policies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249
5.7 Assigning data to be system-managed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250
5.8 Using data classes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251
5.9 Using storage classes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
5.10 Using management classes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255
5.11 Management class functions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257
5.12 Using storage groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258
5.13 Using aggregate backup and recovery support (ABARS) . . . . . . . . . . . . . . . . . . . . . 260
5.14 Automatic Class Selection (ACS) routines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262
5.15 SMS configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264
5.16 SMS control data sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 266
5.17 Implementing DFSMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268
5.18 Steps to activate a minimal SMS configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
5.19 Allocating SMS control data sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271
5.20 Defining the SMS base configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
5.21 Creating ACS routines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 276
5.22 DFSMS setup for z/OS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278
5.23 Starting SMS and activating a new configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . 280
5.24 Control SMS processing with operator commands . . . . . . . . . . . . . . . . . . . . . . . . . . 282
5.25 Displaying the SMS configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284
5.26 Managing data with a minimal SMS configuration . . . . . . . . . . . . . . . . . . . . . . . . . . 285
5.27 Device-independence space allocation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287
5.28 Developing naming conventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289
5.29 Setting the low-level qualifier (LLQ) standards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291
5.30 Establishing installation standards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 292
5.31 Planning and defining data classes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293
5.32 Data class attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 294
5.33 Use data class ACS routine to enforce standards . . . . . . . . . . . . . . . . . . . . . . . . . . 295
5.34 Simplifying JCL use. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296
5.35 Allocating a data set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297
5.36 Creating a VSAM cluster . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299
5.37 Retention period and expiration date . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301
5.38 SMS PDSE support. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 302
5.39 Selecting data sets to allocate as PDSEs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 303
5.40 Allocating new PDSEs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304
5.41 System-managed data types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305
5.42 Data types that cannot be system-managed. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307
5.43 Interactive Storage Management Facility (ISMF) . . . . . . . . . . . . . . . . . . . . . . . . . . . 309
5.44 ISMF: Product relationships . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310
5.45 ISMF: What you can do with ISMF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312
5.46 ISMF: Accessing ISMF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 314
5.47 ISMF: Profile option. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315
5.48 ISMF: Obtaining information about a panel field . . . . . . . . . . . . . . . . . . . . . . . . . . . . 316
5.49 ISMF: Data set option . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318
5.50 ISMF: Volume Option . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319
5.51 ISMF: Management Class option . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 320
5.52 ISMF: Data Class option . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321
5.53 ISMF: Storage Class option . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 322
5.54 ISMF: List option . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323
Chapter 6. Catalogs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325

6.1 Catalogs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 326
vi ABCs of z/OS System Programming Volume 3

6.2 The basic catalog structure (BCS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 328
6.3 The VSAM volume data set (VVDS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 330
6.4 Catalogs by function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 332
6.5 Using aliases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335
6.6 Catalog search order. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337
6.7 Defining a catalog and its aliases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339
6.8 Using multiple catalogs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342
6.9 Sharing catalogs across systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343
6.10 Listing a catalog . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345
6.11 Defining and deleting data sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347
6.12 DELETE command enhancement with z/OS V1R11 . . . . . . . . . . . . . . . . . . . . . . . . 351
6.13 Backup procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353
6.14 Recovery procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355
6.15 Checking the integrity on an ICF structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357
6.16 Protecting catalogs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359
6.17 Merging catalogs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 361
6.18 Splitting a catalog . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 363
6.19 Catalog performance. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365
6.20 F CATALOG,REPORT,PERFORMANCE command . . . . . . . . . . . . . . . . . . . . . . . . 367
6.21 Catalog address space (CAS). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 369
6.22 Working with the catalog address space. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 371
6.23 Fixing temporary catalog problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 373
6.24 Enhanced catalog sharing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375
Chapter 7. DFSMS Transactional VSAM Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . 377

7.1 VSAM share options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 378
7.2 Base VSAM buffering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 380
7.3 Base VSAM locking. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 382
7.4 CICS function shipping before VSAM RLS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 383
7.5 VSAM record-level sharing introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 384
7.6 VSAM RLS overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 385
7.7 Data set sharing under VSAM RLS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 387
7.8 Buffering under VSAM RLS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 388
7.9 VSAM RLS locking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 390
7.10 VSAM RLS/CICS data set recovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 392
7.11 Transactional recovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 394
7.12 The batch window problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 395
7.13 VSAM RLS implementation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 396
7.14 Coupling Facility structures for RLS sharing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 397
7.15 Update PARMLIB with VSAM RLS parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . 400
7.16 Define sharing control data sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 402
7.17 Update SMS configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 405
7.18 Update data sets with log parameters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 408
7.19 The SMSVSAM address space. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 410
7.20 Interacting with VSAM RLS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 412
7.21 Backup and recovery of CICS VSAM data sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . 415
7.22 Interpreting RLSDATA in an IDCAMS LISTCAT output . . . . . . . . . . . . . . . . . . . . . . 417
7.23 DFSMStvs introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 419
7.24 Overview of DFSMStvs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 421
7.25 DFSMStvs use of z/OS RRMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 423
7.26 Atomic updates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 425
7.27 Unit of work and unit of recovery. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 426
7.28 DFSMStvs logging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 427
Contents vii
7.29 Accessing a data set with DFSMStvs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 429
7.30 Application considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 430
7.31 DFSMStvs logging implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 432
7.32 Prepare for logging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 433
7.33 Update PARMLIB with DFSMStvs parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 436
7.34 The DFSMStvs instance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 438
7.35 Interacting with DFSMStvs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 439
7.36 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 442
Chapter 8. Storage management hardware . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 445

8.1 Overview of DASD types. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 446
8.2 Redundant array of independent disks (RAID) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 448
8.3 Seascape architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 450
8.4 Enterprise Storage Server (ESS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 453
8.5 ESS universal access . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 455
8.6 ESS major components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 456
8.7 ESS host adapters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 457
8.8 FICON host adapters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 458
8.9 ESS disks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 460
8.10 ESS device adapters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 462
8.11 SSA loops . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 464
8.12 RAID-10 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 466
8.13 Storage balancing with RAID-10 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 468
8.14 ESS copy services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 469
8.15 ESS performance features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 472
8.16 IBM TotalStorage DS6000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 474
8.17 IBM TotalStorage DS8000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 477
8.18 DS8000 hardware overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 479
8.19 Storage systems LPARs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 481
8.20 IBM TotalStorage Resiliency Family . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 483
8.21 TotalStorage Expert product highlights . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 485
8.22 Introduction to tape processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 487
8.23 SL and NL format . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 489
8.24 Tape capacity - tape mount management. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 491
8.25 TotalStorage Enterprise Tape Drive 3592 Model J1A. . . . . . . . . . . . . . . . . . . . . . . . 493
8.26 IBM TotalStorage Enterprise Automated Tape Library 3494 . . . . . . . . . . . . . . . . . . 495
8.27 Introduction to Virtual Tape Server (VTS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 497
8.28 IBM TotalStorage Peer-to-Peer VTS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 499
8.29 Storage area network (SAN) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 501
Related publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 503

IBM Redbooks publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 503
Other publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 503
Online resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 504
How to get IBM Redbooks publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 504
Help from IBM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 504
viii ABCs of z/OS System Programming Volume 3

Notices
This information was developed for products and services offered in the U.S.A.
IBM may not offer the products, services, or features discussed in this document in other 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.
The following paragraph does not apply to the United Kingdom or any other country 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.
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.
This information contains 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 contains 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.
Copyright IBM Corp. 2010. All rights reserved. ix

Trademarks
IBM, the IBM logo, and ibm.com are trademarks or registered trademarks of International Business Machines
Corporation in the United States, other countries, or both. These and other IBM trademarked terms are
marked on their first occurrence in this information with the appropriate symbol ( or ), indicating US
registered or common law trademarks owned by IBM at the time this information was published. Such
trademarks may also be registered or common law trademarks in other countries. A current list of IBM
trademarks is available on the Web at http://www.ibm.com/legal/copytrade.shtml
The following terms are trademarks of the International Business Machines Corporation in the United States,
other countries, or both:
AIX i5/OS RETAIN
AS/400 IBM RS/6000
CICS IMS S/390
DB2 iSeries System i
DS6000 Language Environment System Storage
DS8000 Magstar System z
Enterprise Storage Server OS/390 Tivoli
ESCON OS/400 TotalStorage
eServer Parallel Sysplex VTAM
FICON POWER5 z/Architecture
FlashCopy PowerPC z/OS
GDPS PR/SM z/VM
Geographically Dispersed Parallel pSeries z9
Sysplex RACF zSeries
Hiperspace Redbooks
HyperSwap Redbooks (logo)
The following terms are trademarks of other companies:
Novell, the Novell logo, and the N logo are registered trademarks of Novell, Inc. in the United States and other
countries.
ACS, Interchange, and the Shadowman logo are trademarks or registered trademarks of Red Hat, Inc. in the
U.S. and other countries.
SAP, and SAP logos are trademarks or registered trademarks of SAP AG in Germany and in several other
countries.
Java, and all Java-based trademarks are trademarks of Sun Microsystems, Inc. in the United States, other
countries, or both.
Microsoft, Windows NT, Windows, and the Windows logo are trademarks of Microsoft Corporation in the
United States, other countries, or both.
Intel, Intel logo, Intel Inside logo, and Intel Centrino logo are trademarks or registered trademarks of Intel
Corporation or its subsidiaries in the United States and other countries.
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.
x ABCs of z/OS System Programming Volume 3

Preface
The ABCs of z/OS System Programming is a thirteen-volume collection that provides an

introduction to the z/OS operating system and the hardware architecture. Whether you are a
beginner or an experienced system programmer, the ABCs collection provides the
information that you need to start your research into z/OS and related subjects. The ABCs
collection serves as a powerful technical tool to help you become more familiar with z/OS in
your current environment, or to help you evaluate platforms to consolidate your e-business
applications.
This edition is updated to z/OS Version 1 Release 1.
The contents of the volumes are:
Volume 1: Introduction to z/OS and storage concepts, TSO/E, ISPF, JCL, SDSF, and z/OS
delivery and installation
Volume 2: z/OS implementation and daily maintenance, defining subsystems, JES2 and
JES3, LPA, LNKLST, authorized libraries, Language Environment, and SMP/E
Volume 3: Introduction to DFSMS, data set basics, storage management hardware and
software, VSAM, System-Managed Storage, catalogs, and DFSMStvs
Volume 4: Communication Server, TCP/IP and VTAM
Volume 5: Base and Parallel Sysplex, System Logger, Resource Recovery Services (RRS),
Global Resource Serialization (GRS), z/OS system operations, Automatic Restart
Management (ARM), Geographically Dispersed Parallel Sysplex (GPDS)
Volume 6: Introduction to security, RACF, Digital certificates and PKI, Kerberos,

cryptography and z990 integrated cryptography, zSeries firewall technologies, LDAP,
Enterprise Identity Mapping (EIM), and firewall technologies
Volume 7: Printing in a z/OS environment, Infoprint Server and Infoprint Central
Volume 8: An introduction to z/OS problem diagnosis
Volume 9: z/OS UNIX System Services
Volume 10: Introduction to z/Architecture, zSeries processor design, zSeries connectivity,

LPAR concepts, HCD, and HMC
Volume 11: Capacity planning, performance management, RMF, and SMF
Volume 12: WLM
Volume 13: JES3
The team who wrote this book

This book was produced by a team of specialists from around the world working at the
International Technical Support Organization, Poughkeepsie Center.
Copyright IBM Corp. 2010. All rights reserved. xi

Paul Rogers is a Consulting IT Specialist at the International Technical Support
Organization, Poughkeepsie Center and has worked for IBM for more than 40 years. He
writes extensively and teaches IBM classes worldwide on various aspects of z/OS, JES3,
Infoprint Server, and z/OS UNIX. Before joining the ITSO 20 years ago, Paul worked in the
IBM Installation Support Center (ISC) in Greenford, England, providing OS/390 and JES
support for IBM EMEA and the Washington Systems Center in Gaithersburg, Maryland.
Redelf Janssen is an IT Architect in IBM Global Services ITS in IBM Germany. He holds a
degree in Computer Science from University of Bremen and joined IBM Germany in 1988. His
areas of expertise include IBM zSeries, z/OS and availability management. He has written
IBM Redbooks publications on OS/390 Releases 3, 4, and 10, and z/OS Release 8.
Andre Otto is a z/OS DFSMS SW service specialist at the EMEA Backoffice team in
Germany. He has 12 years of experience in the DFSMS, VSAM and catalog components.
Andre holds a degree in Computer Science from the Dresden Professional Academy.
Rita Pleus is an IT Architect in IBM Global Services ITS in IBM Germany. She has 21 years
of IT experience in a variety of areas, including systems programming and operations
management. Before joining IBM in 2001, she worked for a German S/390 customer. Rita
holds a degree in Computer Science from the University of Applied Sciences in Dortmund.
Her areas of expertise include z/OS, its subsystems, and systems management.
Alvaro Salla is an IBM retiree who worked for IBM for more than 30 years in large systems.
He has co-authored many IBM Redbooks publications and spent many years teaching S/360
to S/390. He has a degree in Chemical Engineering from the University of Sao Paulo, Brazil.
Valeria Sokal is an MVS system programmer at an IBM customer. She has 16 years of
experience as a mainframe systems programmer.
The fourth edition was updated by Paul Rogers.
Now you can become a published author, too!

Here's an opportunity to spotlight your skills, grow your career, and become a published
author - all at the same time! Join an ITSO residency project and help write a book in your
area of expertise, while honing your experience using leading-edge technologies. Your efforts
will help to increase product acceptance and customer satisfaction, as you expand your
network of technical contacts and relationships. Residencies run from two to six weeks in
length, and you can participate either in person or as a remote resident working from your
home base.
Find out more about the residency program, browse the residency index, and apply online at:
ibm.com/redbooks/residencies.html
Comments welcome
Your comments are important to us!
We want our books to be as helpful as possible. Send us your comments about this book or
other IBM Redbooks in one of the following ways:
Use the online Contact us review Redbooks form found at:
ibm.com/redbooks
xii ABCs of z/OS System Programming Volume 3

Send your comments in an e-mail to:
redbooks@us.ibm.com
Mail your comments to:
IBM Corporation, International Technical Support Organization
Dept. HYTD Mail Station P099
2455 South Road
Poughkeepsie, NY 12601-5400
Stay connected to IBM Redbooks

Find us on Facebook:
http://www.facebook.com/pages/IBM-Redbooks/178023492563?ref=ts
Follow us on twitter:
http://twitter.com/ibmredbooks
Look for us on LinkedIn:
http://www.linkedin.com/groups?home=&gid=2130806
Explore new Redbooks publications, residencies, and workshops with the IBM Redbooks
weekly newsletter:
https://www.redbooks.ibm.com/Redbooks.nsf/subscribe?OpenForm
Stay current on recent Redbooks publications with RSS Feeds:
http://www.redbooks.ibm.com/rss.html
Preface xiii
xiv ABCs of z/OS System Programming Volume 3
1
Chapter 1. DFSMS introduction

This chapter gives a brief overview of the Data Facility Storage Management Subsystem
(DFSMS) and its primary functions in the z/OS operating system. DFSMS comprises a suite
of related data and storage management products for the z/OS system. DFSMS is now an
integral element of the z/OS operating system.
DFSMS is an operating environment that helps automate and centralize the management of
storage based on the policies that your installation defines for availability, performance,
space, and security.
The heart of DFSMS is the Storage Management Subsystem (SMS). Using SMS, the storage
administrator defines policies that automate the management of storage and hardware
devices. These policies describe data allocation characteristics, performance and availability
goals, backup and retention requirements, and storage requirements for the system.
DFSMS is an exclusive element of the z/OS operating system and is a software suite that
automatically manages data from creation to expiration.
DFSMSdfp is a base element of z/OS. DFSMSdfp is automatically included with z/OS.

DFSMSdfp performs the essential data, storage, and device management functions of the
system. DFSMSdfp and DFSMShsm provide disaster recovery functions such as Advanced
Copy Services and aggregate backup and recovery support (ABARS).
The other elements of DFSMSDFSMSdss, DFSMShsm, DFSMSrmm, and

DFSMStvsare optional features that complement DFSMSdfp to provide a fully integrated
approach to data and storage management. In a system-managed storage environment,
DFSMS automates and centralizes storage management based on the policies that your
installation defines for availability, performance, space, and security. With the optional
features enabled, you can take full advantage of all the functions that DFSMS offers.
Copyright IBM Corp. 2010. All rights reserved. 1

1.1 Introduction to DFSMS
IBM 3494
Tape Library dsname.f.data Tape

Data Set
DASD
System Programmer
Figure 1-1 Introduction to data management
Understanding DFSMS
Data management is the part of the operating system that organizes, identifies, stores,
catalogs, and retrieves all the data information (including programs) that your installation
uses. DFSMS is an exclusive element of the z/OS operating system. DFSMS is a software
suite that automatically manages data from creation to expiration.
DFSMSdfp helps you store and catalog information about DASD, optical, and tape devices so
that it can be quickly identified and retrieved from the system. DFSMSdfp provides access to
both record- and stream-oriented data in the z/OS environment. The z/OS operating system
enables you to efficiently manage e-business workloads and enterprise transactions 24 hours
a day. DFSMSdfp is automatically included with z/OS. It performs the essential data, storage,
and device management functions of the system.
Systems programmer
As a systems programmer, you can use DFSMS data management to:
Allocate space on DASD and optical volumes
Automatically locate cataloged data sets
Control access to data
Transfer data between the application program and the medium
Mount magnetic tape volumes in the drive
2 ABCs of z/OS System Programming Volume 3

1.2 Data facility storage management subsystem
NFS
IBM workstations
IBM System z
dfp
tvs NFS
Storage dss DFSMS
Hierarchy
rmm
hsm
P690
Figure 1-2 Data Facility Storage Management Subsystem
DFSMS components
DFSMS is an exclusive element of the z/OS operating system. DFSMS is a software suite
that automatically manages data from creation to expiration. The following elements comprise
DFSMS:
DFSMSdfp, a base element of z/OS
DFSMSdss, an optional feature of z/OS
DFSMShsm, an optional feature of z/OS
DFSMSrmm, an optional feature of z/OS
DFSMStvs, an optional feature of z/OS
DFSMSdfp Provides storage, data, program, and device management. It is comprised of
components such as access methods, OPEN/CLOSE/EOV routines, catalog
management, DADSM (DASD space control), utilities, IDCAMS, SMS, NFS,
ISMF, and other functions.
DFSMSdss Provides data movement, copy, backup, and space management functions.
DFSMShsm Provides backup, recovery, migration, and space management functions. It
invokes DFSMSdss for certain of its functions.
DFSMSrmm Provides management functions for removable media such as tape cartridges
and optical media.
DFSMStvs Enables batch jobs and CICS online transactions to update shared VSAM
data sets concurrently.
Chapter 1. DFSMS introduction 3

Network File System
The Network File System (NFS) is a distributed file system that enables users to access UNIX
files and directories that are located on remote computers as though they were local. NFS is
independent of machine types, operating systems, and network architectures.
Importance of DFSMS elements

The z/OS operating system enables you to efficiently manage e-business workloads and
enterprise transactions 24 hours a day. DFSMSdfp is automatically included with z/OS.
DFSMSdfp performs the essential data, storage, and device management functions of the
system. DFSMSdfp and DFSMShsm provide disaster recovery functions such as Advanced
Copy Services and aggregate backup and recovery support (ABARS).
The other elements of DFSMSDFSMSdss, DFSMShsm, DFSMSrmm, and

DFSMStvscomplement DFSMSdfp to provide a fully-integrated approach to data and
storage management. In a system-managed storage environment, DFSMS automates and
centralizes storage management based on the policies that your installation defines for
availability, performance, space, and security. With these optional features enabled, you can
take full advantage of all the functions that DFSMS offers.

1.3 DFSMSdfp component
DFSMSdfp provides the following functions:

Managing storage
Managing data
Using access methods, commands, and utilities
Managing devices
Tape mount management
Distributed data access
Advanced copy services
Object access method (OAM)
Figure 1-3 DFSMSdfp functions
DFSMSdfp component
DFSMSdfp provides storage, data, program, and device management. It is comprised of
components such as access methods, OPEN/CLOSE/EOV routines, catalog management,
DADSM (DASD space control), utilities, IDCAMS, SMS, NFS, ISMF, and other functions.
Managing storage
The storage management subsystem (SMS) is a DFSMSdfp facility designed for automating
and centralizing storage management. SMS automatically assigns attributes to new data
when that data is created. SMS automatically controls system storage and assigns data to
the appropriate storage device. ISMF panels allow you to specify these data attributes.
For more information about ISMF, see 5.43, Interactive Storage Management Facility (ISMF)
on page 309.
Managing data
DFSMSdfp organizes, identifies, stores, catalogs, shares, and retrieves all the data that your
installation uses. You can store data on DASD, magnetic tape volumes, or optical volumes.
Using data management, you can complete the following tasks:
Allocate space on DASD and optical volumes
Automatically locate cataloged data sets
Control access to data
Transfer data between the application program and the medium
Mount magnetic tape volumes in the drive

Using access methods, commands, and utilities
DFSMSdfp manages the organization and storage of data in the z/OS environment. You can
use access methods with macro instructions to organize and process a data set or object.
Access method services commands manage data sets, volumes, and catalogs. Utilities
perform tasks such as copying and moving data. You can use system commands to display
and set SMS configuration parameters, use DFSMSdfp callable services to write advanced
application programs, and use installation exits to customize DFSMS.
Managing devices with DFSMSdfp

You need to use the Hardware Configuration Definition (HCD) to define I/O devices to the
operating system, and to control these devices. DFSMS manages DASD, storage control
units, magnetic tape devices, optical devices, and printers. You can use DFSMS functions to
manage many device types, but most functions apply specifically to one type or one family of
devices.

Tape mount management is a methodology for improving tape usage and reducing tape
costs. This methodology involves intercepting selected tape data set allocations through the
SMS automatic class selection (ACS) routines and redirecting them to a direct access storage
device (DASD) buffer. Once on DASD, you can migrate these data sets to a single tape or
small set of tapes, thereby reducing the overhead associated with multiple tape mounts.
Distributed data access with DFSMSdfp

In the distributed computing environment, applications must often access data residing on
other computers in a network. Often, the most effective data access services occur when
applications can access remote data as though it were local data.
Distributed FileManager/MVS is a DFSMSdfp client/server product that enables remote

clients in a network to access data on z/OS systems. Distributed FileManager/MVS provides
workstations with access to z/OS data. Users and applications on heterogeneous client
computers in your network can take advantage of system-managed storage on z/OS, data
sharing, and data security with RACF.
z/OS UNIX System Services (z/OS UNIX) provides the command interface that interactive
UNIX users can use. z/OS UNIX allows z/OS programs to directly access UNIX data.
Advanced copy services

Advanced copy services includes remote and point-in-time copy functions that provide
backup and recovery of data. When used before a disaster occurs, Advanced Copy Services
provides rapid backup of critical data with minimal impact to business applications. If a
disaster occurs to your data center, Advanced Copy Services provides rapid recovery of
critical data.
Object access method

Object access method (OAM) provides storage, retrieval, and storage hierarchy management
for objects. OAM also manages storage and retrieval for tape volumes that are contained in
system-managed libraries.

1.4 DFSMSdss component
DFSMSdss provides the following functions:

Data movement and replication
Space management
Data backup and recovery
Data set and volume conversion
Distributed data management
FlashCopy feature with Enterprise Storage Server
(ESS)
SnapShot feature with RAMAC Virtual Array (RVA)
Concurrent copy
Figure 1-4 DFSMSdss functions
DFSMSdss component
DFSMSdss is the primary data mover for DFSMS. DFSMSdss copies and moves data to help
manage storage, data, and space more efficiently. It can efficiently move multiple data sets
from old to new DASD. The data movement capability that is provided by DFSMSdss is useful
for many other operations, as well. You can use DFSMSdss to perform the following tasks.
Data movement and replication

DFSMSdss lets you move or copy data between volumes of like and unlike device types. If
you create a backup in DFSMSdss, you can copy a backup copy of data. DFSMSdss also can
produce multiple backup copies during a dump operation.
Space management
DFSMSdss can reduce or eliminate DASD free-space fragmentation.
Data backup and recovery

DFSMSdss provides you with host system backup and recovery functions at both the data set
and volume levels. It also includes a stand-alone restore program that you can run without a
host operating system.
Data set and volume conversion

DFSMSdss can convert your data sets and volumes to system-managed storage. It can also
return your data to a non-system-managed state as part of a recovery procedure.

Distributed data management
DFSMSdss saves distributed data management (DDM) attributes that are associated with a
specific data set and preserves those attributes during copy and move operations.
DFSMSdss also offers the FlashCopy feature with Enterprise Storage Server (ESS) and
the SnapShot feature with RAMAC Virtual Array (RVA). FlashCopy and SnapShot function
automatically, work much faster than traditional data movement methods, and are well-suited
for handling large amounts of data.
Concurrent copy
When it is used with supporting hardware, DFSMSdss also provides concurrent copy
capability. Concurrent copy lets you copy or back up data while that data is being used. The
user or application program determines when to start the processing, and the data is copied
as though no updates have occurred.

1.5 DFSMSrmm component
DFSMSdmm provides the following functions:

Library management
Shelf management
Volume management
Data set management
Figure 1-5 DFSMSrmm functions
DFSMSrmm component
DFSMSrmm manages your removable media resources, including tape cartridges and reels.
It provides the following functions.
Library management
You can create tape libraries, or collections of tape media associated with tape drives, to
balance the work of your tape drives and help the operators that use them.
DFSMSrmm can manage the following devices:

A removable media library, which incorporates all other libraries, such as:
System-managed manual tape libraries.
System-managed automated tape libraries. Examples of automated tape libraries
include:
IBM TotalStorage
Enterprise Automated Tape Library (3494)
IBM TotalStorage Virtual Tape Servers (VTS)
Non-system-managed or traditional tape libraries, including automated libraries such as a
library under Basic Tape Library Support (BTLS) control.

Shelf management
DFSMSrmm groups information about removable media by shelves into a central online
inventory, and keeps track of the volumes residing on those shelves. DFSMSrmm can
manage the shelf space that you define in your removable media library and in your storage
locations.
Volume management
DFSMSrmm manages the movement and retention of tape volumes throughout their life
cycle.
Data set management

DFSMSrmm records information about the data sets on tape volumes. DFSMSrmm uses the
data set information to validate volumes and to control the retention and movement of those
data sets.

1.6 DFSMShsm component
DFSMShsm provides the following functions:
Storage management
Space management
Availability management
Figure 1-6 DFSMShsm functions
DFSMShsm component
DFSMShsm complements DFSMSdss to provide the following functions.
Storage management
DFSMShsm provides automatic DASD storage management, thus relieving users from
manual storage management tasks.
Space management
DFSMShsm improves DASD space usage by keeping only active data on fast-access storage
devices. It automatically frees space on user volumes by deleting eligible data sets, releasing
overallocated space, and moving low-activity data to lower cost-per-byte devices, even if the
job did not request tape.

DFSMShsm can write multiple output data sets to a single tape, making it a useful tool for
implementing tape mount management under SMS. When you redirect tape data set
allocations to DASD, DFSMShsm can move those data sets to tape, as a group, during
interval migration. This methodology greatly reduces the number of tape mounts on the
system. DFSMShsm uses a single-file format, which improves your tape usage and search
capabilities.

DFSMShsm backs up your data, automatically or by command, to ensure availability if
accidental loss of the data sets or physical loss of volumes occurs. DFSMShsm also allows
the storage administrator to copy backup and migration tapes, and to specify that copies be
made in parallel with the original. You can store the copies onsite as protection from media
damage, or offsite as protection from site damage. DFSMShsm also provides disaster backup
and recovery for user-defined groups of data sets (aggregates) so that you can restore critical
applications at the same location or at an offsite location.
Attention: You must also have DFSMSdss to use the DFSMShsm functions.

1.7 DFSMStvs component
Provide transactional recovery within VSAM

RLS allows batch sharing of recoverable data sets
for read
RLS provides locking and buffer coherency
CICS provides logging and two-phase commit
protocols
Transactional VSAM allows batch sharing of
recoverable data sets for update
Logging provided using the System Logger
Two-phase commit and backout using Recoverable
Resource Management Services (RRMS)
Figure 1-7 DFSMStvs functions
DFSMStvs component
DFSMS Transactional VSAM Services (DFSMStvs) allows you to share VSAM data sets
across CICS, batch, and object-oriented applications on z/OS or distributed systems.
DFSMStvs enables concurrent shared updates of recoverable VSAM data sets by CICS
transactions and multiple batch applications. DFSMStvs enables 24-hour availability of CICS
and batch applications.
VSAM record-level sharing (RLS)

With VSAM RLS, multiple CICS systems can directly access a shared VSAM data set,
eliminating the need to ship functions between the application-owning regions and file-owning
regions. CICS provides the logging, commit, and backout functions for VSAM recoverable
data sets. VSAM RLS provides record-level serialization and cross-system caching. CICSVR
provides a forward recovery utility.
DFSMStvs is built on top of VSAM record-level sharing (RLS), which permits sharing of
recoverable VSAM data sets at the record level. Different applications often need to share
VSAM data sets. Sometimes the applications need only to read the data set. Sometimes an
application needs to update a data set while other applications are reading it. The most
complex case of sharing a VSAM data set is when multiple applications need to update the
data set and all require complete data integrity.
Transaction processing provides functions that coordinate work flow and the processing of
individual tasks for the same data sets. VSAM record-level sharing and DFSMStvs provide

key functions that enable multiple batch update jobs to run concurrently with CICS access to
the same data sets, while maintaining integrity and recoverability.
Recoverable resource management services (RRMS)

RRMS is part of the operating system and comprises registration services, context services,
and recoverable resource services (RRS). RRMS provides the context and unit of recovery
management under which DFSMStvs participates as a recoverable resource manager.

2
Chapter 2. Data set basics

A data set is a collection of logically related data; it can be a source program, a library of
macros, or a file of data records used by a processing program. Data records (also called
logical records) are the basic unit of information used by a processing program. By placing
your data into volumes of organized data sets, you can save and process the data efficiently.
You can also print the contents of a data set, or display the contents on a terminal.
You can store data on secondary storage devices, such as:

A direct access storage device (DASD)
The term DASD applies to disks or to a large amount of magnetic storage media on which
a computer stores data. A volume is a standard unit of secondary storage. You can store
all types of data sets on DASD.
Each block of data on a DASD volume has a distinct location and a unique address, thus
making it possible to find any record without extensive searching. You can store and
retrieve records either directly or sequentially. Use DASD volumes for storing data and
executable programs, including the operating system itself, and for temporary working
storage. You can use one DASD volume for many separate data sets, and reallocate or
reuse space on the volume.
A magnetic tape volume
Only sequential data sets can be stored on magnetic tape. Mountable tape volumes can
reside in an automated tape library. For information about magnetic tape volumes, see
z/OS DFSMS: Using Magnetic Tapes, SC26-7412. You can also direct a sequential data
set to or from spool, a UNIX file, a TSO/E terminal, a unit record device, virtual I/O (VIO),
or a dummy data set.
The Storage Management Subsystem (SMS) is an operating environment that automates the
management of storage. Storage management uses the values provided at allocation time to
determine, for example, on which volume to place your data set, and how many tracks to
allocate for it. Storage management also manages tape data sets on mountable volumes that
reside in an automated tape library. With SMS, users can allocate data sets more easily.
The data sets allocated through SMS are called system-managed data sets or SMS-managed
data sets.

An access method is a DFSMSdfp component that defines the technique that is used to store
and retrieve data. Access methods have their own data set structures to organize data,
macros to define and process data sets, and utility programs to process data sets.
Access methods are identified primarily by the way that they organize the data in the data set.
For example, use the basic sequential access method (BSAM) or queued sequential access
method (QSAM) with sequential data sets. However, there are times when an access method
identified with one organization can be used to process a data set organized in another
manner. For example, a sequential data set (not extended-format data set) created using
BSAM can be processed by the basic direct access method (BDAM), and vice versa. Another
example is UNIX files, which you can process using BSAM, QSAM, basic partitioned access
method (BPAM), or virtual storage access method (VSAM).
This chapter describes various data set basics:

Data set name rules
Data set characteristics
Locating a data set
Volume table of contents (VTOC)
Initializing a volume

2.1 Data sets on storage devices
DASD volume Tape volume
DATASET.SEQ1
DATASET.SEQ DATASET.SEQ2
DATASET.PDS DATASET.SEQ3
DATASET.VSAM
VOLSER=DASD01 VOLSER=SL0001
Figure 2-1 Data sets on volumes
MVS data sets

An MVS data set is a collection of logically related data records stored on one volume or a set
of volumes. A data set can be, for example, a source program, a library of macros, or a file of
data records used by a processing program. You can print a data set or display it on a
terminal. The logical record is the basic unit of information used by a processing program.
Note: As an exception, the z/OS UNIX services component supports Hierarchical File
System (HFS) data sets, where the collection is of bytes and there is not the concept of
logically related data records.
Storage devices
Data can be stored on a magnetic direct access storage device (DASD), magnetic tape
volume, or optical media. As mentioned previously, the term DASD applies to disks or
simulated equivalents of disks. All types of data sets can be stored on DASD, but only
sequential data sets can be stored on magnetic tape. The types of data sets are described in
2.3, DFSMSdfp data set types on page 20.
DASD volumes
Each block of data on a DASD volume has a distinct location and a unique address, making it
possible to find any record without extensive searching. You can store and retrieve records
either directly or sequentially. Use DASD volumes for storing data and executable programs,
Chapter 2. Data set basics 17

including the operating system itself, and for temporary working storage. You can use one
DASD volume for many separate data sets, and reallocate or reuse space on the volume.
The following sections discuss the logical attributes of a data set, which are specified at data
set creation time in:
DCB/ACB control blocks in the application program
DD cards (explicitly, or through the Data Class (DC) option with DFSMS)
In an ACS Data Class (DC) routine (overridden by a DD card)
After the creation, such attributes are kept in catalogs and VTOCs.

2.2 Data set name rules
HARRY.FILE.EXAMPLE.DATA
1 2 3 4
HLQ LLQ
Figure 2-2 Data set name rules
Data set naming rules

Whenever you allocate a new data set, you (or MVS) must give the data set a unique name.
Usually, the data set name is given as the DSNAME keyword in JCL.
A data set name can be one name segment, or a series of joined name segments. Each
name segment represents a level of qualification. For example, the data set name
HARRY.FILE.EXAMPLE.DATA is composed of four name segments. The first name on the left
is called the high-level qualifier (HLQ), the last name on the right is the lowest-level qualifier
(LLQ).
Each name segment (qualifier) is 1 to 8 characters, the first of which must be alphabetic (A to
Z) or national (# @ $). The remaining seven characters are either alphabetic, numeric (0 - 9),
national, a hyphen (-). Name segments are separated by a period (.).
Note: Including all name segments and periods, the length of the data set name must not
exceed 44 characters. Thus, a maximum of 22 name segments can make up a data set
name.

2.3 DFSMSdfp data set types
Data set types supported

VSAM data sets
Non-VSAM data sets
Extended-format data sets
Large format data sets
Basic format data sets
Objects
z/OS UNIX files
Virtual input/output data sets
Figure 2-3 DFSMSdfp data set types supported
DFSMSdfp data set types

The data organization that you choose depends on your applications and the operating
environment. z/OS allows you to use temporary or permanent data sets, and to use several
ways to organize files for data to be stored on magnetic media, as described here.
VSAM data sets

VSAM data sets are formatted differently than non-VSAM data sets. Except for linear data
sets, VSAM data sets are collections of records, grouped into control intervals. The control
interval is a fixed area of storage space in which VSAM stores records. The control intervals
are grouped into contiguous areas of storage called control areas. To access VSAM data
sets, use the VSAM access method. See also 2.4, Types of VSAM data sets on page 22.
Non-VSAM data sets

Non-VSAM data sets are collections of fixed-length or variable-length records, grouped into
blocks, but not in control intervals. To access non-VSAM data sets, use BSAM, QSAM, or
BPAM. See also 2.5, Non-VSAM data sets on page 23.

You can create both sequential and VSAM data sets in extended format on system-managed
DASD; this implies a 32-bye suffix at each physical record. See also 2.6, Extended-format
data sets and objects on page 25.

Large format data sets are sequential data sets that can grow beyond the size limit of 65 535
tracks (4369 cylinders) per volume that applies to other sequential data sets. Large format
data sets can be system-managed or not. They can be accessed using QSAM, BSAM, or
EXCP.
Large format data sets reduce the need to use multiple volumes for single data sets,
especially very large ones such as spool data sets, dumps, logs, and traces. Unlike
extended-format data sets, which also support greater than 65 535 tracks per volume, large
format data sets are compatible with EXCP and do not need to be SMS-managed.
You can allocate a large format data set using the DSNTYPE=LARGE parameter on the DD
statement, dynamic allocation (SVC 99), TSO/E ALLOCATE, or the access method services
ALLOCATE command.

Basic format data sets are sequential data sets that are specified as neither extended-format
nor large-format. Basic format data sets have a size limit of 65 535 tracks (4369 cylinders) per
volume. They can be system-managed or not, and can be accessed using QSAM, BSAM, or
EXCP.
You can allocate a basic format data set using the DSNTYPE=BASIC parameter on the DD
statement, dynamic allocation (SVC 99), TSO/E ALLOCATE, or the access method services
ALLOCATE command, or the data class. If no DSNTYPE value is specified from any of these
sources, then its default is BASIC.
Objects
Objects are named streams of bytes that have no specific format or record orientation. Use
the object access method (OAM) to store, access, and manage object data. You can use any
type of data in an object because OAM does not recognize the content, format, or structure of
the data. For example, an object can be a scanned image of a document, an engineering
drawing, or a digital video. OAM objects are stored either on DASD in a DB2 database, or
on an optical drive, or on an optical or tape storage volume.
The storage administrator assigns objects to object storage groups and object backup
storage groups. The object storage groups direct the objects to specific DASD, optical, or tape
devices, depending on their performance requirements. You can have one primary copy of an
object and up to two backup copies of an object. A Parallel Sysplex allows you to access
objects from all instances of OAM and from optical hardware within the sysplex.
z/OS UNIX files

z/OS UNIX System Services (z/OS UNIX) enables applications and even z/OS to access
UNIX files. Also UNIX applications also can access z/OS data sets. You can use the
hierarchical file system (HFS), z/OS Network File System (z/OS NFS), zSeries File System
(zFS), and temporary file system (TFS) with z/OS UNIX. You can use the BSAM, QSAM,
BPAM, and VSAM access methods to access data in UNIX files and directories. z/OS UNIX
files are byte-oriented, similar to objects.

2.4 Types of VSAM data sets
Types of VSAM data sets
Key-sequenced data set (KSDS)
Entry-sequenced data set (ESDS)
Relative-record data set (RRDS)
Variable relative-record data set (VRRDS)
Linear data set (LDS)
Figure 2-4 VSAM data set types
VSAM data sets

VSAM arranges records by an index key, by a relative byte address, or by a relative record
number. VSAM data sets are cataloged for easy retrieval.

A KSDS VSAM data set contains records in order by a key field and can be accessed by the
key or by a relative byte address. The key contains a unique value, such as an employee
number or part number.

An ESDS VSAM data set contains records in the order in which they were entered and can
only be accessed by relative byte address. An ESDS is similar to a sequential data set.

An RRDS VSAM data set contains records in order by relative-record number and can only
be accessed by this number. Relative records can be fixed length or variable length. VRRDS
is a type of RRDS where the logical records can be variable.

An LDS VSAM data set contains data that can be accessed as byte-addressable strings in
virtual storage. A linear data set does not have imbedded control information that other VSAM
data sets hold.

2.5 Non-VSAM data sets
DSORG specifies the organization of the data set as:

Physical sequential (PS)
Partitioned (PO)
Direct (DA)
PO.DATA.SET
DIRECTORY A B SEQ.DATA.SET1
C
A B MEMBERS SEQ.DATA.SET2
Partitioned Organized
Physical Sequential
(PDS and PDSE)
Figure 2-5 Types of non-VSAM data sets
Data set organization (DSORG)

DSORG specifies the organization of the data set as physical sequential (PS), partitioned
(PO), or direct (DA). If the data set is processed using absolute rather than relative
addresses, you must mark it as unmovable by adding a U to the DSORG parameter (for
example, by coding DSORG=PSU). You must specify the data set organization in the DCB
macro. In addition:
When creating a direct data set, the DSORG in the DCB macro must specify PS or PSU
and the DD statement must specify DA or DAU.
PS is for sequential and extended format DSNTYPE.
PO is the data set organization for both PDSEs and PDSs. DSNTYPE is used to
distinguish between PDSEs and PDSs.
Non-VSAM data sets

Non-VSAM data sets are collections of fixed-length or variable-length records grouped into
physical blocks (a set of logical records). To access non-VSAM data sets, an application can
use BSAM, QSAM, or BPAM. There are several types of non-VSAM data sets, as follows:
Physical sequential data set (PS)

Sequential data sets contain logical records that are stored in physical order. New records are
appended to the end of the data set. You can specify a sequential data set in extended format
or not.

Partitioned data set (PDS)
Partitioned data sets contain a directory of sequentially organized members, each of which
can contain a program or data. After opening the data set, you can retrieve any individual
member without searching the entire data set.
Partitioned data set extended (PDSE)

Partitioned data sets extended contain an indexed, expandable directory of sequentially
organized members, each of which can contain a program or data. You can use a PDSE
instead of a PDS. The main advantage of using a PDSE over a PDS is that a PDSE
automatically reclaims the space released by a previous member deletion, without the need
for a reorganization.

2.6 Extended-format data sets and objects
An extended-format data set supports the following

options:
Compression
Data striping
Extended-addressability
Objects
Use object access method (OAM)
Storage administrator assigns objects
Figure 2-6 Types of extended-format data sets

While sequential data sets have a maximum of 16 extents on each volume, extended-format
sequential data sets have a maximum of 123 extents on each volume. Each extended-format
sequential data set can have a maximum of 59 volumes, so an extended-format sequential
data set can have a maximum of 7257 extents (123 times 59).
An extended-format data set can occupy any number of tracks. On a volume that has more
than 65,535 tracks, a sequential data set cannot occupy more than 65,535 tracks.
An extended-format, striped sequential data set can contain up to 4 GB blocks. The maximum
size of each block is 32 760 bytes.
An extended-format data set supports the following additional functions:

Compression, which reduces the space for storing data and improves I/O, caching, and
buffering performance.
Data striping, which in a sequential processing environment distributes data for one data
set across multiple SMS-managed DASD volumes, improving I/O performance and
reducing the batch window. For example, a data set with 6 stripes is distributed originally
across 6 volumes.
Large data sets with high sequential I/O activity are the best candidates for striped data
sets. Data sets defined as extended-format sequential must be accessed using BSAM or
QSAM, and not EXCP (means no access method is used) or BDAM.

Extended-addressability, which enables you to create a VSAM data set that is larger than
4 GB.
System-managed DASD
You can allocate both sequential and VSAM data sets in extended format on a
system-managed DASD. Extended-format VSAM data sets also allow you to release partial
unused space and to use system-managed buffering (SMB, a fast buffer pool management
technique) for VSAM batch programs. You can select whether to use the primary or
secondary space amount when extending VSAM data sets to multiple volumes.
Objects
Objects are named streams of bytes that have no specific format or record orientation. Use
the object access method (OAM) to store, access, and manage object data. You can use any
type of data in an object because OAM does not recognize the content, format, or structure of
the data. For example, an object can be a scanned image of a document, an engineering
drawing, or a digital video. OAM objects are stored either on DASD in a DB2 database, or on
an optical drive, or on a tape storage volume.
The storage administrator assigns objects to object storage groups and object backup
storage groups. The object storage groups direct the objects to specific DASD, optical, or tape
devices, depending on their performance requirements. You can have one primary copy of an
object, and up to two backup copies of an object.

2.7 Data set striping
Striping is a software implementation that distributes

sequential data sets across multiple 3390 volumes
Data sets across multiple SMS-managed DASD volumes
Improves I/O performance
For example, a data set with 28 stripes is distributed

across 28 volumes and therefore 28 parallel I/Os
All striped data sets must be extended-format data sets
Physical sequential and VSAM data sets
Defining using data class and storage groups
Figure 2-7 Data set striping
Data striping
Sequential data striping can be used for physical sequential data sets that cause I/O
bottlenecks for critical applications. Sequential data striping uses extended-format sequential
data sets that SMS can allocate over multiple volumes, preferably on separate channel paths
and control units, to improve performance. These data sets must reside on 3390 volumes that
are located on the IBM DS8000.
Sequential data striping can reduce the processing time required for long-running batch jobs
that process large, physical sequential data sets. Smaller sequential data sets can also
benefit because of DFSMS's improved buffer management for QSAM and BSAM access
methods for striped extended-format sequential data sets.
A stripe in DFSMS is the portion of a striped data set, such as an extended format data set,
that resides on one volume. The records in that portion are not always logically consecutive.
The system distributes records among the stripes such that the volumes can be read from or
written to simultaneously to gain better performance. Whether it is striped is not apparent to
the application program. Data striping distributes data for one data set across multiple
SMS-managed DASD volumes, which improves I/O performance and reduces the batch
window. For example, a data set with 28 stripes is distributed across 28 volumes.

You can write striped extended-format sequential data sets with the maximum physical block
size for the data set plus control information required by the access method. The access

method writes data on the first volume selected until a track is filled. The next physical blocks
are written on the second volume selected until a track is filled, continuing until all volumes
selected have been used or no more data exists. Data is written again to selected volumes in
this way until the data set has been created. A maximum of 59 stripes can be allocated for a
data set. For striped data sets, the maximum number of extents on a volume is 123.
Physical sequential and VSAM data sets

The sustained data rate (SDR) has an effect only for extended-format data sets. Striping
allows you to spread data across DASD volumes and controllers. The number of stripes is the
number of volumes on which the data set is initially allocated. Striped data sets must be
system-managed and must be in an extended format. When no volumes that use striping are
available, the data set is allocated as nonstriped with EXT=P specified in the data class; the
allocation fails if EXT=R is specified in the data class.
Physical sequential data sets cannot be extended if none of the stripes can be extended. For
VSAM data sets, each stripe can be extended to an available candidate volume if extensions
fail on the current volume.
Data classes
Data class attributes define space and data characteristics of data sets that are normally
specified on JCL DD statements, TSO/E ALLOCATE commands, access method services
(IDCAMS) DEFINE commands, dynamic allocation requests, and ISPF/PDF panels. You can
use data class to allocate sequential and VSAM data sets in extended format for the benefits
of compression (sequential and VSAM KSDS), striping, and large data set sizes (VSAM).
Storage groups
SMS calculates the average preference weight of each storage group using the preference
weights of the volumes that will be selected if the storage group is selected for allocation.
Then, SMS selects the storage group that contains at least as many primary volumes as the
stripe count and has the highest average weight. If there are no storage groups that meet
these criteria, the storage group with the largest number of primary volumes is selected. If
multiple storage groups have the largest number of primary volumes, the one with the highest
average weight is selected. If there are still multiple storage groups that meet the selection
criteria, SMS selects one at random.
For striped data sets, ensure that there are a sufficient number of separate paths to DASD
volumes in the storage group to allow each stripe to be accessible through a separate path.
The maximum number of stripes for physical sequential (PS) data sets is 59. For VSAM data
sets, the maximum number of stripes is 16. Only sequential or VSAM data sets can be
striped.

2.8 Data set striping with z/OS V1R11
Allow striping volume selection to support all current and

future volume preference attributes specified in SMS
constructs
Allow volumes that are above high allocation threshold to be
eligible for selection as secondary volumes
Prefer enabled volumes over quiesced volumes
Prefer normal storage groups over overflow storage groups
Support data set separation function
Support multi-tiered storage group function that honors
storage group sequence derived by ACS
Support availability, accessibility, PAV and other volume
preference attributes that are used in non-striping volume
selection
Figure 2-8 z/OS V1R11 improvements for data set striping
Striping volume selection

Striping volume selection is very similar to conventional volume selection. Volumes that are
eligible for selection are classified as primary and secondary, and assigned a volume
preference weight, based on preference attributes.
Note: This support is invoked when allocating a new striped data set. Volumes are ranked
by preference weight from each individual controller. This support selects the most
preferred storage group that meets or closely meets the target stripe count. This allows
selection from the most preferred volume from individual controllers to meet the stripe
count (try to spread stripes across controllers).
Automatically activate the fast volume selection function to avoid overutilizing system
resources.
High allocation threshold

With V1R11, volumes that have sufficient space for the allocation amount without exceeding
the storage group HIGH THRESHOLD value are eligible for selection as secondary volumes.
Volumes that do not meet all the criteria for the primary volume list are placed on the
secondary list. In z/OS V1R11, the SMS striping volume selection enhancement will try to
make striping allocation function for both VSAM and non-VSAM as close as possible to the
conventional volume selection.

SMS calculates the average preference weight of each storage group using the preference
weights of the volumes that will be selected if the storage group is selected for allocation.
Then, SMS selects the storage group that contains at least as many primary volumes as the
stripe count and has the highest average weight. If there are no storage groups that meet
these criteria, the storage group with the largest number of primary volumes is selected. If
multiple storage groups have the largest number of primary volumes, the one with the highest
average weight is selected. If there are still multiple storage groups that meet the selection
criteria, SMS selects one at random.
Storage group support

Normal storage groups are preferred over overflow storage groups. The storage group
sequence order as specified in the ACS storage group selection routines is supported when a
multi-tiered storage group is requested in the storage class.
After selecting a storage group, SMS selects volumes by their preference weight. Primary
volumes are preferred over secondary volumes because they have a higher preference
weight. Secondary volumes are selected when there is an insufficient number of primary
volumes. If there are multiple volumes with the same preference weight, SMS selects one of
the volumes at random.
Data set separation

Data set separation allows you to designate groups of data sets in which all SMS-managed
data sets within a group are kept separate, on the physical control unit (PCU) level or the
volume level, from all the other data sets in the same group. To use data set separation, you
must create a data set separation profile and specify the name of the profile to the base
configuration. During allocation, SMS attempts to separate the data sets listed in the profile. A
data set separation profile contains at least one data set separation group. Each data set
separation group specifies whether separation is at the PCU or volume level, whether it is
required or preferred, and includes a list of data set names to be separated from each other
during allocation.
Volume preference
Volume preference attributes, such as availability, accessibility, and PAV capability are
supported.
Fast volume selection is supported, regardless of the current specification of the

FAST_VOLSEL parameter. SMS will reject the candidate volumes that do not have sufficient
free space for the stripe when 100 volumes have already been rejected by DADSM for
insufficient space. This is to prevent the striping allocation from overusing the system
resources, because an iteration of volume reselection can consume significant overhead
when there are a large number of candidate volumes.

2.9 Large format data sets
Figure 2-9 Allocating a data set with ISPF option 3.2

Defining large format data sets was introduced with z/OS V1R7. Large format data sets are
physical sequential data sets, with generally the same characteristics as other non-extended
format sequential data sets, but with the capability to grow beyond the basic format size limit
of 65,535 tracks on each volume. (This is about 3,500,000,000 bytes, depending on the block
size.) Large format data sets reduce the need to use multiple volumes for single data sets,
especially very large ones such as spool data sets, dumps, logs, and traces. Unlike
extended-format data sets, which also support greater than 65,535 tracks per volume, large
format data sets are compatible with EXCP and do not need to be SMS-managed.
Data sets defined as large format must be accessed using QSAM, BSAM, or EXCP.
Large format data sets have a maximum of 16 extents on each volume. Each large format
data set can have a maximum of 59 volumes. Therefore, a large format data set can have a
maximum of 944 extents (16 times 59).
A large format data set can occupy any number of tracks, without the limit of 65,535 tracks
per volume. The minimum size limit for a large format data set is the same as for other
sequential data sets that contain data: one track, which is about 56,000 bytes. Primary and
secondary space can both exceed 65,535 tracks per volume.
Large format data sets can be on SMS-managed DASD or non-SMS-managed DASD.

Restriction: The following types of data sets cannot be allocated as large format data
sets:
PDS, PDSE, and direct data sets
Virtual I/O data sets, password data sets, and system dump data sets
Allocating data sets

To process an already existing data set, first allocate it (establish a logical link with it and your
program), then access the data using macros in Assembler or HLL statements to activate the
access method that you have chosen. The allocation of a data set means either or both of two
things:
To set aside (create) space for a new data set on a disk or tape
To establish a logical link between a job step (your program) and any data set using JCL
Figure 2-9 on page 31 shows the creation of a data set using ISPF panel 3.2. Other ways to
create a data set are as follows:
Access method services
You can define VSAM data sets and establish catalogs by using a multifunction services
program called access method services.
TSO ALLOCATE command
You can issue the ALLOCATE command of TSO/E to define VSAM and non-VSAM data
sets.
Using JCL
Any data set can be defined directly with JCL by specifying DSNTYPE=LARGE on the DD
statement.

Basic format data sets are sequential data sets that are specified as neither extended-format
nor large-format. Basic format data sets have a size limit of 65,535 tracks (4,369 cylinders)
per volume. Basic format data sets can be system-managed or not. They can be accessed
using QSAM, BSAM, or EXCP.
You can allocate a basic format data set using the DSNTYPE=BASIC parameter on the DD
statement, dynamic allocation (SVC 99), TSO/E ALLOCATE or the access method services
ALLOCATE command, or the data class. If no DSNTYPE value is specified from any of these
sources, then its default is BASIC.
Virtual input/output data sets

You can manage temporary data sets with a function called virtual input/output (VIO). VIO
uses DASD space and system I/O more efficiently than other temporary data sets.
You can use the BPAM, BSAM, QSAM, BDAM, and EXCP access methods with VIO data
sets. SMS can direct SMS-managed temporary data sets to VIO storage groups.

2.10 Large format data sets and TSO
There are three types of sequential data sets:

Basic format: A traditional data set existing prior to V1R7
that cannot grow beyond 64 K tracks per volume
Large format: A data set (introduced in V1R7) that has
the capability to grow beyond 64 K tracks
Extended format: An extended format data set that must
be DFSMS-managed
With z/OS V1R9:
Updates to the following commands and service ensure
that each can handle large format data sets:
TSO TRANSMIT, RECEIVE
PRINTDS
REXX LISTDSI function
CLIST LISTDSI statement
REXX EXECIO command
CLIST OPENFILE/GETFILE/PUTFILE I/O processing
Figure 2-10 Large format data set enhancement with z/OS V1R9
Sequential data sets

There are three types of sequential data sets, as follows:
Basic format A traditional data set, as existed prior to z/OS V1.7. These data sets
cannot grow beyond 64 K tracks per volume.
Large format A data set (introduced in z/OS v1.7) that has the capability to grow
beyond 64 K tracks but can be very small. The significance is that after
being defined as a large format data set, it can grow to over 64 K tracks
without further intervention. The maximum size is xFFFFFE or
approximately 16 M tracks per volume.
Extended format An extended format data set that must be DFSMS-managed. This
means that it must have a storage class. These data sets can be
striped, and can grow up to xFFFFFFFE tracks per volume.
Using large format data sets with z/OS V1R9

These enhancements are internal, and remove the restriction in z/OS V1R7 and z/OS V1R8
that prevented use of large format data sets.
Updates have been made to the following commands and service to ensure that each can
handle large format data sets:
TSO TRANSMIT, RECEIVE
PRINTDS

REXX LISTDSI function
CLIST LISTDSI statement
REXX EXECIO command
CLIST OPENFILE/GETFILE/PUTFILE I/O processing
Restriction: Types of data sets that cannot be allocated as large format data sets are:
PDS, PDSE, and direct data sets
Virtual I/O data sets, password data sets, and system dump data sets

2.11 IGDSMSxx parmlib member support
Defining format large data set in parmlib (z/OS V1R9)
BLOCKTOKENSIZE(REQUIRE | NOREQUIRE)
REQUIRE: - Requires every open for a large format data

set to use BLOCKTOKENSIZE=LARGE on the DCBE
macro
NOREQUIRE: - Allows applications to access large

format data sets under more conditions without having to
specify BLOCKTOKENSIZE=LARGE on DCBE macro
Services updated for large format data sets
Figure 2-11 Using large format data sets
Using BLOCKTOKENSIZE(REQUIRE)
If your installation uses the default BLOCKTOKENSIZE(REQUIRE) setting in PARMLIB
member IGDSMSxx, you can issue the following command to see the current
BLOCKTOKENSIZE settings, from the MVS console:
D SMS,OPTIONS
Large format data sets with BLOCKTOKENSIZE(REQUIRE)

The following services are updated for large format data sets:
ALLOCATE For a new or existing large format data set, use the
DSNTYPE(LARGE) keyword.
REXX EXECIO Using EXECIO DISKR to read a LARGE format SEQ data set will
work as long as the data set is <64 K trks in size. If data set is
>64 K trks, expect an ABEND213-17. Using EXECIO DISKW to
attempt to write to any LARGE format data set will fail with
ABEND213-15.
CLIST support Using CLIST OPENFILE INPUT/GETFILE and OPENFILE
OUTPUT/PUTFILE, as follows:
Using CLIST OPENFILE INPUT/GETFILE to read a LARGE format
SEQ data set will work as long as the data set is <64 K trks in size.

SEQ data set that is >64 K trks in size will fail with ABEND213-17 from
OPENFILE INPUT.
Using CLIST OPENFILE OUTPUT/PUTFILE to attempt to write to any
LARGE format data set will fail with ABEND213-15 from OPENFILE
OUTPUT.
REXX LISTDSI With the dsname function (CLIST LISTDSI statement), LISTDSI
issued against a large format data set will complete with a REXX
function return code 0. If the data set is <64 K trks in size, LISTDSI will
provide the same accurate information that it does with any BASIC
sequential data set. However, if the data set size is >64 K trks, the
data set space used (SYSUSED) returned by LISTDSI will be
inaccurate.
TRANSMIT/RECEIVE When issued for a large format data set will transmit the data set as
long as it is <64 K trks in size. If the LARGE format data set is >64 K
trks, an ABEND213-15 will occur during the transmit process, and
nothing will be sent.
If you RECEIVE a data set that was sent with TRANSMIT and that
data set is a LARGE format, RECEIVE will be able to receive the data
set as long as it is <64 K trks in size. However, the data set that
RECEIVE allocates and into which RECEIVE writes the data will be a
BASIC format data set.
If you attempt to RECEIVE the data set into a preallocated LARGE
format data set, RECEIVE will ABEND with ABEND213-17, and
nothing will have been received.
PRINTDS Using PRINTDS to print a LARGE format input data set to the spool
will work normally if the data set being printed is <64 K trks. It will fail
with ABEND213-17 if the data set is >64 K trks in size.
In addition, if you are trying to write the PRINTDS output into a
TODATASET, the TODATASET allocated by PRINTDS will be a normal
BASIC type data set. If you preallocate the TODATASET as a LARGE
format data set, and use PRINTDS to write to the preallocated
TODATASET, PRINTDS will fail with ABEND213-15 when trying to
write to a LARGE format data set.
Large format data sets with BLOCKTOKENSIZE(NOREQUIRE)

When BLOCKTOKENSIZE(NOREQUIRE) is set in the active IGDSMSxx parmlib member,
support works as follows:
ALLOCATE ALLOC of a new or existing large format data set will work. Use the
DSNTYPE(LARGE) keyword to allocate a large format data set.
REXX EXECIO Using EXECIO DISKR to read a large format SEQ data set will
work when the data set is <64 K trks in size. Using EXECIO DISKR
to read a LARGE format SEQ data set will also work when the data
set is >64 K trks in size. Using EXECIO DISKW to write to any
LARGE format data set will work.
CLISTs When using CLIST OPENFILE INPUT/GETFILE and OPENFILE
OUTPUT/PUTFILE, support is as follows:
SEQ data set will work when the data set is <64 K trks in size.

SEQ data set will also work when the data set is >64 K trks in size.
Using CLIST OPENFILE OUTPUT/PUTFILE to write to any large
format data set will work.
LISTDSI Support works the same as when BLOCKTOKENSIZE(REQUIRE).
REXX LISTDSI A REXX LISTDSI dsname function or CLIST LISTDSI statement
issued against a large format data set will complete with a REXX
function return code 0.
If the data set is <64 K trks in size, LISTDSI will provide the same
accurate information that it does with any BASIC sequential data set.
However, if the data set size is >64 K trks, the data set space used
(SYSUSED) returned by LISTDSI will be inaccurate.
If LISTDSI issued against a LARGE format data set that is >64 K trks
and the SMSINFO operand is also specified, then LISTDSI completes
with REXX function return code 0, but the space used (SYSUSED)
information will not be correct.
TRANSMIT/RECEIVE Issued for a LARGE format data set will transmit the data set
regardless of whether it is <64 K trks or >64 K trks in size. If you
RECEIVE a data set that was sent with TRANSMIT and that data set
is a LARGE format, RECEIVE will be able to receive the file into a
BASIC data set created by RECEIVE if the transmitted file is <64 K
trks. RECEIVE will be unable to receive all of the transmitted file if it
was >64 K trks in size, because RECEIVE will not get the correct SIZE
information to allocate the received file. An ABENDx37 can occur. In
all cases, any data set allocated by RECEIVE will be BASIC format,
even if it is attempting to receive a large format data set. You can
RECEIVE data into a preallocated LARGE format data set that you
allocate of sufficient size to hold the transmitted data.
PRINTDS Using PRINTDS to print a large format input data set to the spool will
work normally regardless of whether the data set is <64 K trks or
>64 K trks in size. If you attempt to write output to a TODATASET, the
data set allocated by PRINTDS will always be a normal BASIC data
set. This will not be large enough to hold output from a LARGE format
data set that is >64 K trks. An ABENDx37 can occur. If you preallocate
a TODATASET of large format, PRINTDS will be able to successfully
write to the large format TODATASET.

2.12 z/OS UNIX files
Following are the types of z/OS UNIX files:
Hierarchical File System (HFS)
Network File System (NFS)
zSeries File System (zFS)
Temporary File System (TFS)
Figure 2-12 z/OS UNIX files
z/OS UNIX
z/OS UNIX System Services (z/OS UNIX) enables z/OS to access UNIX files. UNIX
applications also can access z/OS data sets. z/OS UNIX files are byte-oriented, similar to
objects. We differentiate between the following types of z/OS UNIX files.
Hierarchical file system (HFS)

You can define on DASD an HFS data set on the z/OS system. Each HFS data set contains a
hierarchical file system. Each hierarchical file system is structured like a tree with subtrees,
and consists of directories and all their related files.
z/OS Network File System (z/OS NFS)

z/OS NFS is a distributed file system that enables users to access UNIX files and directories
that are located on remote computers as though they were local. z/OS NFS is independent of
machine types, operating systems, and network architectures.
zSeries File System (zFS)

A zFS is a UNIX file system that contains one or more file systems in a VSAM linear data set.
zFS is application compatible with HFS and more performance efficient than HFS.
Temporary file system (TFS)

A TFS is stored in memory and delivers high-speed I/O. A systems programmer can use a
TFS for storing temporary files.

2.13 Data set specifications for non-VSAM data sets
DSORG=PS
RECFM=FB
LRECL=80
Data Set
80
80 80
80 BLKSIZE=27920
80 80
DATASET.TEST.SEQ1
Figure 2-13 Data set specifications for non-VSAM data sets
Data set specifications for non-VSAM data set

A non-VSAM data set has several attributes that describe the data set. When you want to
define (create) a new data set, you have to specify those values to tell the system which kind
of data set you want to allocate.
See also z/OS MVS JCL Reference, SA22-7597 for information about the data set
specifications discussed in this section.
Data set organization (DSORG)

DSORG specifies the organization of the data set as physical sequential (PS), partitioned
(PO) for PDS or partitioned PDSE, or direct (DA).
Data set name type (DSNTYPE)

Use the DSNTYPE parameter to specify:
For a partitioned organized (PO) data set, whether it is a:
PDS for a partitioned data set
LIBRARY for a partitioned data set extended (PDSE)
Hierarchical file system if the DSNTYPE is HFS
Large data sets (see 2.16, Volume table of contents (VTOC) on page 45 for more details
about large data sets)

Logical records and blocks
To an application, a logical record is a unit of information (for example, a customer, an
account, a payroll employee, and so on). It is the smallest amount of data to be processed,
and it is comprised of fields that contain information recognized by the processing application.
Logical records, when located in DASD or tape, are grouped into physical records named
blocks (to save space in DASD because of the gaps). Each block of data on a DASD volume
has a distinct location and a unique address (block number, track, and cylinder), thus making
it possible to find any block without extensive sequential searching. Logical records can be
stored and retrieved either directly or sequentially.
DASD volumes are used for storing data and executable programs (including the operating
system itself), and for temporary working storage. One DASD volume can be used for many
separate data sets, and space on it can be reallocated and reused. The maximum length of a
logical record (LRECL) is limited by the physical size of the media used.
Record format (RECFM)

RECFM specifies the characteristics of the logical records in the data set. They can have a:
Fixed length (RECFM=F) - Every record is the same size.
Variable length (RECFM=V) - The logical records can be of varying sizes; every record
has a preceding record descriptor word (RDW) to describe the length of such record.
ASCII variable length (RECFM=D) - Used for ISO/ANSI tape data sets.
Undefined length (RECFM=U) - Permits processing of records that do not conform to the
F or V format.
F, V, or D-format logical records can be blocked (RECFM=FB, VB or DB), which means

several logical records are in the same block.
Spanned records are specified as VS, VBS, DS, or DBS. A spanned record is a logical record
that spans two or more blocks. Spanned records can be necessary if the logical record size is
larger than the maximum allowed block size.
You can also specify the records as fixed-length standard by using FS or FBS, meaning that
there is not an internal short block.
Logical record length (LRECL)

LRECL specifies the length, in bytes, of each record in the data set. If the records are of
variable length or undefined length, LRECL specifies the maximum record length. For input,
the field has no effect for undefined-length (format-U) records.
Block size (BLKSIZE)

BLKSIZE specifies the maximum length, in bytes, of the physical record (block). If the logical
records are format-F, the block size must be an integral multiple of the record length. If the
records are format-V, you must specify the maximum block size. If format-V records are
unblocked, the block size must be 4 bytes (to specify the block length) greater than the record
length (LRECL). For data sets in DASD the maximum block size is 32760. For data sets in
tapes the maximum block size is much larger.
In an extended-format data set, the system adds a 32-byte suffix to each block, which is
transparent to the application program.

System-determined block size
The system can derive the best block size to optimize DASD space, tape, and spooled data
sets (the ones in printing format but stored temporarily in DASD).
Space values
For DASD data sets, you can specify the amount of space required in: logical records, blocks,
records, tracks, or cylinders. You can specify a primary and a secondary space allocation.
When you define a new data set, only the primary allocation value is used to reserve space
for the data set on DASD. Later, when the primary allocation of space is filled, space is
allocated in secondary storage amounts (if specified). The extents can be allocated on other
volumes if the data set was defined as multivolume.
For example, if you allocate a new data set and specify SPACE=(TRK,(2,4)), this initially
allocates two tracks for the data set. As each record is written to the data set and these two
tracks are used up, the system automatically obtains four more tracks. When these four tracks
are used, another four tracks are obtained. The same sequence is followed until the extent
limit for the type of data set is reached.
The procedure for allocating space on magnetic tape devices is not like allocating space on
DASD. Because data sets on magnetic tape devices must be organized sequentially, each
one is located contiguously. All data sets that are stored on a given magnetic tape volume
must be recorded in the same density. See z/OS DFSMS Using Magnetic Tapes, SC26-7412
for information about magnetic tape volume labels and tape processing.

2.14 Locating an existing data set
TSO
MCAT
ALIAS: FPITA
ALIAS: VERA
UCAT
VOLDAT
UCAT
FPIT VTOC
A.DA
TA
FPITA.DATA
FPITA.FILE1
VERA.FILE1
FPITA.DATA
Figure 2-14 Locating an existing data set
Locating a data set

A catalog consists of two separate kinds of data sets: a basic catalog structure (BCS); and a
VSAM volume data set (VVDS). The BCS can be considered the catalog, whereas the VVDS
can be considered an extension of the volume table of contents (VTOC). Following are the
terms used to locate a data set that has been cataloged:
VTOC The volume table of contents is a sequential data set located in each
DASD volume that describes the data set contents of this volume. The
VTOC is used to find empty space for new allocations and to locate
non-VSAM data sets. For all VSAM data sets, and for SMS-managed
non-VSAM data sets, the VTOC is used to obtain information not kept in
the VVDS. See 2.18, VTOC index structure on page 48.
User catalog A catalog is a data set used to locate in which DASD volume the
requested data set is stored; user application data sets are cataloged in
this type of catalog.
Master catalog This has the same structure as a user catalog, but points to system
(z/OS) data sets. It also contains information about the user catalog
location and any alias pointer.
Alias A special entry in the master catalog pointing to a user catalog that
coincides with the HLQ of a data set name. The alias is used to find in
which user catalog the data set location information exists. It means that
the data set with this HLQ is cataloged in that user catalog.

The sequence for locating an existing data set
The MVS system provides a service called LOCATE to read entries in a catalog. When z/OS
tries to locate an existing data set, the following sequence takes place.
When the master catalog is examined:
If it has the searched data set name, the volume information is picked up and the
volume VTOC is used to locate the data set in the specified volume.
If the HLQ is a defined alias in the master catalog, the user catalog is searched. If the
data set name is found, processing proceeds as in a master catalog find.
Finally, the requesting program can access the data set. As you can imagine, it is
impossible to keep track of the location of millions of data sets without the catalog concept.
For detailed information about catalogs refer to Chapter 6, Catalogs on page 325.

2.15 Uncataloged and cataloged data sets
JCL references to data sets in JCL

Uncataloged reference
// DD DSN=PAY.D1,DISP=OLD,UNIT=3390,VOL=SER=MYVOL1
MYVOL1
PAY.D1
CATALOG
Cataloged reference
// DD DSN=PAY.D2,DISP=OLD
PAY.D2
Figure 2-15 Cataloged and uncataloged data sets
Cataloged data sets

When an existing data set is cataloged, z/OS obtains unit and volume information from the
catalog using the LOCATE macro service. However, if the DD statement for a catalog data set
contains VOLUME=SER=serial-number, the system does not look in the catalog; in this case,
you must code the UNIT parameter and volume information.
Uncataloged data sets

When your existing data set is not cataloged, you must know in advance its volume location
and specify it in your JCL. This can be done through the UNIT and VOL=SER, as shown in
Figure 2-15.
See z/OS MVS JCL Reference, SA22-7597 for information about UNIT and VOL parameters.
Note: We strongly recommend that you do not have uncataloged data sets in your
installation because uncataloged data sets can cause problems with duplicate data and
possible incorrect data set processing.

2.16 Volume table of contents (VTOC)
VTOC
A
B C
A
Data sets
B
VTOC Data Set
C (Can be located
after cylinder 0,
track 0.)
Figure 2-16 Volume table of contents (VTOC)
Volume table of contents (VTOC)

The VTOC is a data set that describes the contents of the direct access volume on which it
resides. It is a contiguous data set; that is, it resides in a single extent on the volume and
starts after cylinder 0, track 0, and before track 65,535. A VTOC's address is located in the
VOLVTOC field of the standard volume label. The volume label is described in z/OS DFSMS
Using Data Sets. A VTOC consists of complete tracks.
The VTOC lists the data sets that reside on its volume, along with information about the
location and size of each data set, and other data set attributes. It is created when the volume
is initialized through the ICKDSF utility program.
The VTOC locates data sets on that volume. The VTOC is composed of 140-byte data set
control blocks (DSCBs), of which there are six types shown in Table 2-1 on page 47, that
correspond either to a data set currently residing on the volume, or to contiguous, unassigned
tracks on the volume. A set of assembler macros is used to allow a program or z/OS to
access VTOC information.
IEHLIST utility
The IEHLIST utility can be used to list, partially or completely, entries in a specified volume
table of contents (VTOC), whether indexed or non-indexed. The program lists the contents of
selected data set control blocks (DSCBs) in edited or unedited form.

2.17 VTOC and DSCBs
VTOC
DATA SET A
DSCBs F4 F0 F1 F1 F1
DATA SET B
DATA SET C
Figure 2-17 Data set control block (DSCB)
Data set control block (DSCB)

The VTOC is composed of 140-byte data set control blocks (DSCBs) that point to data sets
currently residing on the volume, or to contiguous, unassigned (free) tracks on the volume
(depending on the DSCB type).
DSCBs also describe the VTOC itself. CVAF routines automatically construct a DSCB when
space is requested for a data set on the volume. Each data set on a DASD volume has one or
more DSCBs (depending on its number of extents) describing space allocation and other
control information such as operating system data, device-dependent information, and data
set characteristics. There are seven kinds of DSCBs, each with a different purpose and a
different format number.
The first record in every VTOC is the VTOC DSCB (format-4). The record describes the
device, the volume the data set resides on, the volume attributes, and the size and contents of
the VTOC data set itself. The next DSCB in the VTOC data set is a free-space DSCB
(format-5) that describes the unassigned (free) space in the full volume. The function of
various DSCBs depends on whether an optional Index VTOC is allocated in the volume. Index
VTOC is a sort of B-tree, to make the search in VTOC faster.
Table 2-1 on page 47 describes the various types of DSCBs, taking into consideration
whether the Index VTOC is in place or not.
In z/OS V1R7 there is a new address space (DEVMAN) containing trace information about
CVAF events.

Table 2-1 DSCBs that can be found in the VTOC
Type Name Function How many
0 Free VTOC Describes unused DSCB records One for every unused 140-byte record
DSCB in the VTOC (contains 140 bytes in the VTOC. The DS4DSREC field of
of binary zeros). To delete a the format-4 DSCB is a count of the
DSCB from the VTOC, a format-0 number of format-0 DSCBs in the
DSCB is written over it. VTOC. This field is not maintained for an
indexed VTOC.
1 Identifier Describes the first three extents One for every data set or data space on
of a data set or VSAM data space. the volume, except the VTOC.
2 Index Describes the indexes of an ISAM One for each ISAM data set (for a
data set. This data set multivolume ISAM data set, a format-2
organization is old, and is not DSCB exists only on the first volume).
supported anymore.
3 Extension Describes extents after the third One for each data set on the volume
extent of a non-VSAM data set or that has more than three extents. There
a VSAM data space. can be as many as 10 for a PDSE, HFS,
extended format data set, or a VSAM
data set component cataloged in an
integrated catalog facility catalog.
PDSEs, HFS, and extended format data
sets can have up to 123 extents per
volume. All other data sets are
restricted to 16 extents per volume. A
VSAM component can have 7257
extents in up to 59 volumes (123 each).
4 VTOC Describes the extent and One on each volume.

contents of the VTOC, and
provides volume and device
characteristics. This DSCB
contains a flag indicating whether
the volume is SMS-managed.
5 Free space On a non-indexed VTOC, One for every 26 non-contagious

describes the space on a volume extents of available space on the
that has not been allocated to a volume for a non- indexed VTOC; for an
data set (available space). For an indexed VTOC, there is only one.
indexed VTOC, a single empty
format-5 DSCB resides in the
VTOC; free space is described in
the index and DS4IVTOC is
normally on.
7 Free space Only one field in the format-7 This DSCB is not used frequently.
for certain DSCB is an intended interface.
device This field indicates whether the
DSCB is a format-7 DSCB. You
can reference that field as
DS1FMTID or DS5FMTID. A
character 7 indicates that the
DSCB is a format-7 DSCB, and
your program is not to modify it.

2.18 VTOC index structure
VOLUME FREE SPACE MAP

LABEL POINTERS TO VTOC
DATA SET ENTRIES
FREE SPACE
LIST OF EMPTY
VTOC INDEX VTOC ENTRIES
VTOC
VVDS
DATA
FREE SPACE
Figure 2-18 VTOC index structure
VTOC index
The VTOC index enhances the performance of VTOC access. The VTOC index is a
physical-sequential data set on the same volume as the related VTOC, created by the
ICKDSF utility program. It consists of an index of data set names in format-1 DSCBs
contained in the VTOC and volume free space information.
Important: An SMS-managed volume requires an indexed VTOC; otherwise, the VTOC

index is highly recommended. For additional information about SMS-managed volumes,
see z/OS DFSMS Implementing System-Managed Storage, SC26-7407.
If the system detects a logical or physical error in a VTOC index, the system disables further
access to the index from all systems that might be sharing the volume. Then, the VTOC
remains usable but with possibly degraded performance.
If a VTOC index becomes disabled, you can rebuild the index without taking the volume offline
to any system. All systems can continue to use that volume without interruption to other
applications, except for a brief pause during the index rebuild. After the system rebuilds the
VTOC index, it automatically re-enables the index on each system that has access to it.
Next, we see more details about the internal implementation of the Index VTOC.

Creating the VTOC and VTOC index
To initialize a volume (prepare for I/O activity), use the Device Support Facilities (ICKDSF)
utility to initially build the VTOC. You can create a VTOC index at that time by using the
ICKDSF INIT command and specifying the INDEX keyword.
You can use ICKDSF to convert a non-indexed VTOC to an indexed VTOC by using the
BUILDIX command and specifying the IXVTOC keyword. The reverse operation can be
performed by using the BUILDIX command and specifying the OSVTOC keyword. For details,
see Device Support Facilities Users Guide and Reference Release 17, GC35-0033, and
z/OS DFSMSdfp Advanced Services, SC26-7400, for more information about that topic.
VTOC index format-1 DSCB

A format-1 DSCB in the VTOC contains the name and extent information of the VTOC index.
The name of the index must be 'SYS1.VTOCIX.xxxxxxxx', where xxxxxxxx conforms to
standard data set naming conventions and is usually the serial number of the volume
containing the VTOC and its index. The name must be unique within the system to avoid ENQ
contention.
VTOC index record (VIR)

Device Support Facilities (ICKDSF) initializes a VTOC index into 2048-byte physical blocks
named VTOC index records (VIRs). VIRs are used in several ways. A VTOC index contains
the following kinds of VIRs:
VTOC index entry record (VIER) identifies the location of format-1 DSCBs and the
format-4 DSCB.
VTOC pack space map (VPSM) identifies the free and allocated space on a volume.
VTOC index map (VIXM) identifies the VIRs that have been allocated in the VTOC index.
VTOC map of DSCBs (VMDS) identifies the DSCBs that have been allocated in the
VTOC.

2.19 Initializing a volume using ICKDSF
//EXAMPLE JOB
//EXEC PGM=ICKDSF
//SYSPRINT DD SYSOUT=A
//SYSIN DD *
INIT UNITADDRESS(0353) NOVERIFY -
VOLID(VOL123)
/*
Figure 2-19 Initializing a volume
Device Support Facilities (ICKDSF)

ICKDSF is a program you can use to perform functions needed for the initialization,
installation, use, and maintenance of DASD volumes. You can also use it to perform service
functions, error detection, and media maintenance. However, due to the virtualization of the
volumes there is no need for executing media maintenance.
Initializing a DASD volume

After you have completed the installation of a device, you must initialize and format the
volume so that it can be used by MVS.
You use the INIT command to initialize volumes. The INIT command writes a volume label (on
cylinder 0, track 0) and a VTOC on the device for use by MVS. It reserves and formats tracks
for the VTOC at the location specified by the user and for the number of tracks specified. If no
location is specified, tracks are reserved at the default location.
If the volume is SMS-managed, the STORAGEGROUP option must be declared, in order to

keep such information (SMS managed) in a format-4 DSCB.
Initializing a volume for the first time in offline mode

In the example in Figure 2-19, a volume is initialized at the minimal level because neither the
CHECK nor VALIDATE parameter is specified (as recommended). Because the volume is
being initialized for the first time, it must be mounted offline (to avoid MVS data set
allocations), and the volume serial number must be specified. Because the VTOC parameter

is not specified, the default VTOC size is the number of tracks in a cylinder minus one. For a
3390, the default location is cylinder 0, track 1 for 14 tracks.
Initializing a volume to be managed in a DFSMS environment

In the following example, a volume that is to be system-managed is initialized. The volume is
initialized in offline mode at the minimal level. The VTOC is placed at cylinder 2, track 1 and
occupies ten tracks. The VTOC is followed by the VTOC index. The STORAGEGROUP
parameter indicates the volume is to be managed in a DFSMS environment.
INIT UNIT(0353) NOVERIFY STORAGEGROUP -
OWNERID(PAYROLL) VTOC(2,1,10) INDEX(2,11,5)
The following example performs an online minimal initialization, and as a result of the
command, an index to the VTOC is created.
// JOB
// EXEC PGM=ICKDSF
//XYZ987 DD UNIT=3390,DISP=OLD,VOL=SER=PAY456
//SYSIN DD *
INIT DDNAME(XYZ987) NOVERIFY INDEX(X'A',X'B',X'2')
/*
ICKDSF stand-alone version

You can run the stand-alone version of ICKDSF under any IBM System z processor. To run
the stand-alone version of ICKDSF, you IPL ICKDSF with a stand-alone IPL tape that you
create under z/OS. This function allows you to initialize volumes without the need for running
an operating system such as z/OS.
Creating an ICKDSF stand-alone IPL tape using z/OS

For z/OS, the stand-alone code is in SYS1.SAMPLIB as ICKSADSF. You can load the
ICKDSF program from a file on tape. The following example can be used to copy the
stand-alone program to an unlabeled tape.
//JOBNAME JOB JOB CARD PARAMETERS

//STEPNAME EXEC PGM=IEBGENER
//SYSIN DD DUMMY,DCB=BLKSIZE=80
//SYSUT1 DD DSNAME=SYS1.SAMPLIB(ICKSADSF),UNIT=SYSDA,
// DISP=SHR,VOLUME=SER=XXXXXX
//SYSUT2 DD DSNAME=ICKDSF,UNIT=3480,LABEL=(,NL),
// DISP=(,KEEP),VOLUME=SER=YYYYYY,
// DCB=(RECFM=F,LRECL=80,BLKSIZE=80)
For details on how to IPL the stand-alone version and to see examples of the commands,
refer to Device Support Facilities Users Guide and Reference Release 17, GC35-0033.

3
Chapter 3. Extended access volumes

z/OS V1R10 introduced extended address volume (EAV), which allowed DASD storage
volumes to be larger than 65,520 cylinders. The space above the first 65,520 cylinders is
referred to as cylinder-managed space. Tracks in cylinder-managed space use extended
addressing space (EAS) techniques to access these tracks. Data sets that can use
cylinder-managed space are referred to as being EAS-eligible.
With z/OS V1R10, only VSAM data sets are EAS-eligible. You can control whether VSAM
data sets can reside in cylinder-managed space by including or excluding EAVs in particular
storage groups. For non-SMS managed data sets, control the allocation to a volume by
specifying a specific VOLSER or esoteric name.
With z/OS V1R11, extended-format sequential data sets are now EAS-eligible. You can
control whether the allocation of EAS-eligible data sets can reside in cylinder-managed space
using both the methods supported in z/OS V1R10 and by using the new EATTR data set
attribute keyword.

3.1 Traditional DASD capacity
D/T3380
2655 Cyl
1770 Cyl
885 Cyl
Model J Model E Model K
D/T3390
10017 Cyl
2226 Cyl 3339 Cyl
1113 Cyl
Model 1 Model 2 Model 3 Model 9

Figure 3-1 Traditional DASD capacity
DASD capacity
Figure 3-1 shows various DASD device types. 3380 devices were used in the 1980s. Capacity
went from 885 to 2,655 cylinders per volume. When storage density increased, new device
types were introduced at the end of the 1980s. Those types were called 3390. Capacity per
volume ranged from 1,113 to 3,339 cylinders. A special device type, model 3390-9 was
introduced to store large amounts of data that needed very fast access. The track geometry
within one device category was (and is) always the same; this means that 3380 volumes have
47,476 bytes per track, and 3390 volumes have 56,664 bytes per track.
Table 3-1 lists further information about DASD capacity.
Table 3-1 DASD capacity

Physical 3380-J 3380-E 3380-K 3390-1 3390-2 3390-3 3390-9
characteristics
Data Cyl/Device 855 1770 2655 1113 2226 3339 10017
Track/Cyl 15 15 15 15 15 15 15
Bytes/Trk 47476 47476 47476 56664 56664 56664 56664
Bytes/Cylinder 712140 712140 712140 849960 849960 849960 849960
MB/Device 630 1260 1890 946 1892 2838 8514

3.2 Large volumes before z/OS V1R10
A "large volume" is larger than a 3390-9
The largest possible volume has 65520 (3390) cylinders
That would be a "3390-54" if it had its own device type
Almost 54 GB
3390-27 3390-54
32760 Cyls 65520 Cyls
Figure 3-2 Large volume support 3390-54
Large volume support

The IBM TotalStorage Enterprise Storage Server (ESS) initially supported custom volumes of
up to 10017 cylinders, the size of the largest standard volume, the 3390 model 9. This was
the limit set by the operating system software. The IBM TotalStorage ESS large volume
support enhancement, announced in November 2001, has now increased the upper limit to
65520 cylinders, approximately 54 GB. The enhancement is provided as a combination of
IBM TotalStorage ESS licensed internal code (LIC) changes and system software changes,
available for z/OS and z/VM.
Today, for example, the IBM Enterprise Storage Server emulates the IBM 3390. On an
emulated disk or on a VM minidisk, the number of cylinders per volume is a configuration
option. It might be less than or greater than the stated number. If so, the number of bytes per
device will differ accordingly. The IBM ESS Model 1750 supports up to 32760 cylinders and
the IBM ESS Model 2107 supports up to 65520 cylinders.
Large volume support is available on z/OS operating systems, the ICKDSF, and DFSORT
utilities.
Large volume support must be installed on all systems in a sysplex prior to sharing data sets
on large volumes. Shared system and application data sets cannot be placed on large
volumes until all system images in a sysplex have large volume support installed.
Chapter 3. Extended access volumes 55

Large volume support design considerations
Benefits of using large volumes can be briefly summarized as follows:
They reduce storage management tasks by allowing you to define and manage smaller
configurations.
They reduce the number of multivolume data sets you have to manage.
They relieve architectural constraints by allowing you to address more data within the
existing 64K subchannel number limit.
The size of the logical volume defined does not have an impact on the performance of the
ESS subsystem. The ESS does not serialize I/O on the basis of logical devices, so an
increase in the logical volume size does not affect the ESS backend performance. Host
operating systems, on the other hand, serialize I/Os against devices. As more data sets
reside on a single volume, there will be greater I/O contention accessing the device. With
large volume support, it is more important than ever to try to minimize contention on the
logical device level. To avoid potential I/O bottlenecks on devices:
Exploit the use of Parallel Access Volumes to reduce IOS queuing on the system level.
Eliminate unnecessary reserves by using WLM in goal mode.
Multiple allegiance will automatically reduce queuing on sharing systems.
Parallel Access Volume (PAV) support is of key importance when implementing large
volumes. PAV enables one MVS system to initiate multiple I/Os to a device concurrently. This
keeps IOSQ times down and performance up even with many active data sets on the same
volume. PAV is a practical must with large volumes. We discourage you from using large
volumes without PAV. In particular, we recommend the use of dynamic PAV and HyperPAV.
As the volume sizes grow larger, more data and data sets will reside on a single S/390 device
address. Thus, the larger the volume, the greater the multi-system performance impact will be
of serializing volumes with RESERVE processing. You need to exploit a GRS Star
configuration and convert all RESERVE's possible into system ENQ requests.

3.3 zArchitecture data scalability
Serialization
Granularity
Multiple allegiance .......................................

reserve
release
Parallel Access Volume (PAV)
HyperPAV
Alias UCBs
Access
visibility
Base UCB Dynamic volume
expansion
3390-9
3390-9
Physical
3390-3 3390-9
volume "3390-54"
3 GB 9 GB 27 GB 54 GB
max cyls: 3339 max cyls: 10017 max cyls: 32760 max cyls: 65520
years
Figure 3-3 zArchitecture data scalability
DASD architecture
In the past decade, as processing power has dramatically increased, great care and
appropriate solutions have been deployed so that the amount of data that is directly
accessible can be kept proportionally equivalent. Over the years DASD volumes have
increased in size by increasing the number of cylinders and thus GB capacity.
However, the existing track addressing architecture has limited growth to relatively small GB
capacity volumes. This has placed increasing strain on the 4-digit device number limit and the
number of UCBs that can be defined. The largest available volume is one with 65,520
cylinders or approximately 54 GB, as shown in Figure 3-3.
Rapid data growth on the z/OS platform is leading to a critical problem for various clients, with
a 37% compound rate of disk storage growth between 1996 and 2007. The result is that this
is becoming a real constraint on growing data on z/OS. Business resilience solutions
(GDPS, HyperSwap, and PPRC) that provide continuous availability are also driving this
constraint.
Serialization granularity
Since the 1960s, shared DASD can be serialized through a sequence of
RESERVE/RELEASE CCWs that are today under the control of GRS, as shown in Figure 3-3.
This was a useful mechanism as long as the volume of data so serialized (the granularity)
was not too great. But whenever such a device grew to contain too much data, bottlenecks
became an issue.

DASD virtual visibility
Traditional S/390 architecture does not allow more than one I/O operation to the same S/390
device because such devices can only handle, physically, one I/O operation at a time.
However, in modern DASD subsystems such as ESS, DS6000, and DS8000, the device
(such as a 3390) is only a logical view. The contents of this logical device are spread in HDA
RAID arrays and in caches. Therefore, it is technically possible to have more than one I/O
operation towards the same logical device. Changes have been made in z/OS (in IOS code),
in the channel subsystem (SAP), and in ESS, DS6000, and DS8000 to allow more than one
I/O operation on the same logical device. This is called parallel I/O, and it is available in two
types:
Multiple allegiance
Parallel access volume (PAV)
This relief builds upon prior technologies that were implemented in part to help reduce the
pressure on running out of device numbers. These include PAV and HyperPAV. PAV alias
UCBs can be placed in an alternate subchannel set (z9 multiple subchannel support).
HyperPAV reduces the number of alias UCBs over traditional PAVs and provides the I/O
throughput required.
Multiple allegiance
Multiple allegiance (MA) was introduced to alleviate the following constraint. It allows
serialization on a limited amount of data within a given DASD volume, which leads to the
possibility of having several (non-overlapping) serializations held at the same time on the
same DASD volume. This is a useful mechanism on which any extension of the DASD volume
addressing scheme can rely. In other terms, multiple allegiance provides finer (than
RESERVE/RELEASE) granularity for serializing data on a volume. It gives the capability to
support I/O requests from multiple systems, one per system, to be concurrently active against
the same logical volume, if they do not conflict with each other. Conflicts occur when two or
more I/O requests require access to overlapping extents (an extent is a contiguous range of
tracks) on the volume, and at least one of the I/O requests involves writing of data.
Requests involving writing of data can execute concurrently with other requests as long as
they operate on non-overlapping extents on the volume. Conflicting requests are internally
queued in the DS8000. Read requests can always execute concurrently regardless of their
extents. Without the MA capability, DS8000 generates a busy indication for the volume
whenever one of the systems issues a request against the volume, thereby causing the I/O
requests to be queued within the channel subsystem (CSS). However, this concurrency can
be achieved as long as no data accessed by one channel program can be altered through the
actions of another channel program.
Parallel access volume

Parallel access volume (PAV) allows concurrent I/Os to originate from the same z/OS image.
Using PAV can provide significant performance enhancements in IBM System z environments
by enabling simultaneous processing for multiple I/O operations to the same logical volume.
PAV was introduced in z/OS V1R6 as a new option that allows to specify PAV capability as
one of the volume selection criteria for SMS-managed data sets assigned to a storage class.
HyperPAV feature
With the IBM System Storage DS8000 Turbo model and the IBM server synergy feature,
the HyperPAV together with PAV, multiple allegiance can dramatically improve performance
and efficiency for System z environments. With HyperPAV technology:
z/OS uses a pool of UCB aliases.

As each application I/O is requested, if the base volume is busy with another I/O:
z/OS selects a free alias from the pool, quickly binds the alias device to the base
device, and starts the I/O.
When the I/O completes, the alias device is used for another I/O on the LSS or is
returned to the free alias pool.
If too many I/Os are started simultaneously:

z/OS queues the I/Os at the LSS level.
When an exposure frees up that can be used for queued I/Os, they are started.
Queued I/O is done within assigned I/O priority.
For each z/OS image within the sysplex, aliases are used independently. WLM is not involved
in alias movement so it does not need to collect information to manage HyperPAV aliases.
Benefits of HyperPAV
HyperPAV has been designed to provide an even more efficient parallel access volume (PAV)
function. When implementing larger volumes, it provides a way to scale I/O rates without the
need for additional PAV alias definitions. HyperPAV exploits FICON architecture to reduce
overhead, improve addressing efficiencies, and provide storage capacity and performance
improvements, as follows:
More dynamic assignment of PAV aliases improves efficiency.
The number of PAV aliases needed might be reduced, taking fewer from the 64 K device
limitation and leaving more storage for capacity use.

3.4 WLM controlling PAVs
WLMs exchange performance information

Goals not met because of IOS queue?
Who can donate an alias?
zSeries zSeries zSeries zSeries

WLM WLM WLM WLM
IOSQ on 100? IOSQ on 100? IOSQ on 100? IOSQ on 100?
Base Alias Alias Alias Alias Base

100 to 100 to 100 to 110 to 110 110
Dynamic PAVs Dynamic PAVs

IBM DS8000
Figure 3-4 WLM controlling PAVs
Workload Manager (WLM)

In the zSeries Parallel Sysplex environments, the z/OS Workload Manager (WLM) controls
where work is run and optimizes the throughput and performance of the total system. The
ESS provides the WLM with more sophisticated ways to control the I/O across the sysplex.
These functions include parallel access to both single-system and shared volumes, and the
ability to prioritize the I/O based upon WLM goals. The combination of these features
significantly improves performance in a wide variety of workload environments.
Parallel Access Volume (PAV)

Parallel Access Volume is one of the original features that the IBM TotalStorage Enterprise
Storage Server brings specifically for z/OS operating systems, helping the zSeries running
applications to concurrently share the same logical volumes.
The ability to do multiple I/O requests to the same volume nearly eliminates IOS queue time
(IOSQ), one of the major components in z/OS response time. Traditionally, access to highly
active volumes has involved manual tuning, splitting data across multiple volumes, and more.
With PAV and the Workload Manager, you can almost forget about manual performance
tuning. WLM manages PAVs across all members of a sysplex, too. The ESS, in conjunction
with z/OS, has the ability to meet the performance requirements on its own.
WLM IOS queuing

As part of the Enterprise Storage Subsystem's implementation of parallel access volumes,
the concept of base addresses versus alias addresses is introduced. While the base address

is the actual unit address of a given volume, there can be many alias addresses assigned to a
base address, and any or all of those alias addresses can be reassigned to a separate base
address. With dynamic alias management, WLM can automatically perform those alias
address reassignments to help work meet its goals and to minimize IOS queuing.
When you specify a yes value on the Service Coefficient/Service Definition Options panel,
you enable dynamic alias management globally throughout the sysplex. WLM will keep track
of the devices used by separate workloads and broadcast this information to other systems in
the sysplex. If WLM determines that a workload is not meeting its goal due to IOS queue time,
then WLM attempts to find alias devices that can be moved to help that workload achieve its
goal. Even if all work is meeting its goals, WLM will attempt to move aliases to the busiest
devices to minimize overall queuing.
Alias assignment
It is not always easy to predict the volumes to have an alias address assigned, and how many.
Your software can automatically manage the aliases according to your goals. z/OS can exploit
automatic PAV tuning if you are using WLM in goal mode. z/OS recognizes the aliases that
are initially assigned to a base during the Nucleus Initialization Program (NIP) phase. WLM
can dynamically tune the assignment of alias addresses. WLM monitors the device
performance and is able to dynamically reassign alias addresses from one base to another if
predefined goals for a workload are not met. WLM instructs IOS to reassign an alias.
WLM goal mode management in a sysplex

WLM keeps track of the devices utilized by the various workloads, accumulates this
information over time, and broadcasts it to the other systems in the same sysplex. If WLM
determines that any workload is not meeting its goal due to IOSQ time, WLM attempts to find
an alias device that can be reallocated to help this workload achieve its goal.
Through WLM, there are two mechanisms to tune the alias assignment:
The first mechanism is goal based. This logic attempts to give additional aliases to a PAV
device that is experiencing IOS queue delays and is impacting a service class period that
is missing its goal. To give additional aliases to the receiver device, a donor device must
be found with a less important service class period. A bitmap is maintained with each PAV
device that indicates the service classes using the device.
The second mechanism is to move aliases to high-contention PAV devices from
low-contention PAV devices. High-contention devices will be identified by having a
significant amount of IOS queue. This tuning is based on efficiency rather than directly
helping a workload to meet its goal.

3.5 Parallel Access Volumes (PAVs)
Applications z/OS Image

do I/O to base UCB 08F1
volumes
UCB 08F0 Storage Server
UCB 0801
Applications UCB 08F3

volumes UCB 0802
Logical Subsystem (LSS) 0800

Alias UA=F1
Applications Alias UA=F0
do I/O to base UCB 08F1 Base UA=01
volumes
UCB 08F0
Alias UA=F3
UCB 0801 Alias UA=F2
Base UA=02

volumes
UCB 0802
z/OS Image
Figure 3-5 Parallel access volumes (PAVs)
Parallel access volume (PAV)

z/OS system IOS maps a device in a unit control block (UCB). Traditionally this I/O device
does not support concurrency, being treated as a single resource, serially used. High I/O
activity towards the same device can adversely affect performance. This contention is worst
for large volumes with many small data sets. The symptom displayed is extended IOSQ time,
where the I/O request is queued in the UCB. z/OS cannot attempt to start more than one I/O
operation at a time to the device.
The ESS and DS8000 support concurrent data transfer operations to or from the same
3390/3380 devices from the same system. A device (volume) accessed in this way is called a
parallel access volume (PAV).
PAV exploitation requires both software enablement and an optional feature on your
controller. PAV support must be installed on each controller. It enables the issuing of multiple
channel programs to a volume from a single system, and allows simultaneous access to the
logical volume by multiple users or jobs. Reads, as well as writes to other extents, can be
satisfied simultaneously. The domain of an I/O consists of the specified extents to which the
I/O operation applies, which corresponds to the extents of the same data set. Writes to the
same domain still have to be serialized to maintain data integrity, which is also the case for
reads and write.
The implementation of N parallel I/Os to the same 3390/3380 device consumes N addresses
in the logical controller, thus decreasing the number of possible real devices. Also, UCBs are

not prepared to allow multiples I/Os due to software product compatibility issues. Support is
then implemented by defining multiple UCBs for the same device.
The UCBs are of two types:

Base address: This is the actual unit address. There is only one for any volume.
Alias address: Alias addresses are mapped back to a base device address. I/O scheduled
for an alias is executed against the base by the controller. No physical disk space is
associated with an alias address. Alias UCBs are stored above the 16 MB line.
PAV benefits
Workloads that are most likely to benefit from PAV functionality being available include:
Volumes with many concurrently open data sets, such as volumes in a work pool
Volumes that have a high read to write ratio per extent
Volumes reporting high IOSQ times
Candidate data sets types:

Have high read to write ratio
Have many extents on one volume
Are concurrently shared by many readers
Are accessed using media manager or VSAM-extended format (32-byte suffix)
PAVs can be assigned to base UCBs either:

Manually (static) by the installation.
Dynamically, WLM can move aliases UCBs from one base UCB to another base UCB in
order to:
Balance device utilizations
Honor the goal of transactions suffering I/O delays because long IOSQ time. All WLMs
in a sysplex must agree with the movement of aliases UCBs.
However, the dynamic PAV still has issues:

Any change must be decided by all WLMs in a sysplex using XCF communication.
The number of aliases for one device must be equal in all z/OS systems.
Any change implies a dynamic I/O configuration.
To resolve such issues HyperPAV was introduced. With HyperPAV all aliases UCBs are
located in a pool and are used dynamically by IOS.

3.6 HyperPAV feature for DS8000 series
HyperPAV
Reduces the number of PAV-aliases needed per
logical subsystem (LSS)
By an order of magnitude but still maintaining optimal
response times
This is accomplished by no longer statically binding
PAV-aliases to PAV-bases
WLM no longer adjusts the bindings
In HyperPAV mode, PAV-aliases are bound to
PAV-bases only for the duration of a single I/O
operation, then reducing the number of aliases required
per LSS significantly
Figure 3-6 HyperPAV implementation
DS8000 feature
HyperPAV is an optional feature on the DS8000 series, available with the HyperPAV indicator
feature number 0782 and corresponding DS8000 series function authorization (2244-PAV
HyperPAV feature number 7899). HyperPAV also requires the purchase of one or more PAV
licensed features and the FICON/ESCON Attachment licensed feature. The FICON/ESCON
Attachment licensed feature applies only to the DS8000 Turbo Models 931, 932, and 9B2.
HyperPAV allows many DS8000 series users to benefit from enhancements to PAV with
support for HyperPAV.
HyperPAV allows an alias address to be used to access any base on the same control unit
image per I/O base. This capability also allows separate HyperPAV hosts to use one alias to
access separate bases, which reduces the number of alias addresses required to support a
set of bases in a System z environment with no latency in targeting an alias to a base. This
functionality is also designed to enable applications to achieve equal or better performance
than is possible with the original PAV feature alone, while also using the same or fewer z/OS
resources. The HyperPAV capability is offered on z/OS V1R6 and later.

3.7 HyperPAV implementation

do I/O to base z/OS Image P
volumes UCB 08F2

O Storage Server
O
L
UCB 0801 UCB 08F1
Applications
do I/O to base UCB 08F0 Logical Subsystem (LSS) 0800
volumes UCB 0802
Alias UA=F0
Alias UA=F1
Alias UA=F2
Alias UA=F3
Applications z/OS Image
do I/O to base Base UA=01
volumes
UCB 08F0
UCB 0801
Base UA=02
UCB 08F1
Applications
do I/O to base UCB 08F3 P
volumes O
UCB 0802
O
UCB 08F2
L
Figure 3-7 HyperPAV implementation using a pool of aliases
HyperPAV feature
With the IBM System Storage DS8000 Turbo model and the IBM server synergy feature, the
HyperPAV together with PAV, Multiple Allegiance, and support for IBM System z MIDAW
facility can dramatically improve performance and efficiency for System z environments.
With HyperPAV technology:

z/OS uses a pool of UCB aliases.
As each application I/O is requested, if the base volume is busy with another I/O:
z/OS selects a free alias from the pool, quickly binds the alias device to the base
device, and starts the I/O.
When the I/O completes, the alias device is used for another I/O on the LSS or is
returned to the free alias pool.
If too many I/Os are started simultaneously:

z/OS will queue the I/Os at the LSS level.
When an exposure frees up that can be used for queued I/Os, they are started.
Queued I/O is done within assigned I/O priority.

For each z/OS image within the sysplex, aliases are used independently. WLM is not involved
in alias movement. Therefore, it does not need to collect information to manage HyperPAV
aliases.
Note: HyperPAV was introduced and integrated in z/OS V1R9 and is available in z/OS
V1R8 with APAR OA12865.

3.8 Device type 3390 and 3390 Model A
An EAV is a volume with more than EAV

65,520 cylinders
EAV volumes increase the amount of
addressable DASD storage per
volume beyond 65,520 cylinders
32760 Cyl 65520 Cyl

10017 Cyl
3339 Cyl
Model 3 Model 9 Model 9 Model 9 Model A

3 GB 9 GB 27 GB 54 GB 100's of TB
Figure 3-8 EAV volumes
EAV volumes
An extended address volume (EAV) is a volume with more than 65,520 cylinders. An EAV
increases the amount of addressable DASD storage per volume beyond 65,520 cylinders by
changing how tracks on volumes are addressed. The extended address volume is the next
step in providing larger volumes for z/OS. z/OS provided this support first in z/OS V1R10 of
the operating system. Over the years, volumes have grown by increasing the number of
cylinders and thus GB capacity. However, the existing track addressing architecture has
limited the required growth to relatively small GB capacity volumes, which has put pressure
on the 4-digit device number limit. Previously, the largest available volume is one with 65,520
cylinders or approximately 54 GB. Access to the volumes includes the use of PAV, HyperPAV,
and FlashCopy SE (Space-efficient FlashCopy).
With EAV volumes, an architecture is implemented that provides a capacity of hundreds of

terabytes for a single volume. However, the first releases are limited to a volume with 223 GB
or up to 262,668 cylinders.
3390 Model A
A volume of this size has to be configured in the DS8000 as a 3390 Model A. However, a
3390 Model A is not always an EAV. A 3390 Model A is any device configured in the DS8000
to have more than 65,220 cylinders. Figure 3-8 illustrates the 3390 device types.

Note: With the 3390 Model A, the model A refers to the model configured in the DS8000. It
has no association with the 3390A notation in HCD that indicates a PAV-alias UCB in the
z/OS operating system. The Model A was chosen so that it did not imply a particular
device size as previous models 3390-3 and 3390-9 did.

3.9 Extended access volumes (EAV)
Increased z/OS addressable disk storage

Provide constraint relief for applications using large
VSAM data sets - extended-format data sets
3390 Model A: Device can be configured to have from
1 to 268,434,453 cylinders - (architectural maximum)
With z/OS V1R10 - Size limited to 223 GB - 262,668
(Max cylinders)
Managed by the system as a general purpose volume
Works well for applications with large files
PAV and HyperPAV technologies help by allowing I/O
rates to scale as a volume gets larger
Figure 3-9 EAV volumes with z/OS
EAV benefit
The benefit of this support is that the amount of z/OS addressable disk storage is further
significantly increased. This provides relief for customers that are approaching the 4-digit
device number limit by providing constraint relief for applications using large VSAM data sets,
such as those used by DB2, CICS, zFS file systems, SMP/E CSI data sets, and NFS
mounted data sets. This support is provided in z/OS V1R10 and enhanced in z/OS V1R11.
EAV eligible data sets

VSAM KSDS, RRDS, ESDS and linear data sets, both SMS-managed and
non-SMS-managed, and zFS data sets (which are VSAM anyway) are supported beginning
with z/OS V1R10.
Extended format sequential data sets are supported in z/OS V1R11. PDSE, basic and large
sequential, BDAM and PDS data set support is provided for all data set types but will not be
enabled to use the extended addressing space (EAS) of an EAV. When this document
references EAS-eligible data sets, it is referring to VSAM and extended-format sequential
data sets.
3390 Model A
With EAV volumes, an architecture is implemented that provides a capacity of hundreds of
terabytes for a single volume. However, the first releases are limited to a volume with 223 GB
or 262,668 cylinders.

Note the following points regarding extended address volumes:
Only 3390 Model A devices can be EAV.
EAV is supported starting in z/OS V1R10 and further releases.
The size is limited to 223 GB (262,668 cylinders) in z/OS V1R10 and V1R11.
Important: The 3390 Model A as a device can be configured to have from 1 to

268,434,453 cylinders on an IBM DS8000. It becomes an EAV if it has more than 65,520
cylinders defined. With current z/OS releases, this maximum size is not supported.
Using EAV volumes

How an EAV is managed by the system allows it to be a general purpose volume. However,
EAVs work especially well for applications with large files. PAV and HyperPAV technologies
help in both these regards by allowing I/O rates to scale as a volume gets larger.

3.10 Data sets eligible for EAV volumes
In z/OS V1R10, the following VSAM data sets are

EAS-eligible:
All VSAM data types (KSDS, RRDS, ESDS, and
Linear)
This covers DB2, IMS, CICS, zFS and NFS
SMS-managed and non-SMS-managed VSAM data

sets
VSAM data sets on an EAV that were inherited from a
prior physical migration or copy
With z/OS V1R11, extended-format sequential data

sets that are SMS-managed
Figure 3-10 Eligible data sets for EAV volumes
Eligible data sets

EAS-eligible data sets are defined to be those that can be allocated in the extended
addressing space, which is the area on an EAV located above the first 65,520 cylinders. This
is sometimes referred to as cylinder-managed space. All of the following data sets can be
allocated in the base addressing space of an EAV:
SMS-managed VSAM (all types)
Non-SMS VSAM (all types)
zFS data sets (which are VSAM LS)
zFS aggregates are supported in an EAV environment.
zFS aggregates or file systems can reside in track-managed space or
cylinder-managed space (subject to any limitations that DFSMS has).
zFS still has an architected limit of 4 TB for the maximum size of a zFS aggregate.
Database (DB2, IMS) use of VSAM
VSAM data sets inherited from prior physical migrations or copies
With z/OS V1R11: Extended-format sequential data sets that are SMS-managed.

EAS non-eligible data sets
An EAS-ineligible data set is a data set that may exist on an EAV but is not eligible to have
extents (through Create or Extend) in the cylinder-managed space. The exceptions to EAS
eligibility are as follows:
Catalogs (BCS (basic catalog structure) and VVDS (VSAM volume data set)
VTOC (continues to be restricted to within the first 64 K-1 tracks)
VTOC index
Page data sets
VSAM data sets with imbed or keyrange attributes
VSAM data sets with incompatible CA sizes
Note: In a future release, various of these data sets may become EAS-eligible. All data set
types, even those listed here, can be allocated in the track-managed space on a device
with cylinder-managed space on an EAV volume. Eligible EAS data sets can be created
and extended anywhere on an EAV. Data sets that are not eligible for EAS processing can
only be created or extended in the track-managed portions of the volume.

3.11 EAV volumes and multicylinder units
EAV Cylinder 262667

Multicylinder Units (MCUs) - 21 cylinders
Value that divides evenly into the 1 GB
cylinder-managed
space
storage segments of an IBM DS8000
These 1 GB segments are the
allocation unit in the IBM DS8000 and
are equivalent to 1113 cylinders
Cylinder 65520
track-managed
space
Cylinder 0
Figure 3-11 EAV and multicylinder units
Multicylinder unit
A multicylinder unit (MCU) is a fixed unit of disk space that is larger than a cylinder. Currently,
on an EAV volume, a multicylinder unit is 21 cylinders and the number of the first cylinder in
each multicylinder unit is a multiple of 21. Figure 3-11 illustrates the EAV and multicylinder
units.
The cylinder-managed space is space on the volume that is managed only in multicylinder
units. Cylinder-managed space begins at cylinder address 65,520. Each data set occupies an
integral multiple of multicylinder units. Space requests targeted for the cylinder-managed
space are rounded up to the next multicylinder unit. The cylinder-managed space only exists
on EAV volumes.
The 21-cylinder value for the MCU is derived from being the smallest unit that can map out
the largest possible EAV volume and stay within the index architecture with a block size of
8,192 bytes, as follows:
It is also a value that divides evenly into the 1 GB storage segments of an IBM DS8000.
These 1 GB segments are the allocation unit in the IBM DS8000 and are equivalent to
1,113 cylinders.
These segments are allocated in multiples of 1,113 cylinders starting at cylinder 65,520.
One of the more important EAV design points is that IBM maintains its commitment to
customers that the 3390 track format and image size, and tracks per cylinders, will remain the

same as for previous 3390 model devices. An application using data sets on an EAV will be
comparable to how it runs today on 3390 numerics kind of models. The extended address
volume has two managed spaces, as shown in Figure 3-11 on page 73:
The track-managed space
The cylinder-managed space
Cylinder-managed space
The cylinder-managed space is the space on the volume that is managed only in
multicylinder units (MCUs). Cylinder-managed space begins at cylinder address 65520. Each
data set occupies an integral multiple of multicylinder units. Space requests targeted for the
cylinder-managed space are rounded up to the next multicylinder unit. The cylinder-managed
space exists only on EAV volumes. A data set allocated in cylinder-managed space may
have its requested space quantity rounded up to the next MCU.
Data sets allocated in cylinder-managed space are described with a new type of data set
control blocks (DSCB) in the VTOC. Tracks allocated in this space will also be addressed
using the new track address. Existing programs that are not changed will not recognize these
new DSCBs and therefore will be protected from seeing how the tracks in cylinder-managed
space are addressed.
Track-managed space
The track-managed space is the space on a volume that is managed in track and cylinder
increments. All volumes today have track-managed space. Track-managed space ends at
cylinder number 65519. Each data set occupies an integral multiple of tracks. The
track-managed space ends at cylinder address 65519. Each data set occupies an integral
multiple of tracks. The track-managed space allows existing programs and physical migration
products to continue to work. Physical copies can be done from a non-EAV to an EAV and
have those data sets accessible.

3.12 Dynamic volume expansion (DVE)
Significantly reduces the complexity of migrating to

larger volumes
Copy service relationships must be removed
Previously, customers must use migration utilities that
require an additional volume for each volume that is
being expanded and require the data to be moved
DVE can expand volumes beyond 65,520 cylinders
Without moving data or application outage
Two methods to dynamically grow a volume:
Use the command-line interface (DSCLI)
Use a Web browser GUI
Figure 3-12 Using dynamic volume expansion to create EAV volumes
Dynamic volume expansion

The IBM System Storage DS8000 series supports dynamic volume expansion, which
increases the capacity of existing zSeries volumes, while the volume remains connected to a
host system.
This capability simplifies data growth by allowing volume expansion without taking volumes
offline. Using DVE significantly reduces the complexity of migrating to larger volumes.
Note: For the dynamic volume expansion function, volumes cannot be in Copy Services
relationships (point-in-time copy, FlashCopy SE, Metro Mirror, Global Mirror, Metro/Global
Mirror, and z/OS Global Mirror) during expansion.
There are two methods to dynamically grow a volume:

Use the command-line interface (DSCLI)
Use a Web browser GUI
Note: All systems must be at the z/OS V1R10 level or above for the DVE feature to be
used when the systems are sharing the Release 4.0 Licensed Internal Microcode updated
DS8000 at an LCU level. Current DS8000 Licensed Internal Microcode is Version
5.4.1.1043.

3.13 Using dynamic volume expansion
More recently, Dynamic Volume Expansion is a

function (available at the IBM DS8000 console) that:
Increases the capacity of existing zSeries volumes
Requires a manual operation for the system
programmer to expand the VTOC size
3390 model 3 to 3390 model 9
3390 model 9 to EAV volume sizes using z/OS V1R10
Note: Volumes cannot be in Copy Services relationships
(point-in-time copy, FlashCopy SE, Metro Mirror, Global Mirror,
Metro/Global Mirror, and z/OS Global Mirror) during expansion.
All systems must be at the z/OS V1R10 level for the DVE feature
to be used when the systems are sharing the Release 4.0
Licensed Internal Microcode updated DS8000 at an LCU level.
Figure 3-13 Using dynamic volume expansion
Using dynamic volume expansion

It is possible to grow an existing volume with the DS8000 with dynamic volume expansion
(DVE). Previously, it was necessary to use migration utilities that require an additional volume
for each volume that is being expanded and require the data to be moved.
A logical volume can be increased in size while the volume remains online to host systems for
the following types of volumes:
3390 model 3 to 3390 model 9
3390 model 9 to EAV volume sizes using z/OS V1R10
Dynamic volume expansion can be used to expand volumes beyond 65,520 cylinders
without moving data or causing an application outage.
Dynamic volume expansion is performed by the DS8000 Storage Manager and can be
requested using its Web GUI. 3390 volumes may be increased in size, for example from a
3390 model 3 to a model 9 or a model 9 to a model A (EAV). z/OS V1R11 introduces an
interface that can be used to make requests for dynamic volume expansion of a 3390 volume
on a DS8000 from the system.
Note: For the dynamic volume expansion function, volumes cannot be in Copy Services
relationships (point-in-time copy, FlashCopy SE, Metro Mirror, Global Mirror, Metro/Global
Mirror, and z/OS Global Mirror) during expansion.

3.14 Command-line interface (DSCLI)
The chckdvol command changes can expand the number

of cylinders on an existing volume
New option -cap new_capacity (optional on DS8000 only)
Specifies cylinders to allocate to the specified volume
3380 volumes cannot be expanded
3390 Model A (DS8000 only), value can be in the range of
1 to 65,520 or 65,667 to 262,668 (increments of 1113)
For 3390 volumes, the -cap parameter value can be in the
range of 1 to 65,520 (849 KB to 55.68 GB)
Change a 3390 Mod 3 device, ID 0860 (actual device
address D860), to a 3390 Mod 9:
chckdvol cap 10017 0860
Figure 3-14 Using the command line interface
Command line interface

The chckdvol command changes the name of a count key data (CKD) base volume. It can be
used to expand the number of cylinders on an existing volume. A new parameter on the
chckdvol command allows a volume to be increased in size, as follows:
-cap new_capacity (Optional and DS8000 only)
This specifies the quantity of CKD cylinders that you want allocated to the specified volume.
3380 volumes cannot be expanded. For 3390 Model A volumes (DS8000 only), the -cap
parameter value can be in the range of 1 to 65,520 (increments of 1) or 65,667 to 262,668
(increments of 1113). For 3390 volumes, the -cap parameter value can be in the range of 1 to
65,520 (849 KB to 55.68 GB).
3390 Model 3 to 3390 Model 9

This example in Figure 3-11 on page 73 shows how to use DSCLI to change the size of a
3390 Mod 3 device, ID 0860 (actual device address D860) to a 3390 Mod 9:
chckdvol cap 10017 0860

3.15 Using Web browser GUI
Need a URL for your own connection to the DS8000

Storage Manager (lpar_IPaddr)
http://lpar_IPaddr:8451/DS8000/Console
Log in
Select --------- Volumes - zSeries
Select --------- Select Storage Image
Select LCU ---- number (xx) - where Device is and

Page number of device
Figure 3-15 Log in to DS8000 to access a device to create an EAV
Access to the DS8000 port

Each DS8000 Fibre Channel card offers four Fibre Channel ports (port speed of 2 or 4 Gbps,
depending on the host adapter). The cable connector required to attach to this card is an LC
type. Each 2 Gbps port independently auto-negotiates to either 2 or 1 Gbps and the 4 Gbps
ports to 4 or 2 Gbps link speed. Each of the 4 ports on one DS8000 adapter can also
independently be either Fibre Channel protocol (FCP) or FICON, though the ports are initially
defined as switched point-to-point FCP. Selected ports will be configured to FICON
automatically based on the definition of a FICON host. Each port can be either FICON or
Fibre Channel Protocol (FCP). The personality of the port is changeable through the DS
Storage Manager GUI. A port cannot be both FICON and FCP simultaneously, but it can be
changed as required.
DS8000 shipped with pre-Release 3 code (earlier than Licensed Machine Code 5.3.xx.xx)
can also establish the communication with the DS Storage Manager GUI using a Web
browser on any supported network-connected workstation by simply entering into the Web
browser the IP address of the HMC and the port that the DS Storage Management server is
listening to:
http://<ip-address>:8451/DS8000
After accessing the GUI through a LOGON, follow these steps to access the device for which
you want to increase the number of cylinders.

3.16 Select volume to increase capacity
Figure 3-16 Select volume to increase capacity
Volume to increase capacity

The 3390 Model A can be defined as new or can be defined when a new EAV volume beyond
65,520 cylinders is specified. The number of cylinders when defining an EAV volume can be
defined as low as 1 cylinder and up to 262,668 cylinders. In addition, expanding a 3390 type
of volume that is below 65,520 cylinders to be larger than 65,520 cylinders converts the
volume from a 3390 standard volume type to a 3390 Model A and thus an expanded address
volume.
To increase the number of cylinders on an existing volume, you can use the Web browser
GUI by selecting the volume (shown as volser MLDF64 in Figure 3-16) and then using the
Select Action pull-down to select Increase Capacity.

3.17 Increase capacity of volumes
Figure 3-17 Screen showing current maximum size available
Screen showing maximum available cylinders

Figure 3-17 is displayed after you select Increase Capacity. A warning message is issued
regarding a possible volume type change to 3390 custom. You can then select Close
Message. Notice that the panel displays the maximum number of cylinders that can be
specified.
After you close the message, the panel in Figure 3-18 on page 81 is displayed where you can
specify the number of cylinders to increase the volume size. When you specify a new capacity
to be applied to the selected volume, specify a value that is between the minimum and
maximum size values that are displayed. Maximum values cannot exceed the amount of total
storage that is available.
Note: Only volumes of type 3390 model 3, 3390 model 9, and 3390 custom, can be
expanded. The total volume cannot exceed the available capacity of the storage image.
Capacity cannot be increased for volumes that are associated with Copy Services
functions.

3.18 Select capacity increase for volume
75000
75000
Figure 3-18 Specifying the number of cylinders to increase the volume size
Specify cylinders to increase volume size

After you close the message, shown in the panel in Figure 3-17 on page 80, Figure 3-18 is
displayed where you can specify the number of cylinders to increase the volume size; 75,000
is shown in Figure 3-18.
When you specify a new capacity to be applied to the selected volumes, specify a value that
is between the minimum and maximum size values that are displayed. Maximum values
cannot exceed the amount of total storage that is available.

3.19 Final capacity increase for volume
Figure 3-19 Capacity increase for volume
Final volume cylinder increase

Since 75,000 cylinders was specified on the panel shown in Figure 3-18 on page 81, the
message displayed in Figure 3-19 is issued when you click OK.
Specify Continue in Figure 3-19, and a requested size of 75,684 is processed for the
expansion of the volume.
Note: Remember that the number of cylinders must be as stated earlier. The reason an
MCU value is 21 cylinders is because it is derived from being the smallest unit that can
map out the largest possible EAV and stay within the index architecture (with a block size
of 8192 bytes). It is also a value that divides evenly into the 1 GB storage segments of a
DS8000. These 1 GB segments are the allocation unit in the DS8000 and are equivalent to
1113 cylinders. Data sets allocated in cylinder-managed space may have their requested
space quantity rounded up to the next MCU.

3.20 VTOC index with EAV volumes
With z/OS V1R10, the index block size is increased

from 2,048 bytes to 8,192 bytes for devices with
cylinder-managed space
This requires a delete of the old VTOC and a build of
a new VTOC
The new block size is recorded in the format-1 DSCB
for the index and is necessary to allow for scaling to
largest sized volumes
The DEVTYPE INFO=DASD macro can be used to
return the actual block size or
Can be determined from examining the format-1 DSCB
of the index data set
Figure 3-20 Build a new VTOC after volume expansion
New VTOC for volume expansion

When a volume is dynamically expanded the VTOC and VTOC index has to be reformatted to
map the additional space. At z/OS V1R10, and earlier, this has to be done manually by the
system programmer, or storage administrator, by submitting an ICKDSF REFVTOC job.
VTOC index
The VTOC index enhances the performance of VTOC access. The VTOC index is a
physical-sequential data set on the same volume as the related VTOC. It consists of an index
of data set names in format-1 DSCBs contained in the VTOC and volume free space
information.
An SMS-managed volume requires an indexed VTOC; otherwise, the VTOC index is highly
recommended. For additional information about SMS-managed volumes, see z/OS DFSMS
Implementing System-Managed Storage, SC26-7407.
Note: You can use the ICKDSF REFORMAT REFVTOC command to rebuild a VTOC index
to reclaim any no longer needed index space and to possibly improve access times.

VTOC index changes
The index block size increased from 2,048 bytes to 8,192 bytes for devices with
cylinder-managed space, as follows:
Contents of index map records increased proportionally (more bits per record, offsets
changed).
Extended header in VIXM for free space statistics.
VPSM, track-managed space. Small (tracks and cylinders) and large (cylinders) unit map.
Same but more bits.
VPSM, cylinder-managed space. Each bit in large unit map represents 21 cylinders (a
MCU).
Programs that access index maps must use existing self-describing fields in the map
records.
Block size is in format-1 DSCB for the VTOC index and returned by DEVTYPE macro.
Default index size computed by ICKDSF; might not be 15 tracks.
This new block size is recorded in the format-1 DSCB for the index and is necessary to allow
for scaling to largest sized volumes. The VTOC index space map (VIXM) has a new bit,
VIMXHDRV, to indicate that new fields exist in the new VIXM extension, as follows:
The VIXM contains a new field for the RBA of the new large unit map and new space
statistics.
The VIXM contains a new field for the minimum allocation unit in cylinders for the
cylinder-managed space. Each extent in the cylinder-managed space must be a multiple
of this on an EAV.
Note: If the VTOC index size is omitted when formatting a volume with ICKDSF and does
not preallocate the index, the default before this release was 15 tracks. In EAV Release 1
(that is, starting with z/OS V1R10), the default size for EAV and non-EAV volumes is
calculated and can be different from earlier releases.
VTOC VTOC Index

Format-4 DSCB VIXM
Format-5 DSCB VPSM
VMDS
Other DSCBs VIER
VIER
VIER
Format-1 DSCB for the VTOC
Index: SYS1.VTOCIX.nnn .
.
Other DSCBs .
VTOC must reside in the "old" space on the volume

3.21 Device Support FACILITY (ICKDSF)
ICKDSF initializes a VTOC index into 2048-byte physical

blocks, or 8192-byte physical blocks on EAV volumes,
named VTOC index records (VIRs)
The DEVTYPE INFO=DASD macro can be used to
return the actual block size or it can be determined from
examining the format-1 DSCB of the index data set
APAR PK56092 for ICKDSF R17 provides support for
the (EAV) function in z/OS V1R10
Changes to the default size of a VTOC index when the
INDEX parameter of the ICKDSF INIT command is
defaulted after you install the PTF for the APAR,
regardless of whether you exploit the EAV function
Figure 3-21 ICKDSF facility to build the VTOC
ICKDSF utility
The ICKDSF utility performs functions needed for the installation, use, and maintenance of
IBM direct-access storage devices (DASD). You can also use it to perform service functions,
error detection, and media maintenance.
The ICKDSF utility is used primarily to initialize disk volumes. At a minimum, this process
involves creating the disk label record and the volume table of contents (VTOC). ICKDSF can
also scan a volume to ensure that it is usable, can reformat all the tracks, can write home
addresses, as well as other functions.
You can use the ICKDSF utility through two methods:

Execute ICKDSF as a job or job step using job control language (JCL). ICKDSF
commands are then entered as part of the SYSIN data for z/OS.
Use Interactive Storage Management Facility (ISMF) panels to schedule ICKDSF jobs.
APAR PK56092
This APAR provides extended address volume (EAV) support up to 262,668 cylinders. If you
define a volume greater than 262,668 cylinders you will receive the following message when
running any ICKDSF command:
ICK30731I X'xxxxxxx' CYLINDER SIZE EXCEEDS MAXIMUM SIZE
SUPPORTED
xxxxxxx will contain the hex value size of the volume you defined.

The documentation is in Device Support Facilities (ICKDSF) User's Guide and Reference
Release 17, GC35-0033.
ICKDSF to build a VTOC

After the expansion of the volume operation completes, you need to use the ICKDSF
REFORMAT REFVTOC command, shown in Figure 3-22, to adjust the volume VTOC to
reflect the additional cylinders. Note that the capacity of a volume cannot be decreased.
//INIT EXEC PGM=ICKDSF,PARM='NOREPLYU'

//IN1 DD UNIT=3390,VOL=SER=ITSO02,DISP=SHR
//SYSPRINT DD SYSOUT=*
//SYSIN DD *
REFORMAT DDNAME(IN1) VERIFY(ITSO02) REFVTOC
/*
Figure 3-22 Build a new VTOC

3.22 Update VTOC after volume expansion
REFORMAT command must be used to rebuild the

VTOC and index structures after a volume has been
expanded - (Method used with V1R10)
The command updates portions of a previously
initialized volume
You can also enter ICKDSF commands with ISMF
If an index exists when you expand the VTOC, it must
be deleted and rebuilt to reflect the VTOC changes
Parameters for index total track size rebuild
EXTINDEX(n)
XINDEX(n)
Figure 3-23 Create a new VTOC after volume expansion
New VTOC after expansion

To prepare a volume for activity by initializing it, use the Device Support Facilities (ICKDSF)
utility to build the VTOC. You can create a VTOC index when initializing a volume by using
the ICKDSF INIT command and specifying the INDEX keyword.
To convert a non-indexed VTOC to an indexed VTOC, use the BUILDIX command with the
IXVTOC keyword. The reverse operation can be performed by using the BUILDIX command
and specifying the OSVTOC keyword.
Parameters for index rebuild

To refresh a volume VTOC and INDEX in its current format, use the ICKDSF command
REFORMAT with the RVTOC keyword. To optionally extend the VTOC and INDEX, use the
ICKDSF command REFORMAT with the EXTVTOC and EXTINDEX keywords.
Run REFORMAT NEWVTOC(cc,hh,n|ANY,n) to expand the VTOC. The new VTOC will be
allocated on the beginning location cc,hh with total size of n tracks. Overlay between the new
and old VTOC is not allowed. If cc,hh is omitted, ICKDSF will locate the new VTOC at the first
eligible location on the volume other than at the location of the old VTOC where free space
with n tracks is found. n must be greater than the old VTOC size. The volume must be offline
to use the NEWVTOC parameter.
Run REFORMAT EXTVTOC(n) where n is the total size of the new VTOC in tracks. There
must be free space available to allow for contiguous expansion of the VTOC. If there is no

free space available following the current location of the VTOC, an error message is issued. n
must be greater than the old VTOC size.
EXTINDEX(n) and XINDEX(n)

If an index exists when you expand the VTOC, it must be deleted and rebuilt to reflect the
VTOC changes. This parameter is used to specify the total track size of the index to be
rebuilt. For n, substitute the decimal or hexadecimal digits (for example, X'1E') to specify the
total number of tracks for the new index after expansion. If the value for n is less than the
current value, the current value is used.
Restriction: Valid only for MVS online volumes. Valid only when EXTVTOC is specified.

3.23 Automatic VTOC index rebuild - z/OS V1R11
DEVSUPxx parmlib member options

REFVTOC=ENABLE - Enables the automatic
REFVTOC function of the device manager that when
enabled, and a volume expansion is detected, the
device manager causes the volume VTOC to be rebuilt
This allows the newly added space on the volume to be
used by the system
REFVTOC=DISABLE - This is the default value. It
disables the automatic REFVTOC function of the device
manager
With the REFVTOC function disabled, when a volume
expansion is detected: (Use R10 method)
The following message is issued:
IEA019I DDDD,VVVVVV,VOLUME CAPACITY CHANGE,OLD=XXXXXXXX NEW=YYYYYYYY
Figure 3-24 DEVSUPxx parmlib member parameters
VTOC rebuild with z/OS V1R11

With z/OS V1R11, when a volume is increased in size, this is detected by the system, which
then does an automatic VTOC and index rebuild. The system is informed by state change
interrupts (SCIs), which are controlled with new DEVSUPxx parmlib member options as
follows:
REFVTOC=ENABLE
Enables the automatic REFVTOC function of the device manager. With the REFVTOC
function enabled, when a volume expansion is detected, the device manager causes the
volume VTOC to be rebuilt. This allows the newly added space on the volume to be used
by the system.
REFVTOC=DISABLE
This is the default value. It disables the automatic REFVTOC function of the device
manager. With the REFVTOC function disabled, when a volume expansion is detected,
the following message is issued:
IEA019I dev, volser, VOLUME CAPACITY CHANGE,OLD=xxxxxxxx,NEW=yyyyyyyy.
The VTOC is not rebuilt. An ICKDSF batch job must be submitted to rebuild the VTOC
before the newly added space on the volume can be used, as is the case for z/OS V1R10.
Invoke ICKDSF with REFORMAT/REFVTOC to update the VTOC and index to reflect the
real device capacity.

Note: The refresh of the index occurs under protection of an exclusive SYSTEMS ENQ
macro for major name SYSZDMO, minor name DMO.REFVTOC.VOLSER.volser.

3.24 Automatic VTOC rebuild with DEVMAN
Use the F DEVMAN command to communicate with

the device manager address space to rebuild the
VTOC
F DEVMAN,ENABLE(REFVTOC)
F DEVMAN,{DUMP}
{REPORT}
{RESTART}
{END(taskid)}
{ENABLE(feature) }
New with z/OS V1R11
{DISABLE(feature)}
{?|HELP}
Figure 3-25 Rebuild the VTOC for EAV volumes with DEVMAN
New functions for DEVMAN with z/OS V1R11

Following are brief descriptions of the new parameters with the F DEVMAN command in z/OS
V1R11.
END(taskid) Terminates the subtask identified by taskid. The F
DEVMAN,REPORT command displays the taskid for a subtask.
ENABLE(feature name) Enables an optional feature. The supported features are named
as follows:
REFVTOC - Use ICKDSF to automatically
REFORMAT/REFVTOC a volume when it expands.
DATRACE - Capture dynamic allocation diagnostic
messages.
DISABLE(feature name) Disables one of the following optional features:
REFVTOC
DATRACE
?|HELP Displays the F DEVMAN command syntax.

Updating the VTOC following volume expansion
Specifying REFVTOC on the F DEVMAN command uses ICKDSF to automatically
REFORMAT/REFVTOC a volume when it has been expanded. See z/OS MVS System
Commands, SA22-7627.
Specifying REFVTOC={ENABLE | DISABLE} in the DEVSUPxx parmlib member to indicate

whether you want to enable or disable the use of ICKDSF to automatically
REFORMAT/REFVTOC a volume when it is expanded. See z/OS MVS Initialization and
Tuning Reference, SA22-7592.
You can also use the F DEVMAN,ENABLE(REFVTOC) command after an IPL as well to
enable automatic VTOC and index reformatting. However, update the DEVSUPxx parmlib
member to ensure that it remains enabled across subsequent IPLs.
Using DEVMAN
The DEVMAN REPORT display has the following format, as shown in Figure 3-26.
Where:
FMID Displays the FMID level of DEVMAN.
APARS Displays any DEVMAN APARs that are installed (or the word NONE).
OPTIONS Displays the currently enabled options (in the example, REFVTOC is
enabled).
SUBTASKS Lists the status of any subtasks that are currently executing.
F DEVMAN,HELP
DMO0060I DEVICE MANAGER COMMANDS:
**** DEVMAN *********************************************************
* ?|HELP - display devman modify command parameters
* REPORT - display devman options and subtasks
* RESTART - quiesce and restart devman in a new address space
* DUMP - obtain a dump of the devman address space
* END(taskid) - terminate subtask identified by taskid
* ENABLE(feature) - enable an optional feature
* DISABLE(feature)- disable an optional feature
*--------------------------------------------------------------------
* Optional features:
* REFVTOC - automatic VTOC rebuild
* DATRACE - dynamic allocation diagnostic trace
**** DEVMAN *********************************************************
Figure 3-26 Displaying DEVMAN modify parameters

3.25 EAV and IGDSMSxx parmlib member
USEEAV(YES|NO)
Specifies, at the system level, whether SMS can select
an extended address volume during volume selection
processing
Check applies to new allocations and when extending
data sets to a new volume
YES - EAV volumes can be used to allocate new data
sets or to extend existing data sets to new volumes
NO - Default - SMS does not select any EAV during
volume selection
SETSMS USEEAV(YES|NO)
Figure 3-27 Parmlib members for using EAV volumes
Extended address volume selection considerations

The two new SMS IGDSMSxx parmlib member parameters need to be set before you can
use extended address volumes with SMS for volume selection, as follows:
USEEAV
BreakPointValue
IGDSMSxx parmlib member parameters

The introduction and use by the system of an EAV is determined by adding the volumes to
SMS storage groups or adding them to non-SMS managed storage pools (specific or esoteric
volumes). Volumes can be in their dedicated storage group or pool, or mixed with other
volumes. Adding and enabling them to these pools allows the system to allocate VSAM data
sets in cylinder-managed space if the USEEAV parmlib setting is set to YES. When USEEAV
is set to NO, no new data set creates (even non-VSAM) are allowed on any EAV in the
system.
IGDSMSxx parmlib member changes to be able to use EAV volumes

USEEAV(YES|NO) specifies, at the system level, whether SMS can select an extended
address volume during volume selection processing. This check applies to new allocations
and when extending data sets to a new volume.

USEEAV(YES|NO)
YES This means that extended address volumes can be used to allocate new data sets
or to extend existing data sets to new volumes.
NO (Default) This means that SMS does not select any extended address volumes
during volume selection. Note that data sets might still exist on extended address
volumes in either the track-managed or cylinder-managed space of the volume.
Note: You can use the SETSMS command to change the setting of USEEAV without
having to re-IPL. This modified setting is in effect until the next IPL, when it reverts to the
value specified in the IGDSMSxx member of PARMLIB.
To make the setting change permanent, you must alter the value in SYS1.PARMLIB. The
syntax of the operator command is:
SETSMS USEEAV(YES|NO)
SMS requests will not use EAV volumes if the USEEAV setting in the IGDSMSxx parmlib
member is set to NO.
Specific allocation requests are failed. For non-specific allocation requests (UNIT=SYSDA),
EAV volumes are not selected. Messages indicating no space available are returned when
non-EAV volumes are not available.
For non-EAS eligible data sets, all volumes (EAV and non-EAV) are equally preferred (or they
have no preference). This is the same as today, with the exception that extended address
volumes are rejected when the USEEAV parmlib value is set to NO.

3.26 IGDSMSxx member BreakPointValue
BreakPointValue (0- 65520) in cylinders

Value used by SMS in making volume selection
decisions and subsequently by DADSM
If the allocation request is less than the BreakPointValue,
the system prefers to satisfy the request from free space
available from the track-managed space
If the allocation request is equal to or higher than the
BreakPointValue, the system prefers to satisfy the
request from free space available from the
cylinder-managed space
SETSMS BreakPointValue(0-65520)
If the preferred area cannot satisfy the request, both
areas become eligible to satisfy the requested space
Figure 3-28 BreakPointValue parameter
Using the BreakPointValue

How the system determines which managed space is to be used is based on the derived
BreakPointValue for the target volume of the allocation. EAV volumes are preferred over
non-EAV volumes where space >=BreakPointValue (there is no EAV volume preference
where space < BreakPointValue).
BreakPointValue (0- 65520)

BreakPointValue (0- 65520) is used by SMS in making volume selection decisions and
subsequently by DADSM. If the allocation request is less than the BreakPointValue, the
system prefers to satisfy the request from free space available from the track-managed
space.
If the allocation request is equal to or higher than the BreakPointValue, the system prefers to
satisfy the request from free space available from the cylinder-managed space. If the
preferred area cannot satisfy the request, both areas become eligible to satisfy the requested
space amount.
Note: The BreakPointValue is only used to direct placement on an expanded address

volume.

Generally, for VSAM data set allocation requests that are equal to or larger than the
BreakPointValue, SMS prefers extended address volumes for:
Non-VSAM allocation requests
VSAM allocation requests that are smaller than the BreakPointValueSMS does not
have a preference.
The default is 10 cylinders.
Note: You can use the SETSMS command to change the setting of BreakPointValue
without having to re-IPL. This modified setting is in effect until the next IPL when it reverts
to the value specified in the IGDSMSxx member of PARMLIB. To make the setting change
permanent, you must alter the value in SYS1.PARMLIB. The syntax of the operator
command is:
SETSMS BreakPointValue(0-65520)
System use of the BreakPointValue

For an extended address volume, the system and storage group use of the BreakPointValue
(BPV) helps direct disk space requests to cylinder-managed or track-managed space. The
breakpoint value is expressed in cylinders and is used as follows:
When the size of a disk space request is the breakpoint value or more, the system prefers
to use the cylinder-managed space for that extent. This rule applies to each request for
primary or secondary space for data sets that are eligible for the cylinder-managed space.
If cylinder-managed space is insufficient, the system uses the track-managed space or
uses both types of spaces.
When the size of a disk space request is less than the breakpoint value, the system
prefers to use the track-managed space.
If space is insufficient, the system uses the cylinder-managed space or uses both types of
space.

3.27 New EATTR attribute in z/OS V1R11
For all EAS eligible data sets - new data set attribute
EATTR - allow a user to control data set allocation to
have extended attribute DSCBs (format 8 and 9)
Use the EATTR parameter to indicate whether the
data set can support extended attributes
To create such data sets, you can include (EAVs) in
specific storage groups or specify an EAV on the
request or direct the allocation to an esoteric
containing EAV devices
By definition, a data set with extended attributes can
reside in the extended address space (EAS) on an
extended address volume (EAV)
Figure 3-29 EATTR attribute with z/OS V1R11
EATTR attribute
This z/OS v1R11support for extended format sequential data sets includes the EATTR
attribute, which has been added for all data set types to allow a user to control whether a data
set can have extended attribute DSCBs and thus control whether it can be allocated in the
EAS.
EAS-eligible data sets are defined to be those that can be allocated in the extended
addressing space and have extended attributes. This is sometimes referred to as
cylinder-managed space.
DFSMShsm checks the data set level attribute EATTR when performing non-SMS volume
selection. The EATTR data set level attribute specifies whether a data set can have extended
attributes (Format 8 and 9 DSCBs) and optionally reside in EAS on an extended address
volume (EAV). Valid value for the EATTR are NO and OPT.
For more information about the EATTR attribute, see z/OS DFSMS Access Method Services
for Catalogs, SC26-7394.
Using the EATTR attribute

Use the EATTR parameter to indicate whether the data set can support extended attributes
(format-8 and format-9 DSCBs). To create such data sets, you can include extended address
volumes (EAVs) in specific storage groups, or specify an EAV on the request, or direct the
allocation to an esoteric containing EAV devices. By definition, a data set with extended

attributes can reside in the extended address space (EAS) on an extended address volume
(EAV). The EATTR parameter can be used as follows:
It can be specified for non-VSAM data sets as well as for VSAM data sets.
The EATTR value has no effect for DISP=OLD processing, even for programs that might
open a data set for OUTPUT, INOUT, or OUTIN processing.
The value on the EATTR parameter is used for requests when the data set is newly
created.

3.28 EATTR parameter
The EATTR parameter can be used as follows:

It can be specified for non-VSAM data sets as well as for
VSAM data sets
The EATTR value has no effect for DISP=OLD processing,
even for programs that might open a data set for OUTPUT,
INOUT, or OUTIN processing
The value of the EATTR parameter is used for requests
when the data set is newly created
EATTR = OPT - Extended attributes are optional because
data sets can have extended attributes and reside in EAS
(Default)
EATTR = NO - No extended attributes. The data set
cannot have extended attributes (format 8 and 9 DSCBs)
or reside in EAS (Default for non-VSAM data sets)
Figure 3-30 Using the EATTR parameter
Subparameter definitions in JCL

Use the EATTR JCL keyword, EATTR=OPT, to specify that the data set can have extended
attribute DSCBs and can optionally reside in EAS.
EATTR = OPT Extended attributes are optional. The data set can have extended attributes
and reside in EAS. This is the default value for VSAM data sets if
EATTR(OPT) is not specified.
EATTR = NO No extended attributes. The data set cannot have extended attributes
(format-8 and format-9 DSCBs) or reside in EAS. This is the default value
for non-VSAM data sets. This is equivalent to specifying EATTR=NO on the
JCL and is applicable to all data set types.
Note: The EATTR specification is recorded in the format-1 or format-8 DSCBs for all data
set types and volume types and is recorded in the VVDS for VSAM cluster names. EATTR
is listed by IEHLIST, ISPF, ISMF, LISTCAT, and the catalog search interface (CSI).

Other methods for EATTR specification
Different BCP components have adapted to use the EATTR attribute introduced with z/OS
V1R11. This attribute is specifiable using any of the following methods:
AMS DEFINE CLUSTER and ALLOCATE
On the ALLOCATE or DEFINE CLUSTER commands, allocate new data sets with
parameter EATTR(OPT) to specify that the data set can have extended attribute DSCBs
(format-8 and format-9) and can optionally reside in EAS.
ISMF Data Class Define panel
Use the EATTR attribute on the ISMF Data Class Define panel, to define a data set that
can have extended attribute DSCBs and optionally reside in EAS.
ISPF
There is new support in ISPF for displaying and setting the EATTR attribute for data sets
that can reside on extended address volumes (EAVs). ISPF has added support for the
display of extended address volumes (EAV) with the specification of a data set level
attribute, EATTR, using the allocation panel, Option 3.2, as shown in Figure 3-31.
Figure 3-31 ISPF Option 3.2 allocation using EATTR

3.29 EATTR JCL DD statement example
If an EATTR value is not specified for a data set, the

following default data set EATTR values are used:
The default behavior for VSAM data sets is OPT
The default behavior for non-VSAM data sets is NO
A DD statement defines a new data set
Example allocates 10,000 cylinders to the data set
When the CONTIG subparameter is coded, the system
allocates 10,000 contiguous cylinders on the volume
//DD2 DD DSNAME=XYZ12,DISP=(,KEEP),UNIT=SYSALLDA,
// VOLUME=SER=25143,SPACE=(CYL,(10000,100),,CONTIG),
// EATTR=OPT
Figure 3-32 JCL example using the EATTR parameter
Using the EATTR parameter

OPT is the default behavior for VSAM data sets if EATTR(OPT) is not specified. For
non-VSAM data sets the default is that the data set cannot have extended attribute DSCBs
and optionally reside in EAS. This is equivalent to specifying EATTR(NO) on the command.
EATTR JCL example

In Figure 3-32, the DD statement defines a new partitioned data set. The system allocates
10,000 cylinders to the data set, of which one hundred 256-byte records are for a directory.
When the CONTIG subparameter is coded, the system allocates 10 contiguous cylinders on
the volume. EATTR=OPT indicates that the data set might be created with extended
attributes. With this option, the data set can reside in the extended address space (EAS) of
the volume.
Note: The EATTR value has no effect for DISP=OLD processing, even for programs that
might open a data set for OUTPUT, INOUT, or OUTIN processing. The value on the EATTR
parameter is used for requests when the data set is newly created.

3.30 Migration assistance tracker
The EAV migration assistance tracker can help with:

Finding programs that you might need to change if you
want to support extended address volumes (EAVs)
Identify interfaces that access the VTOC - they need to
have EADSCB=OK specified for the following functions:
OBTAIN, CVAFDIR, CVAFDSM, CVAFVSM, CVAFSEQ,
CVAFFILT, OPEN to VTOC, OPEN EXCP
Identify programs using new services as info messages
Identify possible improper use of returned information,
Parsing 28-bit cylinder numbers in output as 16-bit cylinder
numbers as warning messages for the following
commands and functions:
IEHLIST LISTVTOC, IDCAMS LISTCAT, IDCAMS
LISTDATA PINNED, LSPACE, DEVTYPE, IDCAMS
Figure 3-33 Migration assistant tracker functions
Migration tracker program

The EAV migration assistance tracker can help you find programs that you might need to
change if you want to support extended address volumes (EAVs). The EAV migration
assistance tracker is an extension of the console ID tracking facility. Programs identified in
this phase of migration assistance tracking will continue to fail if the system service is issued
for an EAV and you do not specify the EADSCB=OK keyword for them.
DFSMS provides an EAV migration assistance tracker program. The tracking of EAV
migration assistance instances uses the Console ID Tracking facility provided in z/OS V1R6.
The EAV migration assistance tracker helps you to do the following:
Identify select systems services by job and program name, where the invoking programs
might require analysis for changes to use new services. The program calls are identified
as informational instances for possible migration actions. They are not considered errors,
because the services return valid information.
Identify possible instances of improper use of returned information in programs, such as
parsing 28-bit cylinder numbers in output as 16-bit cylinder numbers. These instances are
identified as warnings.
Identify instances of programs that will either fail or run with an informational message if
they run on an EAV. These are identified as programs in error. The migration assistance
tracker flags programs with the following functions: when the target volume of the
operation is non-EAV, and the function invoked did not specify the EADSCB=OK keyword.

Note: Being listed in the tracker report does not imply that there is a problem. It is simply a
way to help you determine what to examine to see whether it is to be changed.
Error detection by the tracker

Each category of errors is explained here:
Identify interfaces with access to the VTOC that are to be upgraded to have EADSCB=OK
specified for the following functions:
OBTAIN
CVAFDIR
CVAFDSM
CVAFVSM
CVAFSEQ
CVAFFILT
OPEN to VTOC
OPEN EXCP
Identify programs to use new services as informational messages.
Identify the possible improper use of returned information, such as parsing 28-bit cylinder
numbers in output as 16-bit cylinder numbers as warning messages for the following
commands and functions:
IEHLIST LISTVTOC, IDCAMS LISTCAT, IDCAMS LISTDATA PINNED
LSPACE, DEVTYPE, IDCAMS DCOLLECT
General information about the tracker

The tracker function allows component code to register tracking information as a text string of
its choosing, of up to 28 characters in length. The tracker records this as a unique instance
and appends additional information to it such as job name, program name, and count of
occurrences, to avoid duplicates.
DFSMS instances tracked by the EAV migration assistance tracker are shown in Figure 3-34.
LSPACE (SVC 78)

DEVTYPE (SVC 24)
IDCAMS LISTDATA PINNED
IEHLIST LISTVTOC
IDCAMS DCOLLECT
IDCAMS LISTCAT
OBTAIN (SVC 27)
CVAFDIR
CVAFSEQ
CVAFDSM
CVAFFILT
CVAFVSM
DCB Open of a VTOC
DCB Open of EAS eligible data set
Figure 3-34 DFSMS instances tracked by the migration tracker

3.31 Migration tracker commands
SETCON command
Used to activate and deactivate the Console ID

Tracking facility
SETCON TR=ON
DISPLAY OPDATA,TRACKING command
Used to display the current status of the console ID

tracking facility, along with any recorded instances of
violations
Figure 3-35 Migration tracker commands
Migration tracker commands usage

The tracking facility can be used with the following commands:
The SETCON command is used to activate and deactivate the Console ID Tracking
facility.
The DISPLAY OPDATA,TRACKING command is used to display the current status of the
Console ID Tracking facility, along with any recorded instances of violations. Sample
output of this command is shown in Figure 3-37 on page 105.
CNIDTRxx parmlib member

The CNIDTRxx parmlib member is used to list violations that have already been identified to
prevent them from being recorded again. An optional CNIDTRxx parmlib member can be
defined to exclude instances from being recorded. The exclusion list is picked up when the
tracker is started or through the SET command. The filters for the exclusion list are the three
items that make an instance unique. You can exclude all SMS instances or just LSPACE
instances, or exclude all instances created by job names that begin with BRS*. There are
many more possibilities. The filters support wildcarding on these three fields.
As events are recorded by the Tracking facility, report the instance to the product owner. After
the event is reported, update the parmlib member so that the instance is no longer recorded
by the facility. In this way, the facility only reports new events.

Tracker exclusion list
The exclusion list prevents instances from being recorded. To identify or use an exclusion list,
use the following operator command, as illustrated in Figure 3-36.
set cnidtr=7t
IEE536I CNIDTR VALUE 7T NOW IN EFFECT
SAMPLE EXCLUSION LIST 7T

* Jobname Pgmname *
* Tracking Information Mask Mask Mask Comments (ignored) *
*----------------------------+--------+--------+----------------------+
|SMS-I:3 LSPACE* |*MASTER*|IEE70110| I SMF CALLS TO LSPACE|
|SMS-I:3 LSPACE* |ALLOCAS |IEFW21SD| VARY DEVICE OFFLINE |
|SMS-I:3 LSPACE* |* |VTDSOIS2| VTDSOIS2 PROG CALLS |
|SMS-I:3 LSPACE* |VTDSOIS1|* | VTDSOIS1 JOB CALLS |
Figure 3-36 Sample tracker exclusion list in the CNIDTR7T parmlib member
Tracking command example

In Figure 3-37, the tracking information column identifies the instance as being
DFSMS-related with the SMS prefix. The additional I, E, or W appended to SMS identifies the
instance as being an informational, error, or warning event. The remaining text in the tracking
information describes the event that was recorded or, for error events, the type of error that
would have occurred if the function were executed on an EAV. The tracking value is a value
unique to the error being recorded. JOBNAME, PROGRAM+OFF, and ASID identify what
was being run at the time of the instance. Only unique instances are recorded. Duplicates are
tracked by the NUM column being incremented. The tracking information, jobname, and
program name fields make up a unique instance.
13.21.19 SYSTEM1 d opdata,tracking

13.21.19 SYSTEM1 CNZ1001I 13.21.19 TRACKING DISPLAY 831
STATUS=ON,ABEND NUM=15 MAX=1000 MEM=7T EXCL=45 REJECT=0
----TRACKING INFORMATION---- -VALUE-- JOBNAME PROGNAME+OFF-- ASID NUM
SMS-E:1 CVAFDIR STAT082 045201 CVAFJBN CVAFPGM 756 28 4
SMS-E:1 CVAFDSM STAT082 045201 CVAFJBN CVAFPGM 556 28 4
SMS-E:1 CVAFFILT STAT086 04560601 CVAFJBN CVAFPGM 456 28 4
SMS-E:1 CVAFSEQ STAT082 045201 CVAFJBN CVAFPGM 656 28 4
SMS-E:1 DADSM OBTAIN C08001 OBTJBN OBTPGM 856 28 4
SMS-E:1 DCB OPEN VSAM 113-44 01 OPENJBN OPENPGM 256 28 4
SMS-E:1 DCB OPEN VTOC 113-48 01 OPENJBN OPENPGM 356 28 4
SMS-I:3 DEVTYPE 02 DEVTJOB DEVTPROG CE5C 11 1
SMS-I:3 IDCAMS DCOLLECT 02 DCOLLECT IDCAMS 1515 28 4
SMS-I:3 LSPACE EXPMSG= 8802 VTDS0IS1 VTDS0IS2 118 28 2
SMS-I:3 LSPACE MSG= 5002 ALLOCAS IEFW21SD 4CE5C 11 2
SMS-I:3 LSPACE MSG= 9002 *MASTER* IEE70110 52F6 01 43
SMS-W:2 IDCAMS LISTDATA PINN 03 LISTDATX IDCAMS E48E 28 2
SMS-W:2 IDCAMS LISTCAT 03 LISTCAT IDCAMS 956 28 4
SMS-W:2 IEHLIST LISTVTOC 03 LISTVTOC IEHLIST 1056 28 4
----------------------------------------------------------------------
TO REPORT THESE INSTANCES, SEND THIS MESSAGE VIA E-MAIL TO
CONSOLES@US.IBM.COM. FOR ADDITIONAL INFORMATION OR TO OBTAIN A CURRENT
EXCLUSION LIST, SEE APAR II13752.
Figure 3-37 Tracker instance report output

4
Chapter 4. Storage management software

DFSMS is an exclusive element of the z/OS operating system and it automatically manages
data from creation to expiration. In this chapter we present the following topics:
The DFSMS utility programs to assist you in organizing and maintaining data
The major DFSMS access methods
The data set organizations
An overview of the elements that comprise DFSMS:
DFSMSdfp, a base element of z/OS
DFSMSdss, an optional feature of z/OS
DFSMShsm, an optional feature of z/OS
DFSMSrmm, an optional feature of z/OS
z/OS DFSORT

4.1 Overview of DFSMSdfp utilities
IEBCOMPR: Compare records in SEQ/PDS(E)

IEBCOPY: Copy/Merge/Compress/Manage PDS(E)
IEBDG: Test data generator
IEBEDIT: Selectively copy job steps
IEBGENER: Sequential copy, generate data sets
IEBIMAGE: Create printer image
IEBISAM: Create, copy, backup, print ISAM data set
IEBPTPCH: Print or punch SEQ/PDS(E)
IEBUPDTE: Create/modify SEQ/PDS(E)
IEHINITT: Write standard labels on tape volumes
IEHLIST: List VTOC/PDS(E) entries
IEHMOVE: Move or copy collections of data
IEHPROGM: Build, maintain system control data
IFHSTATR: Formats SMF records type 21 (ESV data)
Figure 4-1 DFSMSdfp utilities
DFSMSdfp utilities
Utilities are programs that perform commonly needed functions. DFSMS provides utility
programs to assist you in organizing and maintaining data. There are system and data set
utility programs that are controlled by JCL, and utility control statements.
The base JCL and certain utility control statements necessary to use these utilities are
provided in the major discussion of the utility programs in this chapter. For more details and to
help you find the program that performs the function you need, see Guide to Utility Program
Functions in z/OS DFSMSdfp Utilities, SC26-7414.
System utility programs

System utility programs are used to list or change information related to data sets and
volumes, such as data set names, catalog entries, and volume labels. Most functions that
system utility programs can perform are accomplished more efficiently with other programs,
such as IDCAMS,ISPF/PDF, ISMF, or DFSMSrmm.
Table 4-1 on page 109 lists and describes system utilities. Programs that provide functions
which are better performed by newer applications (such as ISMF, ISPF/PDF or DFSMSrmm
or DFSMSdss) are marked with an asterisk (*) in the table.

Table 4-1 System utility programs
System utility Alternate program Purpose
*IEHINITT DFSMSrmm EDGINERS Write standard labels on tape volumes.
IEHLIST ISMF, PDF 3.4 List system control data.
*IEHMOVE DFSMSdss, IEBCOPY Move or copy collections of data.
IEHPROGM Access Method Services, Build and maintain system control data.
PDF 3.2
*IFHSTATR DFSMSrmm, EREP Select, format, and write information about tape
errors from the IFASMFDP tape.
Data set utility programs

You can use data set utility programs to reorganize, change, or compare data at the data set
or record level. These programs are controlled by JCL statements and utility control
statements.
These utilities allow you to manipulate partitioned, sequential or indexed sequential data sets,
or partitioned data sets extended (PDSEs), which are provided as input to the programs. You
can manipulate data ranging from fields within a logical record to entire data sets. The data
set utilities included in this section cannot be used with VSAM data sets. You use the
IDCAMS utility to manipulate VSAM data set; refer to Invoking the IDCAMS utility program
on page 130.
Table 4-2 lists data set utility programs and their use. Programs that provide functions which
are better performed by newer applications, such as ISMF or DFSMSrmm or DFSMSdss, are
marked with an asterisk (*) in the table.
Table 4-2 Data set utility programs

Data set utility Use
*IEBCOMPR, SuperC, (PDF 3.12) Compare records in sequential or partitioned data sets, or
PDSEs.
IEBCOPY Copy, compress, or merge partitioned data sets or PDSEs; add

RLD count information to load modules; select or exclude
specified members in a copy operation; rename or replace
selected members of partitioned data sets or PDSEs.
IEBDG Create a test data set consisting of patterned data.
IEBEDIT Selectively copy job steps and their associated JOB

statements.
IEBGENER or ICEGENER Copy records from a sequential data set, or convert a data set
from sequential organization to partitioned organization.
*IEBIMAGE Modify, print, or link modules for use with the IBM 3800 Printing
Subsystem, the IBM 3262 Model 5, or the 4284 printer.
*IEBISAM Unload, load, copy, or print an ISAM data set.
IEBPTPCH or PDF 3.1 or 3.6 Print or punch records in a sequential or partitioned data set.
IEBUPDTE Incorporate changes to sequential or partitioned data sets, or

PDSEs.
Chapter 4. Storage management software 109

4.2 IEBCOMPR utility
Example 1
PDSE1 PDSE2
Directory 1 Directory 2
ABCDGL ABCDEFG
HIJKL
Example 2
PDSE1 PDSE2
Directory 1 Directory 2
ABCFHIJ ABFGHIJ
Figure 4-2 IEBCOMPR utility example
IEBCOMPR utility
IEBCOMPR is a data set utility used to compare two sequential data sets, two partitioned
data sets (PDS), or two PDSEs, at the logical record level, to verify a backup copy. Fixed,
variable, or undefined records from blocked or unblocked data sets or members can also be
compared. However, you should not use IEBCOMPR to compare load modules.
Two sequential data sets are considered equal (that is, are considered to be identical) if:
The data sets contain the same number of records
Corresponding records and keys are identical
Two partitioned data sets or two PDSEs are considered equal if:
Corresponding members contain the same number of records
Note lists are in the same position within corresponding members
Corresponding records and keys are identical
Corresponding directory user data fields are identical
If all these conditions are not met for a specific type of data set, those data sets are
considered unequal. If records are unequal, the record and block numbers, the names of the
DD statements that define the data sets, and the unequal records are listed in a message
data set. Ten successive unequal comparisons stop the job step, unless you provide a routine
for handling error conditions.

Note: Load module partitioned data sets that reside on different types of devices should
not be compared. Under most circumstances, the data sets will not compare as equal.
A partitioned data set or partitioned data set extended can be compared only if all names in
one or both directories have counterpart entries in the other directory. The comparison is
made on members identified by these entries and corresponding user data.
Recommendation: Use the SuperC utility instead of IEBCOMPR. SuperC is part of

ISPF/PDF and the High Level Assembler Toolkit Feature. SuperC can be processed in the
foreground as well as in batch, and its report is more useful.
Examples of comparing data sets

As mentioned, partitioned data sets or PDSEs can be compared only if all the names in one
or both of the directories have counterpart entries in the other directory. The comparison is
made on members that are identified by these entries and corresponding user data.
You can run this sample JCL to compare two cataloged, partitioned organized (PO) data sets:
//DISKDISK JOB ...

// EXEC PGM=IEBCOMPR
//SYSUT1 DD DSN=PDSE1,DISP=SHR
//SYSUT2 DD DSN=PDSE2,DISP=SHR
//SYSIN DD *
COMPARE TYPORG=PO
/*
Figure 4-3 IEBCOMPR sample
Figure 4-2 on page 110 shows several examples of the directories of two partitioned data
sets.
In Example 1, Directory 2 contains corresponding entries for all the names in Directory 1;
therefore, the data sets can be compared.
In Example 2, each directory contains a name that has no corresponding entry in the other
directory; therefore, the data sets cannot be compared, and the job step will be ended.

4.3 IEBCOPY utility
//COPY JOB ...

//JOBSTEP EXEC PGM=IEBCOPY
//OUT1 DD DSNAME=DATASET1,UNIT=disk,VOL=SER=111112,
// DISP=(OLD,KEEP)
//IN6 DD DSNAME=DATASET6,UNIT=disk,VOL=SER=111115,
// DISP=OLD
//IN5 DD DSNAME=DATASET5,UNIT=disk,VOL=SER=111116,
// DISP=(OLD,KEEP)
//SYSUT3 DD UNIT=SYSDA,SPACE=(TRK,(1))
//SYSIN DD *
COPY OUTDD=OUT1
INDD=IN5,IN6
SELECT MEMBER=((B,,R),A)
/*
Figure 4-4 IEBCOPY utility example
IEBCOPY utility
IEBCOPY is a data set utility used to copy or merge members between one or more
partitioned data sets (PDS), or partitioned data sets extended (PDSE), in full or in part. You
can also use IEBCOPY to create a backup of a partitioned data set into a sequential data set
(called an unload data set or PDSU), and to copy members from the backup into a partitioned
data set.
IEBCOPY is used to:

Make a copy of a PDS or PDSE
Merge partitioned data sets (except when unloading)
Create a sequential form of a PDS or PDSE for a backup or transport
Reload one or more members from a PDSU into a PDS or PDSE
Select specific members of a PDS or PDSE to be copied, loaded, or unloaded
Replace members of a PDS or PDSE
Rename selected members of a PDS or PDSE
Exclude members from a data set to be copied, unloaded, or loaded (except on
COPYGRP)
Compress a PDS in place
Upgrade an OS format load module for faster loading by MVS program fetch

Copy and reblock load modules
Convert load modules in a PDS to program objects in a PDSE when copying a PDS to a
PDSE
Convert a PDS to a PDSE, or a PDSE to a PDS
Copy to or from a PDSE data set a member and its aliases together as a group
(COPYGRP)
In addition, IEBCOPY automatically lists the number of unused directory blocks and the
number of unused tracks available for member records in the output partitioned data set.
INDD statement
This statement specifies the names of DD statements that locate the input data sets. When
an INDD=appears in a record by itself (that is, not with a COPY keyword), it functions as a
control statement and begins a new step in the current copy operation.
INDD=[(]{DDname|(DDname,R)}[,...][)]
R specifies that all members to be copied or loaded from this input data set are to replace
any identically named members on the output partitioned data set.
OUTDD statement
This statement specifies the name of a DD statement that locates the output data set.
OUTDD=DDname
SELECT statement
This statement selects specific members to be processed from one or more data sets by
coding a SELECT statement to name the members. Alternatively, all members but a specific
few can be designated by coding an EXCLUDE statement to name members not to be
processed.

4.4 IEBCOPY: Copy operation
DATA.SET5
DATA.SET1 DATA.SET1
Directory
AC
Unsued
Directory Member C Directory
ABF Unused ABCDF
A
Member F Available Member F
A DATA.SET6 A
Unused Unused
B B
Directory
Available D
BCD
Member B C
D
C After Copy
Before Copy Before Compress
Available
Figure 4-5 IEBCOPY copy operation
Copy control command example

In Figure 4-5, two input partitioned data sets (DATA.SET5 and DATA.SET6) are copied to an
existing output partitioned data set (DATA.SET1). In addition, all members on DATA.SET6 are
copied; members on the output data set that have the same names as the copied members
are replaced. After DATA.SET6 is processed, the output data set (DATA.SET1) is compressed
in place. Figure 4-5 shows the input and output data sets before and after copy processing.
The compress process is shown in Figure 4-7 on page 116. Figure 4-6 shows the job that is
used to copy and compress partitioned data sets.
//COPY JOB ...

//JOBSTEP EXEC PGM=IEBCOPY
//INOUT1 DD DSNAME=DATA.SET1,UNIT=disk,VOL=SER=111112,DISP=(OLD,KEEP)
//IN5 DD DSNAME=DATA.SET5,UNIT=disk,VOL=SER=111114,DISP=OLD
//IN6 DD DSNAME=DATA.SET6,UNIT=disk,VOL=SER=111115,
// DISP=(OLD,KEEP)
//SYSIN DD *
COPY OUTDD=INOUT1,INDD=(IN5,(IN6,R),INOUT1)
/*
Figure 4-6 IEBOPY with copy and compress

COPY control statement
In the control statement, note the following:
INOUT1 DD defines a partitioned data set (DATA.SET1) that contains three members (A,
B, and F).
IN5 DD defines a partitioned data set (DATA.SET5) that contains two members (A and C).
IN6 DD defines a partitioned data set (DATA.SET6), that contains three members (B, C,
and D).
SYSUT3 and SYSUT4 DD define temporary spill data sets. One track is allocated for each
DD statement on a disk volume.
SYSIN DD defines the control data set, which follows in the input stream. The data set
contains a COPY statement.
COPY indicates the start of the copy operation. The OUTDD operand specifies
DATA.SET1 as the output data set.
COPY processing
Processing occurs as follows:
1. Member A is not copied from DATA.SET5 into DATA.SET1 because it already exists on
DATA.SET1 and the replace option was not specified for DATA.SET5.
2. Member C is copied from DATA.SET5 to DATA.SET1, occupying the first available space.
3. All members are copied from DATA.SET6 to DATA.SET1, immediately following the last
member. Members B and C are copied even though the output data set already contains
members with the same names because the replace option is specified on the data set
level.
The pointers in the DATA.SET1 directory are changed to point to the new members B and C.
Thus, the space occupied by the old members B and C is unused.

4.5 IEBCOPY: Compress operation
DATA.SET1 DATA.SET1
Directory Directory
ABCDF ABCDF
Member F Member F
A A
Unused Compress B
B D
D C
C Available
After Copy,and After Copy,and

Before Compress After Compress
Figure 4-7 IEBCOPY compress operation
IEBCOPY compress operation

A partitioned data set will contain unused areas (sometimes called gas) where a deleted
member or the old version of an updated member once resided. This unused space is only
reclaimed when a partitioned data set is copied to a new data set, or after a
compress-in-place operation successfully completes. It has no meaning for a PDSE and is
ignored if requested.
The simplest way to request a compress-in-place operation is to specify the same ddname for
both the OUTDD and INDD parameters of a COPY statement.
Example
In our example in 4.4, IEBCOPY: Copy operation on page 114, the pointers in the
DATA.SET1 directory are changed to point to the new members B and C. Thus, the space
occupied by the old members B and C is unused. The members currently on DATA.SET1 are
compressed in place as a result of the copy operation, thereby eliminating embedded unused
space. However, be aware that a compress-in-place operation may bring risk to your data if
something abnormally disrupts the process.

4.6 IEBGENER utility
//COPY JOB ...

//STEP1 EXEC PGM=IEBGENER
//SYSUT1 DD DSNAME=INSET,DISP=SHR
//SYSUT2 DD DSNAME=OUTPUT,DISP=(,CATLG),
// SPACE=(CYL,(1,1)),DCB=*.SYSUT1
//SYSIN DD DUMMY
Figure 4-8 IEBGENER utility
Using IEBGENER
IEBGENER copies records from a sequential data set or converts sequential data sets into
members of PDSs or PDSEs. You can use IEBGENER to:
Create a backup copy of a sequential data set, a member of a partitioned data set or
PDSE, or a UNIX System Services file such as an HFS file.
Produce a partitioned data set or PDSE, or a member of a partitioned data set or PDSE,
from a sequential data set or a UNIX System Services file.
Expand an existing partitioned data set or PDSE by creating partitioned members and
merging them into the existing data set.
Produce an edited sequential or partitioned data set or PDSE.
Manipulate data sets containing double-byte character set data.
Print sequential data sets or members of partitioned data sets or PDSEs or UNIX System
Services files.
Re-block or change the logical record length of a data set.
Copy user labels on sequential output data sets.
Supply editing facilities and exits.
Jobs that call IEBGENER have a system-determined block size used for the output data set if
RECFM and LRECL are specified, but BLKSIZE is not specified. The data set is also
considered to be system-reblockable.

Copying to z/OS UNIX
IEBGENER can be used to copy from a PDS or PDSE member to a UNIX file. You also can
use TSO/E commands and other copying utilities such as ICEGENER or BPXCOPY to copy a
PDS or PDSE member to a UNIX file.
In Figure 4-9, the data set in SYSUT1 is a PDS or PDSE member and the data set in
SYSUT2 is a UNIX file. This job creates a macro library in the UNIX directory.
// JOB ....
// EXEC PGM=IEBGENER
//SYSUT1 DD DSN=PROJ.BIGPROG.MACLIB(MAC1),DISP=SHR
//SYSUT2 DD PATH='/u/BIGPROG/macros/special/MAC1',PATHOPTS=OCREAT,
// PATHDISP=(KEEP,DELETE),
// PATHMODE=(SIRUSR,SIWUSR,
// SIRGRP,SIROTH),
// FILEDATA=TEXT
//SYSIN DD DUMMY
Figure 4-9 Job to copy a PDS to a z/OS UNIX file
Note: If you have the DFSORT product installed, you should be using ICEGENER as an
alternative to IEBGENER when making an unedited copy of a data set or member. It may
already be installed in your system under the name IEBGENER. It generally gives better
performance.

4.7 IEBGENER: Adding members to a PDS
Utility control
statements
Sequential Input Existing
define record Expanded
groups
Member B Data Set Data Set
name members
LASTREC Directory Directory

Member D ACEG ACEGBDF
A
C Member A
LASTREC C
E
Member F E
G
G
B
Available D
F
Figure 4-10 Adding members to a PDS using IEBGENER
Adding members to a PDS

You can use IEBGENER to add members to a partitioned data set or PDSE. IEBGENER
creates the members from sequential input and adds them to the data set. The merge
operation, which is the ordering of the partitioned directory, is automatically performed by the
program.
Figure 4-10 shows how sequential input is converted into members that are merged into an
existing partitioned data set or PDSE. The left side of the figure shows the sequential input
that is to be merged with the partitioned data set or PDSE shown in the middle of the figure.
Utility control statements are used to divide the sequential data set into record groups and to
provide a member name for each record group. The right side of the figure shows the
expanded partitioned data set or PDSE.
Note that members B, D, and F from the sequential data set were placed in available space
and that they are sequentially ordered in the partitioned directory.

4.8 IEBGENER: Copying data to tape
MY.DATA IEBGENER
MY.DATA.OUTPUT
Figure 4-11 Copying data to tape
Copying data to tape example

You can use IEBGENER to copy data to tape. The example in Figure 4-11 copies the data set
MY.DATA to an SL cartridge. The data set name on tape is MY.DATA.OUTPUT.
//DISKTOTP JOB ...

//STEP1 EXEC PGM=IEBGENER
//SYSUT1 DD DSNAME=MY.DATA,DISP=SHR
//SYSUT2 DD DSNAME=MY.DATA.OUTPUT,UNIT=3490,DISP=(,KEEP),
// VOLUME=SER=IBM001,LABEL=(1,SL)
//SYSIN DD DUMMY
Figure 4-12 Copying data to tape with IEBGENER
For further information about IEBGENER, refer to z/OS DFSMSdfp Utilities, SC26-7414.

4.9 IEHLIST utility
//VTOCLIST JOB ...

//STEP1 EXEC PGM=IEHLIST
//DD2 DD UNIT=3390,VOLUME=SER=SBOXED,DISP=SHR
//SYSIN DD *
LISTVTOC VOL=3390=SBOXED,INDEXDSN=SYS1.VTOCIX.TOTTSB
/*
Figure 4-13 IEHLIST utility
Using IEHLIST
IEHLIST is a system utility used to list entries in the directory of one or more partitioned data
sets or PDSEs, or entries in an indexed or non-indexed volume table of contents. Any number
of listings can be requested in a single execution of the program.
Listing a PDS or PDSE directory

IEHLIST can list up to ten partitioned data set or PDSE directories at a time.
The directory of a partitioned data set is composed of variable-length records blocked into
256-byte blocks. Each directory block can contain one or more entries that reflect member or
alias names and other attributes of the partitioned members. IEHLIST can list these blocks in
edited and unedited format.
The directory of a PDSE, when listed, will have the same format as the directory of a
partitioned data set.
Listing a volume table of contents (VTOC)

IEHLIST can be used to list, partially or completely, entries in a specified volume table of
contents (VTOC), whether indexed or non-indexed. The program lists the contents of selected
data set control blocks (DSCBs) in edited or unedited form.

4.10 IEHLIST LISTVTOC output
CONTENTS OF VTOC ON VOL SBOXED <THIS VOLUME IS NOT SMS MANAGED>

THERE IS A 2 LEVEL VTOC INDEX
DATA SETS ARE LISTED IN ALPHANUMERIC ORDER
----------------DATA SET NAME--------------- CREATED DATE.EXP FILE TYPE SM
@LSTPRC@.REXX.SDSF.#MESSG#.$DATASET 2007.116 00.000 PARTITIONED
ADB.TEMP.DB8A.C0000002.AN154059.CHANGES 2007.052 00.000 SEQUENTIAL
ADMIN.DB8A.C0000051.AN171805.CHANGES 2007.064 00.000 SEQUENTIAL
ADMIN.DB8A.C0000062.AN130142.IFF 2007.066 00.000 PARTITIONED
ADMIN.DB8A.C0000090.SMAP.T0001 2007.067 00.000 SEQUENTIAL
ADMIN.DB8A.C0000092.AN153552.SHRVARS 2007.067 00.000 PARTITIONED
ADMIN.DB8A.C0000116.ULD.T0001 2007.068 00.000 SEQUENTIAL
ADMIN.DB8A.C0000120.SDISC.T0001 2007.068 00.000 SEQUENTIAL
ADMIN.DB8A.C0000204.AN120807.SHRVARS 2007.074 00.000 PARTITIONED
ADMIN.DB8A.C0000254.SERR.T0001 2007.078 00.000 SEQUENTIAL
BART.CG38V1.EMP4XML.DATA 2004.271 00.000 SEQUENTIAL
BART.MAZDA.IFCID180.FB80 2004.254 00.000 SEQUENTIAL
BART.PM32593.D060310.T044716.DMRABSUM.UNL 2006.072 00.000 SEQUENTIAL
BART.PM32593.D060317.T063831.DMRAPSUM.UNL 2006.079 00.000 SEQUENTIAL
BART.PM32593.D060324.T061425.DMRABSUM.UNL 2006.086 00.000 SEQUENTIAL
THERE ARE 45 EMPTY CYLINDERS PLUS 162 EMPTY TRACKS ON THIS VOLUME
THERE ARE 4238 BLANK DSCBS IN THE VTOC ON THIS VOLUME
THERE ARE 301 UNALLOCATED VIRS IN THE INDEX
Figure 4-14 IEHLIST LISTVTOC output
Obtaining the VTOC listing

Running the job shown in Figure 4-13 on page 121 produces a SYSOUT very similar to that
shown in Figure 4-14.
If you include the keyword FORMAT in the LISTVTOC parameter, you will have more detailed
information about the DASD and about the data sets, and you can also specify the DSNAME
that you want to request information about. If you specify the keyword DUMP instead of
FORMAT, you will get an unformatted VTOC listing.
Note: This information is at the DASD volume level, and does not have any interaction with
the catalog.

4.11 IEHINITT utility
Initializing Tape Cartridges

Create a tape label - (EBCDIC or ASCII)
Consider placement in an authorized library
//LABEL JOB ...

//STEP1 EXEC PGM=IEHINITT
//LABEL1 DD DCB=DEN=2,UNIT=(TAPE,1,DEFER)
//LABEL2 DD DCB=DEN=3,UNIT=(TAPE,1,DEFER)
//SYSIN DD *
LABEL1 INITT SER=TAPE1
LABEL2 INITT SER=001234,NUMBTAPE=2
/*
DFSMSrmm EDGINERS - the newer alternative
Figure 4-15 IEHINITT utility
IEHINITT utility
IEHINITT is a system utility used to place standard volume label sets onto any number of
magnetic tapes mounted on one or more tape units. They can be ISO/ANSI Version 3 or
ISO/ANSI Version 4 volume label sets written in American Standard Code for Information
Interchange (ASCII) or IBM standard labels written in EBCDIC.
Attention: Because IEHINITT can overwrite previously labeled tapes regardless of

expiration date and security protection, IBM recommends that the security administrator
use PROGRAM protection with the following sequence of RACF commands:
RDEFINE PROGRAM IEHINITT ADDMEM(SYS1.LINKLIB//NODPADCHK) UACC(NONE)
PERMIT IEHINITT CLASS(PROGRAM) ID(users or group) ACCESS(READ)
SETROPTS WHEN(PROGRAM) REFRESH
(Omit REFRESH if you did not have this option active previously.)
IEHINITT should be moved into an authorized password-protected private library, and

deleted from SYS1.LINKLIB.
To further protect against overwriting the wrong tape, IEHINITT asks the operator to verify
each tape mount.

Examples of IEHINITT
In the example in Figure 4-16, two groups of serial numbers, (001234, 001235, 001236, and
001334, 001335, 001336) are placed on six tape volumes. The labels are written in EBCDIC
at 800 bits per inch. Each volume labeled is mounted, when it is required, on a single 9-track
tape unit.
//LABEL3 JOB ...

//LABEL DD DCB=DEN=2,UNIT=(tape,1,DEFER)
//SYSIN DD *
LABEL INITT SER=001234,NUMBTAPE=3
LABEL INITT SER=001334,NUMBTAPE=3
/*
Figure 4-16 IEHINITT example to write EBCDIC labels in different densities
In Figure 4-17, serial numbers 001234, 001244, 001254, 001264, 001274, and so forth are
placed on eight tape volumes. The labels are written in EBCDIC at 800 bits per inch. Each
volume labeled is mounted, when it is required, on one of four 9-track tape units.
//LABEL4 JOB ...

//LABEL DD DCB=DEN=2,UNIT=(tape,4,DEFER)
//SYSIN DD *
LABEL INITT SER=001234
/*
Figure 4-17 IEHINITT Place serial number on eight tape volumes
DFSMSrmm EDGINERS utility

The EDGINERS utility program verifies that the volume is mounted before writing a volume
label on a labeled, unlabeled, or blank tape. EDGINERS checks security and volume
ownership, and provides auditing. DFSMSrmm must know that the volume needs to be
labelled. If the labelled volume is undefined, then DFSMSrmm defines it to DFSMSrmm and
can create RACF volume security protection.
Detailed procedures for using the program are described in z/OS DFSMSrmm
Implementation and Customization Guide, SC26-7405.
Note: DFSMSrmm is an optional priced feature of DFSMS. That means that EDGINERS
can only be used when DFSMSrmm is licensed. If DFSMSrmm is licensed, IBM
recommends that you use EDGINERS for tape initialization instead of using IEHINITT.

4.12 IEFBR14 utility
//DATASETS JOB ...

//STEP1 EXEC PGM=IEFBR14
//DD1 DD DSN=DATA.SET1,
// DISP=(OLD,DELETE,DELETE)
//DD2 DD DSN=DATA.SET2,
// DISP=(NEW,CATLG),UNIT=3390,
// VOL=SER=333001,
// SPACE=(CYL,(12,1,1),),
// DCB=(RECFM=FB,LRECL=80)
Figure 4-18 IEFBR14 program
IEFBR14 program
IEFBR14 is not a utility program. It is a two-line program that clears register 15, thus passing
a return code of 0. It then branches to the address in register 14, which returns control to the
system. So in other words, this program is dummy program. It can be used in a step to force
MVS (specifically, the initiator) to process the JCL code and execute functions such as the
following:
Checking all job control statements in the step for syntax
Allocating direct access space for data sets
Performing data set dispositions like creating new data sets or deleting old ones
Note: Although the system allocates space for data sets, it does not initialize the new data
sets. Therefore, any attempt to read from one of these new data sets in a subsequent step
may produce unpredictable results. Also, we do not recommend allocation of multi-volume
data sets while executing IEFBR14.
In the example in Figure 4-18 the first DD statement DD1 deletes old data set DATA.SET1.
The second DD statement creates a new PDS with name DATA.SET2.

4.13 DFSMSdfp access methods
DFSMSdfp provides several access methods for

formatting and accessing data, as follows:
Basic partitioned access method - (BPAM)
Basic sequential access method - (BSAM)
Object access method - (OAM) - OSREQ interface
Queued sequential access method - (QSAM)
Virtual storage access method - (VSAM)
DFSMS also supports the basic direct access
method (BDAM) for coexistence with previous
operating systems
Figure 4-19 DFSMSdfp access methods
Access methods
An access method is a friendly program interface between programs and their data. It is in
charge of interfacing with Input Output Supervisor (IOS), the z/OS code that starts the I/O
operation. An access method makes the physical organization of data transparent to you by:
Managing data buffers
Blocking and de-blocking logical records into physical blocks
Synchronizing your task and the I/O operation (wait/post mechanism)
Writing the channel program
Optimizing the performance characteristics of the control unit (such as caching and data
striping)
Compressing and decompressing I/O data
Executing software error recovery
In contrast to other platforms, z/OS supports several types of access methods and data
organizations.
An access method defines the organization by which the data is stored and retrieved. DFSMS
access methods have their own data set structures for organizing data, macros, and utilities
to define and process data sets. It is an application choice, depending on the type of access
(sequential or random), to allow or disallow insertions and deletions, to pick up the most
adequate access method for its data.

Access methods are identified primarily by the data set organization to which they apply. For
example, you can use the basic sequential access method (BSAM) with sequential data sets.
However, there are times when an access method identified with one data organization type
can be used to process a data set organized in a different manner. For example, a sequential
data set (not an extended format data set) created using BSAM can be processed by the
basic direct access method (BDAM), and vice versa.
Basic direct access method (BDAM)

BDAM arranges records in any sequence your program indicates, and retrieves records by
actual or relative address. If you do not know the exact location of a record, you can specify a
point in the data set where a search for the record is to begin. Data sets organized this way
are called direct data sets.
Optionally, BDAM uses hardware keys. Hardware keys are less efficient than the optional
software keys in VSAM KSDS.
Note: Because BDAM tends to require the use of device-dependent code, it is not a
recommended access method. In addition, using keys is much less efficient than in VSAM.
BDAM is supported by DFSMS only to enable compatibility with other IBM operating
systems.
Basic partitioned access method (BPAM)

BPAM arranges records as members of a partitioned data set (PDS) or a partitioned data set
extended (PDSE) on DASD. You can use BPAM to view a UNIX directory and its files as
though it were a PDS. You can view each PDS, PDSE, or UNIX member sequentially with
BSAM or QSAM. A PDS or PDSE includes a directory that relates member names to
locations within the data set. Use the PDS, PDSE, or UNIX directory to retrieve individual
members. A member is a sequential file contained in the PDS or PDSE data set. When
members contain load modules (executable code in PDS) or program objects (executable
code in PDSE), the directory contains program attributes that are required to load and rebind
the member. Although UNIX files can contain program objects, program management does
not access UNIX files through BPAM.
For information about partitioned organized data set, see 4.22, Partitioned organized (PO)
data sets on page 143, and subsequent sections.
Basic sequential access method

BSAM arranges logical records sequentially in the order in which they are entered. A data set
that has this organization is a sequential data set. Blocking, de-blocking and the I/O
synchronization is done by the application program. This is basic access. You can use BSAM
with the following data types:
z/OS UNIX files
See also 4.28, Sequential access methods on page 154.
Queued sequential access method (QSAM)

QSAM arranges logical records sequentially in the order that they are entered to form
sequential data sets, which are the same as those data sets that BSAM creates. The system
organizes records with other records. QSAM anticipates the need for records based on their
order. To improve performance, QSAM reads these records into main storage before they are
requested. This is called queued access. QSAM blocks and de-blocks logical records into

physical blocks. QSAM also guarantees the synchronization between the task and the I/O
operation. You can use QSAM with the same data types as BSAM. See also 4.28, Sequential
access methods on page 154.
Object Access Method (OAM)

OAM processes very large named byte streams (objects) that have no record boundary or
other internal orientation as image data. These objects can be recorded in a DB2 data base
or on an optical storage volume. For information about OAM, see z/OS DFSMS Object
Access Method Application Programmers Reference, SC35-0425, and z/OS DFSMS Object
Access Method Planning, Installation, and Storage Administration Guide for Object Support,
SC35-0426.
Virtual Storage Access Method (VSAM)

VSAM is an access method that has several ways of organizing data, depending on the
applications needs.
VSAM arranges and retrieves logical records by an index key, relative record number, or
relative byte addressing (RBA). A logical record has an RBA, which is the relative byte
address of its first byte in relation to the beginning of the data set. VSAM is used for direct,
sequential or skip sequential processing of fixed-length and variable-length records on DASD.
VSAM data sets (also named clusters) are always cataloged. There are five types of cluster
organization:
This contains records in the order in which they were entered. Records are added to the
end of the data set and can be accessed sequentially or randomly through the RBA.
This contains records in ascending collating sequence of the contents of a logical record
field called key. Records can be accessed by the contents of such key, or by an RBA.
This contains data that has no record boundaries. Linear data sets contain none of the
control information that other VSAM data sets do. Data in Virtual (DIV) is an optional
intelligent buffering technique that includes a set of assembler macros that provide
buffering access to VSAM linear data sets. See 4.41, VSAM: Data-in-virtual (DIV) on
page 174.
Relative record data set (RRDS)
This contains logical records in relative record number order; the records can be accessed
sequentially or randomly based on this number. There are two types of relative record data
sets:
Fixed-length RRDS: logical records must be of fixed length.
Variable-length RRDS: logical records can vary in length.
A z/OS UNIX file (HFS or zFS) can be accessed as though it were a VSAM entry-sequenced
data set (ESDS). Although UNIX files are not actually stored as entry-sequenced data sets,
the system attempts to simulate the characteristics of such a data set. To identify or access a
UNIX file, specify the path that leads to it.

4.14 Access method services (IDCAMS)
Access method services is a utility to:

Define VSAM data sets
Maintain catalogs
Command interface to manage VSAM and catalogs
Functional commands
Modal commands
Invoke IDCAMS utility program
Using JCL jobs
From a TSO session
From a user program
Figure 4-20 Access method services
Access method services

You can use the access method services utility (also know as IDCAMS) to establish and
maintain catalogs and data sets (VSAM and non-VSAM). It is used mainly to create and
manipulate VSAM data sets. IDCAMS has other functions (such as catalog updates), but it is
most closely associated with the use of VSAM.
Access method services commands

There are two types of access method services commands:
Functional commands Used to request the actual work (for example, defining a data set or
listing a catalog)
Modal commands Allow the conditional execution of functional commands (to make it
look like a language)
All access method services commands have the following general structure:
COMMAND parameters ... [terminator]
The command defines the type of service requested; the parameters further describe the
service requested; the terminator indicates the end of the command statement.
Time Sharing Option (TSO) users can use functional commands only. For more information
about modal commands, refer to z/OS DFSMS Access Method Services for Catalogs,
SC26-7394.

The automatic class selection (ACS) routines (established by your storage administrator) and
the associated SMS classes eliminate the need to use many access method services
command parameters. The SMS environment is discussed in more detail in Chapter 5,
System-managed storage on page 239.
Invoking the IDCAMS utility program

When you want to use an access method services function, enter a command and specify its
parameters. Your request is decoded one command at a time; the appropriate functional
routines perform all services required by that command.
You can call the access method services program in the following ways:
As a job or jobstep
From a TSO session
From within your own program
TSO users can run access method services functional commands from a TSO session as
though they were TSO commands.
For more information, refer to Invoking Access Method Services from Your Program in z/OS
DFSMS Access Method Services for Catalogs, SC26-7394.
As a job or jobstep
You can use JCL statements to call access method services. PGM=IDCAMS identifies the
access method services program, as shown in Figure 4-21.
//YOURJOB JOB YOUR INSTALLATION'S JOB=ACCOUNTING DATA

//STEP1 EXEC PGM=IDCAMS
//SYSIN DD *
access method services commands and their parameters
/*
Figure 4-21 JCL statements to call IDCAMS
From a TSO session

You can use TSO with VSAM and access method services to:
Run access method services commands
Run a program to call access method services
Each time you enter an access method services command as a TSO command, TSO builds
the appropriate interface information and calls access method services. You can enter one
command at a time. Access method services processes the command completely before
TSO lets you continue processing. Except for ALLOCATE, all the access method services
functional commands are supported in a TSO environment.
To use IDCAMS and certain of its parameters from TSO/E, you must update the IKJTSOxx
member of SYS1.PARMLIB. Add IDCAMS to the list of authorized programs (AUTHPGM). For
more information, see z/OS DFSMS Access Method Services for Catalogs, SC26-7394.
From within your own program

You can also call the IDCAMS program from within another program and pass the command
and its parameters to the IDCAMS program.

4.15 IDCAMS functional commands
ALTER: alters attributes of exiting data sets

DEFINE CLUSTER: creates/catalogs VSAM cluster
DEFINE GENERATIONDATAGROUP: defines
GDG data sets
DEFINE PAGESPACE: creates page data sets
EXPORT: exports VSAM DS, AIX or ICF catalog
IMPORT: imports VSAM DS, AIX or ICF catalog
LISTCAT: lists catalog entries
REPRO: copies VSAM, non-VSAM and catalogs
VERIFY: corrects end-of-file information for VSAM
clusters in the catalog
Figure 4-22 Functional commands
IDCAMS functional commands

Table 4-3 lists and describes the functional commands.
Table 4-3 Functional commands

Command Functions
ALLOCATE Allocates VSAM and non-VSAM data sets.
ALTER Alters attributes of data sets, catalogs, tape library entries, and tape
volume entries that have already been defined.
BLDINDEX Builds alternate indexes (AIX) for existing VSAM base clusters.
CREATE Creates tape library entries and tape volume entries.
DCOLLECT Collects data set, volume usage, and migration utility information.
DEFINE ALIAS Defines an alternate name for a user catalog or a non-VSAM data set.
DEFINE ALTERNATEINDEX Defines an alternate index for a KSDS or ESDS VSAM data set.
DEFINE CLUSTER Creates KSDS, ESDS, RRDS, VRRDS and linear VSAM data sets.
DEFINE Defines a catalog entry for a generation data group (GDG).

GENERATIONDATAGROUP
DEFINE NONVSAM Defines a catalog entry for a non-VSAM data set.

Command Functions
DEFINE PAGESPACE Defines an entry for a page space data set.
DEFINE PATH Defines a path directly over a base cluster or over an alternate index
and its related base cluster.
DEFINE USERCATALOG Defines a user catalog.
DELETE Deletes catalogs, VSAM clusters, and non-VSAM data sets.
DIAGNOSE Scans an integrated catalog facility basic catalog structure (BCS) or a

VSAM volume data set (VVDS) to validate the data structures and
detect structure errors.
EXAMINE Analyzes and reports the structural consistency of either an index or

data component of a KSDS VSAM data set cluster.
EXPORT Disconnects user catalogs, and exports VSAM clusters and

ICF catalog information about the cluster.
EXPORT DISCONNECT Disconnects a user catalog.
IMPORT Connects user catalogs, and imports VSAM cluster and its ICF catalogs
information.
IMPORT CONNECT Connects a user catalog or a volume catalog.
LISTCAT Lists catalog entries.
PRINT Used to print VSAM data sets, non-VSAM data sets, and catalogs.
REPRO Performs the following functions:

Copies VSAM and non-VSAM data sets, user catalogs, master
catalogs, and volume catalogs.
Splits ICF catalog entries between two catalogs.
Merges ICF catalog entries into another ICF user or master
catalog.
Merges tape library catalog entries from one volume catalog into
another volume catalog.
SHCDS Lists SMSVSAM recovery associated with subsystems spheres and

controls that allow recovery of a VSAM RLS environment.
VERIFY Causes a catalog to correctly reflect the end of a data set after an error
occurred while closing a VSAM data set. The error might have caused
the catalog to be incorrect.
For a complete description of all AMS commands, see z/OS DFSMS Access Method
Services for Catalogs, SC26-7394.

4.16 AMS modal commands
IF-THEN-ELSE command sequence controls

command execution on the basis of condition codes
returned by previous commands
NULL command specifies no action be taken
DO-END command sequence specifies more than
one functional access method services command
and its parameters
SET command resets condition codes
CANCEL command terminates processing of the
current sequence of commands
PARM command specifies diagnostic aids and
printed output options
Figure 4-23 AMS modal commands
AMS modal commands

With Access Method Services (AMS), you can set up jobs to execute a sequence of modal
commands with a single invocation of IDCAMS. Modal command execution depends on the
success or failure of prior commands.
Using modal commands

Figure 4-23 lists and briefly describes the AMS modal commands. With access method
services, you can set up jobs to execute a sequence of modal commands with a single
invocation of IDCAMS. Modal command execution depends on the success or failure of prior
commands. The access method services modal commands are used for the conditional
execution of functional commands. The commands are:
The IF-THEN-ELSE command sequence controls command execution on the basis of
condition codes
The NULL command causes the program to take no action.
The DO-END command sequence specifies more than one functional access method
services command and its parameters.
The SET command resets condition codes.
The CANCEL command ends processing of the current job step.
The PARM command chooses diagnostic aids and options for printed output.

Note: These commands cannot be used when access method services is run in TSO. See
z/OS DFSMS Access Method Services for Catalogs, SC26-7394, for a complete
description of the AMS modal commands.
Commonly used single job step command sequences

A sequence of commands commonly used in a single job step includes DELETE-DEFINE-REPRO
or DELETE-DEFINE-BLDINDEX, as follows:
You can specify either a data definition (DD) name or a data set name with these
commands.
When you refer to a DD name, allocation occurs at job step initiation. The allocation can
result in a job failure if a command such as REPRO follows a DELETE-DEFINE sequence that
changes the location (volser) of the data set. (Such failures can occur with either
SMS-managed data sets or non-SMS-managed data sets.)
Avoiding potential command sequence failures

To avoid potential failures with a modal command sequence in your IDCAMS job, perform
either one of the following tasks:
Specify the data set name instead of the DD name.
Or, use a separate job step to perform any sequence of commands (for example, REPRO,
IMPORT, BLDINDEX, PRINT, or EXAMINE) that follow a DEFINE command.

4.17 DFSMS Data Collection Facility (DCOLLECT)
DCOLLECT functions
Capacity planning
Active data sets
VSAM clusters
Migrated data sets
Backed-up data sets
SMS configuration information
Figure 4-24 Data Collection Facility
DFSMS Data Collection Facility (DCOLLECT)

The DFSMS Data Collection Facility (DCOLLECT) is a function of access method services.
DCOLLECT collects data in a sequential file that you can use as input to other programs or
applications.
The IDCAMS DCOLLECT command collects DASD performance and space occupancy data in
a sequential file that you can use as input to other programs or applications.
An installation can use this command to collect information about:

Active data sets: DCOLLECT provides data about space use and data set attributes and
indicators on the selected volumes and storage groups.
VSAM data set information: DCOLLECT provides specific information relating to VSAM
data sets residing on the selected volumes and storage groups.
Volumes: DCOLLECT provides statistics and information about volumes that are selected
for collection.
Inactive data: DCOLLECT produces output for DFSMShsm-managed data (inactive data
management), which includes both migrated and backed up data sets.
Migrated data sets: DCOLLECT provides information about space utilization and data
set attributes for data sets migrated by DFSMShsm.
Backed up data sets: DCOLLECT provides information about space utilization and
data set attributes for every version of a data set backed up by DFSMShsm.

Capacity planning: Capacity planning for DFSMShsm-managed data (inactive data
management) includes the collection of both DASD and tape capacity planning.
DASD Capacity Planning: DCOLLECT provides information and statistics for volumes
managed by DFSMShsm (ML0 and ML1).
Tape Capacity Planning: DCOLLECT provides statistics for tapes managed by
DFSMShsm.
SMS configuration information: DCOLLECT provides information about the SMS
configurations. The information can be from either an active control data set (ACDS) or a
source control data set (SCDS), or the active configuration.
Data is gathered from the VTOC, VVDS, and DFSMShsm control data set for both managed
and non-managed storage. ISMF provides the option to build the JCL necessary to execute
DCOLLECT.
DCOLLECT example
With the sample JCL shown in Figure 4-25 you can gather information about all volumes
belonging to storage group STGGP001.
//COLLECT2 JOB ...

//OUTDS DD DSN=USER.DCOLLECT.OUTPUT,
// STORCLAS=LARGE,
// DSORG=PS,
// DCB=(RECFM=VB,LRECL=644,BLKSIZE=0),
// SPACE=(1,(100,100)),AVGREC=K,
// DISP=(NEW,CATLG,KEEP)
//SYSIN DD *
DCOLLECT -
OFILE(OUTDS) -
STORAGEGROUP(STGGP001) -
NODATAINFO
/*
Figure 4-25 DCOLLECT job to collect information about all volumes in on storage group
Many clients run DCOLLECT on a time-driven basis triggered by automation. In such a

scenario, the peak hours should be avoided due to the heavy access rate caused by
DCOLLECT in VTOC and catalogs.

4.18 Generation data group (GDG)
VOLABC
User catalog ABC.GDG.G0001V00
ABC.GDG.G0003V00
GDG base: ABC.GDG ABC.GDG.G0004V00
GDS: oldest
ABC.GDG.G0001V00 (-4)
ABC.GDG.G0002V00 (-3)
ABC.GDG.G0003V00 (-2)
VOLDEF
ABC.GDG.G0004V00 (-1)
ABC.GDG.G0002V00
ABC.GDG.G0005V00 ( 0)
Limit = 5 newest
ABC.GDG.G0005V00
Figure 4-26 Generation data group (GDG)
Generation data group

Generation data group (GDG) is a catalog function that makes it easier to process data sets
with the same type of data but in different update levels. For example, the same data set is
produced every day but with different data. Then, you can catalog successive updates or
generations of these related data sets. They are called generation data groups (GDG). Each
data set within a GDG is called a generation data set (GDS) or generation.
Within a GDG, the generations can have like or unlike DCB attributes and data set
organizations. If the attributes and organizations of all generations in a group are identical,
the generations can be retrieved together as a single data set.
Generation data sets can be sequential, PDSs, or direct (BDAM). Generation data sets
cannot be PDSEs, UNIX files, or VSAM data sets. The same GDG may contain SMS and
non-SMS data sets.
There are usability benefits to grouping related data sets using a function such as GDS. For
example, the catalog management routines can refer to the information in a special index
called a generation index in the catalog, and as a result:
All data sets in the group can be referred to by a common name.
z/OS is able to keep the generations in chronological order.
Outdated or obsolete generations can be automatically deleted from the catalog by z/OS.
Another benefit is the ability to reference a new generation using the same JCL.

A GDS has sequentially ordered absolute and relative names that represent its age. The
catalog management routines use the absolute generation name in the catalog. Older data
sets have smaller absolute numbers. The relative name is a signed integer used to refer to the
most current (0), the next to most current (-1), and so forth, generation. See also 4.20,
Absolute generation and version numbers on page 141 and 4.21, Relative generation
numbers on page 142 for more information about this topic.
A generation data group (GDG) base is allocated in a catalog before the GDSs are
cataloged. Each GDG is represented by a GDG base entry. Use the access method services
DEFINE command to allocate the GDG base (see also 4.19, Defining a generation data
group on page 139).
The GDG base is a construct that exists only in a user catalog, it does not exist as a data set
on any volume. The GDG base is used to maintain the generation data sets (GDS), which are
the real data sets.
The number of GDSs in a GDG depends on the limit you specify when you create a new
GDG in the catalog.
GDG example
In our example in Figure 4-26 on page 137, the limit is 5. That means, the GDG can hold a
maximum of five GDSs. Our data set name is ABC.GDG. Then, you can access the GDSs by
their relative names; for example, ABC.GDG(0) corresponds to the absolute name
ABC.GDG.G0005V00. ABC.GDG(-1) corresponds to generation ABC.GDG.G0004V00, and
so on. The relative number can also be used to catalog a new generation (+1), which will be
generation number 6 with an absolute name of ABC.GDG.G0006V00. Because the limit is 5,
the oldest generation (G0001V00) is rolled-off if you define a new one.
Rolled in and rolled off

When a GDG contains its maximum number of active generation data sets in the catalog,
defined in the LIMIT parameter, and a new GDS is rolled in at the end-of-job step, the oldest
generation data set is rolled off from the catalog. If a GDG is defined using DEFINE
GENERATIONDATAGROUP EMPTY and is at its limit, then when a new GDS is rolled in, all the
currently active GDSs are rolled off.
The parameters you specify on the DEFINE GENERATIONDATAGROUP IDCAMS command determine
what happens to rolled-off GDSs. For example, if you specify the SCRATCH parameter, the GDS
is scratched from VTOC when it is rolled off. If you specify the NOSCRATCH parameter, the
rolled-off generation data set is re-cataloged as rolled off and is disassociated with its
generation data group.
GDSs can be in a deferred roll-in state if the job never reached end-of-step or if they were
allocated as DISP=(NEW,KEEP) and the data set is not system-managed. However, GDSs in
a deferred roll-in state can be referred to by their absolute generation numbers. You can use
the IDCAMS command ALTER ROLLIN to roll in these GDSs.
For further information about generation data groups, see z/OS DFSMS: Using Data Sets,
SC26-7410.

4.19 Defining a generation data group
//DEFGDG1 JOB ...

//GDGMOD DD DSNAME=ABC.GDG,DISP=(,KEEP),
// SPACE=(TRK,(0)),UNIT=DISK,VOL=SER=VSER03, Model DSCB
// DCB=(RECFM=FB,BLKSIZE=2000,LRECL=100)
//SYSIN DD *
DEFINE GENERATIONDATAGROUP - VSER03
(NAME(ABC.GDG) -
}
VTOC
EMPTY -
NOSCRATCH - A
LIMIT(255)) B ABC.GDG C
/*
A
C
available space
Figure 4-27 Defining a GDG
Defining a generation data group

The DEFINE GENERATIONDATAGROUP command creates a catalog entry for a generation data
group (GDG).
Figure 4-28 shows the JCL to define a GDG.
//DEFGDG1 JOB ...

//SYSIN DD *
DEFINE GENERATIONDATAGROUP -
(NAME(ABC.GDG) -
NOEMPTY -
NOSCRATCH -
LIMIT(5) )
/*
Figure 4-28 JCL to define a GDG catalog entry
The DEFINE GENERATIONDATAGROUP command defines a GDG base catalog entry GDG01.

The parameters are:
NAME This specifies the name of the ABC.GDG. Each GDS in the group will have the
name ABC.GDG.GxxxxVyy, where xxxx is the generation number and yy is the
version number. See Absolute generation and version numbers on page 141.
NOEMPTY This specifies that only the oldest generation data set is to be uncataloged
when the maximum is reached (recommended).
EMPTY This specifies that all GDSs in the group are to be uncataloged when the group
reaches the maximum number of data sets (as specified by the LIMIT
parameter) and one more GDS is added to the group.
NOSCRATCH This specifies that when a data set is uncataloged, its DSCB is not to be
removed from its volume's VTOC. Therefore, even if a data set is uncataloged,
its records can be accessed when it is allocated to a job step with the
appropriate JCL DD statement.
LIMIT This specifies that the maximum number of GDG data sets in the group is 5.
This parameter is required.
Figure 4-29 shows a generation data set defined within the GDG by using JCL statements.
//DEFGDG2 JOB ...

//STEP1 EXEC PGM=IEFBR14
//GDGDD1 DD DSNAME=ABC.GDG(+1),DISP=(NEW,CATLG),
// SPACE=(TRK,(10,5)),VOL=SER=VSER03,
// UNIT=DISK
//SYSIN DD *
/*
Figure 4-29 JCL to define a generation data set
The job DEFGDG2 allocates space and catalogs a GDG data set in the newly-defined GDG.
The job control statement GDGDD1 DD specifies the GDG data set in the GDG.
Creating a model DSCB

As Figure 4-27 on page 139 shows, you can also create a model DSCB with the same name
as the GDG base. You can provide initial DCB attributes when you create your model;
however, you need not provide any attributes now. Because only the attributes in the data set
label are used, allocate the model data set with SPACE=(TRK,0) to conserve direct access
space. You can supply initial or overriding attributes creating and cataloging a generation.
Only one model DSCB is necessary for any number of generations. If you plan to use only
one model, do not supply DCB attributes when you create the model. When you subsequently
create and catalog a generation, include necessary DCB attributes in the DD statement
referring to the generation. In this manner, any number of GDGs can refer to the same model.
The catalog and model data set label are always located on a direct access volume, even for
a magnetic tape GDG.
Restriction: You cannot use a model DSCB for system-managed generation data sets.

4.20 Absolute generation and version numbers
Generation Data Set 1, Version 0,

A.B.C.G0001V00 in generation data group A.B.C
Generation Data Set 9, Version 1,

A.B.C.G0009V01 in generation data group A.B.C
Figure 4-30 Absolute generation and version numbers
Absolute generation and version numbers

An absolute generation and version number is used to identify a specific generation of a
GDG. The same GDS may have different versions, which are maintained by your installation.
The version number allows you to perform normal data set operations without disrupting the
management of the generation data group. For example, if you want to update the second
generation in a three-generation group, then replace generation 2, version 0, with
generation 2, version 1. Only one version is kept for each generation.
The generation and version number are in the form GxxxxVyy, where xxxx is an unsigned
four-digit decimal generation number (0001 through 9999) and yy is an unsigned two-digit
decimal version number (00 through 99). For example:
A.B.C.G0001V00 is generation data set 1, version 0, in generation data group A.B.C.
A.B.C.G0009V01 is generation data set 9, version 1, in generation data group A.B.C.
The number of generations and versions is limited by the number of digits in the absolute
generation name; that is, there can be 9,999 generations. Each generation can have 100
versions. The system automatically maintains the generation number.
You can catalog a generation using either absolute or relative numbers. When a generation is
cataloged, a generation and version number is placed as a low-level entry in the generation
data group. To catalog a version number other than V00, you must use an absolute
generation and version number.

4.21 Relative generation numbers
A.B.C.G0005V00 = A.B.C(-1)
READ/UPDATE OLD GDS
A.B.C.G0006V00 = A.B.C(0)
DEFINE NEW GDS
A.B.C.G0007V00 = A.B.C(+1)
Figure 4-31 Relative generation numbers
Relative generation numbers

As an alternative to using absolute generation and version numbers when cataloging or
referring to a generation, you can use a relative generation number. To specify a relative
number, use the generation data group name followed by a negative integer, a positive
integer, or a zero (0), enclosed in parentheses; for example, A.B.C(-1), A.B.C(+1), or
A.B.C(0).
The value of the specified integer tells the operating system what generation number to
assign to a new generation data set, or it tells the system the location of an entry representing
a previously cataloged old generation data set.
When you use a relative generation number to catalog a generation, the operating system
assigns an absolute generation number and a version number of V00 to represent that
generation. The absolute generation number assigned depends on the number last assigned
and the value of the relative generation number that you are now specifying. For example, if in
a previous job generation, A.B.C.G0006V00 was the last generation cataloged, and you
specify A.B.C(+1), the generation now cataloged is assigned the number G0007V00.
Though any positive relative generation number can be used, a number greater than 1 can
cause absolute generation numbers to be skipped for a new generation data set. For
example, if you have a single step job and the generation being cataloged is a +2, one
generation number is skipped. However, in a multiple step job, one step might have a +1 and
a second step a +2, in which case no numbers are skipped. The mapping between relative
and absolute numbers is kept until the end of the job.

4.22 Partitioned organized (PO) data sets
(PO) data set (PDS)
PO.DATA.SET
DIRECTORY
A
B C
A B
MEMBERS
C
Figure 4-32 Partitioned organized (PO) data set
Partitioned organized (PO) data sets

Partitioned data sets are similar in organization to a library and are often referred to this way.
Normally, a library contains a great number of books, and sorted directory entries are used
to locate them.
In a partitioned organized data set, the books are called members, and to locate them, they
are pointed to by entries in a directory, as shown in Figure 4-32.
The members are individual sequential data sets and can be read or written sequentially, after
they have been located by directory. Then, the records of a given member are written or
retrieved sequentially.
Partitioned data sets can only exist on DASD. Each member has a unique name, one to eight
characters in length, and is stored in a directory that is part of the data set.
The main benefit of using a PO data set is that, without searching the entire data set, you can
retrieve any individual member after the data set is opened. For example, in a program library
(always a partitioned data set) each member is a separate program or subroutine. The
individual members can be added or deleted as required.
There are two types of PO data sets:


4.23 PDS data set organization
Advantages of the PDS organization:

Easier management: Processed by member or a
whole. Members can be concatenated and processed
as sequential files
Space savings: Small members fit in one DASD track
Good usability: Easily accessed via JCL, ISPF, TSO
Required improvements for the PDS organization:
Release space when a member is deleted without the
need to compress
Expandable directory size
Improved directory and member integrity
Better performance for directory search
Improving sharing facilities
Figure 4-33 PDS data organization

A PDS is stored in only one DASD device. It is divided into sequentially organized members,
each described by one or more directory entries. It is an MVS data organization that offers
such useful features as:
Easier management, which makes grouping related data sets under a single name and
managing MVS data easier. Files stored as members of a PDS can be processed
individually, or all the members can be processed as a unit.
Space savings so that small members fit in just one DASD track.
Good usability so that members of a PDS can be used as sequential data sets, and they
can be concatenated to sequential data sets. They are also easy to create with JCL, or
ISPF, and they are easy to manipulate with ISPF utilities or TSO commands.
However, there are requirements for improvement regarding PDS organization:

There is no mechanism to reuse the area that contained a deleted or rewritten member.
This unused space must be reclaimed by the use of the IEBCOPY utility function called
compression. See IEBCOPY utility on page 112 for information about this utility.
Directory size is not expandable, causing an overflow exposure. The area for members
can grow using secondary allocations. However, this is not true for the directory.

A PDS has no mechanism to prevent a directory from being overwritten if a program
mistakenly opens it for sequential output. If this happens, the directory is destroyed and all
the members are lost.
Also, PDS DCB attributes can be easily changed by mistake. If you add a member whose
DCB characteristics differ from those of the other members, you change the DCB
attributes of the entire PDS, and all the old members become unusable. In this case there
is a workaround in the form of code to correct the problem.
Better directory search time: Entries in the directory are physically ordered by the collating
sequence of the names in the members they are pointing to. Any inclusion may cause the
full rearrange of the entries.
There is also no index to the directory entries. The search is sequential using a CKD
format. If the directory is big, the I/O operation takes more time. To minimize this, a strong
directory buffering has been used (library lookaside (LLA)) for load modules in z/OS.
Improved sharing facilities: To update or create a member of a PDS, you need exclusive
access to the entire data set.
All these improvements require almost total compatibility, at the program level and the user
level, with the old PDS.
Allocating space for a PDS

To allocate a PDS, specify PDS in the DSNTYPE parameter and the number of directory
blocks in the SPACE parameter, in either the JCL or the SMS data class. You must specify the
number of directory blocks, or the allocation fails.
If your data set is large, or if you expect to update it extensively, it might be best to allocate a
large space. A PDS cannot occupy more than 65,535 tracks and cannot extend beyond one
volume. If your data set is small or is seldom changed, let SMS calculate the space
requirements to avoid wasted space or wasted time used for recreating the data set.
Space for the directory is expressed in 256 byte blocks. Each block contains from 3 to 21
entries, depending on the length of the user data field. If you expect 200 directory entries,
request at least 10 blocks. Any unused space on the last track of the directory is wasted
unless there is enough space left to contain a block of the first member.
The following DD statement defines a new partitioned data set:

//DD2 DD DSNAME=OTTO.PDS12,DISP=(,CATLG),
// SPACE=(CYL,(5,2,10),,CONTIG)
The system allocates five cylinders to the data set, of which ten 256-byte records are for a
directory. Since the CONTIG subparameter is coded, the system allocates 10 contiguous
cylinders on the volume. The secondary allocation is two cylinders, which is needed when the
data set needs to expand beyond the five cylinders primary allocation.

4.24 Partitioned data set extended (PDSE)
CREATION
DATA CLASS CONSTRUCT
SMS
VOLUME + DSNTYPE=LIBRARY
// DD DSNTYPE=LIBRARY
CONVERSION
DFDSS CONVERT (PDS I PDSE)
USE BSAM, QSAM, BPAM
Figure 4-34 PDSE structure

Partitioned data set extended (PDSE) is a type of data set organization that improves the
partition data set (PDS) organization. It has an improved indexed directory structure and a
different member format. You can use PDSE for source (programs and text) libraries, macros,
and program object (the name of executable code when loaded in PDSE) libraries.
Logically, a PDSE directory is similar to a PDS directory. It consists of a series of directory

records in a block. Physically, it is a set of pages at the front of the data set, plus additional
pages interleaved with member pages. Five directory pages are initially created at the same
time as the data set.
New directory pages are added, interleaved with the member pages, as new directory entries
are required. A PDSE always occupies at least five pages of storage.
The directory is like a KSDS index structure (KSDS is covered in 4.34, VSAM key sequenced
cluster (KSDS) on page 166), making a search much faster. It cannot be overwritten by being
opened for sequential output.
If you try to add a member with DCB characteristics that differs from the rest of the members,
you will get an error.
There is no longer a need for a PDSE data set to be SMS-managed.

Advantages of PDSEs
PDSE advantages when compared with PDS are:
The size of a PDSE directory is flexible and can expand to accommodate the number of
members stored in it (the size of a PDS directory is fixed at allocation time).
PDSE members are indexed in the directory by member name. This eliminates the need
for time-consuming sequential directory searches holding channels for a long time.
The logical requirements of the data stored in a PDSE are separated from the physical
(storage) requirements of that data, which simplifies data set allocation.
PDSE automatically reuses space, without the need for an IEBCOPY compress. A list of
available space is kept in the directory. When a PDSE member is updated or replaced, it is
written in the first available space. This is either at the end of the data set, or in a space in
the middle of the data set marked for reuse. For example, by moving or deleting a PDSE
member, you free space that is immediately available for the allocation of a new member.
This makes PDSEs less susceptible to space-related abends than are PDSs. This space
does not have to be contiguous. The objective of the space reuse algorithm is to avoid
extending the data set unnecessarily.
The number of PDSE members stored in the library can be large or small without concern
for performance or space considerations.
You can open a PDSE member for output or update without locking the entire data set.
The sharing control is at the member level, not the data set level.
The ability to update a member in place is possible with PDSs and PDSEs. But with
PDSEs, you can extend the size of members and the integrity of the library is maintained
while simultaneous changes are made to separate members within the library.
The maximum number of extents of a PDSE is 123; the PDS is limited to 16.
PDSEs are device-independent because they do not contain information that depends on
location or device geometry.
All members of a PDSE are re-blockable.
PDSEs can contain program objects built by the program management binder that cannot
be stored in PDSs.
You can share PDSEs within and across systems using PDSESHARING(EXTENDED) in
the IGDSMSxx member in the SYS1.PARMLIB. Multiple users are allowed to read PDSE
members while the data set is open. You can extend the sharing to enable multiple users
on multiple systems to concurrently create new PDSE members and read existing
members (see also SMS PDSE support on page 302).
For performance reasons, directories can be buffered in dataspaces and members can be
buffered in hiperspaces by using the MSR parameter in the SMS storage class.
Replacing a member without replacing all of its aliases deletes all aliases.
An unlimited number of tracks per volume are now available for PDSEs, that is, more than
the previous limit of 65535.
Restriction: You cannot use a PDSE for certain system data sets that are opened in the
IPL/NIP time frame.

4.25 PDSE enhancements
PDSE, two address spaces

SMXC in charge of PDSE serialization
SYSBMAS, the owner of DS and HS buffering
z/OS V1R6 combines both in a single address space
called SMSPDSE and improves the following:
Reducing excessive ECSA usage
Reducing re-IPLs due to system hangs in failure or
CANCEL situation
Providing tools for monitoring and diagnosis through
VARY SMS,PDSE,ANALYSIS command
Finally, a restartable SMSPDSE1 in charge of all
allocated PDSEs, except the ones in the LNKLST
controlled by SMSPDSE
Figure 4-35 PDSE enhancements
PDSE enhancements
Recent enhancements have made PDSEs more reliable and available, correcting a few
problems that caused IPLs due to a hang, deadlock, or out-of-storage condition.
Originally, in order to implement PDSE, two system address spaces were introduced:
SMXC, in charge of PDSE serialization.
SYSBMAS, the owner of the data space and hiperspace buffering.
z/OS V1R6 combines SMXC and SYSBMAS to a single address space called SMSPDSE.
This improves overall PDSE usability and reliability by:
Reducing excessive ECSA usage (by moving control blocks into the SMSPDSE address
space)
Reducing re-IPLs due to system hangs in failure or CANCEL situations
Providing storage administrators with tools for monitoring and diagnosis through VARY
SMS,PDSE,ANALYSIS command (for example, determining which systems are using a
particular PDSE)
However, the SMSPDSE address space is usually non-restartable because of the eventual
existence of perennial PDSEs data sets in the LNKLST concatenation. Then, any hang
condition can cause an unplanned IPL. To fix this, we have a new AS, the restartable
SMSPDSE1, which is in charge of all allocated PDSEs except the ones in the LNKST.

SMSPDSE1 is created if the following parameters are defined in IGDSMSxx:
PDSESHARING(EXTENDED)
PDSE_RESTARTABLE_AS(YES)
z/OS V1R8 enhancements

With z/OS V1R8, the following improvements for PDSEs are introduced:
SMSPDSE and SMSPDSE1 support 64-bit addressing, allowing more concurrently
opened PDSE members.
The buffer pool can be retained beyond PDSE close, improving performance due to less
buffer pool deletions and creation. This is done by setting
PDSE1_BUFFER_BEYOND_CLOSE(YES) in the IGDSMSxx member in parmlib.
You can dynamically change the amount of virtual storage allocated to hiperspaces for
SMSPDSE1 through the SETSMS PDSE1HSP_SIZE(nnn) command.

4.26 PDSE: Conversion
Using DFSMSdss Using IEBCOPY
COPY DATASET(INCLUDE - //INPDS DD DISP=SHR,

(MYTEST.**) - DSN=USER.PDS.LIBRARY
BY(DSORG = PDS)) - //PDSE DD DSN=USER.PDSE.LIBRARY,
INDY(SMS001) - DISP=OLD
OUTDY(SMS002) - //SYSIN DD *
CONVERT(PDSE(**))- COPY OUTDD=OUTPDSE
RENAMEU(MYTEST2) - INDD=INPDS
DELETE SELECT MEMBER=(A,B,C)
Figure 4-36 Converting PDS to PDSE
Converting a PDS data set to a PDSE

You can use IEBCOPY or DFSMSdss COPY to convert PDS to PDSE, as shown in Figure 4-36.
We recommend using DFSMSdss.
You can convert the entire data set or individual members, and also back up and restore
PDSEs. By using the DFSMSdss COPY function with the CONVERT and PDS keywords, you
can convert a PDSE back to a PDS. This is especially useful if you need to prepare a PDSE
for migration to a site that does not support PDSEs. When copying members from a PDS load
module library into a PDSE program library, or vice versa, the system invokes the program
management binder component.
Many types of libraries are candidates for conversion to PDSE, including:

PDSs that are updated often, and that require frequent and regular reorganization
Large PDSs that require specific device types because of the size of allocation
Converting PDSs to PDSEs is beneficial, but be aware that certain data sets are unsuitable
for conversion to, or allocation as, PDSEs because the system does not retain the original
block boundaries.
Using DFSMSdss
In Figure 4-36, the DFSMSdss COPY example converts all PDSs with the high-level qualifier
of MYTEST on volume SMS001 to PDSEs with the high-level qualifier of MYTEST2 on

volume SMS002. The original PDSs are then deleted. If you use dynamic allocation, specify
INDY and OUTDY for the input and output volumes. However, if you define the ddnames for
the volumes, use the INDD and OUDD parameters.
Using IEBCOPY
To copy one or more specific members using IEBCOPY, as shown in Figure 4-36 on
page 150, use the SELECT control statement. In this example, IEBCOPY copies members A,
B, and C from USER.PDS.LIBRARY to USER.PDSE.LIBRARY.
For more information about DFSMSdss, see z/OS DFSMSdss Storage Administration Guide,
SC35-0423, and z/OS DFSMSdss Storage Administration Reference, SC35-0424.

4.27 Program objects in a PDSE
Functions to create, update, execute, and access

program objects in PDSEs
Load module format
Binder replaces linkage editor
Program fetch
DESERV internal interface function AMASPZAP
Set of utilities such as IEWTPORT, which builds
transportable programs from program objects and
vice versa
Coexistence between PDS and PDSE load module
libraries in the same system
Figure 4-37 Program objects in PDSE
Program objects in PDSE

Program objects are created automatically when load modules are copied into a PDSE.
Likewise, program objects are automatically converted back to load modules when they are
copied into a partitioned data set. Note that certain program objects cannot be converted into
load modules because they use features of program objects that do not exist in load modules.
A load module is an executable program loaded by the binder in a PDS. A program object is
an executable program loaded by the binder in a PDSE.
Load module format

For accessing a PDS directory or member, most PDSE interfaces are indistinguishable from
PDS interfaces. However, PDSEs have a different internal format, which gives them increased
usability. Each member name can be eight bytes long. The primary name for a program
object can be eight bytes long. Alias names for program objects can be up to 1024 bytes long.
The records of a given member of a PDSE are written or retrieved sequentially. You can use a
PDSE in place of a PDS to store data, or to store programs in the form of program objects. A
program object is similar to a load module in a PDS. A load module cannot reside in a PDSE
and be used as a load module. One PDSE cannot contain a mixture of program objects and
data members.
The binder
The binder is the program that processes the output of language translators and compilers
into an executable program (load module or program object). It replaced the linkage editor

and batch loader. The binder converts the object module output of language translators and
compilers into an executable program unit that can either be stored directly into virtual storage
for execution or stored in a program library (PDS or PDSE).
MVS program fetch

Most of the loading functions are transparent to the user. When the executable code is in a
library, the program management loader component knows whether the program being
loaded is a load module or a program object by the source data set type:
If the program is being loaded from a PDS, it calls IEWFETCH (integrated as part of the
loader) to do what it has always done.
If the program is being loaded from a PDSE, a new routine is called to bring in the program
using data-in-virtual (DIV). The loading is done using special loading techniques that can
be influenced by externalized options.
The program management loader increases the services of the program fetch component by
adding support for loading program objects. The program management loader reads both
program objects and load modules into virtual storage and prepares them for execution. It
relocates any address constants in the program to point to the appropriate areas in virtual
storage and supports 24-bit, 31-bit, and 64-bit addressing ranges. All program objects loaded
from a PDSE are page-mapped into virtual storage. When loading program objects from a
PDSE, the loader selects a loading mode based on the module characteristics and
parameters specified to the binder when you created the program object. You can influence
the mode with the binder FETCHOPT parameter. The FETCHOPT parameter allows you to
select whether the program is completely preloaded and relocated before execution, or
whether pages of the program can be read into virtual storage and relocated only when they
are referenced during execution.
DESERV directory service

The directory service DESERV supports both PDS and PDSE libraries. You can issue
DESERV in your Assembler program for either PDS or PDSE directory access, but you must
pass the DCB address. It does not default to a predefined search order, as does BLDL (the
old directory service).
Members of a PDSE program library cannot be rewritten, extended, or updated in place.

When updating program objects in a PDSE program library, the AMASPZAP service aid
invokes the program management binder, which creates a new version of the program rather
than updating the existing version in place.
IEWTPORT utility
The transport utility (IEWTPORT) is a program management service with very specific and
limited function. It obtains (through the binder) a program object from a PDSE and converts it
into a transportable program file in a sequential (nonexecutable) format. It also reconstructs
the program object from a transportable program file and stores it back into a PDSE (through
the binder).
PDS and PDSE coexistence

You can use a PDSE in place of a PDS to store data, or to store programs in the form of
program objects. A program object is similar to a load module in a PDS. A load module
cannot reside in a PDSE and be used as a load module. One PDSE cannot contain a mixture
of program objects and data members. PDSEs and PDSs are processed using the same
access methods (BSAM, QSAM, BPAM) and macros, but you cannot use EXCP because of
the data set's internal structures.

4.28 Sequential access methods
Sequential access data organization

Physical sequential
Extended format
Compressed data sets
Data striped data sets
z/OS UNIX files
These organizations are accessed by the
sequential access methods:
Queued sequential access method (QSAM)
Basic sequential access method (BSAM)
Figure 4-38 Sequential access methods
Access methods
An access method defines the technique that is used to store and retrieve data. Access
methods have their own data set structures to organize data, macros to define and process
data sets, and utility programs to process data sets. Access methods are identified primarily
by the data set organization. For example, use the basic sequential access method (BSAM)
or queued sequential access method (QSAM) with sequential data sets. However, there are
times when an access method identified with one organization can be used to process a data
set organized in a different manner.
Physical sequential
There are two sequential access methods, basic sequential access method (BSAM) and
queued sequential access method (QSAM) and just one sequential organization. Both
methods access data organized in a physical sequential manner; the physical records
(containing logical records) are stored sequentially in the order in which they are entered.
An important performance item in sequential access is buffering. If you allow enough buffers,
QSAM is able to minimize the number of SSCHs by packaging together in the same I/O
operation (through CCW command chaining) the data transfer of many physical blocks. This
function decreases considerably the total amount of I/O connect time. Another key point is the
look-ahead function for reads, that is, reading in advance records that are not yet required by
the application program.

Extended format data sets
A special type of this organization is the extended format data set. Extended format data sets
have a different internal storage format from sequential data sets that are not extended (fixed
block with a 32-byte suffix). This storage format gives extended format data sets additional
usability and availability characteristics, as explained here:
They can be allocated in the compressed format (can be referred to as a compressed
format data set). A compressed format data set is a type of extended format data set that
has an internal storage format that allows for data compression.
They allow data striping, that is, a multivolume sequential file where data can be accessed
in parallel.
They are able to recover from padding error situations.
They can use the system managed buffering (SMB) technique.
Extended format data sets must be SMS-managed and must reside on DASD. You cannot
use an extended format data set for certain system data sets.
z/OS UNIX files

Another type of this organization is the Hierarchical File System (HFS). HFS files are
POSIX-conforming files that reside in an HFS data set. They are byte-oriented rather than
record-oriented, as are MVS files. They are identified and accessed by specifying the path
leading to them. Programs can access the information in HFS files through z/OS UNIX
system calls, such as open(pathname), read(file descriptor), and write(file descriptor).
Programs can also access the information in HFS files through the MVS BSAM, QSAM, and
VSAM (Virtual Storage Access Method) access methods. When using BSAM or QSAM, an
HFS file is simulated as a multi-volume sequential data set. When using VSAM, an HFS file is
simulated as an ESDS. Note the following point about HFS data sets:
They are supported by standard DADSM create, rename, and scratch.
They are supported by DFSMShsm for dump/restore and migrate/recall if DFSMSdss is
used as the data mover.
They are not supported by IEBCOPY or the DFSMSdss COPY function.
QSAM and BSAM

BSAM arranges records sequentially in the order in which they are entered. A data set that
has this organization is a sequential data set. The user organizes records with other records
into blocks. This is basic access. You can use BSAM with the following data types:
Basic format sequential data sets, which before z/OS V1R7 were known as sequential
data sets, or more accurately as non-extended-format sequential data sets
Large format sequential data sets
z/OS UNIX files
QSAM arranges records sequentially in the order that they are entered to form sequential
data sets, which are the same as those data sets that BSAM creates. The system organizes
records with other records. QSAM anticipates the need for records based on their order. To
improve performance, QSAM reads these records into storage before they are requested.
This is called queued access. You can use QSAM with the following data types:
Basic format sequential data sets before z/OS V1R7, which were known as sequential
data sets or more accurately as non-extended-format sequential data sets
Large format sequential data sets
z/OS UNIX files

4.29 z/OS V1R9 QSAM - BSAM enhancements
DFSMS provides the following enhancements to

BSAM and QSAM:
Long-term page fixing for BSAM data buffers with the

FIXED=USER parameter
BSAM and QSAM support for the MULTACC

parameter
QSAM support for the MULTSDN parameter
Figure 4-39 QSAM and BSAM enhancements with z/OS V1R9
Long-term page fixing for BSAM data buffers

To improve performance, in z/OS V1.9 BSAM allows certain calling programs to specify that
all their BSAM data buffers have been page fixed. This specification frees BSAM from the
CPU-time intensive work of fixing and freeing the data buffers itself. The only restrictions are:
The calling program must be APF authorized, or be in system key or supervisor state.
The format of the data set must be either basic format, large format, PDS, or extended
format.
The DCBE macro option FIXED=USER must be coded to specify that the calling program
has done its own page fixing and indicates that the user has page fixed all BSAM data buffers.
Note: Compressed format data sets are not supported.
BSAM and QSAM support for MULTACC

In z/OS V1R9, the MULTACC parameter of the DCBE macro is expanded. This is done to
optimize performance for tape data sets with BSAM, and to support QSAM with optimized
performance for both tape and DASD data sets. The calculations that are used to optimize the
performance for BSAM with DASD data sets are also enhanced. When dealing with a tape
data set, OPEN supports MULTACC for BSAM and QSAM.

BSAM support
For BSAM in z/OS V1R9, if you code a nonzero MULTACC value, OPEN calculates a default
number of READ or WRITE requests that you are suggesting the system queue more
efficiently. OPEN calculates the number of BLKSIZE-length blocks that can fit within 64 KB,
then multiplies that value by the MULTACC value. If the block size exceeds 32 KB, then OPEN
uses the MULTACC value without modification (this can happen only if you are using LBI, the
large block interface). The system then tries to defer starting I/O requests until you have
issued this number of READ or WRITE requests for the DCB. BSAM will never queue (defer)
more READ or WRITE requests than the NCP value set in OPEN.
For BSAM, it will work as documented for DASD:

If you code a nonzero MULTACC value, OPEN will calculate a default number of read or
write requests that you are suggesting the system queue more efficiently.
The system will try to defer starting I/O requests until you have issued this many read or
write requests for the DCB.
Note: BSAM will never queue or defer more read or write requests than the number of
channel programs (NCP) value set in OPEN.
QSAM support
For QSAM in z/OS V1R9, if you code a nonzero MULTACC value, OPEN calculates a default
number of buffers that you are suggesting the system queue more efficiently. OPEN
calculates the number of BLKSIZE-length blocks that can fit within 64 KB, then multiplies that
value by the MULTACC value. If the block size exceeds 32 KB, then OPEN uses the
MULTACC value without modification (this can happen only if you are using LBI, the large
block interface). The system then tries to defer starting I/O requests until that number of
buffers has been accumulated for the DCB. QSAM will never queue (defer) more buffers than
the BUFNO value that is in effect.
Note: If you code a nonzero MULTACC value, OPEN will calculate a default number of
buffers that you are suggesting the system queue more efficiently.
QSAM support for MULTSDN parameter

You can use the MULTSDN parameter of the DCBE macro with QSAM. In previous releases,
QSAM ignored the MULTSDN parameter. This new support for MULTSDN allows the system
to calculate a more efficient default value for DCB's BUFNO parameter, and reduces the
situations where you need to specify a BUFNO value. The user can use MULTSDN to give a
hint to OPEN so it can calculate a better default value for QSAM BUFNO instead of 1, 2, or 5.
The user will not have to be dependent on device information such as blocks per track or
number of stripes. QSAM accepts a MULTSDN value for the following data sets:
Tape data sets
DASD data sets of the following types:
Basic format
Large format
Extended format (non-compressed)
PDS
For these supported data set types, the system uses MULTSDN to calculate a more efficient
value for BUFNO when the following conditions are true:
The MULTSDN value is not zero.
DCBBUFNO has a value of zero after completion of the DCB OPEN exit routine.
The data set block size is available.

4.30 Virtual storage access method (VSAM)
Catalog
Management
VSAM Data set types:
KSDS
ESDS
CATALOG
LDS
RRDS
Fixed Length
Variable Length
Record
Management
VSAM DATA SETS
Figure 4-40 Virtual storage access method
Virtual storage access method (VSAM)

VSAM is a DFSMSdfp component used to organize data and maintain information about the
data in a catalog. VSAM arranges records by an index key, relative record number, or relative
byte addressing. VSAM is used for direct or sequential processing of fixed-length and
variable-length records on DASD. Data that is organized by VSAM is cataloged for easy
retrieval and is stored in one of five types of data sets.
There are two major parts of VSAM:

Catalog management - The catalog contains information about the data sets.
Record management - VSAM can be used to organize records into five types of data sets,
that is, the VSAM access method function:
Key-sequenced data sets (KSDS) contain records in ascending collating sequence.
Records can be accessed by a field, called a key, or by a relative byte address.
Entry-sequenced data sets (ESDS) contain records in the order in which they were
entered. Records are added to the end of the data set.
Linear data sets (LDS) contain data that has no record boundaries. Linear data sets
contain none of the control information that other VSAM data sets do. Linear data sets
must be cataloged in a catalog.
Relative record data sets (RRDS) contains records in relative record number order, and
the records can be accessed only by this number. There are two types of relative

record data sets, as follows:
Fixed-length RRDS: The records must be of fixed length.
Variable-length RRDS (VRRDS): The records can vary in length.
VSAM data sets

The primary difference among these types of data sets is the way in which their records are
stored and accessed. VSAM arranges records by an index key, by relative byte address, or by
relative record number. Data organized by VSAM must be cataloged and is stored in one of
five types of data sets, depending on an application designer option.
z/OS UNIX files can be accessed as though they are VSAM entry-sequenced data sets
(ESDS). Although UNIX files are not actually stored as entry-sequenced data sets, the
system attempts to simulate the characteristics of such a data set. To identify or access a
UNIX file, specify the path that leads to it.
Any type of VSAM data set can be in extended format. Extended-format data sets have a
different internal storage format than data sets that are not extended. This storage format
gives extended-format data sets additional usability characteristics and possibly better
performance due to striping. You can choose that an extended-format key-sequenced data
set be in the compressed format. Extended-format data sets must be SMS managed. You
cannot use an extended-format data set for certain system data sets.

4.31 VSAM terminology
Logical record
Unit of application information in a VSAM data set
Designed by the application programmer
Can be of fixed or variable size
Divided into fields, one of them can be a key
Physical record
Control interval
Control area
Component
Cluster
Sphere
Figure 4-41 VSAM terminology
Logical record
A logical record is a unit of application information used to store data in a VSAM cluster. The
logical record is designed by the application programmer from the business model. The
application program, through a GET, requests that a specific logical record be moved from the
I/O device to memory in order to be processed. Through a PUT, the specific logical record is
moved from memory to an I/O device. A logical record can be of a fixed size or a variable size,
depending on the business requirements.
The logical record is divided into fields by the application program, such as the name of the
item, code, and so on. One or more contiguous fields can be defined as a key field to VSAM,
and a specific logical record can be retrieved directly by its key value.
Logical records of VSAM data sets are stored differently from logical records in non-VSAM
data sets.
Physical record
A physical record is device-dependent and is a set of logical records moved during an I/O
operation by just one CCW (Read or Write). VSAM calculates the physical record size in
order to optimize the track space (to avoid many gaps) at the time the data set is defined. All
physical records in VSAM have the same length. A physical record is also referred to as a
physical block or simply a block. VSAM may have control information along with logical
records in a physical record.

Control interval (CI)
VSAM stores records in control intervals. A control interval is a continuous area of direct
access storage that VSAM uses to store data records and control information that describes
the records. Whenever a record is retrieved from direct access storage, the entire control
interval containing the record is read into a VSAM I/O buffer in virtual storage. The desired
record is transferred from the VSAM buffer to a user-defined buffer or work area.
Control area (CA)

The control intervals in a VSAM data set are grouped together into fixed-length contiguous
areas of direct access storage called control areas. A VSAM data set is actually composed of
one or more control areas. The number of control intervals in a control area is fixed by VSAM.
The maximum size of a control area is one cylinder, and the minimum size is one track of
DASD storage. When you specify the amount of space to be allocated to a data set, you
implicitly define the control area size. Refer to 4.32, VSAM: Control interval (CI) on
page 162, for more information.
Component
A component in systems with VSAM is a named, cataloged collection of stored records, such
as the data component or index component of a key-sequenced file or alternate index. A
component is a set of CAs. It is the VSAM terminology for an MVS data set. A component has
an entry in the VTOC. An example of a component can be the data set containing only data
for a KSDS VSAM organization.
Cluster
A cluster is a named structure consisting of a group of related components. VSAM data sets
can be defined with either the DEFINE CLUSTER command or the ALLOCATE command. The
cluster is a set of components that have a logical binding between them. For example, a
KSDS cluster is composed of the data component and the index component. The concept of
cluster was introduced to make the JCL to access VSAM more flexible. If you want to access
a KSDS normally, just use the clusters name on a DD card. Otherwise, if you want special
processing with just the data, use the data component name on the DD card.
Sphere
A sphere is a VSAM cluster and its associated data sets. The cluster is originally defined with
the access method services ALLOCATE command, the DEFINE CLUSTER command, or through
JCL. The most common use of the sphere is to open a single cluster. The base of the sphere
is the cluster itself.

4.32 VSAM: Control interval (CI)
3 bytes
Control Interval Format 4 bytes
C
R R R
FREE SPACE I
LR1 LR2 LR
LRnn D D D
D
Fn F2 F1
F
Control information fields

LR = Logical record
RDF = Record definition field
CIDF = Control interval definition field
Contigous records of
the same size
LR2 LR4 LR5 C

LR1 LR3 FREE R R R R
I
100 100 100 LR
150
n 100 SPACE D D D D
F3 D
bytes bytes bytes bytes bytes F4 F2 F1
F
Figure 4-42 Control interval format
Control interval (CI)

The control interval is a concept that is unique to VSAM. A CI is formed by one or several
physical records (usually just one). It is the fundamental building block of every VSAM file. A
CI is a contiguous area of direct access storage that VSAM uses to store data records and
control information that describes the records. A CI is the unit of information that VSAM
transfers between the storage device and the main storage during one I/O operation.
Whenever a logical record is requested by an application program, the entire CI containing
the logical record is read into a VSAM I/O buffer in virtual storage. The desired logical record
is then transferred from the VSAM buffer to a user-defined buffer or work area (if in move
mode).
Based on the CI size, VSAM calculates the best size of the physical block in order to better
use the 3390/3380 logical track. The CI size can be from 512 bytes to 32 KB. A CI contents
depends on the cluster organization. A KSDS consists of:
Logical records stored from the beginning to the end of the CI.
Free space, for data records to be inserted into or lengthened.
Control information, which is made up of two types of fields:
One control interval definition field (CIDF) per CI. CIDF is a 4-byte field. CIDF contains
information about the amount and location of free space in the CI.

Several record definition fields (RDF) describing the logical records. RDF is a 3-byte
field and describes the length of logical records. For fixed length records there are two
RDFs, one with the length, and the other with how many records are the same length.
The size of CIs can vary from one component to another, but all the CIs within the data or
index component of a particular cluster data set must be of the same length. The CI
components and properties may vary, depending on the data set organization. For example,
an LDS does not contain CIDFs and RDFs in its CI. All of the bytes in the LDS CI are data
bytes.
Spanned records
Spanned records are logical records that are larger than the CI size. They are needed when
the application requires very long logical records. To have spanned records, the file must be
defined with the SPANNED attribute at the time it is created. Spanned records are allowed to
extend across or span control interval boundaries, but not beyond control area limits. The
RDFs describe whether the record is spanned or not.
A spanned record always begins on a control interval boundary, and fills one or more control
intervals within a single control area. A spanned record does not share the CI with any other
records; in other words, the free space at the end of the last segment is not filled with the next
record. This free space is only used to extend the spanned record.
Control area (CA)

Control area is also a concept unique to VSAM. A CA is formed by two or more CIs put
together into fixed-length contiguous areas of direct access storage. A VSAM data set is
composed of one or more CAs. In most cases, a CA is the size of a 3390/3380 cylinder. The
minimum size of a CA is one track. The CA size is implicitly defined when you specify the size
of a data set at data set definition.
CAs are needed to implement the concept of splits. The size of a VSAM file is always a
multiple of the CA size and VSAM files are extended in units of CAs.
Splits
CI splits and CA splits occur as a result of data record insertions (or increasing the length of
an already existing record) in KSDS and VRRDS organizations. If a logical record is to be
inserted (in key sequence) and there is not enough free space in the CI, the CI is split.
Approximately half the records in the CI are transferred to a free CI provided in the CA, and
the record to be inserted is placed in the original CI.
If there are no free CIs in the CA and a record is to be inserted, a CA split occurs. Half the CIs
are sent to the first available CA at the end of the data component. This movement creates
free CIs in the original CA, then the record to be inserted causes a CI split.

4.33 VSAM data set components
Cluster
HDR 67 95
Index
HDR 38 67 HDR 95 Set
Index
Component
H H H
D 7 11 14 21 30 38 D 43 50 54 57 64 67 D 71 75 78 85 92 95 Sequence
R R R Set
2 5 7 39 41 43 68 69
8 9 44 45 46 72 73 74
Record
12 13 14 51 53 54 76 77 78
Data
Component 15 16 19 55 56 57 79 80 85
key
22 23 26 58 61 62 86 89
31 35 38 65 66 67 93 94 95 Control
Interval
Control Area Control Area Control Area
Logical Record
Figure 4-43 VSAM data set components
VSAM data set components

A component is an individual part of a VSAM cluster. Each component has a name, an entry
in the catalog, and an entry in the VTOC. There are two types of components, the data
component and the index component. Some VSAM organizations (such as an ESDS, RRDS,
and LDS) have only the data component.
Data component
The data component is the part of a VSAM cluster, alternate index, or catalog that contains
the data records. All VSAM cluster organizations have the data component.
Index component
The index component is a collection of records containing data keys and pointers (relative
byte address, or RBA). The data keys are taken from a fixed defined field in each data logical
record. The keys in the index logical records are compressed (rear and front). The RBA
pointers are compacted. Only KSDS and VRRDS VSAM data set organizations have the
index component.
Using the index, VSAM is able to retrieve a logical record from the data component when a
request is made randomly for a record with a certain key. A VSAM index can consist of more
than one level (binary tree). Each level contains pointers to the next lower level. Because
there are random and sequential types of access, VSAM divides the index component into
two parts: the sequence set, and the index set.

Sequence set
The sequence set is the lowest level of index, and it directly points (through an RBA) to the
data CI in the CA of the data component. Each index logical record:
Occupies one index CI.
Maps one CI in the data component.
Contains pointers and high key information for each data CI.
Contains horizontal pointers from one sequence set CI to the next keyed sequence set CI.
These horizontal pointers are needed because of the possibility of splits, which make the
physical sequence different from the logical collating sequence by key.
Index set
The records in all levels of the index above the sequence set are called the index set. An entry
in an index set logical record consists of the highest possible key in an index record in the
next lower level, and a pointer to the beginning of that index record. The highest level of the
index always contains a single index CI.
The structure of VSAM prime indexes is built to create a single index record at the lowest level
of the index. If there is more than one sequence-set-level record, VSAM automatically builds
another index level.
Cluster
A cluster is the combination of the data component (data set) and the index component (data
set) for a KSDS. The cluster provides a way to treat index and data components as a single
component with its own name. Use of the word cluster instead of data set is recommended.
Alternate index (AIX)

The alternate index is a VSAM function that allows logical records of a KSDS or ESDS to be
accessed sequentially and directly by more than one key field. The cluster that has the data is
called the base cluster, then an AIX cluster is built from the base cluster. Alternate indexes
eliminate the need to store the same data in different sequences in multiple data sets for the
purposes of various applications. Each alternate index is a KSDS cluster consisting of an
index component and a data component.
The records in the AIX index component contain the alternate key and the RBA pointing to the
alternate index data component. The records in the AIX data component contain the alternate
key value itself and all the primary keys corresponding to the alternate key value (pointers to
data in the base cluster). The primary keys in the logical record are in ascending sequence
within an alternate index value.
Any field in the base cluster record can be used as an alternate key. It can also overlap the
primary key (in a KSDS), or any other alternate key. The same base cluster may have several
alternate indexes varying the alternate key. There may be more than one primary key value
per the same alternate key value. For example, the primary key might be an employee
number and the alternate key might be the department name; obviously, the same
department name may have several employee numbers.
AIX cluster is created with IDCAMS DEFINE ALTERNATEINDEX command, then it is populated by
the BLDINDEX command. Before a base cluster can be accessed through an alternate index, a
path must be defined. A path provides a way to gain access to the base data through a
specific alternate index. To define a path, use the DEFINE PATH command. The utility to issue
this command is discussed in 4.14, Access method services (IDCAMS) on page 129.
Sphere
A sphere is a VSAM cluster and its AIX associated clusters data sets.

4.34 VSAM key sequenced cluster (KSDS)
VSAM.KSDS Cluster component

Data component VSAM.KSDS.DATA
VSAM.KSDS.INDEX Index component
Figure 4-44 Key sequenced cluster (KSDS)
VSAM KSDS cluster

In a KSDS, logical records are placed in the data set in ascending collating sequence by key.
The key contains a unique value, which determines the record's collating position in the
cluster. The key must be in the same position (off set) in each record.
The key field must be contiguous and each keys contents must be unique. After it is
specified, the value of the key cannot be altered, but the entire record may be deleted.
When a new record is added to the data set, it is inserted in its logical collating sequence by
key.
A KSDS has a data component and an index component. The index component keeps track
of the used keys and is used by VSAM to retrieve a record from the data component quickly
when a request is made for a record with a certain key.
A KSDS can have fixed or variable length records.
A KSDS can be accessed in sequential mode, direct mode, or skip sequential mode (meaning
that you process sequentially, but directly skip portions of the data set).

4.35 VSAM: Processing a KSDS cluster
Application: GET record with key = 23
HDR 67 95
Index
HDR 38 67 HDR 95 Set
Index
Component
H H H
D 7 11 14 21 26 38 D 43 50 54 57 64 67 D 71 75 78 85 92 95 Sequence
R R R Set
2 5 7 39 41 43 68 69
8 9 44 45 46 72 73 74
12 13 14 51 53 54 76 77 78
Data
Component 15 16 19 55 56 57 79 80 85
22 23 26 58 61 62 86 89
31 35 38 65 66 67 93 94 95
Control
Interval Control Area Control Area Control Area
Application
Figure 4-45 Processing an indexed VSAM cluster: Direct access
Processing a KSDS cluster

A KSDS has an index that relates key values to the relative locations in the data set. This
index is called the prime index. It has two uses:
Locate the collating position when inserting records
Locate records for retrieval
When initially loading a KSDS data set, records must be presented to VSAM in key sequence.
This loading can be done through the IDCAMS VSAM utility named REPRO. The index for a
key-sequenced data set is built automatically by VSAM as the data set is loaded with records.
When a data CI is completely loaded with logical records, free space, and control information,
VSAM makes an entry in the index. The entry consists of the highest possible key in the data
control interval and a pointer to the beginning of that control interval.
When accessing records sequentially, VSAM refers only to the sequence set. It uses a
horizontal pointer to get from one sequence set record to the next record in collating
sequence.
Request for data direct access

When accessing records directly, VSAM follows vertical pointers from the highest level of the
index down to the sequence set to find vertical pointers to the requested logical record.
Figure 4-45 shows how VSAM searches the index when an application issues a GET for a
logical record with key value 23.

The sequence is as follows:
1. VSAM scans the index record in the highest level of the index set for a key that is greater
or equal to 23.
2. The entry 67 points to an index record in the next lower level. In this index record, VSAM
scans for an entry for a key that is higher than or equal to 23.
3. The entry 38 points to the sequence set that maps the CA holding the CI containing the
logical record.
4. VSAM scans the sequence set record with the highest key, searching for a key that is
greater than or equal to 23.
5. The entry 26 points to the data component CI that holds the desired record.
6. VSAM searches the CI for the record with key 23. VSAM finds the logical record and gives
it to the application program.
If VSAM does not find a record with the desired key, the application receives a return code
indicating that the record was not found.

4.36 VSAM entry sequenced data set (ESDS)
CI 1
C
R R
RECORD RECORD RECORD RECORD I
UNUSED SPACE D D
1 2 3 4 D
F F
F
RBA 0
CI 2
C
R R R R
RECORD RECORD RECORD RECORD UNUSED I
D D D D
5 6 7 8 SPACE D
F F F F
F
RBA 4096
CI 3
C
R R
RECORD RECORD I
UNUSED SPACE D D
9 10 D
F F
F
RBA 8192
CI 4
C
I
UNUSED SPACE D
F
RBA 12288
Figure 4-46 VSAM entry sequenced data set (ESDS)
VSAM entry sequenced data set (ESDS)

An entry sequenced data set (ESDS) is comparable to a sequential data set. It contains
fixed-length or variable-length records. Records are sequenced by the order of their entry in
the data set, rather than by a key field in the logical record. All new records are placed at the
end of the data set. An ESDS cluster has only a data component.
Records can be accessed sequentially or directly by relative byte address (RBA). When a
record is loaded or added, VSAM indicates its relative byte address (RBA). The RBA is the
offset of the first byte of the logical record from the beginning of the data set. The first record
in a data set has an RBA of 0; the second record has an RBA equal to the length of the first
record, and so on. The RBA of a logical record depends only on the record's position in the
sequence of records. The RBA is always expressed as a full-word binary integer.
Although an entry-sequenced data set does not contain an index component, alternate
indexes are allowed. You can build an alternate index with keys to keep track of these RBAs.

4.37 VSAM: Typical ESDS processing
Application program:
GET NEXT
CI 1
C
R R
RECORD RECORD RECORD RECORD I
UNUSED SPACE D D
1 2 3 4 D
F F
F
RBA 0
CI 2
C
R R R R
RECORD RECORD RECORD RECORD UNUSED I
D D D D
5 6 7 8 SPACE D
F F F F
F
RBA 4096
CI 3
C
R R
RECORD RECORD I
UNUSED SPACE D D
9 10 D
F F
F
RBA 8192
CI 4
C
I
UNUSED SPACE
D
F
RBA 12288
Figure 4-47 Typical ESDS processing (ESDS)
Typical ESDS processing

For an ESDS, two types of processing are supported:
Sequential access (the most common).
Direct (or random) access requires the program to give the RBA of the record.
Skip sequential is not allowed.
Existing records can never be deleted. If the application wants to delete a record, it must flag
that record as inactive. As far as VSAM is concerned, the record is not deleted. Records can
be updated, but without length change.
ESDS organization is suited for sequential processing with variable records, but in a few read
accesses you need a direct (random) access by key (here using the AIX cluster).

4.38 VSAM relative record data set (RRDS)
C
R R
CI 0 SLOT 1 SLOT 2 SLOT 3 SLOT 4 SLOT 5 D
F
D
F
I
D
F
C
O R R
C
N A CI 1 SLOT 6 SLOT 7 SLOT 8 SLOT 9 SLOT 10 D

F
D
F
I
D
F
T R
R E C
CI 2
R R
I
A SLOT 11 SLOT 12 SLOT 13 SLOT 14 SLOT 15 D D
D
O F F
F
L
C
CI 3
R R
I
SLOT 16 SLOT 17 SLOT 18 SLOT 19 SLOT 20 D
F
D
F
D
F
C
R R
F
D
F
I
D
F
C
O R R
C

F
D
F
I
D
F
T R
R E C
CI 2
R R
I
D
O F F
F
L
C
CI 3
R R
I
F
D
F
D
F
Figure 4-48 VSAM relative record data set (RRDS)
Relative record data set

A relative record data set (RRDS) consists of a number of preformed, fixed length slots. Each
slot has a unique relative record number, and the slots are sequenced by ascending relative
record number. Each (fixed length) record occupies a slot, and it is stored and retrieved by the
relative record number of that slot. The position of a data record is fixed; its relative record
number cannot change.
An RRDS cluster has a data component only.
Random load of an RRDS requires a user program implementing such logic.

4.39 VSAM: Typical RRDS processing
Application program:
GET Record 26
C
R R
F
D
F
I
D
F
C
O R R
C

F
D
F
I
D
F
T R
R E C
CI 2
R R
I
D
O F F
F
L
C
CI 3
R R
I
F
D
F
D
F
C
R R
F
D
F
I
D
F
C
O R R
C

F
D
F
I
D
F
T R
R E C
CI 2
R R
I
D
O F F
F
L
C
R R
F
D
F
I
D
F
Figure 4-49 Typical RRDS processing
Processing RRDS data sets

The application program inputs the relative record number of the target record. VSAM is able
to find its location very quickly by using a formula that takes into consideration the geometry
of the DASD device. The relative number is always used as a search argument. For an
RRDS, three types of processing are supported:
Sequential processing.
Skip-sequential processing.
Direct processing; in this case, the randomization routine is supported by the application
program.

4.40 VSAM linear data set (LDS)
CI DATA
C
O
N A CI DATA
T R
R E
A CI DATA
O
L
CI DATA
CI DATA
C
O
N A CI DATA
T R
R E
A CI DATA
O
L
CI DATA
Figure 4-50 VSAM linear data set (LDS)
VSAM linear data set (LDS)

A linear data set is a VSAM data set with a control interval size from 4096 bytes to 32768
bytes, in increments of 4096 bytes. An LDS has no imbedded control information in its CI, that
is, no record definition fields (RDFs) and no control interval definition fields (CIDFs).
Therefore, all LDS bytes are data bytes. Logical records must be blocked and deblocked by
the application program (although logical records do not exist, from the point of view of
VSAM).
IDCAMS is used to define a linear data set. An LDS has only a data component. An LDS data
set is just a physical sequential VSAM data set comprised of 4 KB physical records, but with a
revolutionary buffer technique called data-in-virtual (DIV).
A linear data set is processed as an entry-sequenced data set, with certain restrictions.
Because a linear data set does not contain control information, it cannot be accessed as
though it contained individual records. You can access a linear data set with the DIV macro. If
using DIV to access the data set, the control interval size must be 4096; otherwise, the data
set will not be processed.
When a linear data set is accessed with the DIV macro, it is referred to as the data-in-virtual
object or the data object.
For information about how to use data-in-virtual, see z/OS MVS Programming: Assembler
Services Guide, SA22-7605.

4.41 VSAM: Data-in-virtual (DIV)
DIV Enable users to:

(Address space) Map data set to virtual storage
(Dataspace) Access data by exploiting paging
or algorithms
(Hiperspace)
Figure 4-51 Data-in-virtual (DIV)
Data-in-virtual (DIV)
You can access a linear data set using these techniques:
VSAM
DIV, if the control interval size is 4096 bytes. The data-in-virtual (DIV) macro provides
access to VSAM linear data sets.
Window services, if the control interval size is 4096 bytes.
Data-in-virtual (DIV) is an optional and unique buffering technique used for LDS data sets.
Application programs can use DIV to map a data set (or a portion of a data set) into an
address space, a data space, or a hiperspace. An LDS cluster is sometimes referred to as a
DIV object. After setting the environment, the LDS cluster looks to the application as a table
in virtual storage with no need of issuing I/O requests.
Data is read into main storage by the paging algorithms only when that block is actually
referenced. During RSM page-steal processing, only changed pages are written to the cluster
in DASD. Unchanged pages are discarded since they can be retrieved again from the
permanent data set.
DIV is designed to improve the performance of applications that process large files
non-sequentially and process them with significant locality of reference. It reduces the
number of I/O operations that are traditionally associated with data retrieval. Likely candidates
are large arrays or table files.

4.42 VSAM: Mapping a linear data set
LDS
WINDOW
OFFSET
BLOCK3
BLOCK4 SPAN
BLOCK5
(Address space)
(Dataspace)
or
(Hiperspace)
Figure 4-52 Mapping a linear data set
Mapping a linear data set

To establish a map from a linear data set to a window (a program-provided area in multiples of
4 KB on a 4 KB boundary), the program issues:
DIV IDENTIFY to introduce (allocate) a linear data set to data-in-virtual services.
DIV ACCESS to cause a VSAM open for the data set and indicate access mode
(read/update).
GETMAIN to allocate a window in virtual storage where the LDS will be mapped totally or
in pieces.
DIV MAP to enable the viewing of the data object by establishing an association between
a program-provided area (window) and the data object. The area may be in an address
space, data space, or hiperspace.
No actual I/O is done until the program references the data in the window. The reference will
result in a page fault which causes data-in-virtual services to read the data from the linear
data set into the window.
DIV SAVE can be used to write out changes to the data object. DIV RESET can be used to
discard changes made in the window since the last SAVE operation.

4.43 VSAM resource pool
VSAM resource pool is formed by:

I/O control blocks
Buffer pool (set of equal-sized buffers)
VSAM resource pool can be shared by VSAM
clusters, improving the effectiveness of these
buffers
Four types of VSAM buffering techniques:
Non-shared resource (NSR)
Local shared resource (LSR)
Global shared resource (GSR)
Record-level sharing (RLS)
Figure 4-53 VSAM resource pool
VSAM resource pool

Buffering is one of the key aspects as far as I/O performance is concerned. A VSAM buffer is
a virtual storage area where the CI is transferred during an I/O operation. In VSAM KSDS
there are two types of buffers: buffers for data CIs and buffers for index CIs. A buffer pool is a
set of buffers with the same size. A resource pool is a buffer pool with several control blocks
describing the pool and describing the clusters with CIs in the resource pool.
The objective of a buffer pool is to avoid I/O operations in random accesses (due to re-visiting
data) and to make these I/O operations more efficient in sequential processing, thereby
improving performance.
For more efficient use of virtual storage, buffer pools can be shared among clusters using
locally or globally shared buffer pools. There are four types of resource pool management,
called modes, defined according to the technique used to manage them:
Not shared resources (NSR)
Local shared resources (LSR)
Global shared resources (GSR)
Record-level shared resources (RLS)
These modes can be declared in the ACB macro of the VSAM data set (MACRF keyword)
and are described in the following section.

4.44 VSAM: Buffering modes
Buffers and I/O related

INDEX
User control blocks
ACB
MACRF=
(LSR,NUB) INDEX
Data User ACB
MACRF=
(LSR,NUB)
Data
INDEX
User ACB
MACRF=
(LSR,NUB)
Data
Buffers and I/O related control blocks

associated with a pool (BLDVRP)
Multiple data sets share pooled resources
Figure 4-54 VSAM LSR buffering mode
VSAM buffering modes

The VSAM buffering modes that you can use are:
NSR, LSR, GSR, and RLS
Non-shared resource (NSR)

Non-shared resource (NSR) is the default VSAM buffering technique. It has the following
characteristics:
The resource pool is implicitly constructed at data set open time.
The buffers are not shared among VSAM data sets; only one cluster has CIs in this
resource pool.
Buffers are located in the private area.
For sequential reads, VSAM uses the read-ahead function: when the application finishes
processing half the buffers, VSAM schedules an I/O operation for that half of the buffers.
This continues until a CA boundary is encountered; the application must wait until the last
I/O to the CA is done before proceeding to the next CA. The I/O operations are always
scheduled within CA boundaries.
For sequential writes, VSAM postpones the writes to DASD until half the buffers are filled
by the application. Then VSAM schedules an I/O operation to write that half of the buffers
to DASD. The I/O operations are always scheduled within CA boundaries.

CIs are discarded as soon as they are used, using a sequential algorithm to keep CIs in
the resource pool.
There is dynamic addition of strings. Strings are like cursors; each string represents a
position in the data set for the requested record.
For random access there is not look ahead, but the algorithm is still the sequential one.
NSR is used by high-level languages. Since buffers are managed by a sequential algorithm,
NSR is not the best choice for random processing. For applications using NSR, consider
using system-managed buffering, discussed in 4.45, VSAM: System-managed buffering
(SMB) on page 179.
Local shared resource (LSR)

An LSR resource pool is suitable for random processing and not for sequential processing.
The characteristics of the LSR include that it is:
Shared among VSAM clusters accessed by tasks in the same address space.
Located in the private area and ESO hiperspace. With hiperspace, VSAM buffers are
located in expanded storage to improve the processing of VSAM clusters. With
z/Architecture ESO hiperspaces are mapped in main storage.
A resource pool is explicitly constructed by macro BLDVRP, before the OPEN.
Buffers are managed by the last recently used (LRU) algorithm, that is, the most
referenced CIs are kept in the resource pool. It is very adequate for random processing.
LSR expects that CIs in buffers are re-visited.
Global shared resource (GSR)

GSR is similar to the LSR buffering technique. GSR differs from LSR in the following ways:
The buffer pool is shared among VSAM data sets accessed by tasks in multiple address
spaces in the same z/OS image.
Buffers are located in CSA.
The code using GSR must be in the supervisor state.
Buffers cannot use hiperspace.
The separate index resource pools are not supported for GSR.
GSR is not commonly used by applications, so you should consider the use of VSAM RLS
instead.
Record-level sharing (RLS)

Record-level sharing (RLS) is the implementation of VSAM data sharing. Record-level
sharing is discussed in detail in Chapter 7, DFSMS Transactional VSAM Services on
page 377.
For more information about NSR, LSR, and GSR, refer to 7.2, Base VSAM buffering on
page 380 and also to the IBM Redbooks publication VSAM Demystified, SG24-6105.

4.45 VSAM: System-managed buffering (SMB)
Only for SMS-managed extended format data sets

and NSR buffering mode
RECORD_ACCESS_BIAS in DATACLASS
ACCBIAS subparameter of AMP, in JCL DD
statement
For ACCBIAS equal to SYSTEM, VSAM decisions
based on MACRF parameter of ACB and MSR in
storage class
Optimum number of index and data buffers
For random access, VSAM changes buffering
management algorithm from NSR to LSR
Figure 4-55 System-managed buffering (SMB)
System-managed buffering (SMB)

SMB is a feature of VSAM that was introduced in DFSMS V1R4. SMB enables VSAM to:
Determine the optimum number of index and data buffers
Change the buffer algorithm as declared in the application program in the ACB MACRF
parameter, from NSR (sequential) to LSR (least recently used - LRU)
Usually, SMB allocates many more buffers than are allocated without SMB. Performance
improvements can be dramatic with random access (particularly when few buffers were
available). The use of SMB is transparent from the point of view of the application; no
application changes are needed.
SMB is available to a data set when all the following conditions are met:
It is an SMS-managed data set.
It is in extended format (DSNTYPE = EXT in the data class).
The application opens the data set for NSR processing.
SMB is invoked or disabled through one of the following methods:

1. The Record Access Bias data class field.
2. The ACCBIAS subparameter of AMP in the JCL DD statement. JCL information takes
precedence over data class information.

SMB processing techniques
If all of the required conditions are met, SMB is invoked when option SYSTEM or an SMB
processing technique is used in the fields described previously. SMB is disabled when USER
is entered instead (USER is the default). Since JCL information takes precedence over data
class information, installations can enable or disable SMB for some executions.
The information contained in SMB processing techniques is needed because SMB must
maintain an adequate algorithm for managing the CIs in the resource pool. SMB accepts the
ACB MACRF options when the I/O operation is requested. For this reason, the installation
must accurately specify the processing type, through ACCBIAS options:
Direct Optimized (DO)
SMB optimizes for totally random record access. When this technique is used, VSAM
changes the buffering management from NSR to LSR.
Direct Weighted (DW)
The majority is direct access to records, with some sequential.
Sequential Optimized (SO)
Totally sequential access.
Sequential Weighted (SW)
The majority is sequential access, with some direct access to records.
When SYSTEM is used in JCL or in the data class, SMB chooses the processing technique
based on the MACRF parameter of the ACB.
For more information about the use of SMB, refer to VSAM Demystified, SG24-6105.

4.46 VSAM buffering enhancements with z/OS V1R9
Provide a way to limit the storage SMB DO uses for

a large number of data sets at once, without
changing the JCL for each job step
JCL AMP parameter SMBVSP
Limiting amount of virtual buffer space
Specify data class SMBVSP value using ISMF
JCL AMP keyword (optional)
MSG=SMBBIAS - request a VSAM open message
indicating what SMB access bias actually is used for
a particular component being opened
Issues message IEC161I
Figure 4-56 VSAM buffering enhancements with z/OS V1R9
VSAM system-managed buffering enhancements

The JCL AMP parameter SMBVSP keyword lets you limit the amount of virtual buffer space to
acquire for direct optimized processing when opening a data set. Before z/OS V1R9,
changing that value required editing the JCL statement, which was not practical when running
a batch job. In z/OS V1.9, VSAM provides a simpler and more efficient way of modifying the
SMBVSP value, by specifying it for a data class using ISMF. The system-managed buffering
(SMB) field on the ISMF DATA CLASS DEFINE/ALTER panel lets you specify the value in
kilobytes or megabytes. SMB then uses the specified value for any data set defined to that
data class. With this method, the effect of modifying the SMBVSP keyword is no longer limited
to one single job step, and no longer requires editing individual JCL statements.
JCL AMP keyword

In addition, a new JCL AMP keyword, MSG=SMBBIAS, lets you request a message that
displays the record access bias that is specified on the ACCBIAS keyword or chosen by SMB
in the absence of a user selection. The IEC161I message is issued for each data set that is
opened. The new keyword is optional and the default is to not issue a message. Avoid using
the keyword when a large number of data sets are opened in quick succession.
AMP=('subparameter[,subparameter]...')
AMP='subparameter[,subparameter]...'
//DS1 DD DSNAME=VSAMDATA,AMP=('BUFSP=200,OPTCD=IL,RECFM=FB',
// 'STRNO=6,MSG=SMBBIAS')

MSG=SMBBIAS
When you specify MSG = SMBBIAS in a JCL DD statement, the system issues message
IEC161I to indicate which access bias SMB has chosen. The default is no message.
IEC161I (return code 001)
rc[(sfi)]-ccc,jjj,sss,ddname,dev,volser,xxx,dsname, cat
The first IEC161I (return code 001) message indicates the access bias used by SMB. The sfi
field can be:
DO - Direct Optimized
DW - Direct Weighted
SO - Sequential Optimized
SW - Sequential Weighted
CO - Create optimized
CR - Create Recovery
When you can code MSG=SMBBIAS in your JCL to request a VSAM open message, it
indicates what SMB access bias actually is used for a particular component being opened:
15.00.02 SYSTEM1 JOB00028 IEC161I
001(DW)-255,TESTSMB,STEP2,VSAM0001,,,SMB.KSDS,,
IEC161I SYS1.MVSRES.MASTCAT
15.00.02 SYSTEM1 JOB00028 IEC161I 001(0000002B 00000002 00000000
00000000)-255,TESTSMB,STEP2,
IEC161I VSAMDATA,,,SMB.KSDS,,SYS1.MVSRES.MASTCAT
SMB overview
System-managed buffering (SMB), a feature of DFSMSdfp, supports batch application
processing.
SMB uses formulas to calculate the storage and buffer numbers needed for a specific access
type SMB takes the following actions:
It changes the defaults for processing VSAM data sets. This enables the system to take
better advantage of current and future hardware technology.
It initiates a buffering technique to improve application performance. The technique is one
that the application program does not specify. You can choose or specify any of the four
processing techniques that SMB implements:
Direct Optimized (DO) The DO processing technique optimizes for totally random
record access. This is appropriate for applications that
access records in a data set in totally random order. This
technique overrides the user specification for nonshared
resources (NSR) buffering with a local shared resources
(LSR) implementation of buffering.
Sequential Optimized (SO) The SO technique optimizes processing for record access
that is in sequential order. This is appropriate for backup
and for applications that read the entire data set or a large
percentage of the records in sequential order.

Direct Weighted (DW) The majority is direct processing, some is sequential. DW
processing provides the minimum read-ahead buffers for
sequential retrieval and the maximum index buffers for
direct requests.
Sequential Weighted (SW) The majority is sequential processing, some is direct. This
technique uses read-ahead buffers for sequential requests
and provides additional index buffers for direct requests.
The read-ahead will not be as large as the amount of data
transferred with SO.

4.47 VSAM SMB enhancement with z/OS V1R11
Performance of SMB due to too few index buffers being

allocated to open small data sets that grow in time
SMB (during VSAM OPEN) calculates how to increase
index buffer space
Users now do not have to close and reopen data sets to
obtain more index buffers
Changes to the SMBVSP parameter
SMBVSP=nnK or SMBVSP=nnM
Specifies the amount of virtual buffer space to acquire for
direct optimized processing when opening the data set
nn is 1 to 2048000 kilobytes or 1 to 2048 megabytes
Figure 4-57 VSAM SMB performance enhancement
SMB performance
Performance of System Managed Buffering (SMB) Direct Optimized Access Bias had been
affected adversely when a VSAM data set continued to grow. This is due to the original
allocation for index buffer space becoming increasingly deficient as the data set size
increases. This problem is avoided for data buffer space by using the subparameter SMBVSP
of the JCL AMP parameter. However, for index buffer space, the only way to adjust the index
buffer space to a more appropriate allocation was to close and reopen the data set. Changes
have been made in z/OS V1R11 VSAM to avoid the necessity of closing and reopening the
data set.
SMBVSP parameter
Prior to z/OS V1R11, when SMBVSP was used to specify the amount of storage for SMB
Direct Optimized Access Bias, the value was used by VSAM OPEN to calculate the number of
data buffers (BUFND). The number of index buffers (BUFNI), in contrast,was calculated by
VSAM OPEN based on the current high used CI. That is, it was based upon the data set size
at open time.
With z/OS V1R11, VSAM OPEN calculates the BUFNI to be allocated using 20% of the value
of SMBVSP, or the data set size, if the calculation using it actually yields a higher BUFNI. The
usage in regard to calculating BUFND remains unchanged. Now, as a KSDS grows, provision
can be made for a better storage allocation for both data and index buffers by the use of the
SMBVSP parameter.

SMBVSP is only a parameter for SMB Direct Optimized Access Bias. It can be specified in the
SMS DATACLAS construct or through JCL as shown in Figure 4-58.
//STEP000 EXEC PGM=TVC15

//KSDS0001 DD DSN=PROD.KSDS,DISP=SHR,AMP=(ACCBIAS=DO,SMBVSP=2048M)
Figure 4-58 SMSVSP specification in JCL
Note: For further details about SMB, see z/OS DFSMS Using Data Sets, SC26-7410. For
further details about how to invoke SMB and about specifying Direct Optimized (DO) and
SMBVSP values in a DATACLAS construct, see z/OS DFSMS Storage Administration
Reference (for DFSMSdfp, DFSMSdss, DFSMShsm), SC26-7402. For information about
specification with JCL, see z/OS MVS JCL Reference, SA22-7597.
SMBVSP parameter considerations

You can use the SMBVSP parameter to restrict the size of the pool that is built for the data
component or to expand the size of the pool for the index records. The SMBHWT parameter
can be used to provide buffering in Hiperspace in combination with virtual buffers for the
data component.
The value of this parameter is used as a multiplier of the virtual buffer space for Hiperspace
buffers. This can reduce the size required for an application region, but does have
implications related to processor cycle requirements. That is, all application requests must
orient to a virtual buffer address. If the required data is in a Hiperspace buffer, the data must
be moved to a virtual buffer after stealing a virtual buffer and moving that buffer to a least
recently used (LRU) Hiperspace buffer.

4.48 VSAM enhancements
Data compression for KSDS
Extended addressability
Record level sharing (RLS)
System-managed buffering (SMB)
Data striping and multi-layering
DFSMS data set separation
Free space release
Figure 4-59 VSAM enhancements
VSAM enhancements
The following list presents the major VSAM enhancements since DFSMS V1R2. For the
majority of these functions, extended format is a prerequisite. The enhancements are:
Data compression for KSDS - This is useful for improving I/O mainly for write-once,
read-many clusters.
Extended addressability - This allows data components larger than 4 GB. The limitation
was caused by an RBA field of 4 bytes; RBA now has an 8-byte length.
Record-level sharing (RLS) - This allows VSAM data sharing across z/OS systems in a
Parallel Sysplex.
System-managed buffering (SMB) - This improves the performance of random NSR
processing.
Data stripping and multi-layering - This improves sequential access performance due to
parallel I/Os in several volumes (stripes).
DFSMS data set separation - This allows the allocation of clusters in distinct physical
control units.
Free space release - As with non-VSAM data sets, the free space that is not used at the
end of the data component can be released at deallocation.

4.49 Data set separation
SMS-managed data sets within a group are kept separate

On a physical control unit (PCU) level or
Volume level
From all the other data sets in the same group
Poor performance and single point of failure may occur
when critical data sets are allocated on the same volumes
During data set allocation, SMS attempts to separate data
sets listed in a separation group onto different extent pools
In the storage subsystem and volumes
Facility to separate critical data sets onto different extent
pools and volumes
Reduces I/O contention and single point of failure
Figure 4-60 Data set separation
Data set separation with z/OS V1R11

Data set separation allows you to designate groups of data sets in which all SMS-managed
data sets within a group are kept separate, on the physical control unit (PCU) level or the
volume level, from all the other data sets in the same group.
When allocating new data sets or extending existing data sets to new volumes, SMS volume
selection frequently calls SRM to select the best volumes. Unfortunately, SRM may select the
same set of volumes that currently have the lowest I/O delay. Poor performance or single
points of failure may occur when a set of functional-related critical data sets are allocated onto
the same volumes. SMS provides a function to separate their critical data sets, such as DB2
partitions, onto different volumes to prevent DASD hot spots and reduce I/O contention.
Volume separation groups

This provides a solution by expanding the scope of the data set separation function currently
available at PCU level to the volume level. The user defines the volume separation groups in
a data set separation profile. During data set allocation, SMS attempts to separate data sets
that are specified in the same separation group onto different extent pools and volumes.
This provides a facility for an installation to separate functional-related critical data sets onto
different extent pools and volumes for better performance and to avoid single points of failure.
Important: Use data set separation only for a small set of mission-critical data.

4.50 Data set separation syntax
Data set separation function syntax:

Earlier syntax supports only PCU separation
{SEPARATIONGROUP | SEP}
{FAILLEVEL | FAIL}({PCU|NONE})
{DSNLIST | DSNS | DSN}(data set name[,data set name,....}
New syntax supports both PCU and Volume separation
{SEPARATIONGROUP | SEP}({PCU | VOL})
TYPE({REQ | PREF})
{DSNLIST | DSNS | DSN}(data set name[,data set name,...}
Wildcard characters (* and %) are supported for DSN at
the third level qualifier (for example, A.B.*)
Earlier syntax for PCU separation continues to function
Both earlier and new syntax can coexist in the same separation
profile
Figure 4-61 Defining separation profiles and syntax
Data set profile for separation

To use data set separation, you must create a data set separation profile and specify the
name of the profile to the base configuration. During allocation, SMS attempts to separate the
data sets listed in the profile.
A data set separation profile contains at least one data set separation group. Each data set
separation group specifies whether separation is at the PCU or volume level and whether it is
required or preferred. It also includes a list of data set names to be separated from each other
during allocation.
Restriction: You cannot use data set separation when allocating non-SMS-managed data
sets or during use of full volume copy utilities such as PPRC.
Separation profile
The syntax for the data set separation profiles is defined as follows:
SEPARATIONGROUP(PCU) This indicates that separation is on the PCU level.
SEPARATIONGROUP(VOLUME) This indicates that separation is on the volume level.
VOLUME may be abbreviated as VOL.
TYPE(REQUIRED) This indicates that separation is required. SMS fails the
allocation if the specified data set or data sets cannot be
separated from other data sets on the specified level

(PCU or volume). REQUIRED may be abbreviated as
REQ or R.
TYPE(PREFERRED) This indicates that separation is preferred. SMS allows
the allocation if the specified data set or data sets cannot
be separated from other data sets on the specified level.
SMS allocates the data sets and issues an allocation
message that indicates that separation was not honored
for a successful allocation. PREFERRED may be
abbreviated as PREF or P.
DSNLIST(data-set-name[,,...]) This specifies the names of the data sets that are to be
separated. The data set names must follow the naming
convention described in z/OS MVS JCL Reference,
SA22-7597. You can specify the same data set name in
multiple data set separation groups. Wildcard characters
are supported, beginning with the third qualifier. For
example, you can specify XYZ.TEST.* but not XYZ.*.
The wildcard characters are as follows:
* - This indicates that either a qualifier or one or
more characters within a qualifier can occupy that
position. An asterisk (*) can precede or follow a set
of characters.
** - This indicates that zero (0) or more qualifiers can
occupy that position. A double asterisk (**) cannot
precede or follow any characters; it must be
preceded or followed by either a period or a blank.
% - This indicates that exactly one alphanumeric or
national character can occupy that position. You can
specify up to eight % characters in each qualifier.
Note: If only one data set name is specified with DSNLIST, the data set name must contain
at least one wildcard character.
Earlier syntax
The following earlier form of the syntax for SEPARATIONGROUP is tolerated by z/OS V1R11.
It supports separation at the PCU level only.
SEPARATIONGROUP|SEP
FAILLEVEL|FAIL ({PCU|NONE})
DSNLIST|DSNS|DSN (data-set-name[,data-set-name,...])

4.51 Data facility sort (DFSORT)
Source Data Set

Marie
SORT Carolina
SORT
Paula Ana Ana
Miriam Cassio Carolina
Marie Dovi Cassio
Enete Miriam Dovi
Dovi Enete Enete
Cassio Paula Marie
Carolina Miriam
Ana Paula
Descending order Ascending order
Figure 4-62 DFSORT example
Data facility sort (DFSORT)

The DFSORT licensed program is a high performance data arranger for z/OS users. Using
DFSORT you can sort, merge, and copy data sets using EBCDIC, z/Architecture decimal or
binary keys. It also helps you to analyze data and produce detailed reports using the
ICETOOL utility or the OUTFIL function. DFSORT is an optional feature of z/OS.
DFSORT, together with DFSMS and RACF, form the strategic product base for the evolving
system-managed storage environment. DFSORT is designed to optimize the efficiency and
speed with which operations are completed through synergy with processor, device, and
system features (for example, memory objects, Hiperspace, data space, striping,
compression, extended addressing, DASD and tape device architecture, processor memory,
and processor cache.
DFSORT example
The simple example in Figure 4-62 illustrates how DFSORT merges data sets by combining
two or more files of sorted records to form a single data set of sorted records.
You can use DFSORT to do simple application tasks such as alphabetizing a list of names, or
you can use it to aid complex tasks such as taking inventory or running a billing system. You
can also use DFSORT's record-level editing capability to perform data management tasks.
For most of the processing done by DFSORT, the whole data set is affected. However, certain
forms of DFSORT processing involve only certain individual records in that data set.

DFSORT functions
DFSORT adds the ability to do faster and easier sorting, merging, copying, reporting and
analysis of your business information, as well as versatile data handling at the record, fixed
position/length or variable position/length field, and bit level. While sorting, merging, or
copying data sets, you can also:
Select a subset of records from an input data set. You can include or omit records that
meet specified criteria. For example, when sorting an input data set containing records of
course books from many different school departments, you can sort the books for only one
department.
Reformat records, add or delete fields, and insert blanks, constants, or binary zeros. For
example, you can create an output data set that contains only certain fields from the input
data set, arranged differently.
Sum the values in selected records while sorting or merging (but not while copying). In the
example of a data set containing records of course books, you can use DFSORT to add up
the dollar amounts of books for one school department.
Create multiple output data sets and reports from a single pass over an input data set. For
example, you can create a different output data set for the records of each department.
Sort, merge, include, or omit records according to the collating rules defined in a selected
local.
Alter the collating sequence when sorting or merging records (but not while copying). For
example, you can have the lowercase letters collate after the uppercase letters.
Sort, merge, or copy Japanese data if the IBM Double Byte Character Set Ordering
Support (DBCS Ordering, the 5665-360 Licensed Program, Release 2.0 or an equivalent
product) is used with DFSORT to process the records.
DFSORT and ICETOOL

DFSORT has utilities such as ICETOOL, which is a multipurpose DFSORT utility that uses
the capabilities of DFSORT to perform multiple operations on one or more data sets in a
single step.
Tip: You can use DFSORT's ICEGENER facility to achieve faster and more efficient
processing for applications that are set up to use the IEBGENER system utility. For more
information, see z/OS DFSORT Application Programming Guide, SC26-7523.
DFSORT customization
Specifying the DFSORT customization parameters is a very important task for z/OS system
programmers. Depending on such parameters, DFSORT may use lots of system resources
such as CPU, I/O, and especially virtual storage. The uncontrolled use of virtual storage may
cause IPLs due to the lack of available slots in page data sets. Plan to use the IEFUSI z/OS
exit to control products such as DFSORT.
For articles, online books, news, tips, techniques, examples, and more, visit the z/OS
DFSORT home page:
http://www-1.ibm.com/servers/storage/support/software/sort/mvs

4.52 z/OS Network File System (z/OS NFS)
z/OS
DFSMS AMS
NETWORK
HP/UX
AIX FILE
MVS Data Sets
SYSTEM
SERVER
OMVS
TCP/IP
(z/OS NFS) Network
UNIX Hierarchical File
z/OS
System
DFSMS
NETWORK
FILE
SYSTEM
Other NFS CLIENT
Sun Solaris Client and
Servers
Figure 4-63 DFSMS Network File System
z/OS Network File System (z/OS NFS)

The z/OS Network File System is a distributed file system that enables users to access UNIX
files and directories that are located on remote computers as though they were local. NFS is
independent of machine types, operating systems, and network architectures. Use the NFS
for file serving (as a data repository) and file sharing between platforms supported by z/OS.
Clients and servers

A client is a computer or process that requests services in the network. A server is a
computer or process that responds to a request for service from a client. A user accesses a
service, which allows the use of data or other resources.
Figure 4-63 illustrates the client-server relationship:
The upper center portion shows the DFSMS NFS address space server; the lower portion
shows the DFSMS NFS address space client.
The left side of the figure shows various NFS clients and servers that can interact with the
DFSMS NFS server and client.
In the center of the figure is the Transmission Control Protocol/Internet Protocol (TCP/IP)
network used to communicate between clients and servers.
With the NFS server, you can remotely access z/OS conventional data sets or z/OS UNIX
files from workstations, personal computers, and other systems that run client software for the

Sun NFS version 2 protocols, the Sun NFS version 3 protocols, and the WebNFS protocols
over TCP/IP network.
The z/OS NFS server acts as an intermediary to read, write, create, or delete z/OS UNIX files
and MVS data sets that are maintained on an MVS host system. The remote MVS data sets
or z/OS UNIX files are mounted from the host processor to appear as local directories and
files on the client system.
This server makes the strengths of a z/OS host processorstorage management,

high-performance disk storage, security, and centralized dataavailable to the client
platforms.
With the NFS client you can allow basic sequential access method (BSAM), queued
sequential access method (QSAM), virtual storage access method (VSAM), and z/OS UNIX
users and applications transparent access to data on systems that support the Sun NFS
version 2 protocols and the Sun NFS version 3 protocols.
The Network File System can be used for:

File sharing between platforms
File serving (as a data repository)
Supported clients for the NFS server

The z/OS NFS client supports all servers that implement the server portion of the Sun NFS
Version 2 and Version 3 protocols. The z/OS NFS client does not support NFS version 4.
Tested clients for the z/OS NFS server, using the NFS version 4 protocol, are:
IBM RS/6000 AIX version 5.3
Sun Solaris version 10
Enterprise Linux 4
Windows 2000/XP with Hummingbird Maestro 9 and Maestro 10
Other client platforms should work as well because NFS version 4 is an industry standard
protocol, but they have not been tested by IBM.
NFS client software for other IBM platforms is available from other vendors. You can also
access the NFS server from non-IBM clients that use the NFS version 2 or version 3 protocol,
including:
DEC stations running DEC ULTRIX version 4.4
HP 9000 workstations running HP/UX version 10.20
Sun PC-NFS version 5
Sun workstations running SunOS or Sun Solaris versions 2.5.3
For further information about NFS, refer to z/OS Network File System Guide and Reference,
SC26-7417, and visit:
http://www-1.ibm.com/servers/eserver/zseries/zos/nfs/

4.53 DFSMS optimizer (DFSMSopt)
Figure 4-64 DFSMS Optimizer: Data set performance summary
DFSMS Optimizer (DFSMSopt)

The DFSMS Optimizer gives you the ability to understand how you manage storage today.
With that information, you can make informed decisions about how you should manage
storage in the future.
These are the analyzers you can use:

Performance analyzer
Management class analyzer
I/O trace analyzer
DFSMS Optimizer uses input data from several sources in the system and processes it using
an extract program that merges the data and builds the Optimizer database.
By specifying different filters you can produce reports that help you build a detailed storage
management picture of your enterprise. With the report data, you can use the charting facility
to produce color charts and graphs.
The DFSMS Optimizer provides analysis and simulation information for both SMS and
non-SMS users. The DFSMS Optimizer can help you maximize storage use and minimize
storage costs. It provides methods and facilities for you to:
Monitor and tune DFSMShsm functions such as migration and backup
Create and maintain a historical database of system and data activity

Fine-tune an SMS configuration by performing in-depth analysis of:
Management class policies, including simulations and cost/benefit analysis using your
storage component costs
Storage class policies for SMS data, with recommendations for both SMS and
non-SMS data
High I/O activity data sets, including recommendations for placement and simulation for
cache and expanded storage
Storage hardware performance of subsystems and volumes including I/O rate,
response time, and caching statistics
Simulate potential policy changes and understand the costs of those changes
Produce presentation-quality charts
For more information about the DFSMS Optimizer, see DFSMS Optimizer Users Guide and
Reference, SC26-7047, or visit:
http://www-1.ibm.com/servers/storage/software/opt/

4.54 Data Set Services (DFSMSdss)
//JOB2 JOB accounting information,REGION=nnnnK

//STEP1 EXEC PGM=ADRDSSU
//DASD1 DD UNIT=3390,VOL=(PRIVATE,SER=111111),DISP=OLD
//TAPE DD UNIT=3490,VOL=SER=TAPE02,
// LABEL=(1,SL),DISP=(NEW,CATLG),DSNAME=USER2.BACKUP
//SYSIN DD *
DUMP INDDNAME(DASD1) OUTDDNAME(TAPE) -
DATASET(INCLUDE(USER2.**,USER3.*))
RESTORE...
TAPECNTL...
TSO
Figure 4-65 DFSMSdss backing up and restoring volumes and data sets
Data Set Services (DFSMSdss)

DFSMSdss is a direct access storage device (DASD) data and space management tool.
DFSMSdss works on DASD volumes only in a z/OS environment. You can use DFSMSdss to
do the following:
Copy and move data sets between volumes of like and unlike device types.
Note: Like devices have the same track capacity and number of tracks per cylinder (for
example, 3380 Model D, Model E, and Model K). Unlike DASD devices have different
track capacities (for example, 3380 and 3390), a different number of tracks per cylinder,
or both.
Dump and restore data sets, entire volumes, or specific tracks.

Convert data sets and volumes to and from SMS management.
Compress partitioned data sets.
Release unused space in data sets.
Reduce or eliminate DASD free-space fragmentation by consolidating free space on a
volume.
Implement concurrent copy or a flashcopy in ESS or DS8000 DASD controllers.

Backing up and restoring volumes and data sets with DFSMSdss
You can use the DFSMSdss DUMP command to back up volumes and data sets, and you can
use the DFSMSdss RESTORE command to recover them. You can make incremental
backups of your data sets by specifying a data set DUMP command with RESET and filtering
on the data-set-changed indicator.
The DFSMShsm component of DFSMS provides automated incremental backup, interactive

recovery, and inventory of what it backs up. If DFSMShsm is used, you should use
DFSMSdss for volume backup of data sets not supported by DFSMShsm and for dumping
SYSRES and special volumes such as the one containing the master catalog, as shown in
Figure 4-65 on page 196. If DFSMShsm is not installed, you can use DFSMSdss for all
volume and data set backups.

4.55 DFSMSdss: Physical and logical processing
Physical
or
Logical ?
TSO
Figure 4-66 DFSMSdss physical and logical processing
DFSMSdss: physical and logical processing

Before you begin using DFSMSdss, you should understand the difference between logical
processing and physical processing and how to use data set filtering to select data sets for
processing. DFSMSdss can perform two kinds of processing when executing COPY, DUMP, and
RESTORE commands:
Logical processing operates against data sets independently of physical device format.
Physical processing moves data at the track-image level and operates against volumes,
tracks, and data sets.
Each type of processing offers different capabilities and advantages.
During a restore operation, the data is processed the same way it is dumped because
physical and logical dump tapes have different formats. If a data set is dumped logically, it is
restored logically; if it is dumped physically, it is restored physically. A data set restore
operation from a full volume dump is a physical data set restore operation.

4.56 DFSMSdss: Logical processing
UCAT
DUMP
ABC.FILE
ABC.FILE
TSO
VOLABC
ABC.FILE
DU
MP
01
Figure 4-67 DFSMSdss logical processing
Logical processing
A logical copy, dump, or restore operation treats each data set and its associated information
as a logical entity, and processes an entire data set before beginning the next one.
Each data set is moved by tracks from the source device and is potentially written to the target
device as a set of data records, allowing data movement between devices with different track
and cylinder configurations. Checking of data record consistency is not performed during
dump operation.
DFSMSdss performs logical processing if:

You specify the DATASET keyword with the COPY command. A data set copy is always a
logical operation, regardless of how or whether you specify input volumes.
You specify the DATASET keyword with the DUMP command, and either no input volume is
specified, or LOGINDDNAME, LOGINDYNAM, or STORGRP is used to specify input
volumes.
The RESTORE command is performed, and the input volume was created by a logical dump.
Catalogs and VTOCs are used to select data sets for logical processing. If you do not specify
input volumes, the catalogs are used to select data sets for copy and dump operations. If you
specify input volumes using the LOGINDDNAME, LOGINDYNAM, or STORGRP keywords on
the COPY or DUMP command, DFSMSdss uses VTOCs to select data sets for processing.

When to use logical processing
Use logical processing for the following situations:
Data is copied to an unlike device type.
Logical processing is the only way to move data between unlike device types.
Data that may need to be restored to an unlike device is dumped.
Data must be restored the same way it is dumped.
This is particularly important to bear in mind when making backups that you plan to retain
for a long period of time (such as vital records backups). If a backup is retained for a long
period of time, it is possible that the device type it originally resided on will no longer be in
use at your site when you want to restore it. This means you will have to restore it to an
unlike device, which can be done only if the backup has been made logically.
Aliases of VSAM user catalogs are to be preserved during copy and restore functions.
Aliases are not preserved for physical processing.
Unmovable data sets or data sets with absolute track allocation are moved to different
locations.
Multivolume data sets are processed.
VSAM and multivolume data sets are to be cataloged as part of DFSMSdss processing.
Data sets are to be deleted from the source volume after a successful dump or copy
operation.
Both non-VSAM and VSAM data sets are to be renamed after a successful copy or restore
operation.
You want to control the percentage of space allocated on each of the output volumes for
copy and restore operations.
You want to copy and convert a PDS to a PDSE or vice versa.
You want to copy or restore a data set with an undefined DSORG to an unlike device.
You want to keep together all parts of a VSAM sphere.

4.57 DFSMSdss: Physical processing
DUMP FULL
CACSW3
TSO
DU
MP
0 1
Figure 4-68 DFSMSdss physical processing
Physical processing
Physical processing moves data based on physical track images. Because data movement is
carried out at the track level, only target devices with track sizes equal to those of the source
device are supported. Physical processing operates on volumes, ranges of tracks, or data
sets. For data sets, it relies only on volume information (in the VTOC and VVDS) for data set
selection, and processes only that part of a data set residing on the specified input volumes.
DFSMSdss performs physical processing if:

You specify the FULL or TRACKS keyword with the COPY or DUMP command. This results in
a physical volume or physical tracks operation.
Attention: Take care when invoking the TRACKS keyword with the COPY and RESTORE
commands. The TRACKS keyword should be used only for a data recovery operation.
For example, you can use it to repair a bad track in the VTOC or a data set, or to
retrieve data from a damaged data set. You cannot use it in place of a full-volume or a
logical data set operation. Doing so can destroy a volume or impair data integrity.
You specify the data set keyword on the DUMP command and input volumes with the
INDDNAME or INDYNAM parameter. This produces a physical data set dump.
The RESTORE command is executed and the input volume is created by a physical dump
operation.

When to use physical processing
Use physical processing when:
Backing up system volumes that you might want to restore with a stand-alone DFSMSdss
restore operation.
Stand-alone DFSMSdss restore supports only physical dump tapes.
Performance is an issue.
Generally, the fastest way (measured by elapsed time) to copy or dump an entire volume is
by using a physical full-volume command. This is primarily because minimal catalog
searching is necessary for physical processing.
Substituting one physical volume for another or recovering an entire volume.
With a COPY or RESTORE (full volume or track) command, the volume serial number of the
input DASD volume can be copied to the output DASD volume.
Dealing with I/O errors.
Physical processing provides the capability to copy, dump, and restore a specific track or
range of tracks.
Dumping or copying between volumes of the same device type but different capacity.

4.58 DFSMSdss stand-alone services
DFD
SS
Sta
nd -
A lo
ne
Ta p
e
Figure 4-69 DFSMSdss stand-alone services
DFSMSdss stand-alone services

The DFSMSdss stand-alone restore function is a single-purpose program. It is designed to
allow the system programmer to restore vital system packs during disaster recovery without
relying on an MVS environment.
Stand-alone services can perform either a full-volume restore or a tracks restore from dump
tapes produced by DFSMSdss or DFDSS and offers the following benefits:
Provides user-friendly commands to replace the previous control statements
Supports IBM 3494 and 3495 Tape Libraries, and 3590 Tape Subsystems
Supports IPLing from a DASD volume, in addition to tape and card readers
Allows you to predefine the operator console to be used during stand-alone services
processing
For detailed information about the stand-alone service and other DFSMSdss information,
refer to z/OS DFSMSdss Storage Administration Reference, SC35-0424, and z/OS
DFSMSdss Storage Administration Guide, SC35-0423, and visit:
http://www-1.ibm.com/servers/storage/software/sms/dss/

4.59 Hierarchical Storage Manager (DFSMShsm)
Availability Space
Automatic Backup
Incremental Backup
Figure 4-70 DFSMShsm
Hierarchical Storage Manager (DFSMShsm)

Hierarchical Storage Manager (DFSMShsm) is a disk storage management and productivity
product for managing low activity and inactive data. It provides backup, recovery, migration,
and space management functions as well as full-function disaster recovery support.
DFSMShsm improves disk use by automatically managing both space and data availability in
a storage hierarchy.
Availability management is used to make data available by automatically copying new and
changed data set to backup volumes.
Space management is used to manage DASD space by enabling inactive data sets to be
moved off fast-access storage devices, thus creating free space or new allocations.
DFSMShsm also provides for other supporting functions that are essential to your
installation's environment.
For further information about DFSMShsm, refer to z/OS DFSMShsm Storage Administration
Guide, SC35-0421 and z/OS DFSMShsm Storage Administration Reference, SC35-0422,
and visit:
http://www-1.ibm.com/servers/storage/software/sms/hsm/

4.60 DFSMShsm: Availability management
SMS-managed Non-SMS-managed
Primary and
Storage Groups
Secondary Volumes
(volumes)
User
Catalog
DFSMShsm
Control
Backup Data
Sets
Functions
TAPE
Figure 4-71 DFSMShsm availability management
DFSMShsm backs up your data, automatically or by command, to ensure availability if
accidental loss of the data sets or physical loss of volumes should occur. DFSMShsm also
allows the storage administrator to copy backup and migration tapes, and to specify that
copies be made in parallel with the original. You can store the copies on site as protection
from media damage, or offsite as protection from site damage. DFSMShsm also provides
disaster backup and recovery for user-defined groups of data sets (aggregates) so that you
can restore critical applications at the same location or at an offsite location.
Note: You must also have DFSMSdss to use the DFSMShsm functions.
Availability management ensures that a recent copy of your DASD data set exists. The
purpose of availability management is to ensure that lost or damaged data sets can be
retrieved at the most current possible level. DFSMShsm uses DFSMSdss as a fast data
mover for backups. Availability management automatically and periodically performs functions
that:
1. Copy all the data sets on DASD volumes to tape volumes
2. Copy the changed data sets on DASD volumes (incremental backup) either to other DASD
volumes or to tape volumes
DFSMShsm minimizes the space occupied by the data sets on the backup volume by using
compression and stacking.

Tasks for availability management functions
The tasks for controlling automatic availability management of SMS-managed storage require
adding DFSMShsm commands to the ARCCMDxx parmlib member and specifying attributes
in the SMS storage classes and management classes. It is assumed that the storage classes
and management classes have already been defined.
The attribute descriptions explain the attributes to be added to the previously defined storage
groups and management classes. Similarly, the descriptions of DFSMShsm commands relate
to commands to be added to the ARCCMDxx member of SYS1.PARMLIB.
Two groups of tasks are performed for availability management: dump tasks and backup
tasks. Availability management comprises the following functions:
Aggregate backup and recovery (ABARS)
Automatic physical full-volume dump
Automatic incremental backup
Automatic control data set backup
Command dump and backup
Command recovery
Disaster backup
Expiration of backup versions
Fast replication backup and recovery
Command availability management

Commands cause availability management functions to occur, resulting in the following
conditions:
One data set to be backed up.
All changed data sets on a volume to be backed up.
A volume to be dumped.
A backed-up data set to be recovered.
A volume to be restored from a dump and forward recovered from later incremental
backup versions. Forward recovery is a process of updating a restored volume by
applying later changes as indicated by the catalog and the incremental backup versions.
A volume to be restored from a dump copy.
A volume to be recovered from backup versions.
A specific data set to be restored from a dump volume.
All expired backup versions to be deleted.
A fast replication backup version to be recovered from a copy pool.

4.61 DFSMShsm: Space management
SMS-managed Non-SMS-managed
Primary
Storage Groups
Volumes
(volumes)
User
Catalog
DFSMShsm
Control
Data
Migration Sets
Level 1 DASD
TAPE Migration DASD

Level 2
Figure 4-72 DFSMShsm space management
Space management
Space management is the function of DFSMShsm that allows you to keep DASD space
available for users in order to meet the service level objectives for your system. The purpose
of space management is to manage your DASD storage efficiently. To do this, space
management automatically and periodically performs functions that:
1. Move low activity data sets (using DFSMSdss) from user-accessible volumes to
DFSMShsm volumes
2. Reduce the space occupied by data on both the user-accessible volumes and the
DFSMShsm volumes
DFSMShsm improves DASD space usage by keeping only active data on fast-access storage
devices. It automatically frees space on user volumes by deleting eligible data sets, releasing
overallocated space, and moving low-activity data to lower cost-per-byte devices, even if the
job did not request tape.
Space management functions

The DFSMShsm space management functions are:
Automatic primary space management of DFSMShsm-managed volumes, which includes:
Deletion of temporary data sets
Deletion of expired data sets

Release of unused, over-allocated space
Migration to DFSMShsm-owned migration level 1 (ML1) volumes (compressed)
Automatic secondary space management of DFSMShsm-owned volumes, which includes:
Migration-level cleanup, including deletion of expired migrated data sets and migration
control data set (MCDS) records
Moving migration copies from migration level 1 (ML1) to migration level 2 (ML2)
volumes
Automatic interval migration, initiated when a DFSMShsm-managed volume exceeds a
specified threshold
Automatic recall of user data sets back to DASD volumes, when referenced by the
application
Space management by command
Space-saving functions, which include:
Data compaction and data compression. Compaction provides space savings through
fewer gaps and less control data. Compression provides a more compact way to store
data.
Partitioned data set (PDS) free space compression.
Small data set packing (SDSP) data set facility, which allows small data sets to be
packaged in just one physical track.
Data set reblocking.
It is possible to have more than one z/OS image sharing the same DFSMShsm policy. In this
case one of the DFSMShsm images is the primary host and the others are secondary. The
primary HSM host is identified by HOST= in the HSM startup and is responsible for:
Hourly space checks
During auto backup: CDS backup, backup of ML1 data sets to tape
During auto dump: Expiration of dump copies and deletion of excess dump VTOC copy
data sets
During secondary space management (SSM): Cleanup of MCDS, migration volumes, and
L1-to-L2 migration
If you are running your z/OS HSM images in sysplex (parallel or basic), you can use
secondary host promotion to allow a secondary image to assume the primary image's tasks if
the primary host fails. Secondary host promotion uses XCF status monitoring to execute the
promotion. To indicate a system as a candidate, issue:
SETSYS PRIMARYHOST(YES)
and
SSM(YES)

4.62 DFSMShsm: Storage device hierarchy
Primary Volumes
or Level 0 (ML0)
Migration
Migration Level 1 (ML1)
Level 2 (ML2)
Migration
Level 2 (ML2)
Figure 4-73 Storage device hierarchy
Storage device hierarchy

A storage device hierarchy consists of a group of storage devices that have different costs for
storing data, different amounts of data stored, and different speeds of accessing the data,
processed as follows:
Level 0 volumes A volume that contains data sets that are directly accessible by the
user. The volume may be either DFSMShsm-managed or
non-DFSMShsm-managed.
Level 1 volumes A volume owned by DFSMShsm containing data sets that migrated
from a level 0 volume.
Level 2 volumes A volume under control of DFSMShsm containing data sets that
migrated from a level 0 volume, from a level 1 volume, or from a
volume not managed by DFSMShsm.
DFSMShsm storage device management

DFSMShsm uses the following three-level storage device hierarchy for space management:
Level 0 - DFSMShsm-managed storage devices at the highest level of the hierarchy; these
devices contain data directly accessible to your application.
Level 1 and Level 2 - Storage devices at the lower levels of the hierarchy; level 1 and
level 2 contain data that DFSMShsm has compressed and optionally compacted into a
format that you cannot use. Devices at this level provide lower cost per byte storage and

usually slower response time. Usually L1 is in a cheaper DASD (or the same cost, but with
the gain of compression) and L2 is on tape.
Note: If you have a DASD controller that compresses data, you can skip level 1 (ML1)
migration because the data in L0 is already compacted/compressed.
Defining management class migration attributes

To DFSMShsm, a data set occupies one of two distinct states in storage:
Primary Also known as level 0, the primary state indicates that users can directly
access a data set residing on a volume.
Migrated Users cannot directly access data sets that have migrated from the primary
state to a migrated state. To be accessed, the data sets must be recalled to
primary storage. A migrated data set can reside on either migration level 1
(usually permanently mounted DASD) or migration level 2 (usually tape).
A data set can move back and forth between these two states, and it can move from level 0 to
migration level 2 (and back) without passing through migration level 1. Objects do not migrate.
Movement back to level 0 is known as recall.

4.63 ML1 enhancements with z/OS V1R11
Back up and migrate data sets larger than 64 K to disk

as large format sequential (LFS) data sets
Utilize OVERFLOW volumes for migrated data sets and
backup copies
Specify data set size eligibility for OVERFLOW volumes
and ML1 OVERFLOW volume pool threshold
Larger data sets (>64 K tracks) no longer have to be
directed to tape
Data sets larger than 58 K tracks will be allocated as LFS
data sets regardless of whether they are on a
OVERFLOW or NOOVERFLOW volume
OVERFLOW volumes can be used as a repository for
larger data sets
Customize an installation to exploit the OVERFLOW
volume pool according to a specified environment
Figure 4-74 ML1 enhancements with z/OS V1R11
ML1 enhancements
Beginning in V1R11, DFSMShsm enables ML1 overflow volumes to be selected for migration
processing, in addition to their current use for data set backup processing. DFSMShsm
enables these ML1 overflow volumes to be selected for migration or backup of large data
sets, with the determining size values specified by a new parameter of the SETSYS command.
Use the new ML1OVERFLOW parameter with the subparameter of DATASETSIZE(dssize) to
specify the minimum size that a data set must be in order for DFSMShsm to prefer ML1
overflow volume selection for migration or backup copies.
In addition, DFSMShsm removes the previous ML1 volume restriction against migrating or
backing up a data set whose expected size after compaction (if active and used) is greater
than 65,536 tracks The new limit for backed up or migrated copies is equal to the maximum
size limit for the largest volume available.
ML1OVERFLOW option
ML1OVERFLOW is an optional parameter specifying the following:
The minimum data set size for ML1 OVERFLOW volume preference
The threshold for ML1 OVERFLOW volume capacity for automatic secondary space
management migration from ML1 OVERFLOW to ML2 volumes

DATASSETSIZE(dssize)
This specifies the minimum size, in K bytes, of the data set for which an ML1 OVERFLOW
volume is preferred for migration or backup. This includes invoking inline backup, the HBACKDS
or BACKDS commands, or the ARCHBACK macro.
This parameter has the same meaning when applied to SMS-managed or

non-SMS-managed DASD volumes or data sets. The SETSYS default is 2000000 K bytes.
The DFSMShsm default, if you do not specify this parameter on any SETSYS command, will
use the value of 2000000 K bytes. If the calculated size of a data set being backed up or
migrated is equal to or greater than dssize, then DFSMShsm prefers the OVERFLOW ML1
volume with the least amount of free space that still has enough remaining space for the
data set.
If the calculated size of the data set is less than the minimum size specified in dssize, then
DFSMShsm prefers the NOOVERFLOW ML1 volume with the maximum amount of free
space and least number of users.
OVERFLOW and NOOVERFLOW volumes

DFSMShsm can create backup copies to either ML1 backup OVERFLOW volumes or
NOOVERFLOW volumes. Use the SETSYS ML1OVERFLOW(DATASETSIZE(dssize)) command to
specify the minimum size in K bytes (where K=1024) of the data set for which an ML1
OVERFLOW volume is preferred for the migration or backup copy.
For data sets smaller than 58 K tracks, DFSMShsm allocates a basic sequential format data
set for the backup copy. For data sets larger than 58 K tracks, DFSMShsm allocates a large
format sequential data set for the backup copy.
Basic or large format sequential data sets will prefer OVERFLOW or NOOVERFLOW
volumes based on the SETSYS ML1OVERFLOW(DATASETSIZE(dssize)) value. If there is
not enough free space on the NOOVERFLOW or OVERFLOW volume for a particular backup
copy, then DFSMShsm tries to create the backup on a OVERFLOW or NOOVERFLOW
volume, respectively. If the data set is too large to fit on a single ML1 volume (OVERFLOW or
NOOVERFLOW), then the migration or backup fails.

4.64 DFSMShsm z/OS V1R11 enhancements
Data sets that were larger than 64 K tracks in size can

migrate/backup only to tape
Can now migrate/backup a data set larger than 64 K
tracks
Can go to DASD as large format sequential (LFS) data
sets
OVERFLOW volumes can be only used for backups
and preference is not given to them for data sets
larger in size
OVERFLOW volumes can now be utilized for
migration and backup copies for data sets of an
installation-specified size
Improves free space utilization on ML1 volumes
Figure 4-75 OVERFLOW volumes used for backup
OVERFLOW or NOOVERFLOW use of level 1 volumes

Using these type of volumes for data set migration and backup works as follows:
OVERFLOW and NOOVERFLOW are mutually exclusive optional subparameters of the
MIGRATION parameter that you use to specify how a level 1 volume is considered during
selection for placement of a data set migration or backup version.
OVERFLOW specifies that the volume is considered if either of the following are true:
The data you are migrating or backing up is larger than a given size, as specified on the
SETSYS ML1OVERFLOW(DATASETSIZE(dssize)) command.
DFSMShsm cannot allocate enough space on a NOOVERFLOW volume.
NOOVERFLOW specifies that the volume is considered with other level 1 volumes for
migration data and backup versions of any size.
Defaults: If you are adding a migration volume to DFSMShsm, the default is
NOOVERFLOW. If you are changing the attributes of a volume and do not specify
either subparameter, the overflow attribute is not changed.
Migration and backup to tape

Using the DFSMShsm ML1 enhancements, the installation now can:
Back up and migrate data sets larger than 64 K to disk as large format sequential (LFS)
data sets.

Utilize OVERFLOW volumes for migrated data sets and backup copies.
Specify data set size eligibility for OVERFLOW volumes and ML1 OVERFLOW volume
pool threshold.
Therefore, data sets larger than 64 K tracks no longer have to be directed to tape. Such data
sets larger will be allocated as LFS data sets regardless whether they are on an OVERFLOW
or NOOVERFLOW volume. OVERFLOW volumes can be used as a repository for larger data
sets. You can now customize your installation to exploit the OVERFLOW volume pool
according to a specified environment.
Installation considerations
A coexistence APAR will be required to enable downlevel DFSMShsm to tolerate migrated or
backup copies of LFS format DFSMShsm data sets. APAR OA26330 enables processing of
large format sequential migration and backup copies to be processed on lower level
installations.
Downlevel DFSMShsm installations (pre-V1R11) will be able to recall and recover data sets
from a V1R11 DFSMShsm LFS migration or backup copies. For OVERFLOW volumes on
lower level systems, recalls and recovers will be successful. Migrations from lower level
systems to the V1R11 OVERFLOW volumes will not be allowed because the OVERFLOW
volumes will not be included in the volume selection process.
Large data sets can and will migrate and backup to ML1 DASD. These will be large format
sequential HSM migration data sets on ML1. OVERFLOW volumes will be used for migration
now, in addition to backup. We anticipate that not many installations used OVERFLOW
volumes before but if they were used, then migration actions are needed.

4.65 ML1 and ML2 volumes
DFSMShsm will allocate migration copies and backup

copies for large data sets (>58 K tracks) as large
format physical sequential (LFS) data sets
Migration occurs to either of two levels: migration level
1 or migration level 2
Migration level 1 (ML1) volumes are always DASD
Migration level 2 (ML2) volumes can be either DASD or
tape
The computing center controls which volumes are to
be used as migration volumes
The maximum migration or backup copy size will now be
limited to the size of the ML1 volumes free space
Figure 4-76 ML1 and ML2 volumes
DFSMShsm migration levels

With z/OS V1R11, larger data sets (>58 K tracks) will now be DFSMShsm-managed as large
format sequential data sets when migrated or backed up. OVERFLOW volumes can now be
used for migration and backup. An installation can adjust the values that direct data sets to
OVERFLOW versus NOOVERFLOW volumes and the threshold of the OVERFLOW volume
pool.
If you back up or migrate data sets to ML1 OVERFLOW volumes, you can specify the
percentage of occupied space that must be in the ML1 OVERFLOW volume pool before the
migration of data sets to ML2 volumes occurs during automatic secondary space
management.
Migration volume levels 1 or 2

There is a check to see there is any command level 1 to level 2 running. If yes, the
level 1-to-level 2 migration of automatic secondary space management will not start. There is
no reason to run two level 1-to-level 2 migrations at the same time.

Before automatic secondary space management starts level 1-to-level 2 migration, it performs
the following checks:
It performs a space check on the all ML1 volumes. If any NOOVERFLOW ML1 volume has
an occupancy that is equal to or greater than its high threshold, then all eligible data sets
migrate from the NOOVERFLOW ML1 volume to ML2 volumes.
For OVERFLOW volumes, the percentage of occupied space for each ML1 OVERFLOW
volume will be calculated and then DFSMShsm will compare the average value with the
specified threshold. Secondary space management is started for all eligible ML1
OVERFLOW data sets if the overflow volume pool threshold is met or exceeded.
If you choose to use OVERFLOW volumes, you can specify the average overflow volume
pool capacity threshold at which you want automatic secondary space management to
migrate ML1 OVERFLOW volumes to ML2 volumes by using the SETSYS ML1OVERFLOW
command.
ML1 and ML2 volumes

The default setting for the ML1 OVERFLOW volume pool threshold is 80%. This requires the
occupied space of the entire overflow volume pool reach 80% full before any data is moved
off of OVERFLOW volumes to ML2 volumes.
ML2 volumes can be either DASD or tape. The TAPEMIGRATION parameter of the SETSYS
command specifies what type of ML2 volume is used. The SETSYS command for DFSMShsm
host 2 specifies ML2 migration to tape.
Note: If you want DFSMShsm to perform automatic migration from ML1 to ML2 volumes,
you must specify the thresholds of occupancy parameter (of the ADDVOL command) for the
ML1 volumes.

4.66 Data set allocation format and volume pool determination
Migrate or Backup SETSYS

ML1OVERFLOW,
(DATASETSIZE(200000),
THRESHOLD(80))
2 GB = ~36K tracks
ML1 NOOVERFLOW
VOLUME POOL
ML2
L0
D/S Data Set Size ML1 Copy Format
100 K Tracks Large Format Sequential

ML1 OVERFLOW
40 K Tracks Basic Format Sequential VOLUME POOL
500 Tracks Basic Format Sequential
Figure 4-77 Data set allocation for migration and backup
Migration and backup

With z/OS V1R11, Figure 4-77 shows how three different size data sets are processed when
migrating or backing up to DASD or tape using OVERFLOW and NOOVERFLOW ML1
volumes.
The 500 track data set (top, red arrow), when migrating, is allocated on ML1
NOOVERFLOW volumes as a basic format sequential data set. This is how current HSM
processing works today.
The 40 K track data set (middle, blue arrow) migrates to an OVERFLOW volume because
it is larger than the specified DATASETSIZE parameter and the migration copy will be a
basic format sequential data set because its expected size is smaller than 58 K tracks.
The 100 K tracks data set (bottom, green arrow) are allocated as a large format sequential
data set on a volume in the ML1 OVERFLOW volume pool.
Backing up an individual data set

The ML1OVERFLOW(DATASETSIZE(dssize)) command to specify the minimum size in K bytes
(where K=1024) of the data set for which an ML1 OVERFLOW volume is preferred for the
migration or backup copy.
For data sets smaller than 58 K tracks, DFSMShsm allocates a basic sequential format data
set for the backup copy.

For data sets larger than 58 K tracks, DFSMShsm allocates a large format sequential data set
for the backup copy. Basic or large format sequential data sets will prefer OVERFLOW or
NOOVERFLOW volumes based on the SETSYS ML1OVERFLOW(DATASETSIZE(dssize))
value.
If there is not enough free space on the NOOVERFLOW or OVERFLOW volume for a
particular backup copy, DFSMShsm tries to create the backup on a OVERFLOW or
NOOVERFLOW volume respectively. If the data set is too large to fit on a single ML1 volume
(OVERFLOW or NOOVERFLOW), then the migration or backup fails.
Important: In processing these commands, DFSMShsm first checks the management

class for the data set to determine the value of the ADMIN OR USER COMMAND
BACKUP attribute.
If the value of the attribute is BOTH, a DFSMShsm-authorized user can use either of the
commands, and a non-DFSMShsm-authorized user can use the HBACKDS command to
back up the data set.
If the value of the attribute is ADMIN, a DFSMShsm-authorized user can use either of the
commands to back up the data set, but a non-DFSMShsm-authorized user cannot back up
the data set. If the value of the attribute is NONE, the command backup cannot be done.

4.67 DFSMShsm volume types
Fast
replication
Spill
Level 0
Backup
Migration
Daily Level 1 Migration
Backup Level 2
Aggregate Dump
backup Volumes
Figure 4-78 DFSMShsm volume types
DFSMShsm volume backup

Backing up an individual cataloged data set is performed in the same way as for
SMS-managed data sets. However, to back up individual uncataloged data sets, issue the
following commands:
BACKDS dsname UNIT(unittype) VOLUME(volser)
HBACKDS dsname UNIT(unittype) VOLUME(volser)
The HBACKDS form of the command can be used by either non-DFSMShsm-authorized or

DFSMShsm-authorized users. The BACKDS form of the command can be used only by
DFSMShsm-authorized users. The UNIT and VOLUME parameters are required because
DFSMShsm cannot locate an uncataloged data set without being told where it is.
Volume types
DFSMShsm supports the following volume types:
Level 0 (L0) volumes contain data sets that are directly accessible to you and the jobs you
run. DFSMShsm-managed volumes are those L0 volumes that are managed by the
DFSMShsm automatic functions. These volumes must be mounted and online when you
refer to them with DFSMShsm commands.

Migration level 1 (ML1) volumes are DFSMShsm-supported DASD on which DFSMShsm
maintains your data in DFSMShsm format. These volumes are normally permanently
mounted and online. They can be:
Volumes containing data sets that DFSMShsm migrated from L0 volumes.
Volumes containing backup versions created from a DFSMShsm BACKDS or HBACKDS
command. Backup processing requires ML1 volumes to store incremental backup and
dump VTOC copy data sets, and as intermediate storage for data sets that are backed
up by data set command backup.
Migration level 2 (ML2) volumes are DFSMShsm-supported tape or DASD on which
DFSMShsm maintains your data in DFSMShsm format. These volumes are normally not
mounted or online. They contain data sets migrated from ML1 volumes or L0 volumes.
Daily backup volumes are DFSMShsm-supported tape or DASD on which DFSMShsm
maintains your data in DFSMShsm format. These volumes are normally not mounted or
online. They contain the most current backup versions of data sets copied from L0
volumes. These volumes may also contain earlier backup versions of these data sets.
Spill backup volumes are DFSMShsm-supported tape or DASD on which DFSMShsm
maintains your data sets in DFSMShsm format. These volumes are normally not mounted
or online. They contain earlier backup versions of data sets, which were moved from
DASD backup volumes.
Dump volumes are DFSMShsm-supported tape. They contain image copies of volumes
that are produced by the full volume dump function of DFSMSdss (write a copy of the
entire allocated space of that volume), which is invoked by DFSMShsm.
Aggregate backup volumes are DFSMShsm-supported tape. These volumes are normally
not mounted or online. They contain copies of the data sets of a user-defined group of
data sets, along with control information for those data sets. These data sets and their
control information are stored as a group so that they can be recovered (if necessary) as
an entity by an aggregate recovery process (ABARS).
Fast replication target volumes are contained within SMS copy pool backup storage
groups. They contain the fast replication backup copies of DFSMShsm-managed volumes.
Beginning with z/OS V1R8, fast replication is done with a single command.
New function with z/OS V1R7

With z/OS V1R7, the maximum number of data sets stored by DFSMShsm on tape is one
million; previously, it was 330000.
Also in z/OS V1R7, a new command V SMS,VOLUME is introduced. It allows you to change the
state of the DFSMShsm volumes without having to change and reactivate the SMS
configuration using ISMF.

4.68 DFSMShsm: Automatic space management
HSM.HMIG.ABC.FILE1.T891008.I9012
Level 0 Level 1
10 days
without ABC.FILE1 dsname
any
access
ABC.FILE2
Migrate
ABC.FILE3
Figure 4-79 DFSMShsm automatic space management (migration)
Automatic space management (migration)

Automatic space management prepares DASD space for the addition of new data by freeing
space on the DFSMShsm-managed volumes (L0) and DFSMShsm-owned volumes (ML1).
The functions associated with automatic space management can be divided into two groups,
namely automatic volume space management and automatic secondary space management,
as explained here.Automatic volume space management
Primary Invoked on a daily basis, it cleans L0 volumes by deleting expired and
temporary data sets, releasing allocated and not used space (scratch).
During automatic primary space management, DFSMShsm can
process a maximum of 15 volume migration tasks concurrently.
This activity consists of the deletion of temporary data sets, deletion of
expired data sets, the release of unused and overallocated space, and
migration. Each task processes its own separate user DASD volume.
The storage administrator selects the maximum number of tasks that
can run simultaneously, and specifies which days and the time of day
the tasks are to be performed.
In z/OS V1R8, there is a specific task to scratch data sets. If after that
the free space is still below a threshold, then it moves data sets (under
control of management class) from L0 to ML1/ ML2 volumes.
Interval migration Executed each hour throughout the day, as needed for all storage
groups. In interval migration, DFSMShsm performs a space check on

each DFSMShsm volume being managed. A volume is considered
eligible for interval migration based on the AUTOMIGRATE and
THRESHOLD settings of its SMS storage group.
Automatic interval migration is an option that invokes migration when

DFSMShsm-managed volumes become full during high activity
periods. If the storage administrator chooses this option, DFSMShsm
automatically checks the level of occupancy of all
DFSMShsm-managed volumes periodically.
If the level of occupancy for any volume exceeds a given threshold,

DFSMShsm automatically performs a subset of the space
management functions on the volume. Select a threshold that can be
exceeded only when your installations activity exceeds its usual peak.
For volumes requiring interval migration, DFSMShsm can process up
to 15 volume migration tasks concurrently.
During automatic interval migration on a volume, the expired data sets

are deleted, then the largest eligible data sets are moved first so that
the level of occupancy threshold can be reached sooner. Data sets are
not migrated from ML1 to ML2 volumes during interval migration.
Automatic secondary space management

Automatic secondary space management deletes expired data sets from ML1/ML2, then
moves data sets (under control of the management class) from ML1 to ML2 volumes. It
should complete before automatic primary space management so that the ML1 volumes will
not run out of space. Since z/OS 1.6 there is the possibility of multiple secondary space
management (SSM) tasks.

4.69 DFSMShsm data set attributes
Active copy:
A backup version within the number of backup copies
specified by the management class or SETSYS value
Retained copy:
A backup copy that has rolled-off from being an active
copy, but has not yet met its retention period
Management class retention period:
The maximum number of days to maintain a backup copy
RETAINDAYS new with z/OS V1R11:
The minimum number of days to maintain a backup copy
(this value takes precedence)
Figure 4-80 DFSMShsm management of active and retained copies
Active and retained copies

DFSMShsm maintains backup copies as active backup copies and retained backup copies.
Active copies are the backup copies that have not yet rolled off. Retained copies are the
backup copies that have rolled off from the active copies, but have not yet reached their
retention periods.
With z/OS V1R11, DFSMShsm can maintain a maximum of 100 active copies. DFSMShsm
can maintain more than enough retained copies for each data set to meet all expected
requirements. Active and retained copies are as follows:
Active copies Active copies are the set of backup copies created that have not yet
rolled off. The number of active copies is determined by the SMS
management class or SETSYS value. The maximum number of active
copies will remain 100.
Retained copies Retained copies are the set of backup copies that have rolled off from
the active copies and have not yet reached their retention periods. A
nearly unlimited number of retained copies for each data set can be
maintained.
Management class retention period

The Retention Limit value is a required value that limits the use of retention period (RETPD)
and expiration date (EXPDT) values that are explicitly specified in JCL, are derived from
management class definitions or are explicitly specified in the OSREQ STORE macro. If the

value of a user-specified RETPD or EXPDT is within the limits specified in the Retention Limit
field, it is saved for the data set. For objects, only RETPD is saved.
The default retention limit is NOLIMIT. If you specify zero (0), then a user-specified or data
class-derived EXPDT or RETPD is ignored. If users specify values that exceed the maximum
period, then the retention limit value overrides not only their values but also the expiration
attribute values. The retention limit value is saved. ISMF primes the Retention Limit field with
what you specified the last time.
RETAINDAYS keyword
To specify a retention period for a copy of a backup data set, using the new RETAINDAYS
keyword, you can use one of the following methods:
(H)BACKDS command
ARCHBACK macro
ARCINBAK program
The RETAINDAYS value must be an integer in the range of 0 to 50000, or 99999 (the never
expire value).

4.70 RETAINDAYS keyword
DFSMShsm uses the RETAINDAYS value to determine

when a backup copy is to be rolled-off and during
EXPIREBV processing
The RETAINDAYS keyword specified on the data set
backup request determines how long a data set backup
copy is kept for both SMS-managed and
non-SMS-managed data sets
You can keep an individual backup copy for either shorter
or longer than normal period of time by specifying the
RETAINDAYS keyword on the BACKDS command
RETAINDAYS controls the minimum number of days that
a backup copy of a data set is maintained
Figure 4-81 Use of RETAINDAYS keyword with z/OS V1R11
RETAINDAYS keyword with z/OS V1R11

With z/OS V1R11, each SMS-managed data set's backup versions are processed based on
the value of the RETAINDAYS keyword on the BACKDS command, if specified, and the
attributes specified in the management class associated with the data set (or in the default
management class if the data set is not associated with a management class). If a
management class name is found but there is no definition for the management class,
DFSMShsm issues a message and does not process the cataloged backup versions of that
data set.
DFSMShsm compares the number of backup versions that exist with the value of the
NUMBER OF BACKUPS (DATA SET EXISTS) attribute in the management class. If there are
more versions than requested, the excess versions are deleted if the versions do not have a
RETAINDAYS value or the RETAINDAYS value has been met, starting with the oldest. The
excess versions are kept as excess active versions if they have an un-met RETAINDAYS
value. These excess versions will then be changed to retained backup versions when a new
version is created.
Starting with the now-oldest backup version and ending with the third-newest version,
DFSMShsm calculates the age of the version to determine if the version should be expired. If
a RETAINDAYS value was specified when the version was created, then the age is compared
to the retain days value. If RETAINDAYS was not specified, then the age is compared to the
value of the RETAIN DAYS EXTRA BACKUPS attribute in the management class. If the age
of the version meets the expiration criteria, then the version is expired.

The second-newest version is treated as though it had been created on the same day as the
newest backup version. Therefore, the second-newest version (newest EXTRA backup copy)
is not expired until the number of retention days specified by the RETAINDAYS value or the
number of days specified in RETAIN DAYS EXTRA BACKUPS attribute in the management
class have passed since the creation of the newest backup version. (The management class
value is only used if RETAINDAYS was not specified for the version).
EXPIREBV processing and RETAINDAYS

For EXPIREBV processing, the RETAINDAYS value takes precedence over all existing
criteria when expiring active and retained backup copies for SMS data sets and non-SMS
data sets. EXPIREBV needs to be run to delete backup versions with RETAINDAYS values
that have been met.
The EXPIREBV command is used to delete unwanted backup and expired ABARS versions of
SMS-managed and non-SMS-managed data sets from DFSMShsm-owned storage. The
optional parameters of the EXPIREBV command determine the deletion of the backup versions
of non-SMS-managed data sets. The management class attributes determine the deletion of
backup versions of SMS-managed data sets. The management class fields Retain Extra
Versions and Retain Only Version determine which ABARS versions or incremental backup
versions are deleted. The RETAINDAYS parameter specified on the data set backup request
determines how long a data set backup copy is kept for both SMS-managed and
non-SMS-managed data sets.
BACKDS command
The BACKDS command creates a backup version of a specific data set. When you enter the
BACKDS command, DFSMShsm does not check whether the data set has changed or has met
the requirement for frequency of backup. When DFSMShsm processes a BACKDS
command, it stores the backup version on either tape or the ML1 volume with the most
available space.
With z/OS V1R11, the RETAINDAYS keyword is an optional parameter on the BACKDS
command specifying a number of days to retain a specific backup copy of a data set. If you
specify RETAINDAYS, number of retain days is a required parameter that specifies a
minimum number of days (0-50000) that DFSMShsm retains the backup copy. If you specify
99999, the data set backup version never expires. Any value greater than 50000 (and other
than 99999) causes a failure with an ARC1605I error message. A retain days value of 0
indicates that:
The backup version might expire within the same day that it was created if EXPIREBV
processing takes place or when the next backup version is created,
The backup version is kept as an active copy before roll-off occurs,
The backup version is not managed as a retained copy.

4.71 RETAINDAYS keyword
99999 - If you specify 99999, the data set backup

version never expires.
50000+ - Any value greater than 50000 (and other than
99999) causes a failure with an ARC1605I error message.
0 - A retain days value of 0 indicates that:
The backup version might expire within the same day that it
was created if EXPIREBV processing takes place or when
the next backup version is created.
The backup version is kept as an active copy before roll-off
occurs.
The backup version is not managed as a retained copy.
Restriction: RETAINDAYS applies only to cataloged data sets.
Figure 4-82 RETAINDAYS keyword
RETAINDAYS parameters
The value of RETAINDAYS can be in the range of 0 - 50000, which corresponds to the
maximum of 136 years. If you specify 99999, the data set backup version is treated as never
expire. Any value greater than 50000 (and other than 99999) causes a failure with an error
message ARC1605I. A retain days value of 0 indicates that:
The backup version expires when the next backup copy is created.
The backup version might expire within the same day that it was created if EXPIREBV
processing takes place.
The backup version is kept as an active copy before roll-off occurs.
And the backup version is not managed as a retained copy.
Note: You can use the RETAINDAYS keyword only with cataloged data sets. If you specify
RETAINDAYS with an uncataloged data set, then BACKDS processing fails with the
ARC1378I error message.
EXPIREBV command and RETAINDAYS

When you enter the EXPIREBV command, DFSMShsm checks the retention days for each
active backup copy for each data set, starting with the oldest backup version and ending with
the third newest version.

If the version has a specified retention days value, then DFSMShsm calculates the age of the
version, compares the age to the value of the retention days, and expires the version if it has
met its RETAINDAYS. The second-newest version is treated as though it had been created on
the same day as the newest backup version, and is not expired unless the number of
retention days specified by RETAINDAYS have passed since the creation of the newest
backup version. EXPIREBV does not process the newest backup version until it meets both
the management class retention values, and the RETAINDAYS value.
Note: For non-SMS-managed data sets, the RETAINDAYS value takes precedence over
any of the EXPIREBV parameters.
EXPIREBV processing
During EXPIREBV processing, DFSMShsm checks the retention days for each retained
backup copy. The retained copy is identified as an expired version if it has met its retention
period. The EXPIREBV DISPLAY command displays the backup versions that have met their
RETAINDAYS value. The EXPIREBV EXECUTE command deletes the backup versions that have
met their RETAINDAYS value.
When you enter the EXPIREBV command, DFSMShsm checks the retention days for each
active backup copy for each data set, starting with the oldest backup version and ending with
the third newest version. If the version has a specified retention days value, DFSMShsm
calculates the age of the version, compares the age to the value of the retention days, and
expires the version if it has met its RETAINDAYS.
The second-newest version is treated as though it had been created on the same day as the
newest backup version, and is not expired unless the number of retention days specified by
RETAINDAYS have passed since the creation of the newest backup version. EXPIREBV does
not process the newest backup version until it meets both the management class retention
values, and the RETAINDAYS value.
Restriction: RETAINDAYS applies only to cataloged data sets.

4.72 DFSMShsm: Recall processing
HSM.HMIG.ABC.FILE1.T891008.I9012
Level 1 Level 0
dsname ABC.FILE1
ABC.FILE2
Recall
ABC.FILE3
Figure 4-83 Recall processing
Automatic recall
Using an automatic recall process returns a migrated data set from an ML1 or ML2 volume to
a DFSMShsm-managed volume. When a user refers to the data set, DFSMShsm reads the
system catalog for the volume serial number. If the volume serial number is MIGRAT,
DFSMShsm finds the migrated data set, recalls it to a DFSMShsm-managed volume, and
updates the catalog. The result of the recall process is a data set that resides on a user
volume in a user readable format. The recall can also be requested by a DFSMShsm
command. Automatic recall returns your migrated data set to a DFSMShsm-managed
volume when you refer to it. The catalog is updated accordingly with the real volser.
Recall returns a migrated data set to a user L0 volume. The recall is transparent and the
application does not need to know that it happened or where the migrated data set resides. To
provide applications with quick access to their migrated data sets, DFSMShsm allows up to
15 concurrent recall tasks. RMF monitor III shows delays caused by the recall operation.
The MVS allocation routine discovers that the data set is migrated when, while accessing the
catalog, it finds the word MIGRAT instead of the volser.
Command recall
Command recall returns your migrated data set to a user volume when you enter the HRECALL
DFSMShsm command through an ISMF panel or by directly keying in the command.

ACS routines
For both automatic and command recall, DFSMShsm working with SMS invokes the
automatic class selection (ACS) routines. Data sets that were not SMS-managed at the time
they were migrated may be recalled as SMS-managed data sets. The ACS routines
determine whether the data sets should be recalled as SMS-managed, and if so, the routines
select the classes and storage groups in which the data sets will reside. The system chooses
the appropriate volume for the data sets.
DFSMShsm working without SMS returns a migrated data set to a DFSMShsm-managed

non-SMS level 0 volume with the most free space.

4.73 Removable media manager (DFSMSrmm)
IBM 3494
Virtual Tape Server
Figure 4-84 DFSMSrmm
DFSMSrmm
In your enterprise, you store and manage your removable media in several types of media
libraries. For example, in addition to your traditional tape library (a room with tapes, shelves,
and drives), you might have several automated and manual tape libraries. You probably also
have both onsite libraries and offsite storage locations, also known as vaults or stores.
With the DFSMSrmm functional component of DFSMS, you can manage your removable
media as one enterprise-wide library (single image) across systems. Because of the need for
global control information, these systems must have accessibility to shared DASD volumes.
DFSMSrmm manages your installation's tape volumes and the data sets on those volumes.
DFSMSrmm also manages the shelves where volumes reside in all locations except in
automated tape library data servers.
DFSMSrmm manages all tape media (such as cartridge system tapes and 3420 reels), as
well as other removable media you define to it. For example, DFSMSrmm can record the shelf
location for optical disks and track their vital record status; however, it does not manage the
objects on optical disks.
Library management
DFSMSrmm can manage the following devices:
A removable media library, which incorporates all other libraries, such as:
System-managed manual tape libraries

System-managed automated tape libraries
Non-system-managed or traditional tape libraries, including automated libraries such as a
library under Basic Tape Library Support (BTLS) control.
Examples of automated tape libraries include IBM TotalStorage Enterprise Automated Tape
Library (3494) and IBM TotalStorage Virtual Tape Servers (VTS).
Shelf management
DFSMSrmm groups information about removable media by shelves into a central online
inventory, and keeps track of the volumes residing on those shelves. DFSMSrmm can
manage the shelf space that you define in your removable media library and in your storage
locations.
Volume management
DFSMSrmm manages the movement and retention of tape volumes throughout their life
cycle.
Data set management

DFSMSrmm records information about the data sets on tape volumes. DFSMSrmm uses the
data set information to validate volumes and to control the retention and movement of those
data sets.
For more information about DFSMSrmm, see z/OS DFSMSrmm Guide and Reference,
SC26-7404 and z/OS DFSMSrmm Implementation and Customization Guide, SC26-7405,
and visit:
http://www-1.ibm.com/servers/storage/software/sms/rmm/

4.74 Libraries and locations
IBM 3494
Virtual Tape Server
Figure 4-85 Libraries and locations
Libraries and locations

You decide where to store your removable media based on how often the media is accessed
and for what purpose it is retained. For example, you might keep volumes that are frequently
accessed in an automated tape library data server, and you probably use at least one storage
location to retain volumes for disaster recovery and audit purposes. You might also have
locations where volumes are sent for further processing, such as other data centers within
your company or those of your customers and vendors.
DFSMSrmm automatically records information about data sets on tape volumes so that you
can manage the data sets and volumes more efficiently. When all the data sets on a volume
have expired, the volume can be reclaimed and reused. You can optionally move volumes that
are to be retained to another location.
DFSMSrmm helps you manage your tape volumes and shelves at your primary site and
storage locations by recording information in a DFSMSrmm control data set.

4.75 What DFSMSrmm can manage
Removable media library

System-managed tape libraries
Automated tape libraries
Manual tape libraries
Non-system-managed tape libraries or traditional
tape libraries
Storage locations
Installation-defined
DFSMSrmm built-in
Local
Distant
Remote
Figure 4-86 What DFSMSrmm can manage
What DFSMSrmm can manage

In this section we discuss libraries and storage locations that can be managed by
DFSMSrmm.
Removable media library

A removable media library contains all the tape and optical volumes that are available for
immediate use, including the shelves where they reside. A removable media library usually
includes other libraries:
System-managed libraries, such as automated or manual tape library data servers
Non-system-managed libraries, containing the volumes, shelves, and drives not in an
automated or a manual tape library data server
In the removable media library, you store your volumes in shelves, where each volume
occupies a single shelf location. This shelf location is referred to as a rack number in the
DFSMSrmm TSO subcommands and ISPF dialog. A rack number matches the volumes
external label. DFSMSrmm uses the external volume serial number to assign a rack number
when adding a volume, unless you specify otherwise. The format of the volume serial you
define to DFSMSrmm must be one to six alphanumeric characters. The rack number must be
six alphanumeric or national characters.
System-managed tape library

A system-managed tape library is a collection of tape volumes and tape devices defined in

the tape configuration database. The tape configuration database is an integrated catalog
facility user catalog marked as a volume catalog (VOLCAT) containing tape volumes and tape
library records. A system-managed tape library can be either automated or manual:
An automated tape library is a device consisting of robotic components, cartridge storage
areas (or shelves), tape subsystems, and controlling hardware and software, together with
the set of tape volumes that reside in the library and can be mounted on the library tape
drives. The IBM automated tape libraries are the automated IBM 3494 and IBM 3495
Library Dataservers.
A manual tape library is a set of tape drives and the set of system-managed volumes the
operator can mount on those drives. The manual tape library provides more flexibility,
enabling you to use various tape volumes in a given manual tape library. Unlike the
automated tape library, the manual tape library does not use the library manager. With the
manual tape library, a human operator responds to mount messages that are generated
by the host and displayed on a console. This manual tape library implementation
completely replaces the IBM 3495-M10 implementation. IBM no longer supports the
3495-M10.
You can have several automated tape libraries or manual tape libraries. You use an
installation-defined library name to define each automated tape library or manual tape library
to the system. DFSMSrmm treats each system-managed tape library as a separate location
or destination.
Since z/OS 1.6, a new EDGRMMxx parmlib member OPTION command, together with the
VLPOOL command, allows better support for the client/server environment.
z/OS 1.8 DFSMSrmm introduces an option to provide tape data set authorization
independent of the RACF TAPVOL and TAPEDSN. This option allows you to use RACF
generic DATASET profiles for both DASD and tape data sets.
Non-system-managed tape library

A non-system-managed tape library consists of all the volumes, shelves, and drives not in an
automated tape library or manual tape library. You might know this library as the traditional
tape library that is not system-managed. DFSMSrmm provides complete tape management
functions for the volumes and shelves in this traditional tape library. Volumes in a
non-system-managed library are defined by DFSMSrmm as being shelf-resident.
All tape media and drives supported by z/OS are supported in this environment. Using
DFSMSrmm, you can fully manage all types of tapes in a non-system-managed tape library,
including 3420 reels, 3480, 3490, and 3590 cartridge system tapes.
Storage location
Storage locations are not part of the removable media library because the volumes in storage
locations are not generally available for immediate use. A storage location is comprised of
shelf locations that you define to DFSMSrmm. A shelf location in a storage location is
identified by a bin number. Storage locations are typically used to store removable media that
are kept for disaster recovery or vital records. DFSMSrmm manages two types of storage
locations: installation-defined storage locations and DFSMSrmm built-in storage locations.
You can define an unlimited number of installation-defined storage locations, using any
eight-character name for each storage location. Within the installation-defined storage
location, you can define the type or shape of the media in the location. You can also define
the bin numbers that DFSMSrmm assigns to the shelf locations in the storage location. You
can request DFSMSrmm shelf-management when you want DFSMSrmm to assign a specific
shelf location to a volume in the location.

You can also use the DFSMSrmm built-in storage locations: LOCAL, DISTANT, and
REMOTE. Although the names of these locations imply their purpose, they do not mandate
their actual location. All volumes can be in the same or separate physical locations.
For example, an installation can have the LOCAL storage location onsite as a vault in the
computer room, the DISTANT storage location can be a vault in an adjacent building, and the
REMOTE storage location can be a secure facility across town or in another state.
DFSMSrmm provides shelf-management for storage locations so that storage locations can
be managed at the shelf location level.

4.76 Managing libraries and storage locations
IBM 3494
Virtual Tape Server
RMM CDS
Figure 4-87 DFSMSrmm: managing libraries and storage locations
Managing libraries and storage locations

DFSMSrmm records the complete inventory of the removable media library and storage
locations in the DFSMSrmm control data set, which is a VSAM key-sequenced data set. In
the control data set, DFSMSrmm records all changes made to the inventory (such as adding
or deleting volumes), and also keeps track of all movement between libraries and storage
locations. DFSMSrmm manages the movement of volumes among all library types and
storage locations. This lets you control where a volumeand hence, a data setresides, and
how long it is retained.
DFSMSrmm helps you manage the movement of your volumes and retention of your data
over their full life, from initial use to the time they are retired from service. Among the
functions DFSMSrmm performs for you are:
Automatically initializing and erasing volumes
Recording information about volumes and data sets as they are used
Expiration processing
Identifying volumes with high error levels that require replacement
To make full use of all of the DFSMSrmm functions, you specify installation setup options and
define retention and movement policies. DFSMSrmm provides you with utilities to implement
the policies you define. Since z/OS 1.7, we have DFSMSrmm enterprise enablement that
allows high-level languages to issue DFSMSrmm commands through Web services.

z/OS V1R8 enhancements
DFSMSrmm helps you manage the shelves in your tape library and storage locations,
simplifying the tasks of your tape librarian. When you define a new volume in your library, you
can request that DFSMSrmm shelf-manage the volume by assigning the volume a place on
the shelf. You also have the option to request a specific place for the volume. Your shelves are
easier to use when DFSMSrmm manages them in pools. Pools allow you to divide your
shelves into logical groups where you can store volumes. For example, you can have a
different pool for each system that your installation uses. You can then store the volumes for
each system together in the same pool.
You can define shelf space in storage locations. When you move volumes to a storage
location where you have defined shelf space, DFSMSrmm checks for available shelf space
and then assigns each volume a place on the shelf if you request it. You can also set up
DFSMSrmm to reuse shelf space in storage locations.

5
Chapter 5. System-managed storage

As your business expands, so does your need for storage to hold your applications and data,
and so does the cost of managing that storage. Storage cost includes more than the price of
the hardware, with the highest cost being the people needed to perform storage management
tasks.
If your business requires transaction systems, the batch window can also be a high cost.
Additionally, you must pay for staff to install, monitor, and operate your storage hardware
devices, for electrical power to keep each piece of storage hardware cool and running, and for
floor space to house the hardware. Removable media, such as optical and tape storage, cost
less per gigabyte (GB) than online storage, but they require additional time and resources to
locate, retrieve, and mount.
To allow your business to grow efficiently and profitably, you need to find ways to control the
growth of your information systems and use your current storage more effectively.
With these goals in mind, in this chapter we present:

A description of the z/OS storage-managed environment
An explanation of the benefits of using a system-managed environment
An overview of how SMS manages a storage environment based on installation policies
A description of how to set up a minimal SMS configuration and activate a DFSMS
subsystem
A description of how to manage data using a minimal SMS configuration
A description of how to use Interactive Storage Management Facility (ISMF), an interface
for defining and maintaining storage management policies

5.1 Storage management
Availability
Space
Security
Performance
dfp
dss rmm
DFSMS
IBM
3494
ISMF
VTS
hsm tvs
Figure 5-1 Managing storage with DFSMS
Storage management
Storage management involves data set allocation, placement, monitoring, migration, backup,
recall, recovery, and deletion. These activities can be done either manually or by using
automated processes.
Managing storage with DFSMS

The Data Facility Storage Management Subsystem (DFSMS) comprises the base z/OS
operating system and performs the essential data, storage, program, and device
management functions of the system. DFSMS is the central component of both
system-managed and non-system-managed storage environments.
The DFSMS software product, together with hardware products and installation-specific
requirements for data and resource management, comprises the key to system-managed
storage in a z/OS environment.
The heart of DFSMS is the Storage Management Subsystem (SMS). Using SMS, the storage
administrator defines policies that automate the management of storage and hardware
devices. These policies describe data allocation characteristics, performance and availability
goals, backup and retention requirements, and storage requirements for the system. SMS
governs these policies for the system and the Interactive Storage Management Facility
(ISMF) provides the user interface for defining and maintaining the policies.

5.2 DFSMS and DFSMS environment
DFSMS + z/OS
RACF + + DFSORT
DFSMS Environment for z/OS
Figure 5-2 SMS environment
DFSMS functional components

DFSMS is a set of products, and one of these products, DSFMSdfp, is mandatory for running
z/OS. DFSMS comprises the base z/OS operating system, where DFSMS performs the
essential data, storage, program, and device management functions of the system. DFSMS is
the central component of both system-managed and non-system-managed storage
environments.
DFSMS environment
The DFSMS environment consists of a set of hardware and IBM software products which
together provide a system-managed storage solution for z/OS installations.
DFSMS uses a set of constructs, user interfaces, and routines (using the DFSMS products)
that allow the storage administrator to better manage the storage system. The core logic of
DFSMS, such as the Automatic Class Selection (ACS) routines, ISMF code, and constructs,
is located in DFSMSdfp. DFSMShsm and DFSMSdss are involved in the management class
construct.
In this environment, the Resource Access Control Facility (RACF) and Data Facility Sort
(DFSORT) products complement the functions of the base operating system. RACF provides
resource security functions, and DFSORT adds the capability for faster and more efficient
sorting, merging, copying, reporting, and analyzing of business information.
The DFSMS environment is also called the SMS environment.
Chapter 5. System-managed storage 241

5.3 Goals and benefits of system-managed storage
Simplified data allocation
Improved allocation control
Improved I/O performance management
Automated DASD space management
Automated tape/optical space management
Improved data availability management
Simplified conversion of data to different device types
Figure 5-3 Benefits of system-managed storage
Benefits of system-managed storage

With the Storage Management Subsystem (SMS), you can define performance goals and
data availability requirements, create model data definitions for typical data sets, and
automate data backup. Based on installation policies, SMS can automatically assign those
services and data definition attributes to data sets when they are created. IBM storage
management-related products determine data placement, manage data backup, control
space usage, and provide data security.
Goals of system-managed storage

The goals of system-managed storage are:
To improve the use of the storage media (for example, by reducing out-of-space abends
and providing a way to set a free-space requirement).
To reduce the labor involved in storage management by centralizing control, automating
tasks, and providing interactive controls for storage administrators.
To reduce the user's need to be concerned with the physical details, performance, space,
and device management. Users can focus on using data, instead of on managing data.
Simplified data allocation

System-managed storage enables users to simplify their data allocations. For example,
without using the Storage Management Subsystem, a z/OS user has to specify the unit and
volume on which the system is to allocate the data set. The user also has to calculate the

amount of space required for the data set in terms of tracks or cylinders. This means the user
has to know the track size of the device that will contain the data set.
With system-managed storage, users can allow the system to select the specific unit and
volume for the allocation. They can also specify size requirements in terms of megabytes or
kilobytes. This means the user does not need to know anything about the physical
characteristics of the devices in the installation.
Improved allocation control

System-managed storage enables you to set a requirement for free space across a set of
direct access storage device (DASD) volumes. You can then provide adequate free space to
avoid out-of-space abends. The system automatically places data on a volume containing
adequate free space. DFSMS offers relief to avoid out-of-space conditions. You can also set a
threshold for scratch tape volumes in tape libraries, to ensure enough cartridges are available
in the tape library for scratch mounts.
Improved Input/Output (I/O) performance management

System-managed storage enables you to improve DASD I/O performance across the
installation. At the same time, it reduces the need for manual tuning by defining performance
goals for each class of data. You can use cache statistics recorded in System Management
Facility (SMF) records to help evaluate performance. You can also improve sequential
performance by using extended sequential data sets. The DFSMS environment makes the
most effective use of the caching abilities of storage controllers.
Automated DASD space management

System-managed storage enables you to automatically reclaim space that is allocated to old
and unused data sets or objects. You can define policies that determine how long an unused
data set or object will be allowed to reside on primary storage (storage devices used for your
active data). You can have the system remove obsolete data by migrating the data to other
DASD, tape, or optical volumes, or you can have the system delete the data. You can also
release allocated but unused space that is assigned to new and active data sets.
Automated tape space management

Mount Management System-managed storage lets you fully use the capacity of your tape
cartridges and automate tape mounts. Using tape mount management (TMM) methodology,
DFSMShsm can fill tapes to their capacity. With 3490E, 3590, 3591, and 3592 tape devices,
Enhanced Capacity Cartridge System Tape, recording modes such as 384-track and EFMT1,
and the improved data recording capability, you can increase the amount of data that can be
written on a single tape cartridge.
Tape System-managed storage lets you exploit the device technology of new devices without
having to change the JCL UNIT parameter. In a multi-library environment, you can select the
drive based on the library where the cartridge or volume resides. You can use the IBM
TotalStorage Enterprise Automated Tape Library (3494 or 3495) to automatically mount tape
volumes and manage the inventory in an automated tape library. Similar functionality is
available in a system-managed manual tape library. If you are not using SMS for tape
management, you can still access the IBM TotalStorage Enterprise Automated Tape Library
(3494 or 3495) using Basic Tape Library Storage (BTLS) software.
Automated optical space management

System-managed storage enables you to fully use the capacity of your optical cartridges and
to automate optical mounts. Using a 3995 Optical Library Dataserver, you can automatically
mount optical volumes and manage the inventory in an automated optical library.

Improved data availability management
System-managed storage enables you to provide separate backup requirements to data
residing on the same DASD volume. Thus, you do not have to treat all data on a single
volume the same way.
You can use DFSMShsm to automatically back up your various types of data sets and use
point-in-time copy to maintain access to critical data sets while they are being backed up.
Concurrent copy, virtual concurrent copy, SnapShot, and FlashCopy, along with
backup-while-open, have an added advantage in that they avoid invalidating a backup of a
CICS VSAM KSDS due to a control area or control interval split.
You can also create a logical group of data sets, so that the group is backed up at the same
time to allow recovery of the application defined by the group. This is done with the aggregate
backup and recovery support (ABARS) provided by DFSMShsm.
Simplified conversion of data to other device types

System-managed storage enables you to move data to new volumes without requiring users
to update their job control language (JCL). Because users in a DFSMS environment do not
need to specify the unit and volume which contains their data, it does not matter to them if
their data resides on a specific volume or device type. This allows you to easily replace old
devices with new ones.
You can also use system-determined block sizes to automatically reblock physical sequential
and partitioned data sets that can be reblocked.

5.4 Service level objectives
What performance objectives are required by data
When and how to back up data
Whether data sets should be kept available for use

during backup or copy
How to manage backup copies kept for disaster

recovery
What to do with data that is obsolete or seldom

used
Figure 5-4 Service level objectives
Service level objectives

To allow your business to grow efficiently and profitably, you want to find ways to control the
growth of your information systems and use your current storage more effectively.
In an SMS-managed storage environment, your enterprise establishes centralized policies for

how to use your hardware resources. These policies balance your available resources with
your users' requirements for data availability, performance, space, and security.
The policies defined in your installation represent decisions about your resources, such as:
What performance objectives are required by the applications accessing the data
Based on these objectives, you can try to better exploit cache data striping. By tracking
data set I/O activities, you can make better decisions about data set caching policies and
improve overall system performance. For object data, you can track transaction activities
to monitor and improve OAM's performance.
When and how to back up data - incremental or total
Determine the backup frequency, the number of backup versions, and the retention period
by consulting user group representatives. Be sure to consider whether certain data
backups need to be synchronized. For example, if the output data from application A is
used as input for application B, you must coordinate the backups of both applications to
prevent logical errors in the data when they are recovered.

Whether data sets are to be kept available for use during backup or copy
You can store backup data sets on DASD or tape (this does not apply to objects). Your
choice depends on how fast the data needs to be recovered, media cost, operator cost,
floor space, power requirements, air conditioning, the size of the data sets, and whether
you want the data sets to be portable.
How to manage backup copies kept for disaster recovery - locally or in a vault
Back up related data sets in aggregated tapes. Each application is to have its own,
self-contained aggregate of data sets. If certain data sets are shared by two or more
applications, you might want to ensure application independence for disaster recovery by
backing up each application that shares the data. This is especially important for shared
data in a distributed environment.
What to do with data that is obsolete or seldom used
Data is obsolete when it has exceeded its expiration dates and is no longer needed. To
select obsolete data for deletion using DFSMSdss, issue the DUMP command and the
DELETE parameter, and force OUTDDNAME to DUMMY.
The purpose of a backup plan is to ensure the prompt and complete recovery of data. A
well-documented plan identifies data that requires backup, the levels required, responsibilities
for backing up the data, and methods to be used.

5.5 Implementing SMS policies
ACS Routines
g em ent Cl
n a as
s
Ma
Which are the
services?
Sto
a Class
rage Class
What does it Data What is the
look like? Set service level?
D a t
Where is it
placed?
St
o ra g
e Group
Figure 5-5 Creating SMS policies
Implementing SMS policies

To implement a policy for managing storage, the storage administrator defines classes of
space management, performance, and availability requirements for data sets. The storage
administrator uses:
Data class Data classes are used to define model allocation characteristics for
data sets.
Storage class Storage classes are used to define performance and availability goals.
Management class Management classes are used to define backup and retention
requirements.
Storage group Storage groups are used to create logical groups of volumes to be
managed as a unit.
ACS routines Automatic Class Selection (ACS) routines are used to assign class
and storage group definitions to data sets and objects.
For example, the administrator can define one storage class for data entities requiring high
performance, and another for those requiring standard performance. Then, the administrator
writes Automatic Class Selection (ACS) routines that use naming conventions or other criteria
of your choice to automatically assign the classes that have been defined to data as that data
is created. These ACS routines can then be validated and tested.

When the ACS routines are started and the classes (also referred to as constructs) are
assigned to the data, SMS uses the policies defined in the classes to apply to the data for the
life of the data. Additionally, devices with various characteristics can be pooled together into
storage groups, so that new data can be automatically placed on devices that best meet the
needs for the data.
DFSMS facilitates all of these tasks by providing menu-driven panels with the Interactive
Storage Management Facility (ISMF). ISMF panels make it easy to define classes, test and
validate ACS routines, and perform other tasks to analyze and manage your storage. Note
that many of these functions are available in batch through the NaviQuest tool.

5.6 Monitoring SMS policies
Monitor DASD use
Monitor data set performance
Decide when to consolidate free space on DASD
Set policies for DASD or tape
Use reports to manage your removable media
Figure 5-6 Monitoring your SMS policies
Monitoring SMS policies

After storage administrators have established the installation's service levels and
implemented policies based on those levels, they can use DFSMS facilities to see if the
installation objectives have been met. Information on past use can help to develop more
effective storage administration policies and manage growth effectively. The DFSMS
Optimizer feature can be used to monitor, analyze, and tune the policies.

5.7 Assigning data to be system-managed
DASD Optical Tape
(1) (2)
Assign Storage Not
Data Set Class (SC)
Not applicable
system-managed
(3)
Object Stored Stored Stored
(4)
Define OAM Assign
Assign System
Volume Group (SG)
Storage Groups Storage Group
(SG) (SG)
Figure 5-7 How to be system-managed
How to be system-managed
Using SMS, you can automate storage management for individual data sets and objects, and
for DASD, optical, and tape volumes. Figure 5-7 shows how a data set, object, DASD volume,
tape volume, or optical volume becomes system-managed. The numbers shown in
parentheses are associated with the following notes:
1. A DASD data set is system-managed if you assign it a storage class. If you do not assign a
storage class, the data set is directed to a non-system-managed DASD or tape volume,
one that is not assigned to a storage group.
2. You can assign a storage class to a tape data set to direct it to a system-managed tape
volume. However, only the tape volume is considered system-managed, not the data set.
3. Objects are also known as byte-stream data, and this data is used in specialized
applications such as image processing, scanned correspondence, and seismic
measurements. Object data typically has no internal record or field structure and, after it is
written, the data is not changed or updated. However, the data can be referenced many
times during its lifetime. Objects are processed by OAM. Each object has a storage class;
therefore, objects are system-managed. The optical or tape volume on which the object
resides is also system-managed.
4. Tape volumes are added to tape storage groups in tape libraries when the tape data set is
created.

5.8 Using data classes
JCL IDCAMS DYNALLOC Record and Space Attributes

Key Length and Offset
Record Format
Record Length
Record Organization
TSO Allocate Space (Primary, Secondary, Avg
ISPF/PDF Rec, Avg Value)
Volume and VSAM Attributes
Compaction
Control Interval Size
ACS Media Type and Recording
ROUTINE Technology
Percent Free Space
Data Retention Period or Expiration Date
Share Options (Cross Region,
Class Cross System)
Volume Count
Data Sets Attributes
Backup-While-Open
Data Set Name Type
Allocation Extended Addressability
Extended Format
Non-System-Managed Initial Load (Speed, Recovery)
System-Managed Log and Logstream ID
Volumes Volumes Record Access Bias
Reuse
Space Constrait Relief and Reduce
Space Up to %
Spanned/Nospanned
Figure 5-8 Using data classes
Using data classes

A data class is a collection of allocation and space attributes that you define. It is used when
data sets are created. You can simplify data set allocation for the users by defining data
classes that contain standard data set allocation attributes. You can use data classes with
both system-managed and non-system-managed data sets. However, a variety of data class
characteristics, like extended format, are only available for system-managed data sets.
Data class attributes define space and data characteristics that are normally specified on JCL
DD statements, TSO/E ALLOCATE command, IDCAMS DEFINE commands, and dynamic
allocation requests. For tape data sets, data class attributes can also specify the type of
cartridge and recording method, and if the data is to be compacted. Users then need only
specify the appropriate data classes to create standardized data sets.
You can assign a data class through:

The DATACLAS parameter of a JCL DD statement, ALLOCATE or DEFINE commands.
Data class ACS routine to automatically assign a data class when the data set is being
created. For example, data sets with the low-level qualifiers LIST, LISTING, OUTLIST, or
LINKLIST are usually utility output data sets with similar allocation requirements, and can
all be assigned the same data class.
You can override various data set attributes assigned in the data class, but you cannot
change the data class name assigned through an ACS routine.

Even though data class is optional, we usually recommend that you assign data classes to
system-managed and non-system-managed data. Although the data class is not used after
the initial allocation of a data set, the data class name is kept in the catalog entry for
system-managed data sets for future reference.
Note: The data class name is not saved for non-system-managed data sets, although the
allocation attributes in the data class are used to allocate the data set.
For objects on tape, we recommend that you do not assign a data class through the ACS
routines. To assign a data class, specify the name of that data class on the SETOAM command.
If you change a data class definition, the changes only affect new allocations. Existing data
sets allocated with the data class are not changed.

5.9 Using storage classes
Data Set Requirements
Performance Storage Availability

Requirement Class Requirement
Storage Management Subsystem Mapping
9393 RVA 3990 3990 Model

Storage Model 3 or 6
Control 3 or 6 Storage
or Storage Control
ESS Control with
with cache
cache
3390-like 3390 / RAMAC

3390 / RAMAC Dual
pair
Figure 5-9 Choosing volumes that meet availability requirements
Using storage classes

A storage class is a collection of performance goals and availability requirements that you
define. The storage class is used to select a device to meet those goals and requirements.
Only system-managed data sets and objects can be assigned to a storage class. Storage
classes free users from having to know about the physical characteristics of storage devices
and manually placing their data on appropriate devices.
Some of the availability requirements that you specify to storage classes (such as cache and
dual copy) can only be met by DASD volumes attached through one of the following storage
control units or a similar device:
3990-3 or 3990-6
RAMAC Array Subsystem
Enterprise Storage Server (ESS)
DS6000 or DS8000
Figure 5-9 shows storage control unit configurations and their storage class attribute values.
With a storage class, you can assign a data set to dual copy volumes to ensure continuous
availability for the data set. With dual copy, two current copies of the data set are kept on
separate DASD volumes (by the control unit). If the volume containing the primary copy of the
data set is damaged, the companion volume is automatically brought online and the data set
continues to be available and current. Remote copy is the same, with the two volumes in
distinct control units (generally remote).

You can use the ACCESSIBILITY attribute of the storage class to request that concurrent
copy be used when data sets or volumes are backed up.
You can specify an I/O response time objective with storage class by using the millisecond
response time (MSR) parameter. During data set allocation, the system attempts to select the
closest available volume to the specified performance objective. Also along the data set life,
through the use MSR, DFSMS dynamically uses the cache algorithms as DASD Fast Write
(DFW) and Inhibit Cache Load (ICL) in order to reach the MSR target I/O response time. This
DFSMS function is called dynamic cache management.
To assign a storage class to a new data set, you can use:

The STORCLAS parameter of the JCL DD statement, ALLOCATE or DEFINE command
Storage class ACS routine
For objects, the system uses the performance goals you set in the storage class to place the
object on DASD, optical, or tape volumes. The storage class is assigned to an object when it
is stored or when the object is moved. The ACS routines can override this assignment.
Note: If you change a storage class definition, the changes affect the performance service
levels of existing data sets that are assigned to that class when the data sets are
subsequently opened. However, the definition changes do not affect the location or
allocation characteristics of existing data sets.

5.10 Using management classes
SPACE
Expiration
BACKUP
Migration/Object System-Managed
Transition Volume
GDG Storage
Management Management Management
Class Subsystem
Data DFSMShsm
Management and
Requirements DFSMSdss
DFSMShsm-Owned
Figure 5-10 Using management classes
Using management classes

A management class is a collection of management attributes that you define. The attributes
defined in a management class are related to:
Expiration date
Migration criteria
GDG management
Backup of data set
Object Class Transition Criteria
Aggregate backup
Management classes let you define management requirements for individual data sets, rather
than defining the requirements for entire volumes. All the data set functions described in the
management class are executed by DFSMShsm and DFSMSdss programs. Figure 5-11 on
page 257 shows the sort of functions an installation can define in a management class.
To assign a management class to a new data set, you can use:

The MGMTCLAS parameter of the JCL DD statement, ALLOCATE or DEFINE command
The management class ACS routine to automatically assign management classes to new
data sets
The ACS routine can override the management class specified in JCL, ALLOCATE or DEFINE
command.You cannot override management class attributes through JCL or command
parameters.

If you do not explicitly assign a management class to a system-managed data set or object,
the system uses the default management class. You can define your own default management
class when you define your SMS base configuration.
Note: If you change a management class definition, the changes affect the management
requirements of existing data sets and objects that are assigned that class. You can
reassign management classes when data sets are renamed.
For objects, you can:

Assign a management class when it is stored, or
Assign a new management class when the object is moved, or
Change the management class by using the OAM Application Programming Interface
(OSREQ CHANGE function)
The ACS routines can override this assignment for objects.

5.11 Management class functions
Allow early migration for old generations of GDG

Delete selected old/unused data sets from DASD
volumes
Release allocated but unused space from data sets
Migrate unused data sets to tape or DASD volumes
Specify how often to back up data sets, and whether
concurrent copy should be used for backups
Specify how many backup versions to keep for data sets
Specify how long to save backup versions
Specify the number of versions of ABARS to keep and
how to retain those versions
Establish the expiration date/transition criteria for objects
Indicate if automatic backup is needed for objects
Figure 5-11 Management class functions
Management class functions

By classifying data according to management requirements, an installation can define unique
management classes to fully automate data set and object management. For example:
Control the migration of CICS user databases, DB2 user databases and archive logs.
Test systems and their associated data sets.
IMS archive logs.
Specify that DB2 image copies, IMS image copies and change accumulation logs be
written to primary volumes and then migrated directly to migration level 2 tape volumes.
For objects, define when an object is eligible for a change in its performance objectives or
management characteristics. For example, after a certain number of days an installation
might want to move an object from a high performance DASD volume to a slower optical
volume.
Management class can also be used to specify that the object is to have a backup copy
made when the OAM Storage Management Component (OSMC) is executing.
When changing a management class definition, the changes affect the management
requirements of existing data sets and objects that are assigned to that class.

5.12 Using storage groups
SMS-managed
VIO PRIMARY LARGE
OBJECT
TAPE Storage
Groups
OBJECT BACKUP
DFSMShsm-owned
DB2 IMS CICS
Non-system-managed
Migration SYSTEM UNMOVABLE TAPE
Migration
Level 2,
Level 1
Backup,
Dump
Figure 5-12 Grouping storage volumes for specific purposes
Storage groups
A storage group is a collection of storage volumes and attributes that you define. The
collection can be a group of:
System paging volumes
DASD volumes
Tape volumes
Optical volumes
Combination of DASD and optical volumes that look alike
DASD, tape, and optical volumes treated as a single object storage hierarchy
Storage groups, along with storage classes, help reduce the requirement for users to
understand the physical characteristics of the storage devices which contain their data.
In a tape environment, you can also use tape storage groups to direct a new tape data set to
an automated or manual tape library.
DFSMShsm uses various storage group attributes to determine whether the volumes in the
storage group are eligible for automatic space or availability management.
Figure 5-12 shows an example of how an installation can group storage volumes according to
their objective. In this example:
SMS-managed DASD volumes are grouped into storage groups so that primary data sets,
large data sets, DB2 data, IMS data, and CICS data are all separated.

The VIO storage group uses system paging volumes for small temporary data sets.
The TAPE storage groups are used to group tape volumes that are held in tape libraries.
The OBJECT storage group can span optical, DASD, and tape volumes.
The OBJECT BACKUP storage group can contain either optical or tape volumes within
one OAM invocation.
Some volumes are not system-managed.
Other volumes are owned by DFSMShsm for use in data backup and migration.
DFSMShsm migration level 2 tape cartridges can be system-managed if you assign them
to a tape storage group.
Note: A storage group is assigned to a data set only through the storage group ACS
routine. Users cannot specify a storage group when they allocate a data set, although they
can specify a unit and volume.
Whether or not to honor a users unit and volume request is an installation decision, but we
recommend that you discourage users from directly requesting specific devices. It is more
effective for users to specify the logical storage requirements of their data by storage and
management class, which the installation can then verify in the ACS routines.
For objects, there are two types of storage groups, OBJECT and OBJECT BACKUP. An
OBJECT storage group is assigned by OAM when the object is stored; the storage group
ACS routine can override this assignment. There is only one OBJECT BACKUP storage
group, and all backup copies of all objects are assigned to this storage group.
SMS volume selection
SMS determines which volumes are used for data set allocation by developing a list of all
volumes from the storage groups assigned by the storage group ACS routine. Volumes are
then either removed from further consideration or flagged as the following:
Primary Volumes online, below threshold, that meet all the specified criteria in the
storage class.
Secondary Volumes that do not meet all the criteria for primary volumes.
Tertiary When the number of volumes in the storage group is less than the number of
volumes that are requested.
Rejected Volumes that do not meet the required specifications. They are not
candidates for selection.
SMS starts volume selection from the primary list; if no volumes are available, SMS selects
from the secondary; and, if no secondary volumes are available, SMS selects from the
tertiary list.
SMS interfaces with the system resource manager (SRM) to select from the eligible volumes
in the primary list. SRM uses device delays as one of the criteria for selection, and does not
prefer a volume if it is already allocated in the jobstep. This is useful for batch processing
when the data set is accessed immediately after creation.
SMS does not use SRM to select volumes from the secondary or tertiary volume lists. It uses
a form of randomization to prevent skewed allocations in instances such as when new
volumes are added to a storage group, or when the free space statistics are not current on
volumes.
For a striped data set, when multiple storage groups are assigned to an allocation, SMS
examines each storage group and selects the one that offers the largest number of volumes
attached to unique control units. This is called control unit separation. After a storage group
has been selected, SMS selects the volumes based on available space, control unit
separation, and performance characteristics if they are specified in the assigned storage
class.

5.13 Using aggregate backup and recovery support (ABARS)
DataSet
DataSet
Data and control

information
DataSet
Figure 5-13 ABARS
Aggregate backup and recovery support (ABARS)

Aggregate backup and recovery support, also called application backup application recovery
support, is a command-driven process to back up and recover any user-defined group of data
sets that are vital to your business. An aggregate group is a collection of related data sets and
control information that has been pooled to meet a defined backup or recovery strategy. If a
disaster occurs, you can use these backups at a remote or local site to recover critical
applications.
The user-defined group of data sets can be those belonging to an application, or any
combination of data sets that you want treated as a separate entity. Aggregate processing
enables you to:
Back up and recover data sets by application, to enable business to resume at a remote
site if necessary
Move applications in a non-emergency situation in conjunction with personnel moves or
workload balancing
Duplicate a problem at another site
You can use aggregate groups as a supplement to using management class for applications
that are critical to your business. You can associate an aggregate group with a management
class. The management class specifies backup attributes for the aggregate group, such as
the copy technique for backing up DASD data sets on primary volumes, the number of

aggregate versions to retain, and how long to retain versions. Aggregate groups simplify the
control of backup and recovery of critical data sets and applications.
Although SMS must be used on the system where the backups are performed, you can
recover aggregate groups to systems that are not using SMS, provided that the groups do not
contain data that requires that SMS be active, such as PDSEs. You can use aggregate groups
to transfer applications to other data processing installations, or to migrate applications to
newly-installed DASD volumes. You can transfer the application's migrated data, along with its
active data, without recalling the migrated data.

5.14 Automatic Class Selection (ACS) routines
New Data Set Data Class

Allocations ACS Routine
Non-System-Managed
Volume
DFSMSdss or Storage
DFSMShsm Storage Class Class
Conversion of ACS Routine
Existing Data Sets not
Storage assigned
Class Assigned
Management Class
ACS Routine System-Managed
Volume
Storage Group
ACS Routine
Figure 5-14 Using ACS routines
Using Automatic Class Selection routines

You use automatic class selection (ACS) routines to assign classes (data, storage, and
management) and storage group definitions to data sets, database data, and objects. You
write ACS routines using the ACS language, which is a high-level programming language.
Once written, you use the ACS translator to translate the routines to object form so they can
be stored in the SMS configuration.
The ACS language contains a number of read-only variables, which you can use to analyze
new data allocations. For example, you can use the read-only variable &DSN to make class
and group assignments based on data set or object collection name, or &LLQ to make
assignments based on the low-level qualifier of the data set or object collection name.
With z/OS V1R6, you can use a new ACS routine read-only security label variable,
&SECLABL, as input to the ACS routine. A security label is a name used to represent an
association between a particular security level and a set of security categories. It indicates
the minimum level of security required to access a data set protected by this profile.
Note: You cannot alter the value of read-only variables.
You use the four read-write variables to assign the class or storage group you determine for
the data set or object, based on the routine you are writing. For example, you use the
&STORCLAS variable to assign a storage class to a data set or object.

For a detailed description of the ACS language and its variables, see z/OS DFSMSdfp
Storage Administration Reference, SC26-7402.
For each SMS configuration, you can write as many as four routines: one each for data class,
storage class, management class, and storage group. Use ISMF to create, translate, validate,
and test the routines.
Processing order of ACS routines

Figure 5-14 on page 262 shows the order in which ACS routines are processed. Data can
become system-managed if the storage class routine assigns a storage class to the data, or if
it allows a user-specified storage class to be assigned to the data. If this routine does not
assign a storage class to the data, the data cannot reside on a system-managed volume.
Because data allocations, whether dynamic or through JCL, are processed through ACS
routines, you can enforce installation standards for data allocation on system-managed and
non-system-managed volumes. ACS routines also enable you to override user specifications
for data, storage, and management class, and requests for specific storage volumes.
You can use the ACS routines to determine the SMS classes for data sets created by the
Distributed FileManager/MVS. If a remote user does not specify a storage class, and if the
ACS routines decide that the data set is not to be system-managed, then the Distributed
FileManager/MVS terminates the creation process immediately and returns an error reply
message to the source. Therefore, when you construct your ACS routines, consider the
potential data set creation requests of remote users.

5.15 SMS configuration
An SMS configuration is made of

Set of data class, storage class, management class
and storage group
Optical library and drive definitions
Tape library definitions
ACS routines to assign classes
Aggregate group definitions
SMS base configuration
Figure 5-15 Defining the SMS configuration
SMS configuration
An SMS configuration is composed of:
A set of data class, management class, storage class, and storage group
ACS routines to assign the classes and groups
Optical library and drive definitions
Tape library definitions
Aggregate group definitions
SMS base configuration, that contains information such as:
Default management class
Default device geometry
The systems in the installation for which the subsystem manages storage
The SMS configuration is stored in SMS control data sets, which are VSAM linear data sets.
You must define the control data sets before activating SMS. SMS uses the following types of
control data sets:
Source Control Data Set (SCDS)
Active Control Data Set (ACDS)
Communications Data Set (COMMDS)

SMS complex (SMSplex)
A collection of systems or system groups that share a common configuration is called an
SMS complex. All systems in an SMS complex share an ACDS and a COMMDS. The
systems or system groups that share the configuration are defined to SMS in the SMS base
configuration. The ACDS and COMMDS must reside on a shared volume, accessible for all
systems in the SMS complex.

5.16 SMS control data sets
COMMDS ACDS
SCDS
SMS
Figure 5-16 SMS control data sets
SMS control data sets

SMS stores its class and group definitions, translated ACS routines, and system information
in three control data sets.
Source Control Data Set (SCDS)

The Source Control Data Set (SCDS) contains SMS classes, groups, and translated ACS
routines that define a single storage management policy, called an SMS configuration. You
can have several SCDSs, but only one can be used to activate the SMS configuration.
You use the SCDS to develop and test but, before activating a configuration, retain at least
one prior configuration if you need to regress to it because of error. The SCDS is never used
to manage allocations.
Active Control Data Set (ACDS)

The ACDS is the system's active copy of the current SCDS. When you activate a
configuration, SMS copies the existing configuration from the specified SCDS into the ACDS.
By using copies of the SMS classes, groups, volumes, optical libraries, optical drives, tape
libraries, and ACS routines rather than the originals, you can change the current storage
management policy without disrupting it. For example, while SMS uses the ACDS, you can:
Create a copy of the ACDS
Create a backup copy of an SCDS

Modify an SCDS
Define a new SCDS
We recommend that you have extra ACDSs in case a hardware failure causes the loss of your
primary ACDS. It must reside on a shared device, accessible to all systems, to ensure that
they share a common view of the active configuration. Do not have the ACDS reside on the
same device as the COMMDS or SCDS. Both the ACDS and COMMDS are needed for SMS
operation across the complex. Separation protects against hardware failure. Also create a
backup ACDS in case of hardware failure or accidental data loss or corruption.
Communications Data Set (COMMDS)

The Communications Data Set (COMMDS) data set contains the name of the ACDS and
storage group volume statistics. It enables communication between SMS systems in a
multisystem environment. The COMMDS also contains space statistics, SMS status, and
MVS status for each system-managed volume.
The COMMDS must reside on a shared device accessible to all systems. However, do not
allocate it on the same device as the ACDS. Create a spare COMMDS in case of a hardware
failure or accidental data loss or corruption. SMS activation fails if the COMMDS is
unavailable.

5.17 Implementing DFSMS
Enabling the Activating the

System-Managed Storage Management
Software base Subsystem
Managing
temporary
data
Managing Optimizing tape usage Managing

permanent object
data data
Managing tape
volumes
Figure 5-17 SMS implementation phases
Implementing DFSMS
You can implement SMS to fit your specific needs. You do not have to implement and use all
of the SMS functions. Rather, you can implement the functions you are most interested in
first. For example, you can:
Set up a storage group to only exploit the functions provided by extended format data sets,
such as striping, system-managed buffering (SMB), partial release, and so on.
Put data in a pool of one or more storage groups and assign them policies at the storage
group level to implement DFSMShsm operations in stages.
Exploit VSAM record level sharing (RLS).
DFSMS implementation phases

There are five major DFSMS implementation phases:
Enabling the software base
Activating the storage management subsystem
Managing temporary data
Managing permanent data
Managing tape data
In this book, we present an overview of the steps needed to activate, and manage data with,
a minimal SMS configuration, without affecting your JCL or data set allocations. To implement
DFSMS in your installation, however, see z/OS DFSMS Implementing System-Managed
Storage, SC26-7407.

5.18 Steps to activate a minimal SMS configuration
Allocate the SMS control data sets

Define the GRS resource names for the SMS
control data sets
Define the system group
Define a minimal SMS configuration:
Create the SCDS base data set
Create classes, storage groups and respective ACS
routines
Define the SMS subsystem to z/OS

Start SMS and activate the SMS configuration
Figure 5-18 Steps to activate a minimal SMS configuration
Steps to activate a minimal SMS configuration

Activating a minimal configuration lets you experience managing an SMS configuration
without affecting your JCL or data set allocations. This establishes an operating environment
for the storage management subsystem, without data sets becoming system-managed.
The minimal SMS configuration consists of the following elements:

A base configuration
A storage class definition
A storage group containing at least one volume
A storage class ACS routine
A storage group ACS routine
All of these elements are required for a valid SMS configuration, except for the storage class
ACS routine.
The steps needed to activate the minimal configuration are presented in Figure 5-18. When
implementing DFSMS, beginning by implementing a minimal configuration allows you to:
Gain experience with ISMF applications for the storage administrator, because you use
ISMF applications to define and activate the SMS configuration.

Gain experience with the operator commands that control operation of resources
controlled by SMS.
Learn how the SMS base configuration can affect allocations for non-system-managed
data sets. The base configuration contains installation defaults for data sets:
For non-system-managed data sets, you can specify default device geometry to ease the
conversion from device-dependent space calculations to the device-independent method
implemented by SMS.
For system-managed data sets, you can specify a default management class to be used
for data sets that are not assigned a management class by your management class ACS
routine.
Use simplified JCL.
Implement allocation standards, because you can develop a data class ACS routine to
enforce your standards.

5.19 Allocating SMS control data sets
//ALLOC EXEC PGM=IDCAMS

//SYSIN DD *
DEFINE CLUSTER(NAME(YOUR.OWN.SCDS) LINEAR VOLUME(D65DM1) -
TRK(25 5) SHAREOPTIONS(2,3)) -
DATA(NAME(YOUR.OWN.SCDS.DATA))
DEFINE CLUSTER(NAME(YOUR.OWN.ACDS) LINEAR VOLUME(D65DM2) -

DATA(NAME(YOUR.ACDS.DATA))
DEFINE CLUSTER(NAME(YOUR.OWN.COMMDS) LINEAR VOLUME(D65DM3) -

DATA(NAME(YOUR.OWN.COMMDS.DATA))
Figure 5-19 Using IDCAMS to create SMS control data sets
Calculating the SCDS and ACDS sizes

The size of the ACDS and SCDS may allow constructs for up to 32 systems. Be sure to
allocate sufficient space for the ACDS and SCDS, because insufficient ACDS size can cause
errors such as failing SMS activation. See z/OS DFSMSdfp Storage Administration
Reference, SC26-7402, for the formula used to calculate the appropriate SMS control data
set size.
Calculating the COMMDS size

The size of the communications data set (COMMDS) increased in DFSMS 1.3, because the
amount of space required to store system-related information for each volume increased. To
perform a precise calculation of the COMMDS size, use the formula provided in z/OS
DFSMSdfp Storage Administration Reference, SC26-7402.
Defining the control data sets

After you have calculated their respective sizes, define the SMS control data sets using
access method services. The SMS control data sets are VSAM linear data sets and you
define them using the IDCAMS DEFINE command, as shown in Figure 5-19. Because these
data sets are allocated before SMS is activated, space is allocated in tracks. Allocations in
KBs or MBs are only supported when SMS is active.
Specify SHAREOPTIONS(2,3) only for the SCDS. This lets one update-mode user operate
simultaneously with other read-mode users between regions.

Specify SHAREOPTIONS(3,3) for the ACDS and COMMDS. These data sets must be shared
between systems that are managing a shared DASD configuration in a DFSMS environment.
Define GRS resource names for active SMS control data sets
If you plan to share SMS control data sets between systems, consider the effects of multiple
systems sharing these data sets. Access is serialized by the use of RESERVE, which locks
out access to the entire device volume from other systems until the RELEASE is issued by the
task using the resource. This is undesirable, especially when there are other data sets on the
volume.
A RESERVE is issued when SMS is updating:

The COMMDS with space statistics at the expiration time interval specified in the
IGDSMSxx PARMLIB member
The ACDS due to changes in the SMS configuration
Place the resource name IGDCDSXS in the RESERVE conversion RNL as a generic entry to
convert the RESERVE/RELEASE to an ENQueue/DEQueue. This minimizes delays due to
contention for resources and prevents deadlocks associated with the VARY SMS command.
Important: If there are multiple SMS complexes within a global resource serialization
complex, be sure to use unique COMMDS and ACDS data set names to prevent false
contention.
For information about allocating COMMDS and ACDS data set names, see z/OS DFSMS
Implementing System-Managed Storage, SC26-7407.

5.20 Defining the SMS base configuration
Security
definitions
Storage ACS
Group Storage
Class
Validade Validate
Translate SCDS
ISMF Base Configuration

Storage Class Definition
Storage group containing at
least one volume
Storage group ACS routine
Figure 5-20 Minimal SMS configuration
Protecting the DFSMS environment

Before defining the SMS base configuration, you have to protect access to the SMS control
data sets, programs, and functions. For example, various functions in ISMF are related only to
storage administration tasks and you must protect your storage environment from
unauthorized access. You can protect the DFSMS environment with RACF.
RACF controls access to the following resources:

System-managed data sets
SMS control data sets
SMS functions and commands
Fields in the RACF profile
SMS classes
ISMF functions
For more information, see z/OS DFSMSdfp Storage Administration Reference, SC26-7402.
Defining the system group

A system group is a group of systems within an SMS complex that have similar connectivity to
storage groups, libraries, and volumes. When a Parallel Sysplex name is specified and used
as a system group name, the name applies to all systems in the Parallel Sysplex except for
those systems defined as part of the Parallel Sysplex that are explicitly named in the SMS
base configuration. The system group is defined using ISMF when defining the base
configuration.

Defining the SMS base configuration
After creating the SCDS data set with IDCAMS and setting up the security to the DFSMS
environment, you use the ISMF Control Data Set option to define the SMS base
configuration, which contains information such as:
Default management class
Default device geometry
The systems in the installation for which SMS manages storage using that configuration
To define a minimal configuration, you must do the following:

Define a storage class.
Define a storage group containing at least one volume. (The volume does not have to
exist, as long as you do not direct allocations to either the storage group or the volume.)
Create their respective ACS routines.
Defining a data class, a management class, and creating their respective ACS routines are
not required for a valid SCDS. However, because of the importance of the default
management class, we recommend that you include it in your minimal configuration.
For a detailed description of SMS classes and groups, see z/OS DFSMS Implementing
System-Managed Storage, SC26-7407.
The DFSMS product tape contains a set of sample ACS routines. The appendix of z/OS
DFSMSdfp Storage Administration Reference, SC26-7402, contains sample definitions of the
SMS classes and groups that are used in the sample ACS routines. The starter set
configuration can be used as a model for your own SCDS. For a detailed description of base
configuration attributes and how to use ISMF to define its contents, see z/OS DFSMSdfp
Storage Administration Reference, SC26-7402.
Defining the storage class

You must define at least one storage class name to SMS. Because a minimal configuration
does not include any system-managed volumes, no performance or availability information
need be contained in the minimal configuration's storage class. Specify an artificial storage
class, NONSMS. This class is later used by the storage administrator to create
non-system-managed data sets on an exception basis.
In the storage class ACS routine, the &STORCLAS variable is set to a null value to prevent
users from coding a storage class in JCL before you want to have system-managed data sets.
You define the class using ISMF. Select Storage Class in the primary menu. Then you can
define the class, NONSMS, in your configuration in one of two ways:
Select option 3 Define in the Storage Class Application Selection panel. The CDS Name
field must point to the SCDS you are building.
Select option 1 Display in the Storage Class Application Selection panel. The CDS Name
field must point to the starter set SCDS. Then, in the displayed panel, use the COPY line
operator to copy the definition of NONSMS from the starter set SCDS to your own SCDS.
Defining the storage group

You must define at least one pool storage group name to SMS, and at least one volume serial
number to this storage group. A storage group with no volumes defined is not valid. This
volume serial number is to be for a nonexistent volume to prevent the occurrence of JCL
errors from jobs accessing data sets using a specific volume serial number.

Defining a non-existent volume lets you activate SMS without having any system-managed
volumes. No data sets are system-managed at this time. This condition provides an
opportunity to experiment with SMS without any risk to your data.
Define a storage group (for example, NOVOLS) in your SCDS. A name like NOVOLS is useful
because you know it does not contain valid volumes.
Defining the default management class

Define a default management class and name it STANDEF to correspond with the entry in the
base configuration. We recommend that you specifically assign all system-managed data to a
management class. If you do not supply a default, DFSMShsm uses two days on primary
storage, and 60 days on migration level 1 storage, as the default.
No management classes are assigned when the minimal configuration is active. Definition of
this default is done here to prepare for the managing permanent data implementation phase.
The management class, STANDEF, is defined in the starter set SCDS. You can copy its
definition to your own SCDS in the same way as the storage class, NONSMS.

5.21 Creating ACS routines
Storage group ACS routine
Figure 5-21 Sample ACS routines for a minimal SMS configuration
Creating ACS routines

After you define the SMS classes and group, develop their respective ACS routines. For a
minimal SMS configuration, in the storage class ACS routine, you assign a null storage class,
as shown in the sample storage class ACS routine in Figure 5-21. The storage class ACS
routine ensures that the storage class read/write variable is always set to null. This prevents
users from externally specifying a storage class on their DD statements (STORCLAS
keyword), which causes the data set to be system-managed before you are ready.
The storage group ACS routine will never run if a null storage class is assigned. Therefore, no
data sets are allocated as system-managed by the minimal configuration. However, you must
code a trivial one to satisfy the SMS requirements for a valid SCDS. After you have written the
ACS routines, use ISMF to translate them into executable form.
Follow these steps to create a data set that contains your ACS routines:
1. If you do not have the starter set, allocate a fixed-block PDS or PDSE with LRECL=80 to
contain your ACS routines. Otherwise, start with the next step.
2. On the ISMF Primary Option Menu, select Automatic Class Selection to display the ACS
Application Selection panel.
3. Select option 1 Edit. When the next panel is shown, enter in the Edit panel the name of
the PDS or PDSE data set you want to create to contain your source ACS routines.

Translating the ACS routines
The translation process checks the routines for syntax errors and converts the code into an
ACS object. If the code translates without any syntax errors, then the ACS object is stored in
the SCDS. For translate:
1. From the ISMF ACS Application Selection Menu panel, select 2 Translate.
2. Enter your SCDS data set name, the PDS or PDSE data set name containing the ACS
source routine, and a data set name to hold the translate output listing. When the listing
data set does not exist, it is created automatically.
Validating the SCDS

When you validate your SCDS, you verify that all classes and groups assigned by your ACS
routines are defined in the SCDS. To validate the SCDS:
1. From the ISMF Primary Option Menu panel, select Control Data Set and press Enter.
2. Enter your SCDS data set name and select 4 Validate.
For more information, see z/OS DFSMS: Using the Interactive Storage Management Facility,
SC26-7411.

5.22 DFSMS setup for z/OS
IEASYSxx IEFSSNyy IGDSMSzz

..... SMS,ID=zz .....
SSN=yy JES2,... SMS ACDS(ds1)
SMS=zz ..... .....
ACS Storage
Routines Group
Data Mgmt. Storage

Class Class Class
System Managed Storage Volumes
Figure 5-22 DFSMS setup for z/OS
DFSMS setup for z/OS

In preparation for starting SMS, update the following PARMLIB members to define SMS to
z/OS:
IEASYSxx Verify the suffix of the IEFSSNyy in use and add the SMS=xx parameter,
where xx is the IGDSMS member name suffix.
IEFSSNyy You can activate SMS only after you define the SMS subsystem to z/OS. To
define SMS to z/OS, you must place a record for SMS in the IEFSSNxx
PARMLIB member.
IEFSSNxx defines how z/OS activates the SMS address space. You can
code an IEFSSNxx member with keyword or positional parameters, but not
both. We recommend using keyword parameters. We recommend that you
place the SMS record before the JES2 record in IEFSSNxx in order to start
SMS before starting the JES2 subsystem.
IGDSMSzz For each system in the SMS complex, you must create an IGDSMSxx
member in SYS1.PARMLIB. The IGDSMSzz member contains SMS
initialization control information. The suffix has a default value of 00.
Every SMS system must have an IGDSMSzz member in SYS1.PARMLIB that specifies a
required ACDS and COMMDS control data set pair. This ACDS and COMMDS pair is used if
the COMMDS of the pair does not point to another COMMDS.

If the COMMDS points to another COMMDS, the referenced COMMDS is used. This
referenced COMMDS might contain the name of an ACDS that is separate from the one
specified in the IGDSMSzz. If so, the name of the ACDS is obtained from the COMMDS
rather than from the IGDSMSzz to ensure that the system is always running under the most
recent ACDS and COMMDS.
If the COMMDS of the pair refers to another COMMDS during IPL, it means a more recent
COMMDS has been used. SMS uses the most recent COMMDS to ensure that you cannot
IPL with a down-level configuration.
The data sets that you specify for the ACDS and COMMDS pair must be the same for every
system in an SMS complex. Whenever you change the ACDS or COMMDS, update the
IGDSMSzz for every system in the SMS complex so that it specifies the same data sets.
IGDSMSzz has many parameters. For a complete description of the SMS parameters, see
z/OS MVS Initialization and Tuning Reference, SA22-7592, and z/OS DFSMSdfp Storage
Administration Reference, SC26-7402.

5.23 Starting SMS and activating a new configuration
Starting At IPL, with SMS subsystem

defined
SMS
Later, SET SMS=zz, with SMS
subsystem defined
Activating a new
SMS configuration
Figure 5-23 Starting SMS and activating a new SMS configuration
Starting SMS
To start SMS, which starts the SMS address space, use either of these methods:
With SMS=xx defined in IEASYSxx and SMS defined as a valid subsystem, IPL the
system. This starts SMS automatically.
Or, with SMS defined as a valid subsystem to z/OS, IPL the system. Start SMS later, using
the SET SMS=yy MVS operator command.
For detailed information, see z/OS DFSMSdfp Storage Administration Reference,

SC26-7402.
Activating a new SMS configuration

Activating a new SMS configuration means to copy the configuration from SCDS to ACDS
and to the SMS address space. The SCDS itself is never considered active. Attempting to
activate an ACDS that is not valid results in an error message.
You can manually activate a new SMS configuration in two ways. Note that SMS must be
active before you use one of these methods:
1. Activating an SMS configuration from ISMF:
From the ISMF Primary Option Menu panel, select Control Data Set.
In the CDS Application Selection panel, enter your SCDS data set name and select 5
Activate, or enter the ACTIVATE command on the command line.

2. Activating an SMS configuration from the operator console:
From the operator console, enter the command:
SETSMS {ACDS(YOUR.OWN.ACDS)} {SCDS(YOUR.OWN.SCDS)}
Activating the configuration means that information is brought into the SMS address space
from the ACDS.
To update the current ACDS with the contents of an SCDS, specify the SCDS parameter
only.
If you want to both specify a new ACDS and update it with the contents of an SCDS, enter
the SETSMS command with both the ACDS and SCDS parameters specified.
The ACTIVATE command, which runs from the ISMF CDS application, is equivalent to the
SETSMS operator command with the SCDS keyword specified.
If you use RACF, you can enable storage administrators to activate SMS configurations from
ISMF by defining the facility STGADMIN.IGD.ACTIVATE.CONFIGURATION and issuing
permit commands for each storage administrator.

5.24 Control SMS processing with operator commands
SETSMS - SET SMS=xx - VARY SMS - DISPLAY SMS - DEVSERV

DEVSERV P,CF00,9
IEE459I 16.14.23 DEVSERV PATHS 614
UNIT DTYPE M CNT VOLSER CHPID=PATH STATUS
RTYPE SSID CFW TC DFW PIN DC-STATE CCA DDC ALT CU-TYPE
CF00,33903 ,O,000,ITSO02,80=+ 81=+ 82=+ 83=+
2105 89D7 Y YY. YY. N SIMPLEX 00 00 2105
CF01,33903 ,O,000,TOTSP8,80=+ 81=+ 82=+ 83=+
2105 89D7 Y YY. YY. N SIMPLEX 01 01 2105
CF02,33903 ,O,000,NWCF02,80=+ 81=+ 82=+ 83=+
2105 89D7 Y YY. YY. N SIMPLEX 02 02 2105
CF03,33903 ,O,000,O38CAT,80=+ 81=+ 82=+ 83=+
2105 89D7 Y YY. YY. N SIMPLEX 03 03 2105
CF04,33903 ,O,000,O35CAT,80=+ 81=+ 82=+ 83=+
2105 89D7 Y YY. YY. N SIMPLEX 04 04 2105
CF05,33903 ,O,000,WITP00,80=+ 81=+ 82=+ 83=+
2105 89D7 Y YY. YY. N SIMPLEX 05 05 2105
CF06,33903 ,O,000,MQ531A,80=+ 81=+ 82=+ 83=+
2105 89D7 Y YY. YY. N SIMPLEX 06 06 2105
CF07,33903 ,O,000,NWCF07,80=+ 81=+ 82=+ 83=+
2105 89D7 Y YY. YY. N SIMPLEX 07 07 2105
CF08,33903 ,O,000,TOTMQ5,80=+ 81=+ 82=+ 83=+
2105 89D7 Y YY. YY. N SIMPLEX 08 08 2105
************************ SYMBOL DEFINITIONS ************************
O = ONLINE + = PATH AVAILABLE
Figure 5-24 SMS operator commands
Controlling SMS processing using operator commands

The DFSMS environment provides a set of z/OS operator commands to control SMS
processing. The VARY, DISPLAY, DEVSERV, and SET commands are MVS operator commands
that support SMS operation.
SETSMS This command changes a subset of SMS parameters from the
operator console without changing the active IGDSMSxx PARMLIB
member. For example, you can use this command to activate a new
configuration from an SCDS. The MVS operator must use SETSMS to
recover from ACDS and COMMDS failures.
For an explanation about how to recover from ACDS and COMMDS
failures, see z/OS DFSMSdfp Storage Administration Reference,
SC26-7402.
SET SMS=zz This command starts SMS, if it has not already been started, and is
defined as a valid MVS subsystem. The command also:
Changes options set on the IGDSMSxx PARMLIB member
Restarts SMS if it has terminated
Updates the SMS configuration
Table 5-1 on page 283 lists the differences between the SETSMS and
SET SMS commands.

VARY SMS This command changes storage group, volume, library, or drive status.
You can use this command to:
Limit new allocations to a volume or storage group
Enable a newly-installed volume for allocations
DISPLAY SMS Refer to Displaying the SMS configuration on page 284.
DEVSERV This command displays information for a device. Use it to display the
status of extended functions in operation for a given volume that is
attached to a cache-capable 3990 storage control. An example of the
output of this command is shown in Figure 5-24 on page 282.
Table 5-1 Comparison of SETSMS and SET SMS commands

Difference SET SMS=xx SETSMS
When and how to use the Initializes SMS parameters and Changes SMS parameters only
command. starts SMS if SMS is defined when SMS is running.
but not started at IPL. Changes
SMS parameters when SMS is
running.
Where the parameters are IGDSMSxx PARMLIB member. At the console.

entered.
What default values are Default values are used for No default values. Parameters
available. non-specified parameters. non-specified remain
unchanged.
For more information about operator commands, see z/OS MVS System Commands,
SA22-7627.

5.25 Displaying the SMS configuration
D SMS,SG(STRIPE),LISTVOL
IGD002I 16:02:30 DISPLAY SMS 581
STORGRP TYPE SYSTEM= 1 2 3 4

STRIPE POOL + + + +
VOLUME UNIT SYSTEM= 1 2 3 4 STORGRP NAME

MHLV11 D D D D STRIPE
MHLV15 + + + + STRIPE
SBOX28 6312 + + + + STRIPE
SBOX29 6412 + + + + STRIPE
SBOX3K 6010 + + + + STRIPE
SBOX3L 6108 + + + + STRIPE
SBOX3M 6204 + + + + STRIPE
SBOX3N 6309 + + + + STRIPE
SBOX30 6512 + + + + STRIPE
SBOX31 6013 + + + + STRIPE
***************************** LEGEND *****************************
. THE STORAGE GROUP OR VOLUME IS NOT DEFINED TO THE SYSTEM
+ THE STORAGE GROUP OR VOLUME IS ENABLED
- THE STORAGE GROUP OR VOLUME IS DISABLED
* THE STORAGE GROUP OR VOLUME IS QUIESCED
D THE STORAGE GROUP OR VOLUME IS DISABLED FOR NEW ALLOCATIONS ONLY
Q THE STORAGE GROUP OR VOLUME IS QUIESCED FOR NEW ALLOCATIONS ONLY
> THE VOLSER IN UCB IS DIFFERENT FROM THE VOLSER IN CONFIGURATION
SYSTEM 1 = SC63 SYSTEM 2 = SC64 SYSTEM 3 = SC65
SYSTEM 4 = SC70
Figure 5-25 Displaying the SMS configuration
Displaying the SMS configuration

You can display the SMS configuration in two ways:
Using ISMF Control Data Set, enter ACTIVE in the CDS Name field and select 1 Display.
The DISPLAY SMS operator command shows volumes, storage groups, libraries, drives,
SMS configuration information, SMS trace parameters, SMS operational options, OAM
information, OSMC information, and cache information. Enter this command to:
Confirm that the system-managed volume status is correct
Confirm that SMS starts with the proper parameters
The DISPLAY SMS command can be used in various variations. To learn about the full
functionality of this command, see z/OS MVS System Commands, SA22-7627.

5.26 Managing data with a minimal SMS configuration
Device-independence space allocation
Use ISMF to manage volumes
Use simplified JCL to allocate data sets
Manage expiration date
Establish installation standards and use data class

ACS routine to enforce them
Manage data set allocation
Use PDSE data sets
Figure 5-26 Managing data with minimal SMS configuration
Managing data allocation

After the SMS minimal configuration is active, your installation can exploit SMS capabilities
that give you experience with SMS and help you plan the DFSMS full exploitation with
system-managed data set implementation.
Inefficient space usage and poor data allocation cause problems with space and performance
management. In a DFSMS environment, you can enforce good allocation practices to help
reduce a variety of these problems. The following section highlights how to exploit SMS
capabilities.
Using data classes to standardize data allocation

You can define data classes containing standard data set allocation attributes. Users then
only need to use the appropriate data class names to create standardized data sets. To
override values in the data class definition, they can still provide specific allocation
parameters.
Data classes can be determined from the user-specified value on the DATACLAS parameter
(DD card, TSO Alloc, Dynalloc macro), from a RACF default, or by ACS routines. ACS
routines can also override user-specified or RACF default data classes.
You can override a data class attribute (not the data class itself) using JCL or dynamic
allocation parameters. DFSMS usually does not change values that are explicitly specified,
because doing so alters the original meaning and intent of the allocation. There is an

exception. If it is clear that a PDS is being allocated (DSORG=PO or DSNTYPE=PDS is
specified), and no directory space is indicated in the JCL, then the directory space from the
data class is used.
Users cannot override the data class attributes of dynamically-allocated data sets if you use
the IEFDB401 user exit.
For additional information about data classes see also 5.8, Using data classes on page 251.
For sample data classes, descriptions, and ACS routines, see z/OS DFSMS Implementing

5.27 Device-independence space allocation

//SYSPRINT DD SYSOUT=* SMS
//SYSIN DD *
ALLOCATE
DSNAME('FILE.PDSE') -
NEW -
DSNTYPE(LIBRARY)
SUB
FILE.PDSE
VOLSMS
Figure 5-27 Device independence
Ensuring device independence

The base configuration contains a default unit that corresponds to a DASD esoteric (such as
SYSDA). Default geometry for this unit is specified in bytes/track and tracks/cylinder for the
predominant device type in the esoteric. If users specify the esoteric, or do not supply the
UNIT parameter for new allocations, the default geometry converts space allocation requests
into device-independent units, such as KBs and MBs. This quantity is then converted back
into device-dependent tracks based on the default geometry.

During allocation, DFSMSdfp assists you to assign a block size that is optimal for the device.
When you allow DFSMSdfp to calculate the block size for the data set, you are using a
system-determined block size. System-determined block sizes can be calculated for
system-managed and non-system-managed primary storage, VIO, and tape data sets.
The use of system-determined block size provides:

Device independence, because you do not need to know the track capacity to allocate
efficiently
Space usage optimization
I/O performance improvement
Simplifies JCL, because you do not need to code BLKSIZE
You take full advantage of system-managed storage when you allow the system to place data
on the most appropriate device in the most efficient way, when you use system-managed data

sets (see Figure 5-27 on page 287). In the DFSMS environment, you control volume selection
through the storage class and storage group definitions you create, and by ACS routines. This
means that users do not have to specify volume serial numbers with the VOL=SER
parameter, or code a specific device type with the UNIT= parameter on their JCL. Due to a
large number of volumes in one installation, the volume selection SMS routine can take a long
time. A fast volume selection routine, which improves the best-fit and not-best-fit algorithms,
is introduced in z/OS 1.8.
When converting data sets for use in DFSMS, users do not have to remove these parameters
from existing JCL because volume and unit information can be ignored with ACS routines.
(However, you should work with users to evaluate UNIT and VOL=SER dependencies before
conversion).
If you keep the VOL=SER parameter for a non-SMS volume, but you are trying to access a
system-managed data set, then SMS might not find the data set. All SMS data sets (the ones
with a storage class) must reside on a system-managed volume.

5.28 Developing naming conventions
Setting the high-level qualifier standard
First character Second character Remaining characters

---------------------------------------------------------------------------------------------------------------------
Type of user Type of data Project name, code, or
userid
---------------------------------------------------------------------------------------------------------------------
A - Accounting Support P - Production data Example:
D - Documentation D - Development data
E - Engineering T - Test data 3000 = Project code
F - Field Support M - Master data
M - Marketing Support U - Update data
P - Programming W - Work data
$ - TSO userid
---------------------------------------------------------------------------------------------------------------------
Figure 5-28 Setting data set HLQ conventions
Developing a data set naming convention

Whenever you allocate a new data set, you (or the operating system) must give the data set a
unique name. Usually, the data set name is given as the dsname in JCL. A data set name can
be one name segment, or a series of joined name segments; see also 2.2, Data set name
rules on page 19.
You must implement a naming convention for your data sets. Although a naming convention is
not a prerequisite for DFSMS conversion, it makes more efficient use of DFSMS. You can
also reduce the cost of storage management significantly by grouping data that shares
common management requirements. Naming conventions are an effective way of grouping
data. They also:
Simplify service-level assignments to data
Facilitate writing and maintaining ACS routines
Allow data to be mixed in a system-managed environment while retaining separate
management criteria
Provide a filtering technique useful with many storage management products
Simplify the data definition step of aggregate backup and recovery support
Most naming conventions are based on the HLQ and LLQ of the data name. Other levels of
qualifiers can be used to identify generation data sets and database data. They can also be
used to help users to identify their own data.

Using a high level qualifier (HLQ)
Use the HLQ to identify the owner or owning group of the data, or to indicate data type.
Do not embed information that is subject to frequent change in the HLQ, such as department
number, application location, output device type, job name, or access method. Set a standard
within the HLQ. Figure 5-28 on page 289 shows examples of naming standards.

5.29 Setting the low-level qualifier (LLQ) standards
LLQ naming standards:

ExpDays Max Migrate Cmd/ Retain Retain
Low-Level Non- Ret Partial Days Auto No.GDG Backup Backup Days Day
Qualifier usage Period Release Non-usage Migrate Primary Freqcy Versns OnlyBUP ExtraBUP
ASM...... NOLIM NOLIM YES 15 BOTH -- 0 5 1100 120
CLIST.... NOLIM NOLIM YES 15 BOTH -- 0 5 1100 120
COB*..... NOLIM NOLIM YES 15 BOTH -- 0 5 1100 120
CNTL..... NOLIM NOLIM YES 15 BOTH -- 0 5 1100 120
DATA..... 400 400 YES 15 BOTH -- 2 2 400 60
*DATA.... 400 400 YES 15 BOTH -- 2 2 400 60
FOR*..... NOLIM NOLIM YES 15 BOTH -- 0 5 1100 120
INCL*.... NOLIM NOLIM YES 15 BOTH -- 0 5 1100 120
INPUT.... 400 400 YES 15 BOTH -- 2 2 1100 120
ISPROF... 400 400 YES 30 BOTH -- 0 2 60 30
JCL...... NOLIM NOLIM YES 15 BOTH -- 0 5 1100 120
LIST*.... 2 2 YES NONE NONE -- NONE NONE -- --
*LIST.... 2 2 YES NONE NONE -- NONE NONE -- --
LOAD*.... 400 400 YES 15 BOTH -- 1 2 -- --
MACLIB... 400 400 YES 15 BOTH -- 1 2 400 60
MISC..... 400 400 YES 15 BOTH -- 2 2 400 60
NAMES.... NOLIM NOLIM YES 15 BOTH -- 0 5 1100 120
OBJ*..... 180 180 YES 7 BOTH -- 3 1 180 30
PLI...... NOLIM NOLIM YES 15 BOTH -- 0 5 1100 120
Figure 5-29 Setting the LLQ standards
Setting the low-level qualifier (LLQ) standards

The LLQ determines the contents and storage management processing of the data. You can
use LLQs to identify data requirements for:
Migration (data sets only)
Backup (data sets and objects)
Archiving (data sets)
Retention or expiration (data sets and objects)
Class transitions (objects only)
Release of unused space (data sets only)
Mapping storage management requirements to data names is especially useful in a

system-managed environment. In an environment without storage groups, data with differing
requirements is often segregated onto separate volumes that are monitored and managed
manually. LLQ data naming conventions allow data to be mixed together in a
system-managed environment and still retain the separate management criteria.
Figure 5-29 shows examples of how you can use LLQ naming standards to indicate the
storage management processing criteria.
The first column lists the LLQ of a data name. An asterisk indicates where a partial qualifier
can be used. For example, LIST* indicates that only the first four characters of the LLQ must
be LIST; valid qualifiers include LIST1, LISTING, and LISTOUT. The remaining columns show
the storage management processing information for the data listed.

5.30 Establishing installation standards
Based on user needs
Improve service to users
Better transition to SMS-managed storage
Use service level agreement
Figure 5-30 Establishing installation standards
Establishing installation standards

Establishing standards such as naming conventions and allocation policies helps you to
manage storage more efficiently and improves service to your users. With them, your
installation is better prepared to make a smooth transition to system-managed storage.
Negotiate with your user group representatives to agree on the specific policies for the
installation, how soon you can implement them, and how strongly you enforce them.
You can simplify storage management by limiting the number of data sets and volumes that
cannot be system-managed.
Service level agreement

Document negotiated policies in a service level agreement. We recommend that you develop
the following documents before you start to migrate permanent data to system management:
A storage management plan that documents your storage administration group's strategy
for meeting storage requirements
Formal service level agreements that describe the services that you agree to provide

5.31 Planning and defining data classes
DC C
DC B
DC A
DATA CLASS ATTRIBUTES
DATA SET TYPE
RECORD LENGTH
BLOCKSIZE
SPACE REQUIREMENTS
EXPIRATION DATE
VSAM ATTRIBUTES
Sources of Data Class:

What does it look like? User Data Class defined (DD)
ACS DC routine (*)
RACF default
Figure 5-31 Planning and defining data class
Planning and defining data classes

After you establish your installations standards, use your service level agreement (SLA) for
reference when planning your data classes. SLAs identify users current allocation practices
and their requirements. For example:
Based on user requirements, create a data class to allocate standard control libraries.
Create a data class to supply the default value of a parameter, so that users do not have to
specify a value for that parameter in the JCL or dynamic allocation.
Have data class names indicate the type of data to which they are assigned, which makes it
easier for users to identify the template they need to use for allocation.
You define data classes using the ISMF data class application. Users can access the Data
Class List panel to determine which data classes are available and the allocation values that
each data class contains.
Figure 5-32 on page 294 contains information that can help in this task. For more information
about planning and defining data classes, see z/OS DFSMSdfp Storage Administration
Reference, SC26-7402.

5.32 Data class attributes
Data class name and data class description (DC)

Data set organization (RECORG) and data set name type
(DSNTYPE)
Record format (RECFM) and logical record length (LRECL)
Key length (KEYLEN) and offset (KEYOFF)
Space attributes (AVGREC, AVE VALUE, PRIMARY,
SECONDARY, DIRECTORY)
Retention period or expiration date (RETPD or EXPDT)
Number of volumes the data set can span (VOLUME COUNT)
Allocation amount when extending VSAM extended data set
Control interval size for VSAM data components (CISIZE DATA)
Percentage of control interval or control area free space (%
FREESPACE)
VSAM share options (SHAREOPTIONS)
Compaction option for data sets (COMPACTION)
Tape media (MEDIA TYPE)
Figure 5-32 Data class attributes
Data class attributes

You can specify the data class space attributes to control DASD space waste; for example:
The primary space value is to specify the total amount of space initially required for output
processing. The secondary allocation allows automatic extension of additional space as
the data set grows and does not waste space by overallocating the primary quantity. You
can also use data class space attributes to relieve users of the burden of calculating how
much primary and secondary space to allocate.
The COMPACTION attribute specifies whether data is to be compressed on DASD if the
data set is allocated in the extended format. The COMPACTION attribute alone also allows
you to use the improved data recording capability (IDRC) of your tape device when
allocating tape data sets. To use the COMPACTION attribute, the data set must be
system-managed, because this attribute demands an extended format data set.
The following attributes are used for tape data sets only:
MEDIA TYPE allows you to select the mountable tape media cartridge type.
RECORDING TECHNOLOGY allows you to select the format to use when writing to
that device.
The read-compatible special attribute indicator in the tape device selection information
(TDSI) allows an 18-track tape to be mounted on a 36-track device for read access.
The attribute increases the number of devices that are eligible for allocation when you
are certain that no more data will be written to the tape.
For detailed information about specifying data class attributes, see z/OS DFSMSdfp Storage

5.33 Use data class ACS routine to enforce standards
Examples of standards to be enforced:

Prevent extended retention or expiration periods
Prevent specific volume allocations, unless
authorized
You can control allocations to spare, system,
database, or other volumes
Require valid naming conventions for permanent
data sets
Figure 5-33 Using data class (DC) ACS routine to enforce standards
Using data class (DC) ACS routine to enforce standards

After you start DFSMS with the minimal configuration, you can use data class ACS routine
facilities to automate or simplify storage allocation standards if you:
Use manual techniques to enforce standards
Plan to enforce standards before implementing DFSMS
Use DFSMSdfp or MVS installation exits to enforce storage allocation standards
The data class ACS routine provides an automatic method for enforcing standards because it
is called for system-managed and non-system-managed data set allocations. Standards are
enforced automatically at allocation time, rather than through manual techniques after
allocation.
Enforcing standards optimizes data processing resources, improves service to users, and
positions you for implementing system-managed storage. You can fail requests or issue
warning messages to users who do not conform to standards. Consider enforcing the
following standards in your DFSMS environment:
Prevent extended retention or expiration periods.
Prevent specific volume allocations, unless authorized. For example, you can control
allocations to spare, system, database, or other volumes.
Require valid naming conventions before implementing DFSMS system management for
permanent data sets.

5.34 Simplifying JCL use
DEFINING A NEW DATA SET

(SMS)
//DD1 DD DSN=PAY.D3,
// DISP=(NEW,CATLG)
STORAGE MANAGEMENT SUBSYSTEM
MVS/ESA
Figure 5-34 Using SMS capabilities to simplify JCL
Use simplified JCL

After you define and start using data classes, several JCL keywords can help you simplify the
task of creating data sets and also make the allocation process more consistent. It is also
possible to allocate VSAM data sets through JCL without IDCAMS assistance.
For example, with the use of data classes, you have less use for the JCL keywords UNIT,
DCB, and AMP. When you start using system-managed data sets, you do not need to use the
JCL VOL keyword.
JCL keywords used in the DFSMS environment

You can use JCL keywords to create VSAM and non-VSAM data sets. For a detailed
description of the keywords and their use, see z/OS MVS JCL Users Guide, SA22-7598.
In the following sections, we present sample jobs exemplifying the use of JCL keywords
when:
Creating a sequential data set
Creating a VSAM cluster
Specifying a retention period
Specifying an expiration date

5.35 Allocating a data set
//NEWDATA DD DSN=FILE.SEQ1,
// DISP=(,CATLG),
// SPACE=(50,(5,5)),AVGREC=M,
// RECFM=VB,LRECL=80
FILE.SEQ1
Figure 5-35 Allocating a sequential data set
Creating and allocating data sets

Many times the words create and allocate, when applied to data sets, are used in MVS as
synonyms. However, they are not.
To create (on DASD) means to assign a space in VTOC to be used for a data set
(sometimes create implies cataloging the data set). A data set is created in response to
the DD card DISP=NEW in JCL.
To allocate means to establish a logical relationship between the request for the use of the
data set within the program (through the use of a DCB or ACB) and the data set itself in
the device where it is located.
Figure 5-35 shows an example of JCL used to create a data set in a system-managed
environment.
These are characteristics of the JCL in a system-managed environment:

The LRECL and RECFM parameters are independent keywords. This makes it easier to
override individual attributes that are assigned default values by the data class.
In the example, the SPACE parameter is coded with the average number of bytes per
record (50), and the number of records required for the primary data set allocation (5 M)
and secondary data set allocation (5 M). These are the values that the system uses to
calculate the least number of tracks required for the space allocation.

The AVGREC attribute indicates the scale factor for the primary and secondary allocation
values. In the example, an AVGREC value of M indicates that the primary and secondary
values of 5 are each to be multiplied by 1 048 576.
For system-managed data sets, the device-dependent volume serial number and unit
information is no longer required, because the volume is assigned within a storage group
selected by the ACS routines.
Overriding data class attributes with JCL

In a DFSMS environment, the JCL to allocate a data set is simpler and has no
device-dependent keywords.
Table 5-2 lists the attributes a user can override with JCL.
Table 5-2 Data class attributes that can be overridden by JCL

JCL DD statement keyword Use for
RECORG,KEYLEN,KEYOFF Only VSAM
RECFM Sequential (PO or PS)
LRECL,SPACE,AVGREC, RETPD or EXPDT, All data set types

VOLUME (volume count)
DSNTYPE PDS or PDSE
For more information about data classes refer to 5.8, Using data classes on page 251 and
5.32, Data class attributes on page 294.
As previously mentioned, in order to use a data class, the data set does not have to be
system-managed. An installation can take advantages of a minimal SMS configuration to
simplify JCL use and manage data set allocation.
For information about managing data allocation, see z/OS DFSMS: Using Data Sets,
SC26-7410.

5.36 Creating a VSAM cluster
//VSAM DD DSN=NEW.VSAM,
// DISP=(,CATLG),
// SPACE=(1,(2,2)),AVGREC=M,
// RECORG=KS,KEYLEN=17,KEYOFF=6,
// LRECL=80
NEW.VSAM
NEW.VSAM.DATA
NEW.VSAM.INDEX
Figure 5-36 Creating a VSAM data class
Creating a VSAM data set using JCL

In the DFSMS environment, you can create temporary and permanent VSAM data sets using
JCL by using either of the following:
The RECORG parameter of the JCL DD statement
A data class
You can use JCL DD statement parameters to override various data class attributes; see
Table 5-2 on page 298 for those related to VSAM data sets.
Important: Regarding DISP=(OLD,DELETE), in an SMS environment, the VSAM data set

is deleted at unallocation. In a non-SMS environment, the VSAM data set is kept.
A data set with a disposition of MOD is treated as a NEW allocation if it does not already
exist; otherwise, it is treated as an OLD allocation.
For a non-SMS environment, a VSAM cluster creation is only done through IDCAMS. In
Figure 5-36, NEW.VSAM refers to a KSDS VSAM cluster.

Considerations when specifying space for a KSDS
The space allocation for a VSAM entity depends on the level of the entity being allocated:
If allocation is specified at the cluster or alternate index level only, the amount needed for
the index is subtracted from the specified amount. The remainder of the specified amount
is assigned to the data component.
If allocation is specified at the data level only, the specified amount is assigned to data.
The amount needed for the index is in addition to the specified amount.
If allocation is specified at both the data and index levels, the specified data amount is
assigned to data and the specified index amount is assigned to the index.
If secondary allocation is specified at the data level, secondary allocation must be
specified at the index level or the cluster level.
You cannot use certain parameters in JCL when allocating VSAM data sets, although you can
use them in the IDCAMS DEFINE command.

5.37 Retention period and expiration date
//RETAIN DD DSN=DEPTM86.RETPD.DATA,
// DISP=(,CATLG),RETPD=365
//RETAIN DD DSN=DEPTM86.EXPDT.DATA,
// DISP=(,CATLG),EXPDT=2006/013
Figure 5-37 Retention period and expiration date
Managing retention period and expiration date

The RETPD and EXPDT parameters specify retention period and expiration date. They apply
alike to system-managed and non-system-managed data sets. They control the time during
which a data set is protected from being deleted by the system. The first DD statement in
Figure 5-37 protects the data set from deletion for 365 days. The second DD statement in
Figure 5-37 protects the data set from deletion until January 13, 2006.
The VTOC entry for non-VSAM and VSAM data sets contains the expiration date as declared
in the JCL, the TSO ALLOCATE command, the IDCAMS DEFINE command, or in the data class
definition. The expiration date is placed in the VTOC either directly from the date
specification, or after it is calculated from the retention period specification. The expiration
date in the catalog entry exists for information purposes only. If you specify the current date or
an earlier date, the data set is immediately eligible for replacement.
You can use a management class to limit or ignore the RETPD and EXPDT parameters given
by a user. If a user specifies values that exceed the maximum allowed by the management
class definition, the retention period is reset to the allowed maximum. For an expiration date
beyond year 1999 use the following format: YYYY/DDD. For more information about using
management class to control retention period and expiration date, see z/OS DFSMShsm
Storage Administration Guide, SC35-0421.
Important: Expiration dates 99365, or 99366, or 1999/365 or 1999/366 are special dates
and they mean never expires.

5.38 SMS PDSE support
SMS PDSE support
Keyword PDSESHARING in SYS1.PARMLIB

member IGDSMSxx:
NORMAL: allows multiple users to read any member

of a PDSE
EXTENDED: allows multiple users to read any

member or create new members of a PDSE
Activate new IGDSMSxx member with command:

SET SMS ID=xx
Figure 5-38 SMS PDSE support
SMS PDSE support

With the minimal SMS configuration, you can exploit the use of PDSE data sets. A PDSE
does not have to be system-managed. For information about PDSE, refer to 4.24, Partitioned
data set extended (PDSE) on page 146.
If you have DFSMS installed, you can extend PDSE sharing to enable multiple users on
multiple systems to concurrently create new PDSE members and read existing members.
Using the PDSESHARING keyword in the SYS1.PARMLIB member, IGDSMSxx, you can
specify:
NORMAL. This allows multiple users to read any member of a PDSE.
EXTENDED. This allows multiple users to read any member or create new members of a
PDSE.
All systems sharing PDSEs need to be upgraded to DFSMS to use the extended PDSE
sharing capability.
After updating the IGDSMSxx member of SYS1.PARMLIB, you need to issue the SET SMS
ID=xx command for every system in the complex to activate the sharing capability. See also
z/OS DFSMS: Using Data Sets, SC26-7410 for information about PDSE sharing.
Although SMS supports PDSs, consider converting these to the PDSE format. Refer to 4.26,
PDSE: Conversion on page 150 for more information about PDSE conversion.

5.39 Selecting data sets to allocate as PDSEs
The &DSNTYPE ACS read-only variable

controls the allocation:
&DSNTYPE = 'LIBRARY' for PDSEs

&DSNTYPE = 'PDS' for PDSs
&DSNTYPE is not specified
Indicates that the allocation request is provided by the
user through JCL, the TSO/E ALLOCATE command,
or dynamic allocation
Figure 5-39 Selecting a data set to allocate as PDSE
Selecting a data set to allocate as PDSE

As a storage administrator, you can code appropriate ACS routines to select data sets to
allocate as PDSEs and prevent inappropriate PDSs from being allocated or converted to
PDSEs.
By using the &DSNTYPE read-only variable in the ACS routine for data-class selection, you
can control which PDSs are to be allocated as PDSEs. The following values are valid for
DSNTYPE in the data class ACS routines:
&DSNTYPE = 'LIBRARY' for PDSEs.
&DSNTYPE = 'PDS' for PDSs.
&DSNTYPE is not specified. This indicates that the allocation request is provided by the
user through JCL, the TSO/E ALLOCATE command, or dynamic allocation.
If you specify a DSNTYPE value in the JCL, and a separate DSNTYPE value is also specified
in the data class selected by ACS routines for the allocation, the value specified in the data
class is ignored.

5.40 Allocating new PDSEs

//SYSPRINT DD SYSOUT=* SMS
//SYSIN DD *
ALLOCATE
DSNAME('FILE.PDSE') -
NEW -
DSNTYPE(LIBRARY)
SUB
FILE.PDSE
VOLSMS
Figure 5-40 Allocating a PDSE data set
Allocating new PDSEs

You can allocate PDSEs only in an SMS-managed environment. The PDSE data set does not
have to be system-managed. To create a PDSE, use:
DSNTYPE keyword in the JCL, TSO or IDCAMS ALLOCATE command
A data class with LIBRARY in the Data Set Name Type field
You use DSNTYPE(LIBRARY) to allocate a PDSE, or DSNTYPE(PDS) to allocate a PDS.

Figure 5-40 shows IDCAMS ALLOCATE used with the DSNTYPE(LIBRARY) keyword to
allocate a PDSE.
A PDS and a PDSE can be concatenated in JCL DD statements, or by using dynamic

allocation, such as the TSO ALLOCATE command.

5.41 System-managed data types
Temporary Permanent
Data Data
Object Data
Database
System Data
Data
Figure 5-41 System-managed data types
Data set types that can be system-managed

Now that you have experience with SMS using the minimal SMS configuration, you can plan
system-managed data sets implementation. First you need to know which data sets can be
SMS-managed and which data sets cannot be SMS-managed.
These are common types of data that can be system-managed. For details on how these data
types can be system-managed using SMS storage groups, see z/OS DFSMS Implementing
Temporary data Data sets used only for the duration of a job, job step, or terminal
session, and then deleted. These data sets can be cataloged or
uncataloged, and can range in size from small to very large.
Permanent data Data sets consisting of:
Interactive data
TSO user data sets
ISPF/PDF libraries you use during a terminal session
Data sets classified in this category are typically small, and are
frequently accessed and updated.
Batch data Data that is classified as either online-initiated, production, or test.
Data accessed as online-initiated are background jobs that an
online facility (such as TSO) generates.

Production batch refers to data created by specialized applications
(such as payroll), that can be critical to the continued operation of
your business or enterprise.
Test batch refers to data created for testing purposes.
VSAM data Data organized with VSAM, including VSAM data sets that are part of
an existing database.
Large data For most installations, large data sets occupy more than 10 percent of
a single DASD volume. Note, however, that what constitutes a large
data set is installation-dependent.
Multivolume data Data sets that span more than one volume.
Database data Data types usually having varied requirements for performance,
availability, space, and security. To accommodate special needs,
database products have specialized utilities to manage backup,
recovery, and space usage. Examples include DB2, IMS, and CICS
data.
System data Data used by MVS to keep the operating system running smoothly. In
a typical installation, 30 to 50 percent of these data sets are high
performance and are used for cataloging, error recording, and other
system functions.
Because these critical data sets contain information required to find
and access other data, they are read and updated frequently, often by
more than one system in an installation. Performance and availability
requirements are unique for system data. The performance of the
system depends heavily upon the speed at which system data sets
can be accessed. If a system data set such as a master catalog is
unavailable, the availability of data across the entire system and
across other systems can be affected.
Some system data sets can be system-managed if they are uniquely
named. These data sets include user catalogs. Place other system
data sets on non-system managed volumes. The system data sets
which are allocated at MVS system initialization are not
system-managed, because the SMS address space is not active at
initialization time.
Object data Also known as byte-stream data, this data is used in specialized
applications such as image processing, scanned correspondence, and
seismic measurements. Object data typically has no internal record or
field structure and, after it is written, the data is not changed or
updated. However, the data can be referenced many times during its
lifetime.

5.42 Data types that cannot be system-managed
Uncataloged data
Unmovable Data Sets:

Partitioned unmovable (POU)
Sequential unmovable (PSU)
Direct access unmovable (DAU)
Indexed-sequential unmovable (ISU)
Figure 5-42 Data types that cannot be system-managed
Data types that cannot be system-managed

All permanent DASD data under the control of SMS must be cataloged in integrated catalog
facility (ICF) catalogs using the standard search order. The catalogs contain the information
required for locating and managing system-managed data sets.
When data sets are cataloged, users do not need to know which volumes the data sets reside
on when they reference them; they do not need to specify unit type or volume serial number.
This is essential in an environment with storage groups, where users do not have private
volumes.
Some data cannot be system-managed, as described here:

Uncataloged data
Objects, stored in groups called collections, must have their collections cataloged in ICF
catalogs because they, and the objects they contain, are system-managed data. The
object access method (OAM) identifies an object by its collection name and the object's
own name.
An object is described only by an entry in a DB2 object directory. An object collection is
described by a collection name catalog entry and a corresponding OAM collection
identifier table entry. Therefore, an object is accessed by using the object's collection
name and the catalog entry.
When objects are written to tape, they are treated as tape data sets and OAM assigns two
tape data set names to the objects. Objects in an object storage group being written to

tape are stored as a tape data set named OAM.PRIMARY.DATA. Objects in an object
backup storage group being written to tape are stored as a tape data set named
OAM.BACKUP.DATA. Each tape containing objects has only one tape data set, and that
data set has one of the two previous names. Because the same data set name can be
used on multiple object-containing tape volumes, the object tape data sets are not
cataloged.
If you do not already have a policy for cataloging all permanent data, it is a good idea to
establish one now. For example, you can enforce standards by deleting uncataloged data
sets.
Uncataloged data sets
Data sets that are not cataloged in any ICF catalog. To locate such a data set, you need to
know the volume serial number of the volume on which the data set resides. SMS
information is stored in the catalog, that is why uncataloged data sets are not supported.
Data sets in a jobcat or stepcat
Data set LOCATEs using JOBCATs or STEPCATs are not permitted for system-managed
data sets. You must identify the owning catalogs before you migrate these data sets to
system management. The ISPF/PDF SUPERC utility is valuable for scanning your JCL
and identifying any dependencies on JOBCATs or STEPCATs.
Unmovable data sets
Unmovable data sets cannot be system-managed. These data sets include:
Data sets identified by the following data set organizations (DSORGs):
Partitioned unmovable (POU)
Sequential unmovable (PSU)
Direct access unmovable (DAU)
Indexed-sequential unmovable (ISU)
Data sets with user-written access methods
Data sets containing processing control information about the device or volume on
which they reside, including:
Absolute track data that is allocated in absolute DASD tracks or on split cylinders
Location-dependent direct data sets
All unmovable data sets must be identified and converted for use in a system-managed
environment. For information about identifying and converting unmovable data sets, see
z/OS DFSMSdss Storage Administration Guide, SC35-0423.

5.43 Interactive Storage Management Facility (ISMF)
Panel Help
-------------------------------------------------------------------------------
ISMF PRIMARY OPTION MENU - z/OS DFSMS V1 R6
Enter Selection or Command ===>
Select one of the following options and press Enter:

0 ISMF Profile - Specify ISMF User Profile
1 Data Set - Perform Functions Against Data Sets
2 Volume - Perform Functions Against Volumes
3 Management Class - Specify Data Set Backup and Migration Criteria
4 Data Class - Specify Data Set Allocation Parameters
5 Storage Class - Specify Data Set Performance and Availability
6 Storage Group - Specify Volume Names and Free Space Thresholds
7 Automatic Class Selection - Specify ACS Routines and Test Criteria
8 Control Data Set - Specify System Names and Default Criteria
9 Aggregate Group - Specify Data Set Recovery Parameters
10 Library Management - Specify Library and Drive Configurations
11 Enhanced ACS Management - Perform Enhanced Test/Configuration Management
C Data Collection - Process Data Collection Function
L List - Perform Functions Against Saved ISMF Lists
P Copy Pool - Specify Pool Storage Groups for Copies
R Removable Media Manager - Perform Functions Against Removable Media
X Exit - Terminate ISMF
Use HELP Command for Help; Use END Command or X to Exit.
Figure 5-43 ISMF Primary Option Menu panel
Interactive Storage Management Facility

The Interactive Storage Management Facility (ISMF) helps you analyze and manage data and
storage interactively. ISMF is an Interactive System Productivity Facility (ISPF) application.
Figure 5-43 shows the first ISMF panel, the Primary Option Menu.
ISMF provides interactive access to the space management, backup, and recovery services
of the DFSMShsm and DFSMSdss functional components of DFSMS, to the tape
management services of the DFSMSrmm functional component, as well as to other products.
DFSMS introduces the ability to use ISMF to define attributes of tape storage groups and
libraries.
A storage administrator uses ISMF to define the installation's policy for managing storage by
defining and managing SMS classes, groups, and ACS routines. ISMF then places the
configuration in an SCDS. You can activate an SCDS through ISMF or an operator command.
ISMF is menu-driven, with fast paths for many of its functions. ISMF uses the ISPF data-tag
language (DTL) to give its functional panels on workstations the look of common user access
(CUA) panels and a graphical user interface (GUI).

5.44 ISMF: Product relationships
ISPF/PDF
TSO/Extensions (TSO/E), TSO CLISTs, commands
DFSMS
Data Facility SORT (DFSORT)
Resource Access Control Facility (RACF)
IBM NaviQuest for MVS (NaviQuest), 5655-ACS
Figure 5-44 ISMF product relationships
ISMF product relationships

ISMF works with the following products:
Interactive System Productivity Facility/Program Development Facility (ISPF/PDF), which
provides the edit, browse, data, and library utility functions.
TSO/Extensions (TSO/E), TSO CLISTs and commands.
DFSMS, which consists of five functional components: DFSMSdfp, DFSMShsm,
DFSMSdss, DFSMSrmm and DFSMStvs. ISMF is designed to use the space
management and availability management (backup/recovery) functions provided by those
products.
Data Facility SORT (DFSORT), which provides the record-level functions.
Resource Access Control Facility (RACF), which provides the access control function for
data and services.
Device Support Facilities (ICKDSF) to provide the storage device support and analysis
functions.
IBM NaviQuest for MVS 5655-ACS.
NaviQuest is a testing and reporting tool that speeds and simplifies the tasks associated
with DFSMS initial implementation and ongoing ACS routine and configuration
maintenance. NaviQuest assists storage administrators by allowing more automation of
storage management tasks. More information about NaviQuest can be found in the
NaviQuest User's Guide.

NaviQuest provides:
A familiar ISPF panel interface to functions
Fast, easy, bulk test-case creation
ACS routine and DFSMS configuration-testing automation
Storage reporting assistance
Additional tools to aid with storage administration tasks
Batch creation of data set and volume listings
Printing of ISMF LISTs
Batch ACS routine translation
Batch ACS routine validation

5.45 ISMF: What you can do with ISMF
Edit, browse, and sort data set records
Delete data sets and backup copies
Protect data sets by limiting their access
Copy data sets to another migration level
Back up data sets and copy entire volumes,

mountable optical volumes, or mountable tape
volumes
Recall data sets that have been migrated
Figure 5-45 What you can do with ISMF
What you can do with ISMF

ISMF is a panel-driven interface. Use the panels in an ISMF application to:
Display lists with information about specific data sets, DASD volumes, mountable optical
volumes, and mountable tape volumes
Generate lists of data, storage, and management classes to determine how data sets are
being managed
Display and manage lists saved from various ISMF applications
ISMF generates a data list based on your selection criteria. Once the list is built, you can use
ISMF entry panels to perform space management or backup and recovery tasks against the
entries in the list.
As a user performing data management tasks against individual data sets or against lists of
data sets or volumes, you can use ISMF to:
Edit, browse, and sort data set records
Delete data sets and backup copies
Protect data sets by limiting their access
Recover unused space from data sets and consolidate free space on DASD volumes
Copy data sets or DASD volumes to the same device or another device
Migrate data sets to another migration level

Recall data sets that have been migrated so that they can be used
Back up data sets and copy entire volumes for availability purposes
Recover data sets and restore DASD volumes, mountable optical volumes, or mountable
tape volumes
You cannot allocate data sets from ISMF. Data sets are allocated from ISPF, from TSO, or
with JCL statements. ISMF provides the DSUTIL command, which enables users to get to
ISPF and toggle back to ISMF.

5.46 ISMF: Accessing ISMF
Panel Help
----------------------------------------------------------------------------
ISMF PRIMARY OPTION MENU - z/OS DFSMS V1 R8
Select one of the following options and press Enter:

0 ISMF Profile - Specify ISMF User Profile
1 Data Set - Perform Functions Against Data Sets
2 Volume - Perform Functions Against Volumes
3 Management Class - Specify Data Set Backup and Migration Criteria
4 Data Class - Specify Data Set Allocation Parameters
5 Storage Class - Specify Data Set Performance and Availability
6 Storage Group - Specify Volume Names and Free Space Thresholds
7 Automatic Class Selection - Specify ACS Routines and Test Criteria
8 Control Data Set - Specify System Names and Default Criteria
9 Aggregate Group - Specify Data Set Recovery Parameters
10 Library Management - Specify Library and Drive Configurations
11 Enhanced ACS Management - Perform Enhanced Test/Configuration Management
C Data Collection - Process Data Collection Function
L List - Perform Functions Against Saved ISMF Lists
P Copy Pool - Specify Pool Storage Groups for Copies
R Removable Media Manager - Perform Functions Against Removable Media
X Exit - Terminate ISMF
Use HELP Command for Help; Use END Command or X to Exit.
Figure 5-46 ISMF Primary Option Menu panel for storage administrator mode
Accessing ISMF
How you access ISMF depends on your site.
You can create an option on the ISPF Primary Option Menu to access ISMF. Then access
ISMF by typing the appropriate option after the arrow on the Option field, in the ISPF
Primary Option Menu. This starts an ISMF session from the ISPF/PDF Primary Option
Menu.
To access ISMF directly from TSO, use the command:
ISPSTART PGM(DGTFMD01) NEWAPPL(DGT)
There are two Primary Option Menus, one for storage administrators, and another for end
users. Figure 5-46 shows the menu available to storage administrators; it includes additional
applications not available to end users.
Option 0 controls the user mode or the type of Primary Option Menu to be displayed. Refer to
5.47, ISMF: Profile option on page 315 for information about how to change the user mode.
The ISMF Primary Option Menu example assumes installation of DFSMS at the current
release level. For information about adding the DFSORT option to your Primary Option Menu,
see DFSORT Installation and Customization Release 14, SC33-4034.

5.47 ISMF: Profile option
Panel Help
------------------------------------------------------------------
ISMF PROFILE OPTION MENU
Select one of the following options and Press Enter:
0 User Mode Selection

1 Logging and Abend Control
2 ISMF Job Statement
3 DFSMSdss Execute Statement
4 ICKDSF Execute Statement
5 Data Set Print Execute Statement
6 IDCAMS Execute Statement
X Exit
Figure 5-47 ISMF PROFILE OPTION MENU panel
Setting the ISMF profile

Figure 5-47 shows the ISMF Profile Option Menu panel, option 0 from the ISMF Primary
Menu. Use this menu to control the way ISMF runs during the session. You can:
Change the user mode from user to storage administrator, or from storage administrator to
user
Control ISMF error logging and recovery from abends
Define statements for ISMF to use in processing your jobs, such as:
JOB statements
DFSMSdss
Access Method Services (IDCAMS)
PRINT execute statements in your profile
You can select ISMF or ISPF JCL statements for processing batch jobs.

5.48 ISMF: Obtaining information about a panel field
HELP------------------- DATA SET LIST LINE OPERATORS -------------------HELP

COMMAND ===>
Use ENTER to see the line operator descriptions in sequence or choose them
by number:
1 LINE OPERATOR 10 DUMP 20 HRECOVER

Overview 11 EDIT 21 MESSAGE
2 ALTER 12 ERASE 22 RELEASE
3 BROWSE 13 HALTERDS 23 RESTORE
4 CATLIST 14 HBACKDS 24 SECURITY
5 CLIST 15 HBDELETE 25 SORTREC
6 COMPRESS 16 HDELETE 26 TSO Commands and CLISTs
7 CONDENSE 17 HIDE 27 VSAMREC
8 COPY 18 HMIGRATE 28 VTOCLIST
9 DELETE 19 HRECALL
Figure 5-48 Obtaining information using Help command
Using the Help program function key

On any ISMF panel, you can use the ISMF Help program function key (PF key) to obtain
information about the panel you are using and the panel fields. By positioning the cursor in a
specific field, you can obtain detailed information related to that field.
Figure 5-48 shows the panel you reach when you press the Help PF key with the cursor in the
Line Operator field of the panel shown in Figure 5-49 on page 317 where the arrow points to
the data set. The Data Set List Line Operators panel shows the commands available to enter
in that field. If you want an explanation about a specific command, type the option
corresponding to the desired command and a panel is displayed showing information about
the command function.
You can exploit the Help PF key, when defining classes, to obtain information about what you
have to enter in the fields. Place the cursor in the field and press the Help PF key.
To see and change the assigned functions to the PF keys, enter the KEYS command in the
Command field.

Panel List Dataset Utilities Scroll Help
------------------------------------------------------------------------------
DATA SET LIST
Command ===> Scroll ===> HALF
Entries 1-1 of 1
Enter Line Operators below: Data Columns 3-5 of 39
LINE ALLOC ALLOC % NOT

OPERATOR DATA SET NAME SPACE USED USED
---(1)---- ------------(2)------------ --(3)--- --(4)--- -(5)-
----------> ROGERS.SYS1.LINKLIB -------- -------- ---
---------- ------ ----------- BOTTOM OF DATA ----------- ------ ----
Figure 5-49 Data Set List panel

5.49 ISMF: Data set option
Figure 5-50 Data Set List panel using action bars
Data Set Selection Entry panel

When you select option 1 (Data Set) from the ISMF Primary Option Menu, you get to the Data
Set Selection Entry Panel. There you can specify filter criteria to get a list of data sets, as
follows:
2 Generate a new list from criteria below
Data Set Name . . . 'MHLRES2.**'
Figure 5-50 shows the data set list generated for the generic data set name MHLRES2.**.
Data Set List panel

You can use line operators to execute tasks with individual data sets. Use list commands to
execute tasks with a group of data sets. These tasks include editing, browsing, recovering
unused space, copying, migrating, deleting, backing up, and restoring data sets. TSO
commands and CLISTs can also be used as line operators or list commands. You can save a
copy of a data set list and reuse it later.
If ISMF is unable to get certain information required to check if a data set meets the selection
criteria specified, that data set is also to be included in the list. Missing information is
indicated by dashes on the corresponding column.
The Data Fields field shows how many fields you have in the list. You can navigate throughout
these fields using Right and Left PF keys. The figure also shows the use of the actions bar.

5.50 ISMF: Volume Option
Figure 5-51 Volume Selection Entry panel
Volume option
Selecting option 2 (Volume) from the ISMF Primary Option Menu takes you to the Volume List
Selection Menu panel, as follows:
VOLUME LIST SELECTION MENU

Enter Selection or Command ===> ___________________________________________
1 DASD - Generate a List of DASD Volumes

2 Mountable Optical - Generate a List of Mountable Optical Volume
3 Mountable Tape - Generate a List of Mountable Tape Volumes
Selecting option 1 (DASD) displays the Volume Selection Entry Panel, shown in part (1) of
Figure 5-51. Using filters, you can select a Volume List Panel, shown in part (2) of the figure.
Volume List panel

The volume application constructs a list of the type you choose in the Volume List Selection
Menu. Use line operators to do tasks with an individual volume. These tasks include
consolidating or recovering unused space, copying, backing up, and restoring volumes. TSO
commands and CLISTs can also be line operators or list commands. You can save a copy of
a volume list and reuse it later. With the list of mountable optical volumes or mountable tape
volumes, you can only browse the list.

5.51 ISMF: Management Class option
Figure 5-52 Management Class Selection Menu panel
Management Class Application Selection panel

The first panel (1) in Figure 5-52 shows the panel displayed when you select option 3
(Management Class) from the ISMF Primary Option Menu. Use this option to display, modify,
and define options for the SMS management classes. It also constructs a list of the available
management classes.
Management Class List panel

The second panel (2) in Figure 5-52 shows the management class list generated by the filters
chosen in the previous panel, using option 1 (List). Note how many data columns are
available. You can navigate through them using the right and left PF keys.
To view the commands you can use in the LINE OPERATOR field (marked with a circle in the
figure), place the cursor in the field and press the Help PF key.

5.52 ISMF: Data Class option
Figure 5-53 Data Class Application Selection panel
Displaying information about data classes

The first panel (1) in Figure 5-53 is the panel displayed when you choose option 4 (Data
Class) from the ISMF Primary Option Menu. Use this option to define the way data sets are
allocated in your installation.
Data class attributes are assigned to a data set when the data set is created. They apply to
both SMS-managed and non-SMS-managed data sets. Attributes specified in JCL or
equivalent allocation statements override those specified in a data class. Individual attributes
in a data class can be overridden by JCL, TSO, IDCAMS, and dynamic allocation statements.
Data Class List panel

The second panel (2) in Figure 5-53 is the Data Class List generated by the filters specified in
the previous panel.
Entering the DISPLAY line command in the LINE OPERATOR field, in front of a data class name,
displays the information about that data class, without requiring you to navigate using the
right and left PF keys.

5.53 ISMF: Storage Class option
Figure 5-54 Storage Class Application Selection panel
Storage Class Application Selection panel

The first panel (1) in Figure 5-54 shows the Storage Class Application Selection panel that is
displayed when you select option 5 (Storage Class) of the ISMF Primary Option Menu.
The Storage Class Application Selection panel lets the storage administrator specify
performance objectives and availability attributes that characterize a collection of data sets.
For objects, the storage administrator can define the performance attribute Initial Access
Response Seconds. A data set or object must be assigned to a storage class in order to be
managed by DFSMS.
Storage Class List panel

The second panel (2) in Figure 5-54 shows the storage class list generated by the filters
specified in the previous panel.
You can specify the DISPLAY line operator next to any class name on a class list to generate a
panel that displays values associated with that particular class. This information can help you
decide whether you need to assign a new DFSMS class to your data set or object.
If you determine that a data set you own is to be associated with a separate management
class or storage class, and if you have authorization, you can use the ALTER line operator
against a data set list entry to specify another storage class or management class.

5.54 ISMF: List option
Panel List Dataset Utilities Scroll Help

------------------------------------------------------------------------------
DATA SET LIST
Command ===> Scroll ===> HALF
Entries 1-1 of 1
Enter Line Operators below: Data Columns 3-5 of 39
LINE ALLOC ALLOC % NOT

OPERATOR DATA SET NAME SPACE USED USED
---(1)---- ------------(2)------------ --(3)--- --(4)--- -(5)-
ROGERS.SYS1.LINKLIB -------- -------- ---
---------- ------ ----------- BOTTOM OF DATA ----------- ------ ----
Figure 5-55 Saved ISMF Lists panel
ISMF lists
After obtaining a list (data set, data class, or storage class), you can save the list by typing
SAVE listname in the Command panel field. To see the saved lists, use the option L (List) in
the ISMF Primary Option Menu.
The List Application panel displays a list of all lists saved from ISMF applications. Each entry
in the list represents a list that was saved. If there are no saved lists to be found, the ISMF
Primary Option Menu panel is redisplayed with the message that the list is empty.
You can reuse and delete saved lists. From the List Application, you can reuse lists as though
they were created from the corresponding application. You can then use line operators and
commands to tailor and manage the information in the saved lists.
To learn more about the ISMF panel, see z/OS DFSMS: Using the Interactive Storage
Management Facility, SC26-7411.

6
Chapter 6. Catalogs
A catalog is a data set that contains information about other data sets. It provides users with
the ability to locate a data set by name, without knowing where the data set resides. By
cataloging data sets, your users will need to know less about your storage setup. Thus, data
can be moved from one device to another, without requiring a change in JCL DD statements
that refer to an existing data set.
Cataloging data sets also simplifies backup and recovery procedures. Catalogs are the
central information point for data sets; all data sets must be cataloged. In addition, all
SMS-managed data sets must be cataloged.
DFSMS allows you to use catalogs for any type of data set or object. Many advanced
functions require the use of catalogs, for example, the storage management subsystem.
Multiple user catalogs contain information about user data sets, and a single master catalog
contains entries for system data sets and user catalogs.
In z/OS, the component that controls catalogs is embedded in DFSMSdfp and is called
Catalog Management. Catalog Management has one address space for itself named Catalog
Address Space (CAS). This address space is used for buffering and to store control blocks,
together with code.
The modern catalog structure in z/OS is called the integrated catalog facility (ICF). All data
sets managed by the storage management subsystem (SMS) must be cataloged in an ICF
catalog.
Most installations depend on the availability of catalog facilities to run production job streams
and to support online users. For maximum reliability and efficiency, catalog all permanent
data sets and create catalog recovery procedures to guarantee continuous availability in
z/OS.

6.1 Catalogs
ICF BCS VVDS VTOC
structure 1 VVDS
VVDS VTOC
VVDS
VVDS VTOC
VVDS
ICF BCS VVDS VTOC
structure 2 VVDS
VVDS VTOC
VVDS
Figure 6-1 The ICF catalog structure
Catalogs
Catalogs, as mentioned, are data sets containing information about other data sets, and they
provide users with the ability to locate a data set by name, without knowing the volume where
the data set resides. This means that data sets can be moved from one device to another,
without requiring a change in JCL DD statements that refer to an existing data set.
Cataloging data sets also simplifies backup and recovery procedures. Catalogs are the
central information point for VSAM data sets; all VSAM data sets must be cataloged. In
addition, all SMS-managed data sets must be cataloged. Activity towards the catalog is much
more intense in a batch/TSO workload than in a CICS/ DB2 workload, where the majority of
data sets are allocated at CICS/DB2 initialization time.
The integrated catalog facility (ICF) structure

An integrated catalog facility (ICF) catalog is a structure that replaced the former MVS CVOL
catalog. As a catalog, it describes data set attributes and indicates the volumes on which a
data set is located. ICF catalogs are allocated by the catalog address space (CAS), a system
address space for the DFSMSdfp catalog function.
A catalog consists of two separate kinds of data sets:

A basic catalog structure (BCS) - the BCS can be considered the catalog.
A VSAM volume data set (VVDS - the VVDS can be considered an extension of the
volume table of contents (VTOC).

Basic catalog structure (BCS)
The basic catalog structure (BCS) is a VSAM key-sequenced data set. It uses the data set
name of entries to store and retrieve data set information. For VSAM data sets, the BCS
contains volume, security, ownership, and association information. For non-VSAM data sets,
the BCS contains volume, ownership, and association information. When we talk about a
catalog, we usually mean the BCS.
The VVDS can be considered an extension of the volume table of contents (VTOC). The
VVDS is volume-specific, whereas the complexity of the BCS depends on your definitions.
The relationship between the BCS and the VVDS is many-to-many. That is, a BCS can point
to multiple VVDSs and a VVDS can point to multiple BCSs.
VSAM volume data set (VVDS)

The VSAM volume data set (VVDS) is a data set that describes the characteristics of VSAM
and system-managed data sets residing on a given DASD volume; it is part of a catalog.
The VVDS contains VSAM volume records (VVRs) that hold information about VSAM data
sets residing on the volume. The VVDS also contains non-VSAM volume records (NVRs) for
SMS-managed non-VSAM data sets on the volume. If an SMS-managed non-VSAM data set
spans volumes, then only the first volume contains an NVR for that data set.
The system automatically defines a VVDS with 10 tracks primary and 10 tracks secondary
space, unless you explicitly define it.
Chapter 6. Catalogs 327

6.2 The basic catalog structure (BCS)

Index Data component
VTOC VVDS
DSNAME1 DSNAME1 ... VOL001
DSNAME1 VOL001
DSNAME3
VTOC VVDS
. DSNAME2
DSNAMEn
VOL002
.
. VTOC VVDS
DSNAME4 VOL003
DSNAMEn DSNAMEn ... VOL002
DSNAME5
The BCS is a VSAM KSDS
Figure 6-2 Basic catalog structure

The basic catalog structure (BCS) is a VSAM key-sequenced data set. It uses the data set
name of entries to store and retrieve data set information. For VSAM data sets, the BCS
contains volume, security, ownership, and association information. For non-VSAM data sets,
the BCS contains volume, ownership, and association information.
In other words, the BCS portion of the ICF catalog contains the static information about the
data set, the information that rarely changes.
Every catalog consists of one BCS and one or more VVDSs. A BCS does not own a VVDS;
that is, more than one BCS can have entries for a single VVDS. Every VVDS that is
connected to a BCS has an entry in the BCS. For example, Figure 6-2 shows a possible
relationship between a BCS and three VVDSs on three disk volumes.
For non-VSAM data sets that are not SMS-managed, all catalog information is contained
within the BCS. For other types of data sets, there is other information available in the VVDS.
BCS structure
The BCS contains the information about where a data set resides. That can be a DASD
volume, tape, or other storage medium. Related information in the BCS is grouped into
logical, variable-length, spanned records related by key. The BCS uses keys that are the data
set names (plus one character for extensions).

One control interval can contain multiple BCS records. To reduce the number of I/Os
necessary for catalog processing, logically-related data is consolidated in the BCS.
A catalog can have data sets cataloged on any number of volumes. The BCS can have as
many as 123 extents on one volume. One volume can have multiple catalogs on it. All the
necessary control information is recorded in the VVDS residing on that volume.
Master catalog
A configuration of catalogs depends on a master catalog. A master catalog has the same
structure as any other catalog. What makes it a master catalog is that all BCSs are cataloged
in it, as well as certain data sets called system data sets (for instance, SYS1.LINKLIB and
other SYS1 data sets). Master catalogs are discussed in The master catalog on page 332.
Catalogs offer advantages including improved performance, capability, usability, and

maintainability. The catalog information that requires the most frequent updates is physically
located in the VVDS on the same volume as the data sets, thereby allowing faster access. A
catalog request is expedited because fewer I/O operations are needed. Related entries, such
as a cluster and its alternate index, are processed together.

6.3 The VSAM volume data set (VVDS)
Three types of entries in a VVDS

One VSAM volume control record (VVCR)
Contains control information about BCSs which have
data sets on this volume
Multiple VSAM volume records (VVR)
Contain information about the VSAM data sets on
that volume
Multiple non-VSAM volume records (NVR)
Contain information about the non-VSAM data set
on that volume
VVDS is a VSAM entry-sequenced data set (ESDS)
Data set name: SYS1.VVDS.Vvolser
Can be defined explicitly or implicitly
Figure 6-3 The VSAM volume data set
The VSAM volume data set (VVDS)

The VSAM volume data set (VVDS) contains additional catalog information (not contained in
the BCS) about the VSAM and SMS-managed non-VSAM data sets residing on the volume
where the VVDS is located. Every volume containing any VSAM or any SMS-managed data
sets must have a VVDS on it. The VVDS acts as a kind of VTOC extension for certain types of
data sets. A VVDS can have data set information about data sets cataloged in distinct BCSs.
Entry types in the VVDS

There are three types of entries in a VVDS. They are:
VSAM volume control records (VVCR)
First logical record in a VVDS
Contain information for management of DASD space and the names of the BCSs that
have data sets on the volume
VSAM volume records (VVR)
Contain information about a VSAM data set residing on the volume
Number of VVRs varies according to the type of data set and the options specified for
the data set
Also included are data set characteristics, SMS data, extent information
There is one VVR describing the VVDs itself

Non-VSAM volume record (NVR)
Equivalent to a VVR for SMS-managed non-VSAM data sets
Contains SMS-related information
VVDS characteristics
The VVDS is a VSAM entry-sequenced data set (ESDS) that has a 4 KB control interval size.
The hexadecimal RBA of a record is used as its key or identifier.
A VVDS is recognized by the restricted data set name:

SYS1.VVDS.Vvolser
Volser is the volume serial number of the volume on which the VVDS resides.
You can explicitly define the VVDS using IDCAMS, or it is implicitly created after you define
the first VSAM or SMS-managed data set on the volume.
VVDSSPACE keyword
Prior to z/OS V1R7, the default space parameter is TRACKS(10,10), which could be too small
for sites that use custom 3390 volumes (the ones greater than 3390-9). With z/OS V1R7,
there is a new VVDSSPACE keyword of the F CATALOG command, as follows:
F CATALOG,VVDSSPACE(primary,secondary)
An explicitly defined VVDS is not related to any BCS until a data set or catalog object is
defined on the volume. As data sets are allocated on the VVDS volume, each BCS with
VSAM data sets or SMS-managed data sets residing on that volume is related to the VVDS.
VVDSSPACE indicates that the Catalog Address Space are to use the values specified as the
primary and secondary allocation amount in tracks for an implicitly defined VVDS. The default
value is ten tracks for both the primary and secondary values. The specified values are
preserved across a Catalog Address Space restart, but are not preserved across an IPL.

6.4 Catalogs by function
SYSCAT VOLABC
MASTER USER
CATALOG CATALOG
Catalog by function:
Master catalog
User catalog
SYS1.PARMLIB
ABC.DSNAME
SYS1.LINKLIB
DEF.DSNAME
ABC.DSNAME1
SYSRES VOL001
Figure 6-4 The master catalog
Catalogs by function
By function, the catalogs (BCSs) can be classified as master catalog and user catalog. A
particular case of a user catalog is the volume catalog, which is a user catalog containing only
tape library and tape volume entries.
There is no structural difference between a master catalog and a user catalog. What sets a
master catalog apart is how it is used, and what data sets are cataloged in it. For example,
the same catalog can be master in one z/OS and user in the other z/OS.
The master catalog

Each system has one active master catalog. One master catalog can be shared between
various MVS images. It does not have to reside on the system residence volume (the one that
is IPLed).
The master catalog for a system must contain entries for all user catalogs and their aliases
that the system uses. Also, all SYS1 data sets must be cataloged in the master catalog for
proper system initialization.

Important: To minimize update activity to the master catalog, and to reduce the exposure
to breakage, only SYS1 data sets, user catalog connector records, and the aliases pointing
to those connectors are to be in the master catalog.
During a system initialization, the master catalog is read so that system data sets and
catalogs can be located.
Identifying the master catalog for IPL

At IPL, you must indicate the location (volser and data set name) of the master catalog. This
information can be specified in one of two places:
SYS1.NUCLEUS member SYSCATxx (default is SYSCATLG)
SYS1.PARMLIB/SYSn.IPLPARM member LOADxx. This method is recommended.
For more information see z/OS MVS Initialization and Tuning Reference, SA22-7592.
Determine the master catalog on a running system

You can use the IDCAMS LISTCAT command for a data set with a high-level qualifier (HLQ) of
SYS1 to determine the master catalog on a system. Because all data sets with an HLQ of
SYS1 are to be in the master catalog, the catalog shown in the LISTCAT output is the master
catalog.
For information about the IDCAMS LISTCAT command, see also 6.10, Listing a catalog on
page 345.
If you do not want to run an IDCAMS job, you can run LISTCAT as a line command in ISPF
option 3.4. List the SYS1.PARMLIB and type listc ent(/), as shown in Figure 6-5.
Menu Options View Utilities Compilers Help
DSLIST - Data Sets Matching SYS1.PARMLIB Row 1 of 5

Command ===> Scroll ===> PAGE
Command - Enter "/" to select action Message Volume

-------------------------------------------------------------------------------
listc ent(/)1.PARMLIB LISTC RC=0 O37CAT
SYS1.PARMLIB.BKUP SBOX01
SYS1.PARMLIB.INSTALL Z17RB1
SYS1.PARMLIB.OLD00 SBOX01
SYS1.PARMLIB.POK Z17RB1
***************************** End of Data Set list ****************************
Figure 6-5 Example for a LISTCAT in ISPF option 3.4
Note: The forward slash (/) specifies to use the data set name on the line where the
command is entered.
This command produces output similar to the following example:

NONVSAM ------- SYS1.PARMLIB
IN-CAT --- MCAT.SANDBOX.Z17.SBOX00

User catalogs
The difference between the master catalog and the user catalogs is in the function. User
catalogs are to be used to contain information about your installation cataloged data sets
other than SYS1 data sets. There are no set rules as to how many or what size to have; it
depends entirely on your environment.
Cataloging data sets for two unrelated applications in the same catalog creates a single point
of failure for them that otherwise might not exist. Assessing the impact of outage of a given
catalog can help to determine if it is too large or can impact too many applications.

6.5 Using aliases
// DD DSN=PAY.D1
// DD DSN=PAY.D2
// DD DSN=DEPT1.VAC
// DD DSN=DEPT2.VAC
PAY.D1
...
UCAT1
PAY.D1
MCAT PAY.D2
...
ALIAS: PAY
UCAT1 PAY.D2
ALIAS: DEPT1
...
ALIAS: DEPT2
UCAT2
SYS1.LINKLIB UCAT2
SYS1.PARMLIB DEPT1.VAC
... DEPT2.VAC
...
DEPT1.VAC
DEPT2.VAC
...
Figure 6-6 Using aliases
Using aliases
Aliases are used to tell catalog management which user catalog your data set is cataloged in.
First, you place a pointer to an user catalog in the master catalog through the IDCAMS DEFINE
UCAT command. Next, you define an appropriate alias name for a user catalog in the master
catalog. Then, match the high-level qualifier (HLQ) of your data set with the alias. This
identifies the appropriate user catalog to be used to satisfy the request.
In Figure 6-6, all data sets with an HLQ of PAY have their information in the user catalog
UCAT1 because in the master catalog there is an alias PAY pointing to UCAT1.
The data sets with an HLQ of DEPT1 and DEPT2, respectively, have their information in the
user catalog UCAT2 because in the master catalog there are aliases DEPT1 and DEPT2
pointing to UCAT2.
Note: Aliases can also be used with non-VSAM data sets in order to create alternate
names to the same data set. Those aliases are not related to a user catalog.
To define an alias, use the IDCAMS command DEFINE ALIAS. An example is shown in 6.7,
Defining a catalog and its aliases on page 339.

Multilevel aliases
You can augment the standard catalog search order by defining multilevel catalog aliases. A
multilevel catalog alias is an alias of two or more high-level qualifiers. You can define aliases
of up to four high-level qualifiers.
However, the multilevel alias facility is only to be used when a better solution cannot be found.
The need for the multilevel alias facility can indicate data set naming conventions problems.
For more information about the multilevel alias facility, see z/OS DFSMS: Managing Catalogs,
SC26-7409.

6.6 Catalog search order
Start
Standard search order for a LOCATE request
IDCAMS Y Fail if not found

'CATALOG' Search in specified
KEYWORD catalog
?
Y
STEPCAT Search data set in STEPCAT
?
N
Continue if not found
Y
JOBCAT Search/define data set in
? JOBCAT
N
Continue if not found
Valid Y Search the data set in the Fail if not found

ALIAS? related USER catalog
Search the data set in Fail if not found

N
the MASTER catalog
Figure 6-7 Catalog search order for a LOCATE request
Catalog search order

LOCATE is an SVC that calls catalog management asking for a data set name search. Most
catalog searches are based on catalog aliases. Alternatives to catalog aliases are available
for directing a catalog request, specifically the JOBCAT and STEPCAT DD statements; the
CATALOG parameter of access method services; and the name of the catalog. JOBCAT and
STEPCAT are no longer allowed beginning with z/OS V1R7.
Search order for catalogs for a data set define request

For the system to determine where a data set is to be cataloged, the following search order is
used to find the catalog:
1. Use the catalog named in the IDCAMS CATALOG parameter, if coded.
2. If the data set is a generation data set, the catalog containing the GDG base definition is
used for the new GDS entry.
3. If the high-level qualifier is a catalog alias, use the catalog identified by the alias or the
catalog whose name is the same as the high-level qualifier of the data set.
4. If no catalog has been identified yet, use the master catalog.

Defining a cataloged data set
When you specify a catalog in the IDCAMS CATALOG parameter, and you have appropriate
RACF authority to the FACILITY class profile STGADMIN.IGG.DIRCAT, then the catalog you
specify is used. For instance:
DEFINE CLUSTER (NAME(PROD.PAYROLL) CATALOG(SYS1.MASTER.ICFCAT))
This command defines the data set PROD.PAYROLL in catalog SYS1.MASTER.ICFCAT. You
can use RACF to prevent the use of the CATALOG parameter and restrict the ability to define
data sets in the master catalog.
Search order for locating a data set

Base catalog searches on the catalog aliases. When appropriate aliases are defined for
catalogs, the high-level qualifier of a data set name is identical to a catalog alias and identifies
the appropriate catalog to be used to satisfy the request.
However, alternatives to catalog aliases are available for directing a catalog request,
specifically the CATALOG parameter of access method services and the name of the catalog.
The following search order is used to locate the catalog for an already cataloged data set:
1. Use the catalog named in IDCAMS CATALOG parameter, if coded. If the data set is not
found, fail the job.
2. If the data set is a generation data set, the catalog containing the GDG base definition is
used for the new GDS entry.
3. If not found, and the high-level qualifier is an alias for a catalog, search the catalog or if the
high-level qualifier is the name of a catalog, search the catalog. If the data set is not found,
fail the job.
4. Otherwise, search the master catalog.
Note: For SMS-managed data sets, JOBCAT and STEPCAT DD statements are not
allowed and cause a job failure. Also, they are not suggested even for non-SMS data sets,
because they can cause conflicted information. Therefore, do not use them and keep in
mind that they have been phased out starting with z/OS V1R7.
To use an alias to identify the catalog to be searched, the data set must have more than one
data set qualifier.
For information about the catalog standard search order also refer to z/OS DFSMS:
Managing Catalogs, SC26-7409.

6.7 Defining a catalog and its aliases
Master catalog MCAT User catalog UCAT1

Self-describing record Self-describing record
User catalog connector for UCAT1 entry for TEST1.A TEST1.A

Alias TEST1 for UCAT1 entry for TEST1.B TEST2.A
Alias TEST2 for UCAT1 entry for TEST2.A
TEST1.B
data set entry for SYS1.PARMLIB

SYS1.PARMLIB
data set entry for SYS1.LINKLIB SYS1.LINKLIB
1. DEFINE USERCATALOG ICFCAT - UCAT1

2. DEFINE ALIAS - TEST1 and TEST2
3. Create data sets with HLQ of - TEST1 and TEST2
Figure 6-8 JCL to create a basic catalog structure
Defining a catalog
You can use the IDCAMS to define and maintain catalogs. See also 4.14, Access method
services (IDCAMS) on page 129. Defining a master catalog or user catalog is basically the
same.
//DEFCAT JOB ...

//DEFCAT EXEC PGM=IDCAMS
//SYSIN DD *
DEFINE USERCATALOG -
( NAME(OTTO.CATALOG.TEST) -
MEGABYTES(15 15) -
VOLUME(VSF6S4) -
ICFCATALOG -
FREESPACE(10 10) -
STRNO(3) ) -
DATA( CONTROLINTERVALSIZE(4096) -
BUFND(4) ) -
INDEX( BUFNI(4) )
/*
Figure 6-9 Sample JCL to define a BCS

The example in Figure 6-9 on page 339 shows the JCL you can use to define a user catalog.
The catalog defined, OTTO.CATALOG.TEST, is placed on volume VSF6S4, and is allocated
with 15 megabytes primary and secondary space.
Use the access method services command DEFINE USERCATALOG ICFCATALOG to define the
basic catalog structure (BCS) of an ICF catalog. Using this command you do not specify
whether you want to create a user or a master catalog. How to identify the master catalog to
the system is described in 6.4, Catalogs by function on page 332.
A connector entry to this user catalog is created in the master catalog, as the listing in
Figure 6-10 shows.
LISTING FROM CATALOG -- CATALOG.MVSICFM.VVSF6C1

USERCATALOG --- OTTO.CATALOG.TEST
HISTORY
RELEASE----------------2
VOLUMES
VOLSER------------VSF6S4 DEVTYPE------X'3010200F'
VOLFLAG------------PRIME
ASSOCIATIONS--------(NULL)
Figure 6-10 Listing of the user catalog connector entry
The attributes of the user catalog are not defined in the master catalog. They are described in
the user catalog itself and its VVDS entry. This is called the self-describing record. The
self-describing record is given a key of binary zeros to ensure it is the first record in the
catalog. There are no associations (aliases) yet for this user catalog. To create associations,
you need to define aliases.
To define a volume catalog (for tapes), use the parameter VOLCATALOG instead of ICFCATALOG.
See z/OS DFSMS Access Method Services for Catalogs, SC26-7394, for more detail.
Defining a BCS with a model

When you define a BCS or VVDS, you can use an existing BCS or VVDS as a model for the
new one. The attributes of the existing data set are copied to the newly defined data set
unless you explicitly specify another value for an attribute. You can override any of a model's
attributes.
If you do not want to change or add any attributes, you need only supply the entry name of the
object being defined and the MODEL parameter. When you define a BCS, you must also
specify the volume and space information for the BCS.
For further information about using a model, see z/OS DFSMS: Managing Catalogs,
SC26-7409.
Defining aliases
To use a catalog, the system must be able to determine which data sets are to be defined in
that catalog. The simplest way to accomplish this is to define aliases in the master catalog for
the user catalog. Before defining an alias, carefully consider the effect the new alias has on
old data sets. A poorly chosen alias can make other data sets inaccessible.
You can define aliases for the user catalog in the same job in which you define the catalog by
including DEFINE ALIAS commands after the DEFINE USERCATALOG command. You can use
conditional operators to ensure the aliases are only defined if the catalog is successfully
defined. After the catalog is defined, you can add new aliases or delete old aliases.

Catalog aliases are defined only in the master catalog, which contains an entry for the user
catalog. The number of aliases a catalog can have is limited by the maximum record size for
the master catalog.
You cannot define an alias if a data set cataloged in the master catalog has the same
high-level qualifier as the alias. The DEFINE ALIAS command fails with a Duplicate data set
name error. For example, if a catalog is named TESTE.TESTSYS.ICFCAT, you cannot define
the alias TESTE for any catalog.
Use the sample SYSIN for an IDCAMS job in Figure 6-11 to define aliases TEST1 and
TEST2.
DEFINE ALIAS -
(NAME(TEST1) -
RELATE(OTTO.CATALOG.TEST))
DEFINE ALIAS -
(NAME(TEST2) -
RELATE(OTTO.CATALOG.TEST))
Figure 6-11 DEFINE ALIAS example
These definitions result in the following entries in the master catalog (Figure 6-12).
ALIAS --------- TEST1

IN-CAT --- CATALOG.MVSICFM.VVSF6C1
HISTORY
RELEASE----------------2
ASSOCIATIONS
USERCAT--OTTO.CATALOG.TEST
...
ALIAS --------- TEST2
IN-CAT --- CATALOG.MVSICFM.VVSF6C1
HISTORY
RELEASE----------------2
ASSOCIATIONS
USERCAT--OTTO.CATALOG.TEST
Figure 6-12 Listing an ALIAS in the master catalog
Both aliases have an association to the newly defined user catalog. If you now create a new
data set with an HLQ of TEST1 or TEST2, its entry will be directed to the new user catalog.
Also, the listing of the user catalog connector now shows both aliases; see Figure 6-13.
USERCATALOG --- OTTO.CATALOG.TEST

HISTORY
RELEASE----------------2
VOLUMES
VOLSER------------VSF6S4 DEVTYPE------X'3010200F'
VOLFLAG------------PRIME
ASSOCIATIONS
ALIAS----TEST1
ALIAS----TEST2
Figure 6-13 Listing of the user catalog connector entry

6.8 Using multiple catalogs
CATALOG.MASTER
ALIAS=PROD ALIAS=TEST ALIAS=USER1

ALIAS=USER2
CATALOG.PROD CATALOG.TEST CATALOG.USERS
data sets: data sets: data sets:

PROD.** TEST.** USER1.**
USER2.**
Figure 6-14 Using multiple catalogs
Using multiple catalogs

Multiple catalogs on multiple volumes can perform better than fewer catalogs on fewer
volumes. This is because of the interference between requests to the same catalog; for
example, a single shared catalog being locked out by another system in the sysplex. This
situation can occur if another application issues a RESERVE against the volume that has
nothing to do with catalog processing. Another reason can be that there is more competition
to use the available volumes, and thus more I/O can be in progress concurrently.
Tip: Convert all intra-sysplex RESERVES in global ENQs through the conversion RNL.
Independent of the number of catalogs, use the virtual lookaside facility (VLF) for buffering the
user catalog CIs. The master catalog CIs are naturally buffered in the catalog address space
(CAS). Multiple catalogs can reduce the impact of the loss of a catalog by:
Reducing the time necessary to recreate any given catalog
Allowing multiple catalog recovery jobs to be in process at the same time
Recovery from a pack failure is dependent on the total amount of catalog information about a
volume, regardless of whether this information is stored in one catalog or in many catalogs.
When using multiple user catalogs, consider grouping data sets under different high-level
qualifiers. You can then spread them over multiple catalogs by defining aliases for the various
catalogs.

6.9 Sharing catalogs across systems
Cache
MVS 1 Cache
MVS 2
VVR Shared catalogs:

SHAREOPTIONS (3 4) AND
CATALOG Shared device
Guarantee current buffer information
by using the VVR for serialization
Figure 6-15 Sharing catalogs
Sharing catalogs across systems

A shared catalog is a basic catalog structure (BCS) that is eligible to be used by more than
one system. It must be defined with SHAREOPTIONS(3 4), and reside on a shared volume. A
DASD volume is initialized as shared using the MVS hardware configuration definition (HCD)
facility.
Note: The device must be defined as shared to all systems that access it.
If several systems have the device defined as shared and other systems do not, then catalog
corruption will occur. Check with your system programmer to determine shared volumes.
Note that it is not necessary to have the catalog actually be shared between systems; the
catalog address space assumes it is shared if it meets the criteria stated. All VVDSs are
defined as shared. Tape volume catalogs can be shared in the same way as other catalogs.
By default, catalogs are defined with SHAREOPTIONS(3 4). You can specify that a catalog is
not to be shared by defining the catalog with SHAREOPTIONS(3 3). Only define a catalog as
unshared if you are certain it will not be shared. Place unshared catalogs on volumes that
have been initialized as unshared. Catalogs that are defined as unshared and that reside on
shared volumes will become damaged if referred to by another system.

If you need to share data sets across systems, it is advisable that you share the catalogs that
contain these data sets. A BCS catalog is considered shared when both of the following are
true:
It is defined with SHAREOPTIONS (3 4).
It resides on a shared device, as defined at HCD.
Attention: To avoid catalog corruption, define a catalog volume on a shared UCB and set
catalog SHAREOPTIONS to (3 4) on all systems sharing a catalog.
Using SHAREOPTIONS 3 means that VSAM does not issue the ENQ SYSVSAM SYSTEMS
for the catalog; SHAREOPTIONS 4 means that the VSAM buffers need to be refreshed.
You can check whether a catalog is shared by running the operator command:
MODIFY CATALOG,ALLOCATED
A flag in the catalog indicates whether the catalog is shared.
If a catalog is not really shared with another system, move the catalog to an unshared device
or alter its SHAREOPTIONS to (3 3). To prevent potential catalog damage, never place a
catalog with SHAREOPTIONS (3 3) on a shared device.
There is one VVR in a shared catalog that is used as a log by all the catalog management
accessing such catalog. This log is used to guarantee the coherency of each catalog buffer in
each z/OS system.
Guaranteeing current information

For performance reasons, CIs contained in a master catalog are buffered in the catalog
address space (CAS). Optionally (and highly advisable), the user catalogs are buffered in
VLF data space. Thus, before each search in the buffers looking for a match, a simple I/O
operation checking is performed in the log VVR. If the catalog was updated by another
system, the information in the buffer needs to be refreshed by reading the data from the
volume.
The checking also affects performance because, to maintain integrity, for every catalog
access a special VVR in the shared catalog must be read before using the cached version of
the BCS record. This access implies a DASD reserve and I/O operations.
To avoid having I/O operations to read the VVR, you can use enhanced catalog sharing
(ECS). For information about ECS, see 6.24, Enhanced catalog sharing on page 375.
Checking also ensures that the control blocks for the catalog in the CAS are updated. This
occurs if the catalog has been extended or otherwise altered from another system. This
checking maintains data integrity.

6.10 Listing a catalog
Use IDCAMS LISTCAT command to extract

information from the BCS and VVDS for:
Aliases
User catalog connectors in the master catalog
Catalogs self-describing records
VSAM data sets
Non-VSAM data sets
Library entries and volume entries of a volume
catalog
Generation data sets
Alternate index and path for a VSAM cluster
Page spaces
Figure 6-16 Listing a catalog
Requesting information from a catalog

You can list catalog records using the IDCAMS LISTCAT command, or the ISMF line operator
command CATLIST. CATLIST produces the same output as LISTCAT. With z/OS V1R8, the
LISTCAT processing has performance improvements.
You can use the LISTCAT output to monitor VSAM data sets including catalogs. The statistics
and attributes listed can be used to help determine if you reorganize, recreate, or otherwise
alter a VSAM data set to improve performance or avoid problems.
The LISTCAT command can be used in many variations to extract information about a
particular entry in the catalog. It extracts the data from the BCS and VVDS.
LISTCAT examples
LISTCAT examples for monitoring catalogs include:
List all ALIAS entries in the master catalog:
LISTCAT ALIAS CAT(master.catalog.name)
This command provides a list of all aliases that are currently defined in your master
catalog. If you need information only about one specific alias, use the keyword
ENTRY(aliasname) and specify ALL to get detailed information. For sample output of this
command, see Figure 6-12 on page 341.

List a user catalog connector in the master catalog:
LISTCAT ENT(user.catalog.name) ALL
You can use this command to display the volser and the alias associations of a user
catalog as it is defined in the master catalog. For sample output of this command, see
Figure 6-13 on page 341.
List the catalogs self-describing record:
LISTCAT ENT(user.catalog.name) CAT(user.catalog.name) ALL
This gives detailed information about a user catalog, such as attributes, statistics, extent
information, and more. Because the self-describing record is in the user catalog, you must
o specify the name of the user catalog in the CAT statement. If you do not use the CAT
keyword, only the user catalog connector information from the master catalog is listed as
in the previous example. Figure 6-17 shows sample output for this command.
LISTING FROM CATALOG -- CATALOG.MVSICFU.VVSF6C1

CLUSTER ------- 00000000000000000000000000000000000000000000
HISTORY
DATASET-OWNER-----(NULL) CREATION--------2004.260
RELEASE----------------2 EXPIRATION------0000.000
BWO STATUS--------(NULL) BWO TIMESTAMP-----(NULL)
BWO---------------(NULL)
PROTECTION-PSWD-----(NULL) RACF----------------(NO)
ASSOCIATIONS
DATA-----CATALOG.MVSICFU.VVSF6C1
INDEX----CATALOG.MVSICFU.VVSF6C1.CATINDEX
DATA ------- CATALOG.MVSICFU.VVSF6C1
HISTORY
ACCOUNT-INFO-----------------------------------(NULL)
PROTECTION-PSWD-----(NULL) RACF----------------(NO)
ASSOCIATIONS
CLUSTER--00000000000000000000000000000000000000000000
ATTRIBUTES
KEYLEN----------------45 AVGLRECL------------4086 BUFSPACE-----------11776 CISIZE--------------4096
RKP--------------------9 MAXLRECL-----------32400 EXCPEXIT----------(NULL) CI/CA----------------180
BUFND------------------4 STRNO------------------3
SHROPTNS(3,4) SPEED UNIQUE NOERASE INDEXED NOWRITECHK NOIMBED NOREPLICAT
UNORDERED NOREUSE SPANNED NOECSHARE ICFCATALOG
STATISTICS
REC-TOTAL--------------0 SPLITS-CI-------------34 EXCPS------------------0
REC-DELETED--------11585 SPLITS-CA--------------0 EXTENTS----------------1
REC-INSERTED-----------0 FREESPACE-%CI---------10 SYSTEM-TIMESTAMP:
REC-UPDATED------------0 FREESPACE-%CA---------10 X'0000000000000000'
REC-RETRIEVED----------0 FREESPC----------3686400
ALLOCATION
SPACE-TYPE------CYLINDER HI-A-RBA---------3686400
SPACE-PRI--------------5 HI-U-RBA----------737280
SPACE-SEC--------------5
VOLUME
VOLSER------------VSF6C1 PHYREC-SIZE---------4096 HI-A-RBA---------3686400 EXTENT-NUMBER----------1
DEVTYPE------X'3010200F' PHYRECS/TRK-----------12 HI-U-RBA----------737280 EXTENT-TYPE--------X'00'
VOLFLAG------------PRIME TRACKS/CA-------------15
EXTENTS:
LOW-CCHH-----X'00110000' LOW-RBA----------------0 TRACKS----------------75
HIGH-CCHH----X'0015000E' HIGH-RBA---------3686399
INDEX ------ CATALOG.MVSICFU.VVSF6C1.CATINDEX
...
Figure 6-17 Example of a LISTCAT output for a user catalog
Listing a VSAM or non-VSAM data set:

LISTCAT ENT(data.set.name) ALL
The output for a VSAM data set looks the same as in Figure 6-17 (remember, a catalog is
a VSAM data set). For a non-VSAM data set, the output is much shorter.
You can use the LISTCAT command to list information for other catalog entries as well. For
information about LISTCAT, see z/OS DFSMS Access Method Services for Catalogs,
SC26-7394.

6.11 Defining and deleting data sets
//DELDS JOB ...

//SYSIN DD *
DELETE TEST1.A
user catalog UCAT1
/*
Self describing record
entry for TEST1.A
entry for TEST1.B
entry for TEST2.A
TEST1.A TEST1.B ... VTOC
TEST1.A TEST1.B ... VVDS
TEST1.A TEST1.B
Figure 6-18 Deleting a data set
Define data sets

There are many ways to define a data set. To define means to create in VTOC and to
catalog in an ICF catalog (BCS plus VVDS). Examples are using IDCAMS DEFINE CLUSTER to
create a VSAM data set or using a JCL DD statement to define a non-VSAM data set. If you
define a VSAM data set or an SMS-managed non-VSAM data set, then an entry is created in
the BCS, in the VTOC, and in the VVDS.
Since z/OS V1R7, an attempt to define a page data set in a catalog not pointed to by the
running master causes an IDCAMS message, instead of it being executed and causing later
problems.
Delete data sets

To delete a data set implies uncataloging the data set. To scratch a data set means take it
out from the VTOC. You can delete, for example, by running an IDCAMS DELETE job; by using
a JCL DD statement; or by issuing the command DELETE in ISPF 3.4 in front of the data set
name.
The default of the DELETE command is scratch, which means the BCS, VTOC, and VVDS data
set entries are erased. By doing that, the reserved space for this data set on the volume is
released. The data set itself is not overwritten until the freed space is reused by another data
set. You can use the parameter ERASE for an IDCAMS DELETE if you want the data set to be
overwritten with binary zeros for security reasons.

Note: When you delete a data set, the BCS, VVDS, and VTOC entries for the data set are
removed. If you later recover a BCS, there can be BCS entries for data sets which have
been deleted. In this case, the data sets do not exist, and there are no entries for them in
the VVDS or VTOC. To clean up the BCS, delete the BCS entries.
Delete aliases
To simply delete an alias, use the IDCAMS DELETE ALIAS command, specifying the alias you
are deleting. To delete all the aliases for a catalog, use EXPORT DISCONNECT to disconnect the
catalog. The aliases are deleted when the catalog is disconnected. When you again connect
the catalog (using IMPORT CONNECT), the aliases remain deleted.
Delete the catalog entry

To only delete a catalog entry, you can use the DELETE NOSCRATCH command; the VVDS and
VTOC entries are not deleted. The entry deleted can be reinstated with the DEFINE
RECATALOG command, as shown in Figure 6-19.
//DELDS JOB ...

//SYSIN DD *
DELETE TEST1.A NOSCRATCH /* deletes TEST1.A from the BCS */
DEFINE NONVSAM - /* redefines TEST1.A into the BCS */
(NAME(TEST1.A) -
DEVICETYPE(3390)-
VOLUMES(VSF6S3) -
RECATALOG -
)
/*
Figure 6-19 Delete NOSCRATCH, define RECATALOG
Delete VVR or NVR records

When the catalog entry is missing, and the data set remains on the DASD, you can use the
DELETE VVR for VSAM data sets and DELETE NVR for non-VSAM SMS-managed data sets to
remove its entry from the VVDS. You can only use these commands if there is no entry in the
BCS for the data set.
//DELDS JOB ...

//S1 EXEC PGM=IDCAMS
//DD1 DD VOL=SER=VSF6S3,UNIT=3390,DISP=OLD
//SYSIN DD *
DELETE TEST1.A -
FILE(DD1) -
NVR
/*
Figure 6-20 Delete the VVDS entry for a non-VSAM data set
Important: When deleting VSAM KSDS, you must issue a DELETE VVR for each of the
components, the DATA, and the INDEX.

Delete generation data groups
In this example, a generation data group (GDG) base catalog entry is deleted from the
catalog. The generation data sets associated with GDGBASE remain unaffected in the
VTOC, as shown in Figure 6-21.
//DELGDG JOB ...

//SYSIN DD *
DELETE TEST1.GDG -
GENERATIONDATAGROUP -
RECOVERY
/*
Figure 6-21 Delete a GDG base for recovery
The DELETE command with keyword RECOVERY removes the GDG base catalog entry from the
catalog.
Delete an ICF
When deleting an ICF, take care to specify whether you want to delete only the catalog, or if
you want to delete all associated data. The following examples show how to delete a catalog.
Delete with recovery
In Figure 6-22, a user catalog is deleted in preparation for replacing it with an imported
backup copy. The VVDS and VTOC entries for objects defined in the catalog are not
deleted and the data sets are not scratched, as shown in the JCL.
//DELET13 JOB ...

//DD1 DD VOL=SER=VSER01,UNIT=3390,DISP=OLD
//SYSIN DD *
DELETE -
USER.CATALOG -
FILE(DD1) -
RECOVERY -
USERCATALOG
/*
Figure 6-22 Delete a user catalog for recovery
RECOVERY specifies that only the catalog data set is deleted, without deleting the objects
defined in the catalog.
Delete an empty user catalog
In Figure 6-23 on page 350, a user catalog is deleted. A user catalog can be deleted when
it is empty; that is, when there are no objects cataloged in it other than the catalog's
volume. If the catalog is not empty, it cannot be deleted unless the FORCE parameter is
specified.

//DELET6 JOB ...
//SYSIN DD *
DELETE -
USER.CATALOG -
PURGE -
USERCATALOG
/*
Figure 6-23 Delete an empty user catalog
Important: The FORCE parameter deletes all data sets in the catalog. The DELETE command
deletes both the catalog and the catalog's user catalog connector entry in the master
catalog.
Delete a migrated data set

A migrated data set is a data set moved by DFSMShsm to a cheaper storage device to make
room in your primary DASD farm. Catalog management recognizes that a data set is
migrated by the MIGRAT volser in its catalog entry. A migrated data set can be DELETE
SCRATCH (TSO DELETE issues the DFSMShsm HDELETE command) or NOSCRATCH.
Where:
SCRATCH This means that the non-VSAM data set being deleted from the catalog is to
be removed from the VTOC of the volume on which it resides. When
SCRATCH is specified for a cluster, alternate index, page space, or data
space, the VTOC entries for the volumes involved are updated to reflect the
deletion of the object.
NOSCRATCH This means that the non-VSAM data set being deleted from the catalog is to
remain in the VTOC of the volume on which it resides, or that it has already
been scratched from the VTOC. When NOSCRATCH is specified for a
cluster, page space, alternate index, or data space, the VTOC entries for the
volumes involved are not updated.
To execute the DELETE command against a migrated data set, you must have RACF group
ARCCATGP defined. In general to allow certain authorized users to perform these operations
on migrated data sets without recalling them, perform the following steps:
1. Define a RACF catalog maintenance group named ARCCATGP.
ADDGROUP (ARCCATGP)
2. Connect the desired users to that group.
Only when such a user is logged on under group ARCCATGP does DFSMShsm bypass the
automatic recall for UNCATALOG, RECATALOG, and DELETE/NOSCRATCH requests for
migrated data sets. For example, the following LOGON command demonstrates starting a
TSO session under ARCCATGP. For further information about ARCCATGP group, see z/OS
DFSMShsm Implementation and Customization Guide, SC35-0418.
LOGON userid | password GROUP(ARCCATGP)
To delete a migrated data set, but the data set is not recorded in the HSM control data sets,
execute a DELETE NOSCRATCH command for the data set to clean up the ICF catalog.

6.12 DELETE command enhancement with z/OS V1R11
IDCAMS DELETE command is enhanced to include a

new function called DELETE MASK
Allows users to specify the data set name selection
criteria desired with a mask-entry-name and a keyword
MASK
A mask-entry-name (also called as filter key) can have
two consecutive asterisks (**) or one or more (%) signs
Two consecutive asterisks represent zero or more
characters
% sign is the replacement for any character in that same
relative position
MASK keyword is the keyword to turn on the new feature
Figure 6-24 DELETE command enhancement with z/OS V1R11
DELETE command
The DELETE command deletes catalogs, VSAM data sets, non-VSAM data sets, and objects.
With z/OS V1R11, the IDCAMS DELETE command is enhanced to include a new function
called DELETE MASK. It allows users to specify the data set name selection criteria desired
with a mask-entry-name and a keyword MASK. A mask-entry-name (also called as filter
key) can have two consecutive asterisks (**) or one or more percentage signs (%)
The two consecutive asterisks represent zero or more characters, and it is not limited to the
number of levels. For example, A.B.** means all data set names with two or more levels with
A and B as their first and second qualifiers, respectively. The percentage sign is the
replacement for any character in that same relative position. For example, ABCDE matches
the mask-entry A%%DE, but not A%DE.
The MASK keyword is the keyword to turn on the new feature; for example:
DELETE A.B.** MASK
DELETE A.BC.M%%K MASK
NOMASK is the keyword to turn the new function off. The default is NOMASK.
If more than one entry is to be deleted, the list of entrynames must be enclosed in
parentheses. The maximum number of entrynames that can be deleted is 100. If the MASK
keyword is specified, then only one entryname can be specified. This entryname is also
known as the mask filter key.

Note: The command will result in error if there is more than one mask filter key specified in
one command.
DELETE command examples

Following are examples of how generic-level DELETE works given the following data sets:
1) AAA.BBB.CCC.DDD
2) AAA.BBB.CCC.DDD
3) AAA.BBB.CCC.DDD.EEE
4) AAA.BBB.CCC
5) BBB.DDD.AAC.BBC.EEE
6) BBB.DDD.ABC.BBC.EEE
7) BBB.DDD.ADC.BBDD.EEEE
8) BBB.DDD.ADC.BCCD.EEEE
DELETE AAA.* results in the deletion of no data sets.
DELETE AAA.BBB.* results in the deletion of data set #4.
DELETE AAA.BBB.*.DDD results in the selection of data sets #1 and #2.
DELETE AAA.BBB.*.DDD.EEE results in the deletion of data set #3.
DELETE AAA.** results in delete #1 #2 #3 and #4.
DELETE BBB.DDD.** results in delete #5 #6 #7 and #8.
DELETE BBB.DDD.BBC.A%C.BBC.EEE results in delete #5 #6.
DELETE AAA.DDD.ADC.B%%%.EEEE results in delete #7 #8.
Note: When a generic level name is specified, only one qualifier can replace the asterisk
(*). When a generic level name is specified, an asterisk (*) may only represent 1 qualifier of
a data set name. When a filter key of double asterisks (**) is specified with the MASK
parameter, the key may represent multiple qualifiers within a data set name. The double
asterisks (**) may precede or follow a period. It must be preceded or followed by either a
period or a blank.
For masking filter key of percent signs (%), it allows one to eight percent (%) specified in
each qualifier. A data set name ABCDE matches the mask entry-name 'A%09/06/03', but
does not match 'A%DE'.
MASK keyword
The DELETE MASK command allows you to specify many variations of a data set name on a
single deletion, using new wild card characters and rules to give more flexibility in selecting
the data sets to be deleted. Mask can be an asterisk (*), two consecutive asterisks (**) or a
percentage sign (%).

6.13 Backup procedures
Backing up a BCS
IDCAMS EXPORT command
DFSMSdss logical dump command
DFSMShsm BACKDS command
Backing up a VVDS
Backup the full volume
Backup all data sets described in the VVDS
Figure 6-25 Backup procedures for catalogs
Backup procedures
The two parts of an ICF catalog, the BCS and the VVDS, require separate backup
techniques. The BCS can be backed up like any other data set. Only back up the VVDS as
part of a volume dump. The entries in the VVDS and VTOC are backed up when the data sets
they describe are:
Exported with IDCAMS
Logically dumped with DFSMSdss
Backed up with DFSMShsm
Important: Because catalogs are essential system data sets, it is important that you
maintain backup copies. The more recent and accurate a backup copy, the less impact a
catalog outage will have on your installation.
Backing up a BCS
To back up a BCS you can use one of the following methods:
The access method services EXPORT command
The DFSMSdss logical DUMP command
The DFSMShsm BACKDS command

You can later recover the backup copies using the same utility used to create the backup:
The access method services IMPORT command for exported copies
The DFSMSdss RESTORE command for logical dump copies
The DFSMShsm RECOVER command for DFSMShsm backups
The copy created by these utilities is a portable sequential data set that can be stored on a
tape or direct access device, which can be of another device type than the one containing the
source catalog.
When these commands are used to back up a BCS, the aliases of the catalog are saved in
the backup copy. The source catalog is not deleted, and remains as a fully functional catalog.
The relationships between the BCS and VVDSs are unchanged.
You cannot permanently export a catalog by using the PERMANENT parameter of EXPORT.
The TEMPORARY option is used even if you specify PERMANENT or allow it to default.
Figure 6-26 shows you an example for an IDCAMS EXPORT.
//EXPRTCAT JOB ...

//RECEIVE DD DSNAME=CATBACK,UNIT=(TAPE,,DEFER),
// DISP=(NEW,KEEP),VOL=SER=327409,LABEL=(1,SL)
//SYSIN DD *
EXPORT -
USER.CATALOG -
OUTFILE(RECEIVE) -
TEMPORARY
/*
Figure 6-26 JCL to create a backup of a BCS using IDCAMS EXPORT
Note: You cannot use IDCAMS REPRO or other copying commands to create and recover
BCS backups.
Backing up a master catalog

A master catalog can be backed up like any other BCS.Use IDCAMS, DFSMSdss, or
DFSMShsm for the backup. Another way to provide a backup for the master catalog is to
create an alternate master catalog. For information about defining and using an alternate
master catalog, see z/OS DFSMS: Managing Catalogs, SC26-7409.
Also make periodic volume dumps of the master catalog's volume. This dump can later be
used by the stand-alone version of DFSMSdss to restore the master catalog if you cannot
access the volume from another system.
Backing up a VVDS
Do not back up the VVDS as a data set to provide for recovery. To back up the VVDS, back up
the volume containing the VVDS, or back up all data sets described in the VVDS (all VSAM
and SMS-managed data sets). If the VVDS ever needs to be recovered, recover the entire
volume, or all the data sets described in the VVDS.
You can use either DFSMSdss or DFSMShsm to back up and recover a volume or individual
data sets on the volume.

6.14 Recovery procedures
MASTER
CATALOG
LOCK IMPORT
USER
CATALOG
Figure 6-27 Recovery procedures
Recovery procedures
Before you run the recovery procedures mentioned in this section, see 6.23, Fixing
temporary catalog problems on page 373.
Normally, a BCS is recovered separately from a VVDS. A VVDS usually does not need to be
recovered, even if an associated BCS is recovered. However, if you need to recover a VVDS,
and a BCS resides on the VVDSs volume, you must recover the BCS as well. If possible,
export the BCS before recovering the volume, and then recover the BCS from the exported
copy. This ensures a current BCS.
Before recovering a BCS or VVDS, try to recover single damaged records. If damaged
records can be rebuilt, you can avoid a full recovery.
Single BCS records can be recovered using the IDCAMS DELETE and DEFINE commands as
described in 6.11, Defining and deleting data sets on page 347. Single VVDS and VTOC
records can be recovered using the IDCAMS DELETE command and by recovering the data
sets on the volume.
The way you recover a BCS depends on how it was saved (see 6.13, Backup procedures on
page 353). When you recover a BCS, you do not need to delete and redefine the target
catalog unless you want to change the catalog's size or other characteristics, or unless the
BCS is damaged in such a way as to prevent the usual recovery.

Aliases to the catalog can be defined if you use DFSMSdss, DFSMShsm, or if you specify
ALIAS on the IMPORT command. If you have not deleted and redefined the catalog, all existing
aliases are maintained, and any aliases defined in the backup copy are redefined if they are
not already defined.
Lock the BCS before you start recovery so that no one else has access to it while you recover
the BCS. If you do not restrict access to the catalog, users might be able to update the
catalog during recovery or maintenance and create a data integrity exposure. The catalog
also will be unavailable to any system that shares the catalog. You cannot lock a master
catalog.
After you recover the catalog, update the BCS with any changes which have occurred since
the last backup, for example, by running IDCAMS DEFINE RECATALOG for all missing entries.
You can use the access method services DIAGNOSE command to identify certain
unsynchronized entries.
The Integrated Catalog Forward Recovery Utility

You also can use the Integrated Catalog Forward Recovery Utility (ICFRU) to recover a
damaged catalog to a correct and current status. This utility uses SMF records that record
changes to the catalog, updating the catalog with changes made since the BCS was backed
up. The SMF records are used by ICFRU as a login database. Use ICFRU to avoid the loss of
catalog data even after recovery.
Recovery step by step

Follow these steps to recover a BCS using the IDCAMS IMPORT command:
1. If the catalog is used by the job scheduler for any batch jobs, hold the job queue for all job
classes except the one you use for the recovery.
2. Lock the catalog using the IDCAMS ALTER LOCK command.
3. Use ICFRU to create an updated version of your last EXPORT backup.
4. Import the most current backup copy of the BCS (which contains the BCS's aliases as
they existed when the backup was made). For example, use this JCL:
//RECOVER EXEC PGM=IDCAMS
//BACKCOPY DD DSN=BACKUP.SYS1.ICFCAT.PROJECT1,DISP=OLD
//SYSIN DD *
IMPORT INFILE(BACKCOPY) -
OUTDATASET(SYS1.ICFCAT.PROJECT1) -
ALIAS -
LOCK
5. If you did not run step 3, manually update the catalog with the changes made between the
last backup and the time of error, for example by using IDCAMS DEFINE RECATALOG.
6. Use IDCAMS DIAGNOSE and EXAMINE commands to check the contents and integrity of the
catalog (see 6.15, Checking the integrity on an ICF structure on page 357).
7. If the catalog is shared by other systems and was disconnect there for recovery, run
IDCAMS IMPORT CONNECT ALIAS on those systems to reconnect the user catalog to the
master catalog.
8. Unlock the catalog using IDCAMS ALTER UNLOCK.
9. Free the job queue if you put it on hold.
For further information about recovery procedures, see z/OS DFSMS: Managing Catalogs,
SC26-7409. For information about the IDCAMS facility, see z/OS DFSMS Access Method
Services for Catalogs, SC26-7394.

6.15 Checking the integrity on an ICF structure
UCAT1
Index Data component
DSNAME1 DSNAME1 ... VOL001 VOL001
VVDS
DSNAME1 ... UCAT1
DSNAME3 ... UCAT1
DSNAME5 ... UCAT27
VSAM error, catalog error,

. detected by EXAMINE detected by DIAGNOSE
. VVDS
. DSNAME2 ... UCAT1

VOL002
DSNAME4 ... UCAT1
DSNAMEn ... UCAT1
DSNAMEn DSNAMEn ... VOL002
Figure 6-28 Errors in an ICF structure
Two types of errors

There are two types of errors in the ICF structure that cause the need for recovery. They are:
An error in the structural integrity of the BCS or VVDS as VSAM data sets - VSAM error
Errors in the structure of a BCS as a VSAM KSDS or the VVDS as a VSAM ESDS data set
usually mean that the data set is broken (logically or physically). The data set no longer
has a valid structure that the VSAM component can handle. VSAM does not care about
the contents of the records in the BCS or VVDS.
An error within the data structure of a BCS or VVDS - catalog error
The VSAM structure of the BCS or VVDS is still valid. VSAM has no problems accessing
the data set. However, the content of the single records in the BCS or VVDS does not
conform with the catalog standards. The information in the BCS and VVDS for a single
data set can be unsynchronized, thereby making the data set inaccessible.
VSAM errors
Two kinds of VSAM errors that can occur with your BCS or VVDS:
Logical errors
The records on the DASD volume still have valid physical characteristics like record size or
CI size. The VSAM information in those records is wrong, like pointers from one record to
another or the end-of-file information.

Physical errors
The records on the DASD volume are invalid; for example, they are of a wrong length.
Reasons can be an overlay of physical DASD space or wrong extent information for the
data set in the VTOC or VVDS.
When errors in the VSAM structure occur, they are in most cases logical errors for the BCS.
Because the VVDS is an entry-sequenced data set (ESDS), it has no index component.
Logical errors for an ESDS are unlikely.
You can use the IDCAMS EXAMINE command to analyze the structure of the BCS. As
explained previously, the BCS is a VSAM key-sequenced data set (KSDS). Before running the
EXAMINE, run an IDCAMS VERIFY to make sure that the VSAM information is current, and
ALTER LOCK the catalog to prevent update from others while you are inspecting it.
With the parameter INDEXTEST, you analyze the integrity of the index. With parameter
DATATEST, you analyze the data component. If only the index test shows errors, you might
have the chance to recover the BCS by just running an EXPORT/IMPORT to rebuild the index. If
there is an error in the data component, you probably have to recover the BCS as described
in 6.14, Recovery procedures on page 355.
Catalog errors
By catalog errors we mean errors in the catalog information of a BCS or VVDS, or
unsynchronized information between the BCS and VVDS. The VSAM structure of the BCS is
still valid, that is, an EXAMINE returns no errors.
Catalog errors can make a data set inaccessible. Sometimes it is sufficient to delete the
affected entries, sometimes the catalog needs to be recovered (see 6.14, Recovery
procedures on page 355).
You can use the IDCAMS DIAGNOSE command to validate the contents of a BCS or VVDS. You
can use this command to check a single BCS or VVDS and to compare the information
between a BCS and multiple VVDSs.
For various DIAGNOSE examples, see z/OS DFSMS Access Method Services for Catalogs,
SC26-7394.

6.16 Protecting catalogs
DEFINE
RECATALOG
Protecting Catalogs
RACF
STGADMIN profiles in
vsam.rectlg
RACF FACILITY class:
dataset.def
STGADMIN.IDC.DIAGNOSE.CATALOG
STGADMIN.IDC.DIAGNOSE.VVDS dataset.ghi
STGADMIN.IDC.EXAMINE.DATASET
Figure 6-29 Protecting catalogs with RACF
Protecting catalogs
The protection of data includes:
Data security: the safety of data from theft or intentional destruction
Data integrity: the safety of data from accidental loss or destruction
Data can be protected either indirectly, by preventing access to programs that can be used to
modify data, or directly, by preventing access to the data itself. Catalogs and cataloged data
sets can be protected in both ways.
To protect your catalogs and cataloged data, use the Resource Access Control Facility
(RACF) or a similar product.
Authorized program facility (APF) protection

The authorized program facility (APF) limits the use of sensitive system services and
resources to authorized system and user programs.
For information about using APF for program authorization, see z/OS MVS Programming:
Authorized Assembler Services Guide, SA22-7608.
All IDCAMS load modules are contained in SYS1.LINKLIB, and the root segment load
module (IDCAMS) is link-edited with the SETCODE AC(1) attribute. These two characteristics
ensure that access method services executes with APF authorization.

Because APF authorization is established at the job step task level, access method services
is not authorized if invoked by an unauthorized application or terminal monitor program.
RACF authorization checking

RACF provides a software access control measure you can use in addition to or instead of
passwords. RACF protection and password protection can coexist for the same data set.
To open a catalog as a data set, you must have ALTER authority and APF authorization.
When defining an SMS-managed data set, the system only checks to make sure the user has
authority to the data set name and SMS classes and groups. The system selects the
appropriate catalog, without checking the user's authority to the catalog. You can define a
data set if you have ALTER or OPERATIONS authority to the applicable data set profile.
Deleting any type of RACF-protected entry from a RACF-protected catalog requires ALTER
authorization to the catalog or to the data set profile protecting the entry being deleted. If a
non-VSAM data set is SMS-managed, RACF does not check for DASDVOL authority. If a
non-VSAM, non-SMS-managed data set is being scratched, DASDVOL authority is also
checked.
For ALTER RENAME, the user is required to have the following two types of authority:
ALTER authority to either the data set or the catalog
ALTER authority to the new name (generic profile) or CREATE authority to the group
Be sure that RACF profiles are correct after you use REPRO MERGECAT or CNVTCAT on a
catalog that uses RACF profiles. If the target and source catalogs are on the same volume,
the RACF profiles remain unchanged.
Tape data sets defined in an integrated catalog facility catalog can be protected by:
Controlling access to the tape volumes
Controlling access to the individual data sets on the tape volumes
Profiles
To control the ability to perform functions associated with storage management, define
profiles in the FACILITY class whose profile names begin with STGADMIN (storage
administration). For a complete list of STGADMIN profiles, see z/OS DFSMSdfp Storage
Administration Reference, SC26-7402. Examples of profiles include:
STGADMIN.IDC.DIAGNOSE.CATALOG
STGADMIN.IDC.DIAGNOSE.VVDS
STGADMIN.IDC.EXAMINE.DATASET

6.17 Merging catalogs
UCAT1
REPRO MERGECAT UCAT3
UCAT2
Figure 6-30 Merging catalogs
Merging catalogs
You might find it beneficial to merge catalogs if you have many small or seldom-used
catalogs. An excessive number of catalogs can complicate recovery procedures and waste
resources such as CAS storage, tape mounts for backups, and system time performing
backups.
Merging catalogs is accomplished in much the same way as splitting catalogs (see 6.18,
Splitting a catalog on page 363). The only difference between splitting catalogs and merging
them is that in merging, you want all the entries in a catalog to be moved to another catalog,
so that you can delete the obsolete catalog.
Use the following steps to merge two integrated catalog facility catalogs:
1. Use ALTER LOCK to lock both catalogs.
2. Use LISTCAT to list the aliases for the catalog you intend to delete after the merger:
//JOB ...
//DD1 DD DSN=listcat.output,DISP=(NEW,CATLG),
// SPACE=(TRK,(10,10)),
// DCB=(RECFM=VBA,LRECL=125,BLKSIZE=629)
//SYSIN DD *
LISTC ENT(catalog.name) ALL -
OUTFILE(DD1)
/*

3. Use EXAMINE and DIAGNOSE to ensure that the catalogs are error-free. Fix any errors
indicated (see also Checking the integrity on an ICF structure on page 357).
4. Use REPRO MERGECAT without specifying the ENTRIES or LEVEL parameter. The
OUTDATASET parameter specifies the catalog that you are keeping after the two catalogs
are merged. Here is an example:
//MERGE6 JOB ...
//DD1 DD VOL=SER=VSER01,UNIT=DISK,DISP=OLD
// DD VOL=SER=VSER02,UNIT=DISK,DISP=OLD
//SYSIN DD *
REPRO -
INDATASET(USERCAT4) -
OUTDATASET(USERCAT5) -
MERGECAT -
FILE(DD1)
FROMKEY(WELCHA.C.*) TOKEY(WELCHA.E.*)
/*
Important: This step can take a long time to complete. If the MERGECAT job is cancelled,
then all merged entries so far remain in the target catalog. They are not backed out in
case the job fails. See Recovering from a REPRO MERGECAT Failure in z/OS
DFSMS: Managing Catalogs, SC26-7409, for more information about this topic.
Since z/OS V1R7, REPRO MERGECAT provides the capability to copy a range of records from
one user catalog to another. It allows recovery of a broken catalog by enabling you to copy
from one specific key to another specific key just before where the break occurred and
then recover data beginning after the break. Refer to the parameters FROMKEY/TOKEY in the
previous example.
5. Use the listing created in step 2 to create a sequence of DELETE ALIAS and DEFINE ALIAS
commands to delete the aliases of the obsolete catalog, and to redefine the aliases as
aliases of the catalog you are keeping.
The DELETE ALIAS/DEFINE ALIAS sequence must be run on each system that shares the
changed catalogs and uses a separate master catalog.
6. Use DELETE USERCATALOG to delete the obsolete catalog. Specify RECOVERY on the
DELETE command.
7. If your catalog is shared, run the EXPORT DISCONNECT command on each shared system to
remove unwanted user catalog connector entries. If your catalog is shared, run the EXPORT
DISCONNECT command on each shared system to remove unwanted user catalog
connector entries.
8. Use ALTER UNLOCK to unlock the remaining catalog.
You can also merge entries from one tape volume catalog to another using REPRO MERGECAT.
REPRO retrieves tape library or tape volume entries and redefines them in a target tape volume
catalog. In this case, VOLUMEENTRIES needs to be used to correctly filter the appropriate
entries. The LEVEL parameter is not allowed when merging tape volume catalogs.

6.18 Splitting a catalog
UCAT2
UCAT1 REPRO MERGECAT
UCAT3
Figure 6-31 Splitting a catalog
Splitting catalogs
You can split a catalog to create two catalogs or to move a group of catalog entries if you
determine that a catalog is either unacceptably large or that it contains too many entries for
critical data sets.
If the catalog is unacceptably large (that is, a catalog failure leaving too many entries
inaccessible), then you can split the catalog into two catalogs. If the catalog is of an
acceptable size but contains entries for too many critical data sets, then you can simply move
entries from one catalog to another.
To split a catalog or move a group of entries, use the access method services REPRO MERGECAT
command. Use the following steps to split a catalog or to move a group of entries:
1. Use ALTER LOCK to lock the catalog. If you are moving entries to an existing catalog, lock it
as well.
2. If you are splitting a catalog, define a new catalog with DEFINE USERCATALOG LOCK (see also
Defining a catalog and its aliases on page 339).
3. Use LISTCAT to obtain a listing of the catalog aliases you are moving to the new catalog.
Use the OUTFILE parameter to define a data set to contain the output listing (see also
Merging catalogs on page 361).
4. Use EXAMINE and DIAGNOSE to ensure that the catalogs are error-free. Fix any errors
indicated (see also Checking the integrity on an ICF structure on page 357).

5. Use REPRO MERGECAT to split the catalog or move the group of entries. When splitting a
catalog, the OUTDATASET parameter specifies the catalog created in step 2. When
moving a group of entries, the OUTDATASET parameter specifies the catalog which is to
receive the entries.
Use the ENTRIES or LEVEL parameters to specify which catalog entries are to be
removed from the source catalog and placed in the catalog specified in OUTDATASET.
In the following example all entries that match the generic name VSAMDATA.* are moved
from catalog USERCAT4 to USERCAT5.
//MERGE76 JOB ...
//DD1 DD VOL=SER=VSER01,UNIT=DISK,DISP=OLD
//SYSIN DD *
REPRO -
INDATASET(USERCAT4) -
OUTDATASET(USERCAT5) -
ENTRIES(VSAMDATA.*) -
MERGECAT -
FILE(DD1)
/*
Important: This step can take a long time to complete. If the MERGECAT job is cancelled,
all merged entries so far will remain in the target catalog. They are not backed out in
case the job fails. See Recovering from a REPRO MERGECAT Failure in z/OS
DFSMS: Managing Catalogs, SC26-7409, for more information about this topic.
6. Use the listing created in step 3 to create a sequence of DELETE ALIAS and DEFINE ALIAS
commands for each alias. These commands delete the alias from the original catalog, and
redefine them as aliases for the catalog which now contains entries belonging to that alias
name.
The DELETE ALIAS/DEFINE ALIAS sequence must be run on each system that shares the
changed catalogs and uses a separate master catalog.
7. Unlock both catalogs using ALTER UNLOCK.

6.19 Catalog performance
Factors that have influence on the performance:

Main factor: amount of I/Os
Cache catalogs to decrease number of I/Os
Use Enhanced Catalog Sharing
No user data sets in master catalog

Options for DEFINE USERCATALOG
Convert SYSIGGV2 resource to avoid enqueue
contention and deadlocks between systems
Eliminate the use of JOBCAT/STEPCAT
Figure 6-32 Catalog performance
Catalog performance
Performance is not the main consideration when defining catalogs. It is more important to
create a catalog configuration that allows easy recovery of damaged catalogs with the least
amount of system disruption. However, there are several options you can choose to improve
catalog performance without affecting the recoverability of a catalog. Remember that in an
online environment, such as CICS/DB2, the number of data set allocations is minimal and
consequently the catalog activity is low.
Factors affecting catalog performance

The main factors affecting catalog performance are the amount of I/O required for the catalog
and the subsequent amount of time it takes to perform the I/O. These factors can be reduced
by buffering catalogs in special buffer pools used only by CI catalogs. Other factors are the
size and usage of the master catalog, options that were used to define a catalog, and the way
you share catalogs between systems.
Buffering catalogs
The simplest method of improving catalog performance is to use a buffer to maintain catalog
records within CAS private area address space or VLF data space. Two types of buffer are
available exclusively for catalogs. The in-storage catalog (ISC) buffer is contained within the
catalog address space (CAS). The catalog data space buffer (CDSC) is separate from CAS
and uses the z/OS VLF component, which stores the buffered records in a data space. Both
types of buffer are optional, and each can be cancelled and restarted without an IPL.

The two types of buffer are used to keep catalog records in the storage. This avoids I/Os that
are necessary to read the records from DASD. There are several things you need to take into
considerations to decide what kind of buffer to use for which catalog. See z/OS DFSMS:
Managing Catalogs, SC26-7409, for more information about buffering. Another kind of
caching is using enhanced catalog sharing to avoid I/Os to read the catalog VVR. Refer to
6.24, Enhanced catalog sharing on page 375 for more information about this topic.
Master catalog
If the master catalog only contains entries for catalogs, catalog aliases, and system data sets,
the entire master catalog is read into main storage during system initialization. Because the
master catalog, if properly used, is rarely updated, the performance of the master catalog is
not appreciably affected by I/O requirements. For that reason, keep the master catalog small
and do not define user data sets into it.
Options for defining a user catalog

There are several options you can specify when you define a user catalog that have an impact
on the performance. The options are:
STRNO - Specifies the numbers of concurrent requests.
BUFFERSPACE - Required buffer space for your catalog. It is determined by catalog
management, but you can change it.
BUFND - Number of buffers for transmitting data between virtual and DASD. The default is
STRNO+1, but you can change it.
BUFNI - Number of buffers for transmitting index entries between virtual and auxiliary
storage. Default is STRNO+2.
FREESPACE - An adequate value allows catalog updates without an excessive number of
control interval and control area splits
For more information about these values, see z/OS DFSMS Access Method Services for
Catalogs, SC26-7394.
Since z/OS V1R7, a new catalog auto-tuning (every 10-minutes) automatically modifies
temporarily the number of data buffers, index buffers, and VSAM strings for catalogs. When
any modification occurs the message IEC391I is issued telling the new values. This function is
by default enabled, but can be disabled through the F CATALOG,DISABLE(AUTOTUNING).
Convert SYSIGGV2 resource

Catalog management uses the SYSIGGV2 reserve when serializing access to catalogs. The
SYSIGGV2 reserve is used to serialize the entire catalog BCS component across all I/O as
well as to serialize access to specific catalog entries.
If the catalog is shared only within one GRSplex, convert the SYSIGGV2 resource to a global
enqueue to avoid reserves on the volume on which the catalog resides. If you are not
converting SYSIGGV2, you can have ENQ contentions on those volumes and even run into
deadlock situations.
Important: If you share a catalog with a system that is not in the same GRS complex, do
not convert the SYSIGGV2 resource for this catalog. Sharing a catalog outside the
complex requires reserves for the volume on which the catalog resides. Otherwise, you will
break the catalog. For more information, see z/OS MVS Planning: Global Resource
Serialization, SA22-7600.

6.20 F CATALOG,REPORT,PERFORMANCE command
The F CATALOG,REPORT,PERFORMANCE
command can be used to examine the following:
Certain events that occur in the catalog address space
These events represent points at which catalog code

calls some function outside of the catalog component
Such as enqueues, I/O, or allocations
All such events are tracked, except for:
Lock manager requests and GETMAIN/FREEMAIN

activity
Figure 6-33 F CATALOG,REPORT,PERFORMANCE command
F CATALOG,REPORT,PERFORMANCE command
You can use the F CATALOG command to list information about catalogs currently allocated to
the catalog address space. Sometimes you need this information so that you can use another
MODIFY command to close or otherwise manipulate a catalog in cache.
The command displays information about the performance of specific events that catalog
processing invokes. Each line shows the number of times (nnn) that event has occurred since
IPL or the last reset of the statistics by using the F CATALOG,REPORT,PERFORMANCE(RESET), and
the average time for each occurrence (nnn.nnn). The unit of measure of the average time
(unit) is either milliseconds (MSEC), seconds (SEC), or the average shown as hours, minutes,
and seconds (hh:mm:ss.th).
Note: Other forms of the REPORT command provide information about various aspects of
the catalog address space.
The F CATALOG,REPORT,CACHE command also provides rich information about the use of
catalog buffering. This command causes general information about catalog cache status for
all catalogs currently active in the catalog address space to be listed. The report generated
shows information useful in evaluating the catalog cache performance for the listed catalogs.

Catalog report output
This MODIFY command is very important for performance analysis. Try to get familiar with each
meaning to understand what can be done to improve catalog performance. These counters
can be zeroed through the use of a reset command. The F CATALOG,REPORT,PERFORMANCE
command can be used to examine certain events that occur in the catalog address space.
These events represent points at which catalog code calls a function outside of the catalog
component, such as enqueues, I/O, or allocation. All such events are tracked, except for lock
manager requests and GETMAIN/FREEMAIN activity. Figure 6-34 shows a catalog
performance report.
F CATALOG,REPORT,PERFORMANCE
IEC351I CATALOG ADDRESS SPACE MODIFY COMMAND ACTIVE
IEC359I CATALOG PERFORMANCE REPORT
*CAS***************************************************
* Statistics since 23:04:21.14 on 01/25/2010 *
* -----CATALOG EVENT---- --COUNT-- ---AVERAGE--- *
* Entries to Catalog 607,333 3.632 MSEC *
* BCS ENQ Shr Sys 651,445 0.055 MSEC *
* BCS ENQ Excl Sys 302 0.062 MSEC *
* BCS DEQ 1,051K 0.031 MSEC *
* VVDS RESERVE CI 294,503 0.038 MSEC *
* VVDS DEQ CI 294,503 0.042 MSEC *
* VVDS RESERVE Shr 1,482K 0.045 MSEC *
* VVDS RESERVE Excl 113 0.108 MSEC *
* VVDS DEQ 1,482K 0.040 MSEC *
* SPHERE ENQ Excl Sys 49 0.045 MSEC *
* SPHERE DEQ 49 0.033 MSEC *
* CAXWA ENQ Shr 144 0.006 MSEC *
* CAXWA DEQ 144 0.531 MSEC *
* VDSPM ENQ 651,816 0.005 MSEC *
* VDSPM DEQ 651,816 0.005 MSEC *
* BCS Get 63,848 0.095 MSEC *
* BCS Put 24 0.597 MSEC *
* BCS Erase 11 0.553 MSEC *
* VVDS I/O 1,769K 0.625 MSEC *
* VLF Delete Minor 1 0.019 MSEC *
* VLF Define Major 84 0.003 MSEC *
* VLF Identify 8,172 0.001 MSEC *
* RMM Tape Exit 24 0.000 MSEC *
* OEM Tape Exit 24 0.000 MSEC *
* BCS Allocate 142 15.751 MSEC *
* SMF Write 106,367 0.043 MSEC *
* IXLCONN 2 107.868 MSEC *
* IXLCACHE Read 2 0.035 MSEC *
* MVS Allocate 116 19.159 MSEC *
* Capture UCB 39 0.008 MSEC *
* SMS Active Config 2 0.448 MSEC *
* RACROUTE Auth 24,793 0.080 MSEC *
* RACROUTE Define 7 0.066 MSEC *
* Obtain QuiesceLatch 606,919 0.001 MSEC *
* ENQ SYSZPCCB 27,980 0.005 MSEC *
* DEQ SYSZPCCB 27,980 0.003 MSEC *
* Release QuiesceLatch 606,919 0.000 MSEC *
* Capture to Actual 149 0.014 MSEC *
*CAS***************************************************
IEC352I CATALOG ADDRESS SPACE MODIFY COMMAND COMPLETED
Figure 6-34 Catalog performance report

6.21 Catalog address space (CAS)
User CATALOG CATALOG User

Catalog request Catalog request
Address Address Address Address
Space SVC 26 Space Space SVC 26 Space
(CAS) (CAS)
Return data Return data
System 1 System 2
User catalog
Figure 6-35 Catalog address space (CAS)
The catalog address space

Catalog functions are performed in the catalog address space (CAS). The jobname of the
catalog address space is CATALOG.
As soon as a user requests a catalog function (for example, to locate or define a data set), the
CAS gets control to handle the request. When it has finished, it returns the requested data to
the user. A catalog task which handles a single user request is called a service task. To each
user request a service task is assigned. The minimum number of available service tasks is
specified in the SYSCATxx member of SYS1.NUCLEUS (or the LOADxx member of
SYS1.PARMLIB). A table called the CRT keeps track of these service tasks.
The CAS contains all information necessary to handle a catalog request, like control block
information about all open catalogs, alias tables, and buffered BCS records.
During the initialization of an MVS system, all user catalog names identified in the master
catalog, their aliases, and their associated volume serial numbers are placed in tables in
CAS.
You can use the MODIFY CATALOG operator command to work with the catalog address space.
See also 6.22, Working with the catalog address space on page 371.
Since z/OS 1.8, the maximum number of parallel catalog requests is 999, as defined in the
SYSCAT parmlib member. Previously it was 180.

Restarting the catalog address space
Consider restarting the CAS only as a final option before IPLing a system. Never try restarting
CAS unless an IPL is your only other option. A system failure caused by catalogs, or a CAS
storage shortage due to FREEMAIN failures, might require you to use MODIFY
CATALOG,RESTART to restart CAS in a new address space.
Never use RESTART to refresh catalog or VVDS control blocks or to change catalog
characteristics. Restarting CAS is a drastic procedure, and if CAS cannot restart, you will
have to IPL the system.
When you issue MODIFY CATALOG,RESTART, the CAS mother task is abended with abend code
81A, and any catalog requests in process at the time are redriven.
The restart of CAS in a new address space is transparent to all users. However, even when all
requests are redriven successfully and receive a return code of zero (0), the system might
produce indicative dumps. There is no way to suppress these indicative dumps.
Since z/OS 1.6, the F CATALOG has new options:

TAKEDUMP This option causes the CAS to issue an SVCDUMP using the proper
options to ensure that all data needed for diagnosis is available.
RESTART This option prompts the operator for additional information with the
following messages:
IEC363D IS THIS RESTART RELATED TO AN EXISTING CATALOG PROBLEM
(Y OR N)?
If the response to message IEC363D is N, the restart continues; if the
response is Y, another prompt is issued.
IEC364D HAS AN SVC DUMP OF THE CATALOG ADDRESS SPACE ALREADY
BEEN TAKEN (Y OR N)?

6.22 Working with the catalog address space
Use the MODIFY CATALOG command

To list information such as:
Settings
Catalogs currently open in the CAS
Performance statistics
Cache statistics
Service tasks
Module information
To interact with the CAS by:
Changing settings
Ending or abending service tasks
Restarting the CAS
Closing or unallocating catalogs
Figure 6-36 Working with the CAS
Working with the catalog address space

You can use the command MODIFY CATALOG to extract information from the CAS and to interact
with the CAS. This command can be used in many variations. In this section we provide an
overview of parameters to be aware of when maintaining your catalog environment. This
command is further discussed in 6.23, Fixing temporary catalog problems on page 373.
For a discussion about the entire functionality of the MODIFY CATALOG command, see z/OS
DFSMS: Managing Catalogs, SC26-7409.
Examples of the MODIFY CATALOG command:

MODIFY CATALOG,REPORT
The catalog report lists information about the CAS, like the service level, the catalog
address space ID, service tasks limits, and more. Since z/OS 1.7, this command
generates the IEC392I that identifies the top three holders of the CATALOG service tasks,
maybe indicating a lockup.
MODIFY CATALOG,OPEN
This command lists all catalogs that are currently open in the CAS. It shows whether the
catalog is locked or shared, and the type of buffer used for the catalog. A catalog is
opened after an IPL or catalog restart when it is referenced for the first time. It remains
open until it is manually closed by the MODIFY CATALOG command or it is closed because
the maximum number of open catalogs has been reached.

MODIFY CATALOG,LIST
The LIST command shows you all service task that are currently active handling a user
request. Normally the active phase of a service task is only of a short duration, so no tasks
are listed. This command helps you to identify tasks that can be the reason for catalog
problems if performance slows down.
MODIFY CATALOG,DUMPON(rc,rsn,mm,cnt)
Use this command to get a dump of the catalog address space when a specific catalog
error occurs. This catalog error is identified by the catalog return code (rc), the reason
code (rsn) and the name of the module which sets the error (mm). You can also specify, in
the cnt-parameter, for how many of the rc/rsn-combinations a dump is to be taken. The
default is one. The module identifier corresponds to the last two characters in the catalog
module name. For example, the module identifier is A3 for IGG0CLA3. You can substitute
the module name by two asterisks (**) if you do not know the module name or do not care
about it.
MODIFY CATALOG,ENTRY(modulename)
This command lists the storage address, fmid, and the level of a catalog module (csect). If
you do not specify a module name, all catalog csects are listed.
MODIFY CATALOG,REPORT,PERFORMANCE
The output of this command shows you significant catalog performance numbers about
number and duration of ENQs and DEQs for the BCS and VVDS and more. Use the
command to identify performance problems.
MODIFY CATALOG,REPORT,CACHE
The cache report lists cache type and statistics for the single catalogs that are open in the
CAS.
MODIFY CATALOG,RESTART
Use this command to restart the catalog address space. Take this action only in error
situations when the other option is an IPL. See also 6.21, Catalog address space (CAS)
on page 369, for more information about this topic.

6.23 Fixing temporary catalog problems
Fixing temporary catalog problems involves

Rebuilding information of BCS or VVDS control
blocks
Close or unallocate a BCS
Close or unallocate a VVDS
Determining tasks that cause performance problems

Display GRS information about catalog resources
and devices
Display information about catalog service tasks
End service tasks
Figure 6-37 Fixing temporary catalog problems
Fixing temporary catalog problems

This section explains how to rebuild information in the catalog address space, as well as how
to obtain information about and recover from performance slowdowns.
Rebuild information about a BCS or VVDS

Occasionally, the control blocks for a catalog kept in the catalog address space might be
damaged. You might think the catalog is damaged and in need of recovery, when only the
control blocks need to be rebuilt. If the catalog appears damaged, try rebuilding the control
blocks first. If the problem persists, recover the catalog.
Use the following commands to close or unallocate a BCS or VVDS in the catalog address
space. The next access to the BCS or VVDS reopens it and rebuilds the control blocks.
MODIFY CATALOG,CLOSE(catalogname) - Closes the specified catalog but leaves it
allocated.
MODIFY CATALOG,UNALLOCATE(catalogname) - Unallocates a catalog; if you do not specify a
catalog name, then all catalogs are unallocated.
MODIFY CATALOG,VCLOSE(volser) - Closes the VVDS for the specified volser.
MODIFY CATALOG,VUNALLOCATE - Unallocates all VVDSs; you cannot specify a volser, so try
to use VCLOSE first.

Recover from performance slow downs
Sometimes the performance of the catalog address space slows down. Catalog requests can
take a long time or even hang. There can be various reasons for such situations, for example
a volume on which a catalog resides is reserved by another system, thus all requests from
your system to this catalog wait until the volume is released.
Generally, the catalog component uses two resources for serialization:

SYSIGGV2 to serialize on the BCS
SYSZVVDS to serialize on the VVDS
Delays or hangs can occur if the catalog needs one of these resources and it is held already
by someone else, for example by a CAS of another system. You can use the following
commands to display global resource serialization (GRS) data:
D GRS,C - Displays GRS contention data for all resources, who is holding a resource, and
who is waiting.
D GRS,RES=(resourcename) - Displays information for a specific resource.
D GRS,DEV=devicenumber - Displays information about a specific device, such as whether it
is reserved by the system.
Route these commands to all systems in the sysplex to get an overview about hang
situations.
When you have identified a catalog address space holding a resource for a long time, or the
GRS outputs do not show you anything but you have still catalog problems, you can use the
following command to get detailed information about the catalog services task:
MODIFY CATALOG,LIST - Lists the currently active service tasks, their task IDs, duration, and
the job name for which the task is handling the request.
Watch for tasks with long duration time. You can obtain detailed information about a specific
task by running the following command for a specific task ID:
MODIFY CATALOG,LISTJ(taskid),DETAIL - Shows detailed information about a service task,
for example if it is waiting for the completion of an ENQ.
If you identify a long-running task that is in a deadlock situation with another task (on another
system), you can end and redrive the task to resolve the lockout. The following commands
help you to end a catalog service task:
MODIFY CATALOG,END(taskid),REDRIVE - End a service task and redrive it.
MODIFY CATALOG,END(taskid),NOREDRIVE - Permanently end the task without redriving.
MODIFY CATALOG,ABEND(taskid) - Abnormally end a task which cannot be stopped by
using the END parameter.
You can use the FORCE parameter for these commands if the address space that the service
task is operating on behalf of has ended abnormally. Use this parameter only in this case.
You can also try to end the job for which the catalog task is processing a request.
For more information about the MODIFY CATALOG command and fixing temporary catalog
problems, see z/OS DFSMS: Managing Catalogs, SC26-7409.

6.24 Enhanced catalog sharing
Conditions for ECS:
ECSHARING in IDCAMS DEFINE/ALTER

MVS 1
Active connection to ECS CF structure
MVS 2
Activate ECS through MVS command:
MODIFY CATALOG,ECSHR(AUTOADD)
VVR VVR
Coupling
CATALOG
Facility
Figure 6-38 DFSMS enhanced catalog sharing
Enhanced catalog sharing (ECS)

In 6.9, Sharing catalogs across systems on page 343, sharing catalogs across multiple
systems is discussed. A catalog uses its specific VVR to serialize the access from multiple
systems. Reading serialization information from the VVR on the DASD causes a significant
amount of overhead for I/O operations. However, it is better to have this VVR than to discard
the catalog buffers, when in shared catalog mode. This is called VVDS mode.
Most of the overhead associated with shared catalog is eliminated if you use enhanced
catalog sharing (ECS). ECS uses a cache Coupling Facility structure to keep the special
VVR. In addition, the Coupling Facility structure (as defined in CFRM) keeps a copy of
updated records. There is no I/O necessary to read the catalog VVR to verify the updates. In
addition, the eventual modifications are also kept in the Coupling Facility structure, thereby
avoiding more I/O. ECS saves about 50% in elapsed time and provides an enormous
reduction in ENQ/Reserves.
Implementing enhanced catalog sharing

Perform these steps to implement ECS:
1. Define a Coupling Facility cache structure with the name SYSIGGCAS_ECS in the CFRM
couple data set and activate this CFRM policy.
This action connects all ECS-eligible systems to the ECS structure.

2. Define or alter your existing catalogs with the attribute ECSHARING using the IDCAMS
DEFINE or ALTER commands.
The ECSHARING attribute makes a catalog eligible for sharing using the ECS protocol,
as opposed to the VVDS/VVR protocol. The catalog can still be used in VVDS mode.
3. Activate all eligible catalogs for ECS sharing.
You can use the command MODIFY CATALOG,ECSHR(AUTOADD) on one system to activate all
ECS-eligible catalogs throughout the sysplex for ECS sharing. They are automatically
activated at their next reference. You can manually add a catalog to use ECS by running
the command MODIFY CATALOG,ECSHR(ENABLE,catname) where catname is the name of the
catalog you want to add.
Only catalogs that were added are shared in ECS mode. The command MODIFY
CATALOG,ECSHR(STATUS) shows you the ECS status for each catalog, as well as whether it is
eligible and already activated.
Restrictions for ECS mode usage

The following restrictions apply to ECS mode usage:
You cannot use ECS mode from one system and VVDS mode from another system
simultaneously to share a catalog. You will get an error message if you try this.
Important: If you attempt to use a catalog that is currently ECS-active from a system
outside the sysplex, the request might break the catalog.
No more than 1024 catalogs can currently be shared using ECS from a single system.
All systems sharing the catalog in ECS mode must have connectivity to the same
Coupling Facility, and must be in the same global resource serialization (GRS) complex.
When you use catalogs in ECS mode, convert the resource SYSIGGV2 to a SYSTEMS
enqueue. Otherwise, the catalogs in ECS mode will be damaged.
For more information about ECS, see z/OS DFSMS: Managing Catalogs, SC26-7409. For
information about defining Coupling Facility structures, see z/OS MVS Setting Up a Sysplex,
SA22-7625.

7
Chapter 7. DFSMS Transactional VSAM

Services
DFSMS Transactional VSAM Services (DFSMStvs) is an enhancement to VSAM RLS access
that enables multiple batch update jobs and CICS to share access to the same data sets.
DFSMStvs provides two-phase commit and backout protocols, as well as backout logging and
forward recovery logging. DFSMStvs provides transactional recovery directly within VSAM.
As an extension of VSAM RLS, DFSMStvm enables any job or application that is designed for
data sharing to read-share or write-share VSAM recoverable data sets. VSAM RLS provides
a server for sharing VSAM data sets in a sysplex. VSAM RLS uses Coupling Facility-based
locking and data caching to provide sysplex-scope locking and data access integrity.
DFSMStvs adds logging, commit, and backout processing.
To understand DFSMStvs, it is necessary to first review base VSAM information and VSAM
record-level sharing (RLS).
In this chapter we cover the following topics:

Review of base VSAM information and CICS concepts
Introduction to VSAM RLS
Introduction to DFSMStvs

7.1 VSAM share options
Share options are an attribute of the data set

SHAREOPTIONS(crossregion,crosssystem)
SHAREOPTIONS(1,x)
One user can have the data set open for read/write
access or any number of users for read only
SHAREOPTIONS(2,x)
One user can have the data set open for read/write
access and any number of users for read only
SHAREOPTIONS(3,x)
Any number of users can have the data set open for
both read and write access
Figure 7-1 VSAM share options
Share options as a data set attribute

The share options of a VSAM data set are specified as a parameter for an IDCAMS DEFINE
CLUSTER that creates a new data set. They show how a component or cluster can be shared
among users.
SHAREOPTIONS (crossregion,crosssystem)
The cross-region share options specify the amount of sharing allowed among regions within
the same system or multiple systems. Cross-system share options specify how the data set is
shared among systems. Use global resource serialization (GRS) or a similar product to
perform the serialization.
SHAREOPTIONS (1,x)
The data set can be shared by any number of users for read access (open for input), or it can
be accessed by only one user for read/write access (open for output). If the data set is open
for output by one user, a read or read/write request by another user will fail. With this option,
VSAM ensures complete data integrity for the data set. When the data set is already open for
RLS processing, any request to open the data set for non-RLS access will fail.
SHAREOPTIONS (2,x)
The data set can be shared by one user for read/write access, and by any number of users for
read access. If the data set is open for output by one user, another open for output request
will fail, but a request for read access will succeed. With this option, VSAM ensures write

integrity. If the data set is open for RLS processing, non-RLS access for read is allowed.
VSAM provides full read and write integrity for its RLS users, but no read integrity for non-RLS
access.
SHAREOPTIONS (3,x)
The data set can be opened by any number of users for read and write request. VSAM does
not ensure any data integrity. It is the responsibility of the users to maintain data integrity by
using enqueue and dequeue macros. This setting does not allow any type of non-RLS access
while the data set is open for RLS processing.
For more information about VSAM share options, see z/OS DFSMS: Using Data Sets,
SC26-7410.
Chapter 7. DFSMS Transactional VSAM Services 379

7.2 Base VSAM buffering
Three types of base VSAM buffering
NSR - non-shared resources
LSR - local shared resources
GSR - global shared resources
Specified in the VSAM ACB macro:

MACRF=(NSR/LSR/GSR)
Figure 7-2 Base VSAM buffering
Three types of base VSAM buffering

Before VSAM RLS there were only three buffering techniques available for the user that
opens the data set.
NSR - non-shared resources

For NSR, data buffers belong to a particular request parameter list (RPL). When an
application uses an RPL (for example, for a direct GET request), VSAM manages the buffers
as follows:
1. For record 1000, VSAM locates the CI containing the record and reads it into the local
buffer in private storage.
2. If the next GET request is for record 5000, which is in a separate CI from record 1000,
VSAM overwrites the buffer with the new CI.
3. Another GET request for record 1001, which is in the same CI as record 1000, causes
another I/O request for the CI to read it into the buffer, because it had been overlaid by the
second request for record 5000.

LSR - local shared resources
For LSR, data buffers are managed so that the buffers will not be overlaid. Before opening the
data set, the user builds a resource pool in private storage using the BLDVRP macro. The
LSR processing is as follows:
1. A GET request for record 1000 reads in the CI containing the record in one buffer of the
buffer pool.
2. The next GET request for record 5000, which is in a separate CI, will read in the CI in
another, separate buffer in the same buffer pool.
3. A third GET request for record 1001 can be satisfied by using the CI that was read in for
the request for record 1000.
GSR - global shared resources

GSR is the same concept as LSR. The only difference is, with GSR the buffer pool and VSAM
control blocks are built in common storage and can be accessed by any address space in the
system.
For more information about VSAM buffering techniques refer to 4.44, VSAM: Buffering
modes on page 177.
MACRF=(NSR/LSR/GSR)
The Access Method Control block (ACB) describes an open VSAM data set. A subparameter
for the ACB macro is MACRF, in which you can specify the buffering technique to be used by
VSAM. For LSR and GSR, you need to run the BLDVRP macro before opening the data set to
create the resource pool.
For information about VSAM macros, see z/OS DFSMS: Macro Instructions for Data Sets,
SC26-7408.

7.3 Base VSAM locking
Example of base VSAM LSR serialization

scope = single LSR buffer pool
granularity = control interval
ownership = RPL
GET UPD RPL_1 GET UPD RPL_2
(record B) (Record E)
succeeds - locks fails - exclusive

the CI to update control conflict
the record Record A
Record B
Record C
Record D control
interval (CI)
Record E
Record F
Record G
Figure 7-3 Example of LSR serialization
Serialization of base VSAM

Base VSAM serializes on a CI level. Multiple users attempting to access the same CI to read
different records either defer on the CI or are returned an exclusive control conflict error by
VSAM. Large CIs with many records per CI, or applications that repeatedly access the same
CI, can have a performance impact due to retrying of exclusive control conflict errors or
waiting on a deferred request.
Example of VSAM LSR serialization

In the example in Figure 7-3, two RPLs need to read and write different records in the same
CI. RPL_1 obtains access to the buffer containing the CI first and locks it for update
processing against Record B. RPL_2 fails with an exclusive control conflict error on this CI
although it needs to update a different record.

7.4 CICS function shipping before VSAM RLS
CICS CICS CICS

AOR FOR AOR
CICS CICS
AOR AOR
CICS CICS
AOR AOR
VSAM
CICS CICS
AOR AOR
System 1 System n
AOR = Application Owning Region

FOR = File Owning Region
Figure 7-4 CICS function shipping before VSAM RLS
CICS function shipping

Prior to VSAM RLS, a customer information control system (CICS) VSAM data set was
owned and directly accessed by one single CICS. Shared access across CICS Application
Owning Regions (AORs) to a single VSAM data set was provided by CICS function shipping.
With function shipping, one CICS File Owning Region (FOR) accesses the VSAM data sets
on behalf of other CICS regions (see Figure 7-4).
Problems
There are a couple of problems with this kind of CICS configuration:
CICS FOR is a single point of failure.
Multiple system performance is not acceptable.
Lack of scalability.
Over time the FORs became a bottleneck because CICS environments became increasingly
complex. CICS required a solution to have direct shared access to VSAM data sets from
multiple CICSs.

7.5 VSAM record-level sharing introduction
CICS VSAM RLS VSAM RLS CICS

AOR instance 1 instance n AOR
CICS CICS
AOR AOR
CICS CICS
AOR AOR
coupling facility
CICS CICS
AOR AOR
System 1 System n
Figure 7-5 Parallel Sysplex CICS with VSAM RLS
Why VSAM record-level sharing (RLS)

The solution for the problems inherent in CICS function shipping is VSAM record-level
sharing. This is a major extension to base VSAM, and although it was designed for use by
CICS, it can be used by any application. It provides an environment where multiple CICSs can
directly access a shared VSAM data set (Figure 7-5).
VSAM record-level sharing (RLS) is a method of access to your existing VSAM files that
provides full read and write integrity at the record level to any number of users in your Parallel
Sysplex.
Benefits of VSAM RLS

The benefits of VSAM RLS are:
Enhances cross-system data sharing - scope is sysplex
Improves performance and availability in CICS and also non-CICS VSAM environments
Provides data protection after a system failure
Provides automation for data recovery
Provides full read/write integrity to your existing VSAM files; the user does not need to
serialize using ENQ/DEQ macros
Allows CICS to register as a recoverable subsystem, which will automate recovery
processing as well as protect the data records to be recovered

7.6 VSAM RLS overview
VSAM RLS enables multiple address spaces on

multiple systems to access recoverable VSAM data
sets at the same time
VSAM RLS involves support from multiple products
Level of sharing is determined by whether the data
set is recoverable or not
Coupling Facility is used for sharing
Supported data set types:
Variable length relative-record data set (VRRDS)
Figure 7-6 VSAM RLS overview
Multiple access on recoverable data sets

VSAM RLS is a data set access mode that enables multiple address spaces, CICS
application-owning regions on multiple systems, and batch jobs to access recoverable VSAM
data sets at the same time.
With VSAM RLS, multiple CICS systems can directly access a shared VSAM data set,
eliminating the need to ship functions between the application-owning regions and file-owning
regions. CICS provides the logging, commit, and backout functions for VSAM recoverable
data sets. VSAM RLS provides record-level serialization and cross-system caching. CICSVR
provides a forward recovery utility.
Multiple products involved

VSAM RLS processing involves support from multiple products:
CICS Transaction Server
CICS VSAM Recovery (CICSVR)
DFSMS
Level of sharing
The level of sharing that is allowed between applications is determined by whether or not a
data set is recoverable; for example:
Both CICS and non-CICS jobs can have concurrent read or write access to
nonrecoverable data sets. There is no coordination between CICS and non-CICS, so data
integrity can be compromised.

Non-CICS jobs can have read-only access to recoverable data sets concurrently with
CICS jobs, which can have read or write access.
Coupling Facility overview

The Coupling Facility (CF) is a shareable storage medium. It is licensed internal code (LIC)
running in a special type of PR/SM logical partition (LPAR) in certain zSeries and S/390
processors. It can be shared by the systems in one sysplex only. A CF makes data sharing
possible by allowing data to be accessed throughout a sysplex with assurance that the data
will not be corrupted and that the data will be consistent among all sharing users.
VSAM RLS uses a Coupling Facility to perform data-set-level locking, record locking, and
data caching. VSAM RLS uses the conditional write and cross-invalidate functions of the
Coupling Facility cache structure, thereby avoiding the need for control interval (CI) level
locking.
VSAM RLS uses the Coupling Facility caches as store-through caches. When a control
interval of data is written, it is written to both the Coupling Facility cache and the direct access
storage device (DASD). This ensures that problems occurring with a Coupling Facility cache
do not result in the loss of VSAM data.
Supported data set types

VSAM RLS supports access to these types of data sets:
Variable-length relative-record data set cluster (VRRDS)
VSAM RLS also supports access to a data set through an alternate index, but it does not
support opening an alternate index directly in RLS mode. Also, VSAM RLS does not support
access through an alternate index to data stored under z/OS UNIX System Services.
Extended format, extended addressability, and spanned data sets are supported with VSAM
RLS. Compression is also supported.
VSAM RLS does not support:

Linear data sets (LDS)
Keyrange data sets
KSDS with an imbedded index (defined with IMBED option)
Temporary data sets
Striped data sets
Catalogs and VVDSs
Keyrange data sets and the IMBED attribute for a KSDS are obsolete. You cannot define new
data sets as keyrange or with an imbedded index anymore. However, there still might be old
data sets with these attributes in your installation.

7.7 Data set sharing under VSAM RLS
Share options are largely ignored under VSAM RLS

Exception: SHAREOPTIONS(2,x)
One user can have the data set open for non-RLS
read/write access and any number of users for
non-RLS read
Or any number of users can have the data set open
for RLS read/write and any number of users for
non-RLS read
Non-CICS access for data sets open by CICS in RLS
mode
Allowed for non-recoverable data sets
Not allowed for recoverable data sets
Figure 7-7 Data set sharing under VSAM RLS
Share options are largely ignored under VSAM RLS

The VSAM share options specification applies only when non-RLS access like NSR, LSR, or
GSR is used. They are ignored when the data set is open for RLS access. Record-level
sharing always assumes multiple readers and writers to the data set. VSAM RLS ensures full
data integrity. When a data set is open for RLS access, non-RLS requests to open the data
set fail.
Exception: SHAREOPTIONS(2,x)
For non-RLS access, SHAREOPTIONS(2,x) are handled as always. One user can have the
data set open for read/write access and multiple users can have it open for read access only.
VSAM does not provide data integrity for the readers.
If the data set is open for RLS access, non-RLS opens for read are possible. These are the
only share options, where a non-RLS request to open the data set will not fail if the data set is
already open for RLS processing. VSAM does not provide data integrity for the non-RLS
readers.
Non-CICS access
RLS access from batch jobs to data sets that are open by CICS depends on whether the data
set is recoverable or not. For recoverable data sets, non-CICS access from other applications
(that do not act as recoverable resource manager) is not allowed.
See 7.10, VSAM RLS/CICS data set recovery on page 392 for details.

7.8 Buffering under VSAM RLS
CICS VSAM RLS VSAM RLS CICS

R/W aka aka R/W
SMSVSAM SMSVSAM
CICS CICS
R/W R/W
1 3 1 4
Batch 1 3 4 Batch
R/O R/O
coupling facility
SMSVSAM SMSVSAM
data space data space
1 2 3 4
System 1 System 2
VSAM data set
MACRF=RLS
Figure 7-8 Buffering under VSAM RLS
New VSAM buffering technique MACRF=RLS

We have already discussed NSR, LSR, and GSR (refer to 7.2, Base VSAM buffering on
page 380). RLS is another method of buffering which you can specify in the MACRF
parameter of the ACB macro. RLS and NSR/LSR/GSR are mutually exclusive.
Bufferpools in the data space

Unlike NSR, LSR, or GSR, the VSAM buffers reside in a data space and not in private or
global storage of a user address space. Each image in the sysplex has one large local buffer
pool in the data space.
The first request for a record after data set open for RLS processing will cause an I/O
operation to read in the CI that contains this record. A copy of the CI is stored into the cache
structure of the Coupling Facility and in the buffer pool in the data space.
Buffer coherency
Buffer coherency is maintained through the use of Coupling Facility (CF) cache structures
and the XCF cross-invalidation function. For the example in Figure 7-8, that means:
1. System 1 opens the VSAM data set for read/write processing.
2. System 1 reads in CI1 and CI3 from DASD; both CIs are stored in the cache structure in
the Coupling Facility.
3. System 2 opens the data set for read processing.

4. System 2 needs CI1 and CI4; CI1 is read from the CF cache, CI4 from DASD.
5. System 1 updates a record in CI1 and CI3; both copies of these CIs in the CF are
updated.
6. XCF notices the change of these two CIs and invalidates the copy of CI1 for System 2.
7. System 2 needs another record from CI1; it notices that its buffer was invalidated and
reads in a new copy of CI1 from the CF.
For further information about cross-invalidation, see z/OS MVS Programming: Sysplex
Services Guide, SA22-7617.
The VSAM RLS Coupling Facility structures are discussed in more detail in 7.14, Coupling
Facility structures for RLS sharing on page 397.

7.9 VSAM RLS locking
Example of VSAM RLS serialization

scope = sysplex
granularity = record
ownership = CICS transaction or batch job
CICS1.Tran1 CICS2.Tran2 CICS3.Tran3 CICS4.Tran4
GET UPD RPL_1 GET UPD RPL_2 GET CR RPL_3 GET NRI RPL_4
(Record B) (Record E) (Record B) (Record B)
Waits for
record lock
Record A
Record B
Record C Record B Record E
Holder Holder (EXCL)
control Record D (EXCL) CICS2.Tran2
interval (CI) CICS1.Tran1
Record E Waiter
Record F (SHARE)
CICS3.Tran3
Record G
VSAM RLS locks
Figure 7-9 Example of VSAM RLS serialization
VSAM RLS serialization

7.3, Base VSAM locking on page 382 presents an example of LSR serialization. The
granularity under LSR, NSR, or GSR is a control interval, whereas VSAM RLS serializes on
a record level. With VSAM RLS, it is possible to concurrently update separate records in the
same control interval. Record locks for UPDATE are always exclusive. Record locks for read
depend on the level of read integrity.
Levels of read integrity

There are three levels of read integrity, as explained here:
NRI (no read integrity)
This level tells VSAM not to obtain a record lock on the record accessed by a GET or
POINT request. This avoids the overhead of record locking. This is sometimes referred to
as a dirty read because the reader might see an uncommitted change made by another
transaction.
Even with this option specified, VSAM RLS still performs buffer validity checking and
refreshes the buffer when the buffer is invalid.
CR (consistent read)
This level tells VSAM to obtain a shared lock on the record that is accessed by a GET or
POINT request. It ensures that the reader does not see an uncommitted change made by
another transaction. Instead, the GET or POINT request waits for the change to be

committed or backed out. The request also waits for the exclusive lock on the record to be
released.
CRE (consistent read explicit)
This level has a meaning similar to that of CR, except that VSAM RLS holds the shared
lock on the record until the end of the unit of recovery, or unit of work. This option is only
available to CICS or DFSMStvs transactions. VSAM RLS does not understand
end-of-transaction for non-CICS or non-DFSMStvs usage.
The type of read integrity is specified either in the ACB macro or in the JCL DD statement:
ACB RLSREAD=NRI/CR/CRE
//dd1 DD dsn=datasetname,RLS=NRI/CR/CRE
Example situation
In our example in Figure 7-9 on page 390 we have the following situation:
1. CICS transaction Tran1 obtains an exclusive lock on Record B for update processing.
2. Transaction Tran2 obtains an exclusive lock for update processing on Record E, which is in
the same CI.
3. Transaction Tran3 needs a shared lock also on Record B for consistent read; it has to wait
until the exclusive lock by Tran1 is released.
4. Transaction Tran4 does a dirty read (NRI); it does not have to wait because in that case,
no lock is necessary.
With NRI, Tran4 can read the record even though it is held exclusively by Tran1. There is no
read integrity for Tran4.
CF lock structure
RLS locking is performed in the Coupling Facility through the use of a CF lock structure
(IGWLOCK00) and the XES locking services.
Contention
When contention occurs on a VSAM record, the request that encountered the contention
waits for the contention to be removed. The lock manager provides deadlock detection. When
a lock request is in deadlock, the request is rejected, resulting in the VSAM record
management request completing with a deadlock error response.

7.10 VSAM RLS/CICS data set recovery
Recoverable data sets

Defined as LOG(UNDO/ALL) in the catalog
UNDO - backout logging performed by CICS
ALL - both backout and forward recovery logging
LOG(ALL) data sets must have a
LOGSTREAMID(forwardrecoverylog) also defined in
the catalog
Non-recoverable data sets
Defined as LOG(NONE) in the catalog
No logging performed by CICS
Figure 7-10 Recoverable data sets
Recoverable data set

VSAM record-level sharing introduces a VSAM data set attribute called LOG. With this
attribute a data set can be defined as recoverable or non-recoverable. A data set whose log
parameter is undefined or NONE is considered non-recoverable. A data set whose log
parameter is UNDO or ALL is considered recoverable. For recoverable data sets, a log of
changed records is maintained to commit and back out transaction changes to a data set.
A data set is considered recoverable if the LOG attribute has one of the following values:
UNDO
The data set is backward recoverable. Changes made by a transaction that does not
succeed (no commit was done) are backed out. CICS provides the transactional recovery.
See also 7.11, Transactional recovery on page 394.
ALL
The data set is both backward and forward recoverable. In addition to the logging and
recovery functions provided for backout (transactional recovery), CICS records the image
of changes to the data set, after they were made. The forward recovery log records are
used by forward recovery programs and products such as CICS VSAM Recovery
(CICSVR) to reconstruct the data set in the event of hardware or software damage to the
data set. This is referred to as data set recovery. For LOG(ALL) data sets, both types of
recovery are provided, transactional recovery and data set recovery.

For LOG(ALL) you need to define a logstream in which changes to the data sets are
logged.
Non-recoverable data sets

A data set whose LOG parameter is undefined or NONE is considered as non-recoverable.
Non-CICS access to recoverable and non-recoverable data sets

VSAM RLS supports non-recoverable files. Non-recoverable means CICS does not do
transactional recovery (logging, commit, backout). VSAM RLS provides record locking and file
integrity across concurrently executing CICS and batch applications. Transactional recovery
is not provided. This is because VSAM RLS does not provide undo logging and two-phase
commit/backout support. Most transactions and batch jobs are not designed to use this form
of data sharing.
Non-CICS read/write access for recoverable data sets that are open by CICS is not allowed.
The recoverable attribute means that when the file is accessed in RLS mode, transactional
recovery is provided. With RLS, the recovery is only provided when the access is through
CICS file control, so RLS does not permit a batch (non-CICS) job to open a recoverable file
for OUTPUT.
Transactional recovery is described in 7.11, Transactional recovery on page 394.

7.11 Transactional recovery
Commited transaction Failed transaction (back out)

Trans1: Trans1:
Read Record 1 Read Record 1
(Lock Record 1) (Lock Record 1)
Record 1 Record 1' Record 1 Record 1'
(Log Record 1) (Log Record 1)
<Modify Record 1> <Modify Record 1>
Write Record 1' Write Record 1'
Read Record 2 ---------------------------------- Failure ------------------------------------
(Lock Record 2) Read Log Record 1
Record 2 Record 2' Record 1 Record 1
(Log Record 2) Re-Lock Record 1
<Modify Record 2> Write Record 1
Write Record 2' Commit
Commit (update log)
(update log) (release locks)
(release locks)
Figure 7-11 Transactional recovery
CICS transactional recovery for VSAM recoverable data sets

During the life of a transaction, its changes to recoverable resources are not seen by other
transactions. The exception is if you are using the no-read integrity (NRI) option. Then you
might see uncommitted changes.
Exclusive locks that VSAM RLS holds on the modified records cause other transactions that
have read-with-integrity requests and write requests for these records to wait. After the
modifying transaction is committed or backed out, VSAM RLS releases the locks and the
other transactions can access the records.
If the transaction fails, its changes are backed out. This capability is called transactional
recovery.
The CICS backout function removes changes made to the recoverable data sets by a
transaction. When a transaction abnormally ends, CICS performs a backout implicitly.
Example
In our example in 7.11, Transactional recovery on page 394, transaction Trans1 is complete
(committed) after Record 1 and Record 2 are updated. Transactional recovery ensures that
either both changes are made or no change is made. When the application requests commit,
both changes are made atomically. In the case of an failure after updating Record 1, the
change to this record is backed out. This applies only for recoverable data sets, not for
non-recoverable ones.

7.12 The batch window problem
Batch window - a period of time in which CICS

access to recoverable data sets is quiesced so
batch jobs can run
Requires taking a backup of the data set
Batch updates are then performed
A forward recovery backup is taken, if needed
When finished, CICS access to the data set is
re-enabled
Figure 7-12 Batch window problem
Batch window
The batch window is a period of time in which online access to recoverable data sets must be
disabled. During this time, no transaction processing can be done. This is normally done
because it is necessary to run batch jobs or other utilities that do not properly support
recoverable data, even if those utilities use also RLS access. Therefore, to allow these jobs or
utilities to safely update the data, it is first necessary to make a copy of the data. In the event
that the batch job or utility fails or encounters an error, this copy can be safely restored and
online access can be re-enabled. If the batch job completes successfully, the updated copy of
the data set can be safely used because only the batch job had access to the data while it
was being updated. Therefore, the data cannot have been corrupted by interference from
online transaction processing.
Quiescing a data set from RLS processing

Before updating a recoverable data set in non-RLS mode, quiesce the data set around the
sysplex. This is to ensure that no RLS access can be done while non-RLS applications are
updating those data sets. The quiesced state is stored in the ICF catalog. After a quiesce has
completed, all CICS files associated with the data set are closed. A quiesced data set can be
opened in non-RLS mode only if no retained locks are presented. Once the data set was
quiesced from RLS processing, it can be opened again in RLS mode only after it is
unquiesced.
See 7.20, Interacting with VSAM RLS on page 412 for information about how to quiesce and
unquiesce a data set.

7.13 VSAM RLS implementation
Update CFRM policy to define lock and cache

structures
Update SYS1.PARMLIB(IGDSMSxx) with RLS
parameters
Define sharing control data sets (SHCDSs)
Update SMS configuration for cache sets
Update data sets with LOG(NONE/UNDO/ALL) and
LOGSTREAMID
Figure 7-13 VSAM RLS configuration changes
VSAM RLS configuration changes

There are a few configuration changes necessary to run VSAM RLS. They include:
Update the Coupling Facility resource manager (CFRM) policy to define lock and cache
structures.
Update SYS1.PARMLIB(IGDSMSxx) with VSAM RLS parameters.
Define new sharing control data sets (SHCDSs).
Update SMS configuration for cache sets and assign them to a storage group.
Update data sets with the attribute LOG(NONE/UNDO/ALL) and optionally assign a
LOGSTREAMID.

7.14 Coupling Facility structures for RLS sharing
Two types of Coupling Facility structures are

needed by VSAM RLS
Lock structure
Maintain record locks and other DFSMSdfp
serializations
Enforce protocol restrictions for VSAM RLS data
sets
Required structure name: IGWLOCK00
Cache structures
Multiple cache structures are possible, at least one
is necessary
Provide level of storage between DASD and local
memory
Figure 7-14 VSAM RLS Coupling Facility structures
Structures in a Coupling Facility

A CF stores information in structures. z/OS recognizes three structure types: cache, list, and
lock. It provides a specific set of services for each of the structure types to allow the
manipulation of data within the structure. VSAM RLS needs list and lock structures for data
sharing and high-speed serialization.
Lock structure
In a Parallel Sysplex, you need only one lock structure for VSAM RLS because only one
VSAM sharing group is permitted. The required name is IGWLOCK00.
The lock structure is used to:

Enforce the protocol restrictions for VSAM RLS data sets
Maintain the record-level locks and other DFSMSdfp serializations
Ensure that the Coupling Facility lock structure has universal connectivity so that it is
accessible from all systems in the Parallel Sysplex that support VSAM RLS.
Tip: For high-availability environments, use a nonvolatile Coupling Facility for the lock
structure. If you maintain the lock structure in a volatile Coupling Facility, a power outage
can cause a failure and loss of information in the Coupling Facility lock structure.

Cache structures
Coupling Facility cache structures provide a level of storage hierarchy between local memory
and DASD cache.
They are also used as system buffer pool with cross-invalidation being done (see 7.8,
Buffering under VSAM RLS on page 388).
Each Coupling Facility cache structure is contained in a single Coupling Facility. You may
have multiple Coupling Facilities and multiple cache structures.
The minimum size of the cache structure is 10 MB.
Sizing the lock and cache structure

For information about sizing the CF lock and cache structure for VSAM RLS, see:
z/OS DFSMStvs Planning and Operation Guide, SC26-7348
z/OS DFSMSdfp Storage Administration Reference, SC26-7402
A sizing tool known as CFSIZER is also available on the IBM Web site at:
http://www-1.ibm.com/servers/eserver/zseries/cfsizer/vsamrls.html
Defining Coupling Facility structures

Use CFRM policy definitions to specify an initial and maximum size for each Coupling Facility
structure. DFSMS uses the initial structure size you specify in the policy each time it connects
to a Coupling Facility cache structure.
//STEP10 EXEC PGM=IXCMIAPU

//SYSABEND DD SYSOUT=A
//SYSIN DD *
DATA TYPE(CFRM) REPORT(YES)
DEFINE POLICY NAME(CFRM01) REPLACE(YES)
STRUCTURE NAME(CACHE01)
SIZE(70000)
INITSIZE(50000)
PREFLIST(CF01,CF02)
STRUCTURE NAME(CACHE02)
SIZE(70000)
INITSIZE(50000)
PREFLIST(CF01,CF02)
STRUCTURE NAME(IGWLOCK00)
SIZE(30000)
INITSIZE(15000)
PREFLIST(CF01,CF02)
/*
Figure 7-15 Example of defining VSAM RLS CF structure
Displaying Coupling Facility structures

You can use the system command DISPLAY XCF to view your currently defined Coupling
Facility structures. An example of XCF display of the lock structure is in Figure 7-16 on
page 399.

D XCF,STR,STRNAME=IGWLOCK00
IXC360I 10.00.38 DISPLAY XCF 337
STRNAME: IGWLOCK00
STATUS: ALLOCATED
TYPE: LOCK
POLICY INFORMATION:
POLICY SIZE : 28600 K
POLICY INITSIZE: 14300 K
POLICY MINSIZE : 0 K
FULLTHRESHOLD : 80
ALLOWAUTOALT : NO
REBUILD PERCENT: 75
DUPLEX : ALLOWED
PREFERENCE LIST: CF1 CF2
ENFORCEORDER : NO
EXCLUSION LIST IS EMPTY
ACTIVE STRUCTURE
----------------
ALLOCATION TIME: 02/24/2005 14:22:56
CFNAME : CF1
COUPLING FACILITY: 002084.IBM.02.000000026A3A
PARTITION: 1F CPCID: 00
ACTUAL SIZE : 14336 K
STORAGE INCREMENT SIZE: 256 K
ENTRIES: IN-USE: 0 TOTAL: 33331, 0% FULL
LOCKS: TOTAL: 2097152
PHYSICAL VERSION: BC9F02FD EDC963AC
LOGICAL VERSION: BC9F02FD EDC963AC
SYSTEM-MANAGED PROCESS LEVEL: 8
XCF GRPNAME : IXCLO001
DISPOSITION : KEEP
ACCESS TIME : 0
NUMBER OF RECORD DATA LISTS PER CONNECTION: 16
MAX CONNECTIONS: 4
# CONNECTIONS : 4
CONNECTION NAME ID VERSION SYSNAME JOBNAME ASID STATE

---------------- -- -------- -------- -------- ---- ----------------
SC63 01 000100B0 SC63 SMSVSAM 0009 ACTIVE
SC64 02 000200C6 SC64 SMSVSAM 000A ACTIVE
SC65 03 000300DD SC65 SMSVSAM 000A ACTIVE
SC70 04 00040035 SC70 SMSVSAM 000A ACTIVE
Figure 7-16 Example of XCF display of structure IGWLOCK00

7.15 Update PARMLIB with VSAM RLS parameters
PARMLIB parameters to support VSAM RLS

RLSINIT
CF_TIME
DEADLOCK_DETECTION
RLS_MaxCfFeatureLevel
RLS_MAX_POOL_SIZE
SMF_TIME
Figure 7-17 PARMLIB parameters to support VSAM RLS
New PARMLIB parameters to support VSAM RLS

The SYS1.PARMLIB member IGDSMSxx includes several parameters that support the
Coupling Facility. With the exception of RLSINIT, these parameters apply across all systems
in the Parallel Sysplex. The parameter values specified for the first system that was activated
in the sysplex are used by all other systems in the sysplex.
The following IGDSMSxx parameters support VSAM RLS:

RLSINIT({NO|YES})
This specifies whether to start the SMSVSAM address space as a part of the system.
initialization
CF_TIME(nnn|3600)
This indicates the number of seconds between recording SMF type 42 records with
subtypes 15, 16, 17, 18, and 19 for the CF (both cache and lock structures).
DEADLOCK_DETECTION(iiii|15,kkkk|4)
This specifies the deadlock detection intervals used by the Storage Management
Locking Services.
iiii - This is the local detection interval, in seconds.
kkkk - This is the global detection interval, which is the number of iterations of the local
detection interval that must be run until the global deadlock detection is invoked.

RLS_MaxCfFeatureLevel({A|Z})
This specifies the method that VSAM RLS uses to determine the size of the data that is
placed in the CF cache structure.
RLS_MAX_POOL_SIZE({nnnn|100})
This specifies the maximum size in megabytes of the SMSVSAM local buffer pool.
SMF_TIME({YES|NO})
This specifies that the SMF type 42 records are created at the SMF interval time, and that
all of the indicated records are synchronized with SMF and RMF data intervals.
For more information about VSAM RLS parameters, see z/OS DFSMSdfp Storage

7.16 Define sharing control data sets
CICS SMSVSAM SMSVSAM CICS

CACHE02
R/W address space address space R/W
CACHE01
CICS CICS
R/W IGWLOCK00 R/W
SMSVSAM Coupling Facility SMSVSAM

data space data space
Batch Batch
R/O R/O
System 1 System 2
Primary Secondary Spare

SHCDS SHCDS SHCDS
Volume1 Volume2 Volume3
SHCDS name: SYS1.DFPSHCDS.qualifier.Vvolser
Figure 7-18 VSAM RLS sharing control data sets
VSAM RLS sharing control data sets

The sharing control data set (SHCDS) is designed to contain the information required for
DFSMS to continue processing with a minimum of unavailable data and no corruption of data
when failures occur. The SCDS data can be either SMS or non-SMS managed.
The SHCDS contains the following:

Name of the CF lock structure in use
System status for each system or failed system instance
Time that the system failed
List of subsystems and their status
List of open data sets using the CF
List of data sets with unbound locks
List of data sets in permit non-RLS state
Defining sharing control data sets

The SHCDS is a logically partitioned VSAM linear data set. Consider the following for the
SHCDS allocation:
At a minimum, define and activate two SHCDSs and at least one spare SHCDS for
recovery purposes to ensure duplexing of your data.

Place the SHCDSs on volumes with global connectivity because VSAM RLS processing is
only available on those systems that currently have access to the active SHCDS.
The SHCDSs must not be shared outside the sysplex.
Use SHAREOPTIONS(3,3).
If SMS-managed, use a storage class with guaranteed space for the SHCDSs.
Use the naming convention: SYS1.DFPSHCDS.qualifier.Vvolser, where:
qualifier is a 1- to 8-character qualifier of your choice.
volser is the volume serial number of the volume on which the data set resides.
All SHCDSs are to be of the same size.
An SHCDS can have extents only on the same volume.
Both the primary and secondary SHCDS contain the same data. With the duplexing of the
data, VSAM RLS ensures that processing can continue in case VSAM RLS loses connection
to one SHCDS or the control data set got damaged. In that case, you can switch the spare
SHCDS to active.
Note: The SMSVSAM address space needs RACF UPDATE authority to

SYS1.DFPSHCDS.
JCL to allocate the SHCDSs

You can use the sample JCL in Figure 7-19 to allocate your SHCDSs.

//SYSIN DD *
DEFINE CLUSTER( -
NAME(SYS1.DFPSHCDS.PRIMARY.VTOTSMA) -
VOLUMES(TOTSMA) -
MEGABYTES(10 10) -
LINEAR -
SHAREOPTIONS(3,3) -
STORAGECLASS(GSPACE))
DEFINE CLUSTER( -
NAME(SYS1.DFPSHCDS.SECONDRY.VTOTCAT) -
VOLUMES(TOTCAT) -
MEGABYTES(10 10) -
LINEAR -
SHAREOPTIONS(3,3) -
DEFINE CLUSTER( -
NAME(SYS1.DFPSHCDS.SPARE.VTOTSMS) -
VOLUMES(TOTSMS) -
MEGABYTES(10 10) -
SHAREOPTIONS(3,3) -
LINEAR -
Figure 7-19 Allocating VSAM RLS SHCDSs
To calculate the size of the sharing control data sets, follow the guidelines provided in z/OS
DFSMSdfp Storage Administration Reference, SC26-7402.

Tip: Place the SHCDSs on separate volumes to maximize availability. Avoid placing
SHCDSs on volumes for which there might be extensive volume reserve activity.
SHCDS operations
Use the following command to activate your newly defined SHCDS for use by VSAM RLS.
For the primary and secondary SHCDS, use:
VARY SMS,SHCDS(SHCDS_name),NEW
For the spare SHCDS, use:
VARY SMS,SHCDS(SHCDS_name),NEWSPARE
To display the used SHCDSs in an active configuration, use:

D SMS,SHCDS
Figure 7-20 displays an example of a SHCDS list in a sysplex.
D SMS,SHCDS
IEE932I 539
IGW612I 17:10:12 DISPLAY SMS,SHCDS
Name Size %UTIL Status Type
WTSCPLX2.VSBOX48 10800Kb 4% GOOD ACTIVE
WTSCPLX2.VSBOX52 10800Kb 4% GOOD ACTIVE
WTSCPLX2.VSBOX49 10800Kb 4% GOOD SPARE
----------------- 0Kb 0% N/A N/A
----------------- 0Kb 0% N/A N/A
----------------- 0Kb 0% N/A N/A
----------------- 0Kb 0% N/A N/A
----------------- 0Kb 0% N/A N/A
----------------- 0Kb 0% N/A N/A
----------------- 0Kb 0% N/A N/A
Figure 7-20 Example of SHCDS display
To delete logically either an active or a spare SHCDS, use:

VARY SMS,SHCDS(SHCDS_name),DELETE
Note: In the VARY SMS,SHCDS commands, the SHCDS name is not fully qualified.
SMSVSAM takes as a default the first two qualifiers, which must always be
SYS1.DFPSHCDS. You must specify only the last two qualifiers as the SHCDS names.

7.17 Update SMS configuration
SMSBase Configuration Information Storage Class Information Coupling Facility
SC=CICS1
Cache Structure Names Cache set name = CACHE01

PUBLIC1
PUBLIC1 CACHE01, CACHE02
PUBLIC2 CACHE02, CACHE03
PAYROLL CACHE03, PAYSTRUC SC=CICS2
Cache set name = CACHE02

PUBLIC2
CF Cache
Structures
SC=CICS3
CACHE03
Cache set name =
PAYROLL
SC=NORLS
PAYSTRUC
Cache set name =
(blank)
No VSAM RLS capability available
Figure 7-21 Example of SMS configuration with cache sets
Update the SMS configuration

For DFSMSdfp to use the CF for VSAM RLS, after defining one or more CF cache structures
to MVS you must also add them in the SMS base configuration.
Define cache sets

In 7.14, Coupling Facility structures for RLS sharing on page 397 we discuss how to define
CF cache structures. You now need to add the CF cache structures to the DFSMS base
configuration. To do so, you need to associate them with a cache set name.
The following steps describe how to define a cache set and how to associate the cache
structures to the cache set:
1. From the ISMF primary option menu for storage administrators, select option 8, Control
Data Set.
2. Select option 7, Cache Update, and make sure that you specified the right SCDS name
(SMS share control data set, do not mix up with SHCDS).
3. Define your CF cache sets (see Figure 7-22 on page 406).

CF CACHE SET UPDATE PAGE 1 OF 1
Command ===>
SCDS Name : SYS1.SMS.SCDS

Define/Alter/Delete CF Cache Sets: ( 001 Cache Sets Currently Defined )
Cache Set CF Cache Structure Names

PUBLIC1 CACHE01 CACHE02
PUBLIC2 CACHE02 CACHE03
PAYROLL CACHE03 PAYSTRUC
F1=Help F2=Split F3=End F4=Return F7=Up F8=Down F9=Swap

F10=Left F11=Right F12=Cursor
Figure 7-22 CF cache update panel in ISMF
SMS storage class changes

After defining your cache sets, you need to assign them to one or more storage classes (SC)
so that data sets associated with those SCs are eligible for VSAM RLS by using Coupling
Facility cache structures.
Follow these steps to assign the CF cache sets:

1. Select option 5, Storage Class, from the ISMF primary option menu for storage
administrators
2. Select option 3, Define, and make sure, that you specified the right SCDS name
3. On the second page of the STORAGE CLASS DEFINE panel enter the name of the
Coupling Facility cache set you defined in the base configuration (see Figure 7-23 on
page 407).
4. On the same panel enter values for direct and sequential weight. The higher the value, the
more important it is that the data be assigned more cache resources.

STORAGE CLASS DEFINE Page 2 of 2
Command ===>
SCDS Name . . . . . : SYS1.SMS.SCDS

Storage Class Name : CICS1
To DEFINE Storage Class, Specify:
Guaranteed Space . . . . . . . . . N (Y or N)
Guaranteed Synchronous Write . . . N (Y or N)
Multi-Tiered SG . . . . . . . . . . (Y, N, or blank)
Parallel Access Volume Capability N (R, P, S, or N)
CF Cache Set Name . . . . . . . . . PUBLIC1 (up to 8 chars or blank)
CF Direct Weight . . . . . . . . . 6 (1 to 11 or blank)
CF Sequential Weight . . . . . . . 4 (1 to 11 or blank)
F1=Help F2=Split F3=End F4=Return F7=Up F8=Down F9=Swap

F10=Left F11=Right F12=Cursor
Figure 7-23 ISMF storage class definition panel, page 2
Note: Be sure to change your Storage Class ACS routines so that RLS data sets are
assigned the appropriate storage class.
More detailed information about setting up SMS for VSAM RLS is in z/OS DFSMSdfp Storage

7.18 Update data sets with log parameters
Two new parameters to support VSAM RLS

LOG(NONE/UNDO/ALL)
NONE - no data set recovery
UNDO - data set is backward recoverable
ALL - data set is backward and forward recoverable
LOGSTREAMID(logstreamname)
Specifies the name of the CICS forward recovery
logstream for data sets with LOG(ALL)
Figure 7-24 New parameters to support VSAM RLS
New parameters to support VSAM RLS

The two new parameters LOG and LOGSTREAMID are stored in the ICF catalog. You can
use the IDCAMS DEFINE and ALTER commands to set the data set LOG attribute and to
assign a log data set name.
Another way to assign the LOG attribute and a LOGSTREAMID is to use a data class that has
those values already defined.
The LOG parameter is described in detail in 7.10, VSAM RLS/CICS data set recovery on
page 392.
Use the LOGSTREAMID parameter to assign a CICS forward recovery log stream to a data
set which is forward recoverable.
JCL to define a cluster with LOG(ALL) and a LOGSTREAMID

The example in Figure 7-25 on page 409 shows you how to define a VSAM data set that is
eligible for RLS processing and that is forward recoverable.

//LABEL JOB ...
//STEP1 EXEX PGM=IDCAMS
//SYSIN DD *
DEFINE CLUSTER (NAME (OTTI.IV4A62A1.RLS) -
CYL (1 1) -
RECSZ (5000 10009) -
CISZ (20000) -
IXD -
KEYS (8 0) -
FREESPACE(40 40) -
SHR (2,3) -
REUSE -
BWO(TYPECICS) -
LOG(ALL) -
LOGSTREAMID(CICS.IV4A62A1.DFHJ05) -
STORAGECLASS(CICS1) -
)
/*
Figure 7-25 Sample JCL to define a recoverable VSAM data set RLS processing
For more information about the IDCAMS DEFINE and ALTER commands, see z/OS DFSMS
Access Method Services for Catalogs, SC26-7394.
JCL to define a log stream

You can use the JCL in Figure 7-26 to define a log stream.
//LABEL JOB ...

//SYSIN DD *
DELETE LOGSTREAM NAME(CICS.IV4A62A1.DFHJ05)
DATA TYPE(LOGR) REPORT(NO)
DEFINE LOGSTREAM
NAME(CICS.IV4A62A1.DFHJ05)
DASDONLY(YES)
STG_SIZE(50)
LS_SIZE(25)
HIGHOFFLOAD(80)
LOWOFFLOAD(0)
/*
Figure 7-26 JCL to define a log stream
For information about the IXCMIAPU utility, see z/OS MVS Setting Up a Sysplex, SA22-7625.

7.19 The SMSVSAM address space
RLS clients: RLS server: RLS server: RLS clients:
subsystem SMSVSAM SMSVSAM subsystem

CICS1 address space address space CICS1
CACHE02
subsystem subsystem
CICS2 CICS2
CACHE01
MMFSTUFF MMFSTUFF
IGWLOCK00
batch job SMSVSAM SMSVSAM batch job
HSM Coupling Facility HSM
data spaces data spaces
System 1 System n
Figure 7-27 SMSVSAM address space
The SMSVSAM address space

SMSVSAM is the MVS job name of the VSAM RLS address space. It is started automatically
at IPL time if RLSINIT(YES) is specified in the IGDSMSxx member of SYS1.PARMLIB.
Another way to start it is by operator command (see 7.20, Interacting with VSAM RLS on
page 412).
The SMSVSAM address space needs to be started on each system where you want to exploit
VSAM RLS. It is responsible for centralizing all processing necessary for cross-system
sharing, which includes one connect per system to XCF lock, cache, and VSAM control block
structures.
The SMSVSAM address space owns two data spaces:

SMSVSAM
This contains VSAM RLS control blocks and a system-wide buffer pool
MMFSTUFF
This collects activity monitoring information that is used to produce SMF records
Terminology
We use the following terms to describe an RLS environment:
RLS server
The SMSVSAM address space is also referred to as the RLS server.

RLS client
Any address space that invokes an RLS function that results in a program call to the
SMSVSAM address space is called an RLS client. Those address spaces can be CICS
regions as well as batch jobs.
Recoverable subsystem
A subsystem is an RLS client space that registers with the SMSVSAM address space as
an address space that will provide transactional and data set recovery. CICS, for example,
is a recoverable subsystem.
Batch job
An RLS client space that does not first register with SMSVSAM as a recoverable
subsystem is called a batch job. An example for such a batch job is HSM.

7.20 Interacting with VSAM RLS
Interacting with VSAM RLS includes

Use of SETSMS commands to change the
IGDSMSxx specifications
Use of VARY SMS commands to start and stop the
RLS server
Use of display commands to get information about
the current VSAM RLS configuration
Activate and display the SHCDS
Quiesce/unquiesce a data set from RLS processing
Changing the size of an XCF structure
List and change recovery information
Figure 7-28 Interacting with VSAM RLS
Interacting with VSAM RLS

This section provides a brief overview of several commands you need to know to monitor and
control the VSAM RLS environment.
SETSMS command
Use the SETSMS command to overwrite the PARMLIB specifications for IGDSMSxx. The syntax
is:
SETSMS CF-TIME(nnn|3600)
DEADLOCK_DETECTION(iiii,kkkk)
RLSINIT
RLS_MAXCFFEATURELEVEL({A|Z})
RLS_MAX_POOL_SIZE(nnnn|100)
SMF_TIME(YES|NO)
For information about these PARMLIB values refer to 7.15, Update PARMLIB with VSAM
RLS parameters on page 400.

VARY SMS command
Use the VARY SMS command to control SMSVSAM processing.
Start the VSAM RLS server address space on a single system.
If not started during IPL because of RLSINIT(NO) in the IGDSMSxx PARMLIB member,
you can start the RLS server manually with:
V SMS,SMSVSAM,ACTIVE
Stop the RLS server address space.
If the SMSVSAM address space fails, it is automatically restarted. If you want to stop the
SMSVSAM address space permanently on a single system, use:
V SMS,SMSVSAM,TERMINATESERVER
Use this command only for specific recovery scenarios that require the SMSVSAM server
to be down and to not restart automatically.
Fallback from RLS processing.
A detailed procedure about falling back from RLS processing is described in z/OS
DFSMSdfp Storage Administration Reference, SC26-7402. Read it before using the
following command:
V SMS,SMSVSAM,FALLBACK
This command is used as the last step in the disablement procedure to fall back from
SMSVSAM processing.
Further VARY SMS commands.
There are additional VARY SMS commands available to interact with VSAM RLS. They are:
V SMS,CFCACHE(cachename),ENABLE|QUIESCE
CFVOL(volid),ENABLE|QUIESCE
MONDS(dsname[,dsname...]),ON|OFF
SHCDS(shcdsname),NEW|NEWSPARE|DELETE
SMSVSAM,SPHERE(spherename),ENABLE
FORCEDELETELOCKSTRUCTURE
Refer to z/OS MVS System Commands, SA22-7627 for information about these
commands.
Display commands
There are a several display commands available to provide RLS-related information.
Display the status of the SMSVSAM address space:
DISPLAY SMS,SMSVSAM{,ALL}
Specify ALL to see the status of all the SMSVSAM servers in the sysplex.
Display information about the Coupling Facility cache structure:
DISPLAY SMS,CFCACHE(CF_cache_structure_name|*)
Display information about the Coupling Facility lock structure IGWLOCK00:
DISPLAY SMS,CFLS
This information includes the lock rate, lock contention rate, false contention rate, and
average number of requests waiting for locks.
Display XCF information for a CF structure:
DISPLAY XCF,STR,STRNAME=structurename
This provides information such as status, type, and policy size for a CF structure.

To learn about other DISPLAY commands, see z/OS MVS System Commands, SA22-7627.
Activate and display the sharing control data sets

Refer to 7.16, Define sharing control data sets on page 402 for information about how to
activate and display the VSAM RLS sharing control data sets.
Quiesce and unquiesce a data set from RLS processing

In 7.12, The batch window problem on page 395 we discuss the reasons for quiescing data
sets from RLS processing.
There are several ways to quiesce/unquiesced a data set.

Use the CICS command:
CEMT SET DSNAME(dsname) QUIESCED|UNQUIESCED
Use an equivalent SPI command in a user program.
Use the system command:
VARY SMS,SMSVSAM,SPHERE(dsname),QUIESCE|ENABLE
The quiesce status of a data set is set in the catalog and is shown in an IDCAMS LISTCAT
output for the data set. See 7.22, Interpreting RLSDATA in an IDCAMS LISTCAT output on
page 417 for information about interpreting LISTCAT outputs.
Changing the size of an XCF structure

Use the following command to change the size of a Coupling Facility structure:
SETXCF START,ALTER,STRNAME=CF_cachestructurename,SIZE=newsize
This new size can be larger or smaller than the size of the current CF cache structure, but it
cannot be larger than the maximum size specified in the CFRM policy. The SETXCF
START,ALTER command will not work unless the structures ALLOW ALTER indicator is set to
YES.
List and change recovery information

You can use the IDCAMS command SHCDS to list and change recovery information kept by the
SMSVSAM server. It also resets VSAM record-level sharing settings in catalogs, allowing for
non-RLS access or fallback from RLS processing. To learn about the single parameters for
this command, see z/OS DFSMS Access Method Services for Catalogs, SC26-7394.
Important: This section simply provides an overview of commands that are useful for
working with VSAM RLS. Before using these commands (other than the DISPLAY
command), read the official z/OS manuals carefully.

7.21 Backup and recovery of CICS VSAM data sets
Backup tool DFSMSdss

Supports backup-while-open processing
BWO type TYPECICS allows to backup the data set
while it is open in CICS
BWO of forward recoverable data sets allows a
recovery tool to take this backup for forward recovery
Recovery of a broken data set
Non-recoverable data sets:
Lost updates, data in the backup can be inconsistent
Backward recoverable data sets
Lost updates, data in the backup are consistent
Forward recoverable data sets
No lost updates, data are consistent
Figure 7-29 Backup and recovery of CICS VSAM data sets
Backup tool DFSMSdss

DFSMSdss supports backup-while-open (BWO) serialization, which can perform backups of
data sets that are open for update for long periods of time. It can also perform a logical data
set dump of these data sets even if another application has them serialized.
Backup-while-open is a better method than using SHARE or TOLERATE(ENQFAILURE) for
dumping CICS VSAM file-control data sets that are in use and open for update.
When you dump data sets that are designated by CICS as eligible for backup-while-open
processing, data integrity is maintained through serialization interactions between:
CICS (database control program)
VSAM RLS
VSAM record management
DFSMSdfp
DFSMSdss
Backup-while-open
In order to allow DFSMSdss to take a backup while your data set is open by CICS, you need
to define the data set with the BWO attribute TYPECICS or assign a data class with this
attribute.
TYPECICS
Use TYPECICS to specify BWO in a CICS environment. For RLS processing, this
activates BWO processing for CICS. For non-RLS processing, CICS determines whether

to use this specification or the specification in the CICS FCT. The BWO type is stored in
the ICF catalog.
Backup-while-open of a forward recoverable data set

If you use the DSS BWO processing for a forward recoverable data set, CICS will log the start
and end of the copy/backup operation. The data set can then be fully recovered from this
backup.
For information about BWO processing, see z/OS DFSMSdss Storage Administration
Reference, SC35-0424.
Recover a CICS VSAM data set

Sometimes it is necessary to recover a data set, for example if it became broken.
Recovery of a non-recoverable data set
Data sets with LOG(NONE) are considered non-recoverable. To recover such a data set,
restore the last backup of the data set. All updates to this data set after the backup was
taken are lost. If the backup was taken after a transaction failed (did not commit), the data
in the backup might be inconsistent.
Recovery of a backward recoverable data set
Data sets with the LOG(NONE) attribute are considered backward recoverable. To recover
such a data set, restore the last backup of the data set. All updates to this data set after
the backup was taken are lost. The data in the backup is consistent.
Recovery of a forward recoverable data set
Data sets with LOG(ALL) and a log stream assigned are forward recoverable. Restore the
last backup of the data set. Then run a tool like CICS VSAM Recovery (CICSVR), which
uses the forward recovery log to redrive all committed updates until the data set became
broken. No updates are lost.

7.22 Interpreting RLSDATA in an IDCAMS LISTCAT output
CLUSTER ------- OTTI.IV4A62A1.RLS

IN-CAT --- CATALOG.MVSICFU.VO260C1
HISTORY
SMSDATA
STORAGECLASS ------CICS1 MANAGEMENTCLASS-CICSRLSM
DATACLASS --------(NULL) LBACKUP ---0000.000.0000
BWO STATUS------00000000 BWO TIMESTAMP---00000 00:00:00.0
BWO-------------TYPECICS
RLSDATA
LOG ------------------ALL RECOVERY REQUIRED --(NO)
VSAM QUIESCED -------(NO) RLS IN USE ---------(YES)
LOGSTREAMID--------------CICS.IV4A62A1.DFHJ05
RECOVERY TIMESTAMP LOCAL-----X'0000000000000000'
RECOVERY TIMESTAMP GMT-------X'0000000000000000'
Figure 7-30 Sample RLSDATA in an IDCAMS LISTCAT output
RLSDATA in an IDCAMS LISTCAT output

The RLSDATA in the output of an IDCAMS LISTCAT job shows you the RLS status of the data
set and recovery information.
You can use the sample JCL in Figure 7-31 to run an IDCAMS LISTCAT job.
//LABEL JOB ...

//SYSIN DD *
LISTC ENT(OTTI.IV4A62A1.RLS) ALL
/*
Figure 7-31 Sample JCL for IDCAMS LISTCAT
RLSDATA
RLSDATA contains the following information:
LOG
This field shows you the type of logging used for this data set. It can be NONE, UNDO or
ALL.

RECOVERY REQUIRED
This field indicates whether the sphere is currently in the process of being forward
recovered.
VSAM QUIESCED
If this value is YES, the data set is currently quiesced from RLS processing so no RLS
access is possible until it is unquiesced (see also 7.20, Interacting with VSAM RLS on
page 412).
RLS IN USE
If this value is YES it means:
The data set was last opened for RLS access.
The data set is not opened for RLS processing but is recoverable and either has
retained locks protecting updates or is in a lost locks state.
Note: If the RLS-IN-USE indicator is on, it does not mean that the data set is currently
in use by VSAM RLS. It simply means that the last successful open was for RLS
processing.
Non-RLS open will always attempt to call VSAM RLS if the RLS-IN-USE bit is on in the
catalog. This bit is a safety net to prevent non-RLS users from accessing a data set
which can have retained or lost locks associated with it.
The RLS-IN-USE bit is set on by RLS open and is left on after close. This bit is only
turned off by a successful non-RLS open or by the IDCAMS SHCDS CFRESET command.
LOGSTREAMID
This value tells you the forward recovery log stream name for this data set if the LOG
attribute has the value of ALL.
RECOVERY TIMESTAMP
The recovery time stamp gives the time the most recent backup was taken when the data
set was accessed by CICS using VSAM RLS.
All LISTCAT keywords are described in Appendix B of z/OS DFSMS Access Method Services
for Catalogs, SC26-7394.

7.23 DFSMStvs introduction
DFSMStvs addresses key requirements from

customers
Objective: Provide transactional recovery within
VSAM
VSAM RLS allows batch sharing of recoverable
data sets for read
VSAM RLS provides locking and buffer coherency
CICS provides logging and two-phase commit
protocols
DFSMStvs allows batch sharing of recoverable data
sets for update
Logging provided using the MVS system logger
Two-phase commit and back out using MVS
recoverable resource management services
Figure 7-32 DFSMS Transactional VSAM Services (DFSMStvs) introduction
Why to use DFSMS Transactional VSAM Services (DFSMStvs)

There following key requirements were expressed by customers running CICS applications in
RLS mode:
Address the batch window problem for CICS and batch sharing of VSAM data sets.
In 7.12, The batch window problem on page 395, we discuss the problem with the batch
window. Customers reported batch windows ranging from two to ten hours. The programs
run during the batch window consisted of both in-house applications and vendor-written
applications.
Allow batch update sharing concurrent with CICS use of recoverable data.
Allow multiple batch update programs to run concurrently.
Allow programs to interact with multiple resource managers such as IMS and DB2.
Allow full transactional read/write access from new Java programs to existing VSAM
data.
Objective of DFSMStvs
The objective of DFSMStvs is to provide transactional recovery directly within VSAM. It is an
extension to VSAM RLS. It allows any job or application that is designed for data sharing to
read/write share VSAM recoverable files.

DFSMStvs is a follow-on project/capability based on VSAM RLS. VSAM RLS supports CICS
as a transaction manager. This provides sysplex data sharing of VSAM recoverable files
when accessed through CICS. CICS provides the necessary unit-of-work management,
undo/redo logging, and commit/back out functions. VSAM RLS provides the underlying
sysplex-scope locking and data access integrity.
DFSMStvs adds logging and commit/back out support to VSAM RLS. DFSMStvs requires
and supports the RRMS (recoverable resource management services) component as the
commit or sync point manager.
DFSMStvs provides a level of data sharing with built-in transactional recovery for VSAM
recoverable files that is comparable with the data sharing and transactional recovery support
for databases provided by DB2 and IMSDB.

7.24 Overview of DFSMStvs
DFSMStvs enhances VSAM RLS to provide data

recovery capabilities such as
Transactional recovery
Data set recovery
DFSMStvs does not perform forward recovery
DFSMStvs uses
RRMS to manage the unit of recovery (UR)
System logger to manage the log streams
Undo log
Shunt log
Forward recovery logs
Log of logs
VSAM RLS manages locking and buffer coherency
Allows atomic commit of changes - all or nothing
DFSMStvs provides peer recovery
Figure 7-33 DFSMStvs overview
Enhancements of VSAM RLS

DFSMStvs enhances VSAM RLS to perform data recovery in the form of:
Transactional recovery (see Transactional recovery on page 394)
Data set recovery (see VSAM RLS/CICS data set recovery on page 392)
Before DFSMStvs, those two types of recovery were only supported by CICS.
CICS performs the transactional recovery for data sets defined with a LOG parameter UNDO
or ALL.
For forward recoverable data sets (LOG(ALL)) CICS also records updates in a log stream for
forward recovery. CICS itself does not perform forward recovery, it performs only logging. For
forward recovery you need a utility like CICS VSAM recovery (CICSVR).
Like CICS, DFSMStvs also provides transactional recovery and logging.
Without DFSMStvs, batch jobs cannot perform transactional recovery and logging. That is the
reason batch jobs were granted only read access to a data set that was opened by CICS in
RLS mode. A batch window was necessary to run batch updates for CICS VSAM data sets.
With DFSMStvs, batch jobs can perform transactional recovery and logging concurrently with
CICS processing. Batch jobs can now update data sets while they are in use by CICS. No
batch window is necessary any more.

Like CICS, DFSMStvs does not perform data set forward recovery.
Components used by DFSMStvs

The following components are involved in DFSMStvs processing:
RRMS to manage the unit of recovery
DFSMStvs uses the recoverable resource management services (RRMS) to manage the
unit of recovery that is needed for transactional recovery. More information about RRMS is
in the next section.
System logger to manage log streams
There are three kinds of logs used by DFSMStvs:
Undo log: Used for transactional recovery (back out processing)
Forward recovery log: Used to log updates against a data set for forward recovery
Shunt log: Used for long running or failed units of recovery
You can also define a log of logs for use to automate forward recovery.
These logs are maintained by the MVS system logger. For information about the various
log types, see 7.28, DFSMStvs logging on page 427.
VSAM RLS - used for record locking and buffering
The need for VSAM RLS is discussed in prior sections.
Atomic commit of changes

DFSMStvs allows atomic commit of changes. That means, if multiple updates to various
records are necessary to complete a transaction, either all updates are committed or none
are, if the transaction does not complete. See also 7.11, Transactional recovery on
page 394 and 7.26, Atomic updates on page 425.
Peer recovery
Peer recovery allows DFSMStvs to recover for a failed DFSMStvs instance to clean up any
work that was left in an incomplete state and to clear retained locks that resulted from the
failure.
For more information about peer recovery, see z/OS DFSMStvs Planning and Operation
Guide, SC26-7348.

7.25 DFSMStvs use of z/OS RRMS
z/OS RRMS:
- Registration services Prepare/Commit
- Context services DFSMStvs
Rollback
- Resource recovery
services (RRS)
Another
Recoverable
Resource Mgr
Another
Recoverable
Resource Mgr
Figure 7-34 DFSMStvs and RRMS
Recoverable resource management services (RRMS)

z/OS provides recoverable resource management services (RRMS) comprising:
Registration services
Context services
Resource recovery services (RRS), which acts as sync point manager
Role of resource recovery services (RRS)

RRS provides the sync point services and is the most important component from a
DFSMStvs use perspective.
DFSMStvs is a recoverable resource manager. It is not a commit or sync point manager.

DFSMStvs interfaces with the z/OS sync point manager (RRS).
When an application issues a commit request directly to z/OS or indirectly through a sync
point manager that interfaces with the z/OS sync point manager, DFSMStvs is invoked to
participate in the two-phase commit process.
Other resource managers (like DB2) whose recoverable resources were modified by the
transaction are also invoked by the z/OS sync point manager, thus providing a commit scope
across the multiple resource managers.
RRS is a system-wide commit coordinator. It enables transactions to update protected

resources managed by many resource managers.

It is RRS that provides the means to implement two-phase commit, but a resource manager
must also use registration services and context services in conjunction with resource
recovery services.
Two-phase commit
The two-phase commit protocol is a set of actions used to make sure that an application
program either makes all changes to the resources represented by a single unit of recovery
(UR), or it makes no changes at all. This protocol verifies that either all changes or no
changes are applied even if one of the elements (such as the application, the system, or the
resource manager) fails. The protocol allows for restart and recovery processing to take place
after system or subsystem failure.
For a discussion of the term unit of recovery, see 7.27, Unit of work and unit of recovery on
page 426.

7.26 Atomic updates
Before After Before After
$200 -$100 $100 $200 $200
+$100 +$100
$700 $800 $700 $800
Transaction to Incomplete
transfer $100 transaction
Figure 7-35 Example of an atomic update
Atomic updates
A transaction is known as atomic when an application changes data in multiple resource
managers as a single transaction, and all of those changes are accomplished through a
single commit request by a sync point manager. If the transaction is successful, all the
changes are committed. If any piece of the transaction is not successful, then all changes are
backed out. An atomic instant occurs when the sync point manager in a two-phase commit
process logs a commit record for the transaction.
Also see 7.11, Transactional recovery on page 394 for information about recovering an
uncompleted transaction.

7.27 Unit of work and unit of recovery
Start of program synchronized implicit
update 1
update 2
commit
} A = unit of recovery
synchronized explicit
update 3
update 4
update 5
} B
commit synchronized explicit
update 6
} C
End of program synchronized implicit
Figure 7-36 Unit of recovery example
Unit of work and unit of recovery

A unit of work (UOW) is the term used in CICS publications for a set of updates that are
treated as an atomic set of changes.
RRS uses unit of recovery (UR) to mean much the same thing. Thus, a unit of recovery is the
set of updates between synchronization points. There are implicit synchronization points at
the start and at the end of a transaction. Explicit synchronization points are requested by an
application within a transaction or batch job. It is preferable to use explicit synchronization for
greater control of the number of updates in a unit of recovery.
Changes to data are durable after a synchronization point. That means that the changes
survive any subsequent failure.
In Figure 7-36 there are three units of recovery, noted as A, B and C. The synchronization
points between the units of recovery are either:
Implicit - At the start and end of the program
Explicit - When requested by commit

7.28 DFSMStvs logging
CICSA system TVS1 CICSB system TVS2

logger logger
...
System 1 System n
CICSA undo log log stream
TVS1 undo log log stream

cache
structures CICSB undo log log stream
TVS2 undo log log stream

list
structures
CICS/DFSMStvs merged
lock structures forward recovery log log
stream
Coupling Facility
Figure 7-37 DFSMStvs logging
DFSMStvs logging
DFSMStvs logging uses the z/OS system logger. The design of DFSMStvs logging is similar
to the design of CICS logging. Forward recovery logstreams for VSAM recoverable files will
be shared across CICS and DFSMStvs. CICS will log changes made by CICS transactions;
DFSMStvs will log changes made by its callers.
The system logger

The system logger (IXGLOGR) is an MVS component that provides a rich set of services that
allow another component or an application to write, browse, and delete log data. The system
logger is used because it can merge log entries from many z/OS images to a single log
stream, where a log stream is simply a set of log entries.
Types of logs
There are various types of logs involved in DFSMStvs (and CICS) logging. They are:
Undo logs (mandatory, one per image) - tvsname.IGWLOG.SYSLOG
The backout or undo log contains images of changed records for recoverable data sets as
they existed prior to being changed. It is used for transactional recovery to back out
uncommitted changes if a transaction failed.
Shunt logs (mandatory, one per image) - tvsname.IGWSHUNT.SHUNTLOG
The shunt log is used when backout requests fail and for long running units of recovery.

Log of logs (optional, shared with CICS and CICSVR) - Default name is
CICSUSER.CICSVR.DFHLGLOG
The log of logs contains copies of log records that are used to automate forward recovery.
Forward recovery logs (optional, shared with CICS) - Name of your choice
The forward recovery log is used for data sets defined with LOG(ALL). It is used for
forward recovery. It needs to be defined in the LOGSTREAMID parameter of the data set.
For more information see 7.10, VSAM RLS/CICS data set recovery on page 392.
The system logger writes log data to log streams. The log streams are put in list structures in
the Coupling Facility (except for DASDONLY log streams).
As Figure 7-37 on page 427 shows, you can merge forward recovery logs for use by CICS
and DFSMStvs. You can also share it by multiple VSAM data sets. You cannot share an undo
log by CICS and DFSMStvs; you need one per image.
For information about how to define log streams and list structures refer to 7.32, Prepare for
logging on page 433.
Note: DFSMStvs needs UPDATE RACF access to the log streams.

7.29 Accessing a data set with DFSMStvs
This table lists the type of OPEN resulting from the

parameters specified
Left side shows the type of data set and type of open
Column headings indicate the RLS option specified
Data Set Type & NRI CR CRE
Type of OPEN
Recoverable VSAM RLS VSAM RLS DFSMStvs

Open for Input
Recoverable DFSMStvs DFSMStvs DFSMStvs
Open for Output
Non-recoverable VSAM RLS VSAM RLS DFSMStvs
Open for Input
Non-recoverable VSAM RLS VSAM RLS DFSMStvs
Open for Output
Figure 7-38 Accessing a data set with DFSMStvs
Data set access with DFSMStvs

In 7.9, VSAM RLS locking on page 390 we discuss the MACRF options NRI, CR, and CRE.
CRE gives DFSMStvs access to VSAM data sets open for input or output. CR or NRI gives
DFSMStvs access to VSAM recoverable data sets only for output.
You can modify an application to use DFSMStvs by specifying RLS in the JCL or the ACB and
having the application access a recoverable data set using either open for input with CRE or
open for output from a batch job.
The table in Figure 7-38 shows in which cases DFSMStvs is invoked.
DFSMStvs is always invoked if:

RLS and CRE is specified in the JCL or MACRF parameter of the ACB macro
CRE is also known as repeatable read.
RLS is specified in the JCL or MACRF parameter of the ACB macro and the data set is
recoverable and it is opened for output (update processing)
This allows DFSMStvs to provide the necessary transactional recovery for the data set.

7.30 Application considerations
Break processing into a series of transactions

Invoke RRS for commit and back out
Modify program/JCL to request transactional VSAM
access
Specify through JCL or ACB
Prevent multiple RPLs from causing intra-UR lock
contention
Handle potential loss of positioning at sync point
Handle all work that is part of one UR under the
same context
Do not use file backup/restore as job restart
technique
Figure 7-39 Application considerations
Application considerations
For an application to participate in transactional recovery, it must first understand the concept
of a transaction. It is not a good idea simply to modify an existing batch job to use DFSMStvs
with no further change, as this causes the entire job to be seen as a single transaction. As a
result, locks would be held and log records would need to exist for the entire life of the job.
This can cause a tremendous amount of contention for the locked resources. It can also
cause performance degradation as the undo log becomes exceedingly large.
Break processing into a series of transactions

To exploit the DFSMStvs capabilities, break down application processing into a series of
transactions. Have the application issue frequent sync points by invoking RRS commit and
backout processing (MVS callable services SRRCMIT and SRRBACK). For information about
RRS, see 7.25, DFSMStvs use of z/OS RRMS on page 423.
RLS and DFSMStvs provide isolation until commit/backout. Consider the following rules:
Share locks on records accessed with repeatable read.
Hold write locks on changed records until the end of a transaction.
Use commit to apply all changes and release all locks.
Information extracted from shared files must not be used across commit/backout for the
following reasons:
Need to re-access the records

Cannot get a record before a sync point and update it after
Do not position to a record before a sync point and access it after
Modify the program or JCL to request DFSMStvs access

You need to change the ACB MACRF or JCL specifications for a data set to request
DFSMStvs access. See the previous section for more information.
Prevent contention within a unit of recovery

If the application uses multiple RPLs, care must be taken in how they are used. Using various
RPLs to access the same record can cause lock contention within a UR.
Handle potential loss of positioning at sync point

The batch application must have a built-in method of tracking its processing position within a
series of transactions. One potential method of doing this is to use a VSAM recoverable file to
track the job's commit position.
Handle all work that is part of one UR under the same context
For information about units of recovery, see 7.27, Unit of work and unit of recovery on
page 426. Reconsider your application to handle work that is part of one unit of recovery
under the same context.
Do not use file backup/restore as a job restart technique

Today's batch applications that update VSAM files in a non-shared environment can create
backup copies of the files to establish a restart/recovery point for the data. If the batch
application fails, the files can be restored from the backup copies, and the batch jobs can be
re-executed. This restart/recovery procedure cannot be used in a data sharing DFSMStvs
environment because restoring to the point-in-time backup erases changes made by other
applications sharing the data set.
Instead, the batch application must have a built-in method of tracking its processing position
within a series of transactions. One potential method of doing this is to use a VSAM
recoverable file to track the jobs commit position. When the application fails, any
uncommitted changes are backed out.
The already-committed changes cannot be backed out, because they are already visible to
other jobs or transactions. In fact, it is possible that the records that were changed by
previously-committed UR were changed again by other jobs or transactions. Therefore, when
the job is rerun, it is important that it determines its restart point and not attempt to redo any
changes it had committed before the failure.
For this reason, it is important that jobs and applications using DFSMStvs be written to
execute as a series of transactions and use a commit point tracking mechanism for restart.

7.31 DFSMStvs logging implementation
Prepare for logging

Update CFRM policy to define structures for the
sytem logger
Define log structures and log streams
LOGSTREAMID
Update SYS1.PARMLIB(IGDSMSxx) with
DFSMStvs parameters
Ensure RACF authorization to the log streams
Update SMS configuration
Figure 7-40 DFSMStvs configuration changes
DFSMStvs configuration changes

To run DFSMStvs you need to have set up VSAM RLS already. DFSMStvs uses the locking
and buffering techniques of VSAM RLS to access the VSAM data sets. Additionally, you need
to make the following changes:
Prepare for logging.
Update CFRM policy to define structures for the system logger.
Define log streams.
Update data sets with LOG(NONE/UNDO/ALL) and LOGSTREAMID.
This is described in more detail in 7.32, Prepare for logging on page 433.
Update SYS1.PARMLIB(IGDSMSxx) with DFSMStvs parameters.
For the necessary PARMLIB changes see 7.33, Update PARMLIB with DFSMStvs
parameters on page 436
Ensure that DFSMStvs has RACF authorization to the log streams.
To provide logging, DFSMStvs needs RACF UPDATE authority to the log streams. See
z/OS DFSMStvs Planning and Operation Guide, SC26-7348, for more information.
Update the SMS configuration.
Recoverable data sets must be SMS managed. Therefore, you must change your SMS
configuration to assign a storage class to those data sets. In addition, you can optionally
change your data class with BWO, LOG, and LOGSTREAMID parameters.

7.32 Prepare for logging
Update CFRM policy to add list structures for use

by DFSMStvs
Update LOGR policy to add Coupling Facility
structures needed for all log streams
Define log streams for use by DFSMStvs as:
Undo log
Shunt log
Log of logs
Forward recovery logs
LOGSTREAMID
Figure 7-41 Prepare for logging
Prepare for logging

You have to define log structures in two places: the Coupling Facility resource management
(CFRM) policy and the system logger LOGR policy. A policy is a couple data set.
Furthermore, you need to define log streams for the various kinds of logs used by DFSMStvs.
If a data set is forward recoverable, you also need to assign a log stream for forward recovery
to this data set.
Update the CFRM policy

The CFRM policy is used to divide Coupling Facility space into structures. The CFRM policy
enables you to define how MVS manages Coupling Facility resources. In 7.14, Coupling
Facility structures for RLS sharing on page 397, we describe how to define CF structures for
use be VSAM RLS. You need to run a similar job to either define a new CFRM policy or to
update an existing CFRM policy with the structures that are used by the logger.
A sample JCL you can use to define a new CFRM policy is shown in Figure 7-42 on
page 434.

//LABEL JOB ...
//SYSIN DD *
DATA TYPE(CFRM) REPORT(YES)
DEFINE POLICY NAME(CFRM02) REPLACE(YES)
STRUCTURE NAME(LOG_IGWLOG_001)
SIZE(10240)
INITSIZE(5120)
PREFLIST(CF01,CF02)
STRUCTURE NAME(LOG_IGWSHUNT_001)
SIZE(10240)
INITSIZE(5120)
PREFLIST(CF01,CF02)
STRUCTURE NAME(LOG_IGWLGLGS_001)
SIZE(10240)
INITSIZE(5120)
PREFLIST(CF01,CF02)
STRUCTURE NAME(LOG_FORWARD_001)
SIZE(10240)
INITSIZE(5120)
PREFLIST(CF01,CF02)
/
Figure 7-42 Example of defining structures in the CFRM policy
Update the LOGR policy

You must also define the Coupling Facility structures in the LOGR policy. The system logger
component manages log streams based on the policy information that installations place in
the LOGR policy.
Multiple log streams can write data to a single Coupling Facility structure. This does not mean
that the log data is merged; the log data stays segregated according to log stream.
Figure 7-43 shows how to define the structures in the LOGR policy.
//LABEL JOB ...

//SYSIN DD *
DATA TYPE(LOGR) REPORT(YES)
DEFINE STRUCTURE NAME(LOG_IGWLOG_001)
LOGSNUM(10) MAXBUFSIZE(64000)
AVGBUFSIZE(4096)
DEFINE STRUCTURE NAME(LOG_IGWSHUNT_001)
AVGBUFSIZE(4096)
DEFINE STRUCTURE NAME(LOG_IGWLGLGS_001)
AVGBUFSIZE(4096)
DEFINE STRUCTURE NAME(LOG_FORWARD_001)
AVGBUFSIZE(4096)
/*
Figure 7-43 Defining structures to the LOGR policy

Define log streams
Each DFSMStvs instance requires its own pair of system logs: a primary (undo) log and a
secondary (shunt) log. You must define these two logs before you can use DFSMStvs.
Additionally, you can define forward recovery logs and a log of logs.
For the various types of log streams that are used by DFSMStvs refer to 7.28, DFSMStvs
logging on page 427. A log stream is a VSAM linear data set which simply contains a
collection of data. To define log streams, you can use the example JCL in Figure 7-44.
//LABEL JOB ...

//SYSIN DD *
DATA TYPE(LOGR) REPORT(YES)
DEFINE LOGSTREAM NAME(IGWTV001.IGWLOG.SYSLOG)
STRUCTURENAME(LOG_IGWLOG_001)
LS_SIZE(1180)
LS_DATACLAS(dataclas) LS_STORCLAS(storclas)
STG_DUPLEX(YES) DUPLEXMODE(COND)
HIGHOFFLOAD(80) LOWOFFLOAD(60)
DIAG(YES)
DEFINE LOGSTREAM NAME(IGWTV001.IGWSHUNT.SHUNTLOG)

STRUCTURE(LOG_IGWSHUNT_001)
LS_SIZE(100)
LS_DATACLAS(dataclas) LS_STORCLAS(storclas)
HIGHOFFLOAD(80) LOWOFFLOAD(0)
DIAG(YES)
/
Figure 7-44 Example of defining log streams
Defining log streams

Each log stream is assigned to a structure previously defined to the LOGR policy. You can
assign multiple log streams to the same structure, for example, for the forward recovery logs.
If you were using CICS forward recovery in the past, you do not need to define new log
streams for forward recovery. CICS and DFSMStvs can share the same forward recovery
logs.
Attention: Log streams are single-extent VSAM linear data sets and need
SHAREOPTIONS(3,3). The default is SHAREOPTIONS(1,3) so you must alter the share
options explicitly by running IDCAMS ALTER.
Update data set attributes

You must update your VSAM data sets with the LOG and LOGSTREAMID parameters in
order to use DFSMStvs. If you were using VSAM RLS already and do not want to change the
kind of recovery, no changes are necessary. See also 7.18, Update data sets with log
parameters on page 408.

7.33 Update PARMLIB with DFSMStvs parameters
SMS ACDS(acds) COMMDS(commds)

INTERVAL(nnn|15) DINTERVAL(nnn|150)
REVERIFY(YES|NO) ACSDEFAULTS(YES|NO)
SYSTEMS(8|32) TRACE(OFF|ON)
SIZE(nnnnnK|M) TYPE(ALL|ERROR)
JOBNAME(jobname|*) ASID(asid|*)
SELECT(event,event....) DESELECT(event,event....)
DSNTYPE(LIBRARY|PDS)
VSAM RLS
RLSINIT(NO|YES) RLS_MAX_POOL_SIZE(nnn|100)
SMF_TIME(NO|YES) CF_TIME(nnn|3600)
BMFTIME(nnn|3600) CACHETIME(nnn|3600)
DEADLOCK_DETECTION(iii|15,kkk|4) RLSTMOUT(nnn|0)
DFSMStvs
SYSNAME(sys1,sys2....) TVSNAME(nnn1,nnn2....)
TV_START_TYPE(WARM|COLD,WARM|COLD...) AKP(nnn|1000,nnn|1000)
LOG_OF_LOGS(logstream) QTIMEOUT(nnn|300)
MAXLOCKS(max|0,incr|0)
Figure 7-45 PARMLIB parameters to support DFSMStvs
New PARMLIB parameters to support DFSMStvs

There are a few new parameters you can specify in the PARMLIB member IGDSMSxx to
support DFSMStvs. Note that DFSMStvs requires VSAM RLS. For a description of the VSAM
RLS parameters, see 7.15, Update PARMLIB with VSAM RLS parameters on page 400.
These are the new parameters for DFSMStvs:

SYSNAME(sysname[,sysname]...)
This identifies the systems on which you want to run DFSMStvs.
This specifies the names of the systems to which the DFSMStvs instance names of the
TVSNAME parameter apply.
TVSNAME(nnn[,nnn]...)
This specifies the identifiers of the DFSMStvs instances on the systems you specified
in the SYSNAME parameter.
If only one TVSNAME is specified, this parameter applies only to the system on which
the PARMLIB member is read; in this case, no SYSNAME is required.
The number of sysnames and the number of tvsnames must be the same.
TV_START_TYPE({WARM|COLD}[,{WARM|COLD}]...)
This specifies the start type for the single DFSMStvs instances as they are listed in the
TVSNAME parameter.

AKP(nnn[,nnn]...)
This specifies the activity keypoint trigger values for the DFSMStvs instances as they
appear in the TVSNAME parameter.
AKP is the number of logging operations between keypoints.
Activity keypointing enables DFSMStvs to delete records from the undo or shunt log
that are no longer involved in active units of recovery.
LOG_OF_LOGS(logstream)
This specifies the name of the log stream that is used as a log of logs.
QTIMEOUT({nnn|300})
This specifies the amount of time the DFSMStvs quiesce exits allow to elapse before
concluding that a quiesce event cannot be completed successfully.
MAXLOCKS({max|0},{incr|0})
max is the maximum number of unique lock requests that a single unit of recovery can
make; when this value is reached, the system will issue a warning message.
incr is an increment value.
After max is reached, each time the number of locks increases by the incr value,
another warning message is issued.
For information about these PARMLIB parameters, see z/OS MVS Initialization and Tuning
Reference, SA22-7592.

7.34 The DFSMStvs instance
CICSA CICSB Batch1 Batch2 CICSC CICSD Batch3 Batch4

R/W R/W R/W R/W R/W R/W R/W R/W
IGWTV001 IGWTV002
SMSVSAM SMSVSAM
System 1 System 2
Coupling Facility
recoverable
data set
Figure 7-46 Sysplex with two active DFSMStvs instances
The DFSMStvs instance

DFSMStvs runs in the SMSVSAM address space as an instance. The name of the
SMSVSAM address space is always IGWTVnnn, where nnn is a unique number per system
as defined for TVSNAME in the PARMLIB member IGDSMSxx. This number can range from
0 to 255. You can define up to 32 DFSMStvs instances, one per system. An example is
IGWTV001.
As soon as an application that does not act as a recoverable resource manager has RLS
access to a recoverable data set, DFSMStvs is invoked (see also 7.29, Accessing a data set
with DFSMStvs on page 429). DFSMStvs calls VSAM RLS (SMSVSAM) for record locking
and buffering. With DFSMStvs built on top of VSAM RLS, full sharing of recoverable files
becomes possible. Batch jobs can now update the recoverable files without first quiescing
CICS' access to them.
As a recoverable resource manager, CICS interacts directly with VSAM RLS.

7.35 Interacting with DFSMStvs
Interacting with DFSMStvs includes

Use of SETSMS commands to change the
IGDSMSxx specifications
Use of SET SMS=xx command to activate a new
IGDSMSxx member
Use of VARY SMS commands
Use of display commands to get information about
the current DFSMStvs configuration
IDCAMS command SHCDS
Figure 7-47 Operator commands to interact with DFSMStvs
Interacting with DFSMStvs

This section provides a brief overview of several commands that were introduced particularly
for displaying and changing DFSMStvs-related information. You can use those commands in
addition to the commands we already described in 7.20, Interacting with VSAM RLS on
page 412.
SETSMS command
Use the SETSMS command to overwrite the PARMLIB specifications for IGDSMSxx. The syntax
is:
SETSMS AKP(nnn|1000)
QTIMEOUT(nnn|300)
MAXLOCKS(max|0,incr|0)
These are the only DFSMStvs PARMLIB specifications you can overwrite using the SETSMS
command. For information about these parameters, see 7.33, Update PARMLIB with
DFSMStvs parameters on page 436.
SET SMS=xx command

The command SET SMS=xx causes the IGDSMSxx member of SYS1.PARMLIB to be read. It
allows you to activate another IGDSMSxx member, where xx specifies the last two digits of
the member. You can change DFSMStvs parameters in IGDSMSxx that you cannot change
with the SETSMS command, and activate the changes afterwards by running SET SMS=xx.

Changes to parameters other than AKP, QTIMEOUT and MAXLOCKS take affect the next
time DFSMStvs restarts.
VARY SMS command

The VARY SMS command has the following functions for DFSMStvs:
Enable, quiesce, or disable one or all DFSMStvs instances:
VARY SMS,TRANVSAM(tvsname|ALL),{QUIESCE|Q}
{DISABLE|D}
{ENABLE|E}
This command is routed to all systems in the sysplex. If you specify ALL, it affects all
DFSMStvs instances. Otherwise, it only affects the instance specified by tvsname. If
QUIESCE is specified, DFSMStvs completes the current work first, but does not accept any
new work. If DISABLE is specified, DFSMStvs stops processing immediately.
Enable, quiesce, or disable DFSMStvs access to a specified logstream:
VARY SMS,LOG(logstream),{QUIESCE|Q}
{DISABLE|D}
{ENABLE|E}
Quiescing or disabling the DFSMStvs undo or shunt logstream is equivalent to quiescing
or disabling DFSMStvs processing respectively. Quiescing or disabling the log of logs has
no influence on DFSMStvs processing. Quiescing or disabling a forward recovery log
causes all attempts to process data sets which use this log stream to fail.
Start or stop peer recovery processing for a failed instance of DFSMStvs:
VARY SMS,TRANVSAM(tvsname),PEERRECOVERY,{ACTIVE}
{ACTIVEFORCE}
{INACTIVE}
This command applies only to the system on which it is issued. That system will then be
responsible for performing all peer recovery processing for a failed DFSMStvs instance.
For a discussion of the term peer recovery, see 7.24, Overview of DFSMStvs on
page 421.
Display command
There are a few display commands you can use to get information about DFSMStvs.
To display common DFSMStvs information:
DISPLAY SMS,TRANVSAM{,ALL}
This command lists information about the DFSMStvs instance on the system were it was
issued. To get information from all systems use ALL. This information includes name and
state of the DFSMStvs instance, values for AKP, start type, and qtimeout, and also the
names, types, and states of the used log streams.
To display information about a particular job that uses DFSMStvs:
DISPLAY SMS,JOB(jobname)
The information about the particular job includes the current job step, the current ID, and
status of the unit of recovery used by this job.
To display information about a particular unit of recovery currently active within the
sysplex:
DISPLAY SMS,URID(urid|ALL)
This command provides information about a particular UR in the sysplex or about all URs
of the system on which this command was issued. If ALL is specified, you do not obtain

information about shunted URs and URs that are restarting. The provided information
includes age and status of the UR, the jobname with which this UR is associated, and the
current step within the job.
To display entries currently contained in the shunt log:
DISPLAY SMS,SHUNTED,{SPHERE(sphere)|
URID(urid|ALL)}
Entries are moved to the shunt log when DFSMStvs is unable to finish processing a sync
point for them. There are several reasons this might occur; an I/O error is one of the
possible causes. Depending on what was specified, you get information for a particular
VSAM sphere, a particular UR, or for all shunted URs in the sysplex.
To display information about logstreams that DFSMStvs is currently using:
DISPLAY SMS,LOG(logstream|ALL)
If ALL is specified, information about all log streams in use is provided from the system on
which the command is issued. The output includes the status and type of the log stream,
the job name and URID of the oldest unit of recovery using the log, and also a list of all
DFSMStvs instances that are using the log.
To display information about a particular data set:
DISPLAY SMS,DSNAME(dsn)
Use this command to display jobs currently accessing the data set using DFSMStvs
access on the systems within the sysplex.
IDCAMS command SHCDS

The IDCAMS command SHCDS was enhanced to display and modify recovery information
related to DFSMStvs. For more information about this command, see z/OS DFSMS Access
Method Services for Catalogs, SC26-7394.
Important: This chapter simply provides an overview of new operator commands to know
to work with DFSMStvs. Before using these commands (other than the DISPLAY
command), read the official z/OS manuals carefully.

7.36 Summary
Summary
Base VSAM
Functions and limitations
VSAM record-level sharing

Functions and limitations
DFSMStvs
Functions
Figure 7-48
Summary
In this chapter we showed the limitations of base VSAM that made it necessary to develop
VSAM RLS. Further, we exposed the limitations of VSAM RLS that were the reason to
enhance VSAM RLS by the functions provided by DFSMStvs.
Base VSAM
VSAM does not provide read or read/write integrity for share options other than 1.
User needs to use enqueue/dequeue macros for serialization.
The granularity of sharing on a VSAM cluster is at the control interval level.
Buffers reside in the address space.
Base VSAM does not support CICS as a recoverable resource manager; a CICS file
owning region is necessary to ensure recovery.
VSAM RLS
Enhancement of base VSAM.
User does not need to serialize; this is done by RLS locking.
Granularity of sharing is record level, not CI level.
Buffers reside in the data space and Coupling Facility.
Supports CICS as a recoverable resource manager (CICS logging for recoverable data
sets); no CICS file owning region is necessary.

Does not act as a recoverable resource manager (provides no logging for recoverable
data sets).
Sharing of recoverable data sets that are in use by CICS with non-CICS applications
like batch jobs is not possible.
DFSMStvs
Enhancement of VSAM RLS.
Uses VSAM RLS to access VSAM data sets.
Acts as a recoverable resource manager, thus provides logging.
Enables non-CICS applications like batch jobs to share recoverable data sets with
CICS.

8
Chapter 8. Storage management hardware

The use of DFSMS requires storage management hardware that includes both direct access
storage device (DASD) and tape device types. In this chapter we provide an overview of both
storage device categories and a brief introduction to RAID technology.
For many years, DASDs have been the most used storage devices on IBM eServer zSeries
systems and their predecessors, delivering the fast random access to data and high
availability that customers have come to expect.
We cover the following types of DASD:

Traditional DASD (such as 3380 and 3390)
DS6000 and DS8000
The era of tapes began before DASD was introduced. During that time, tapes were used as
the primary application storage medium. Today customers use tapes for such purposes as
backup, archiving, or data transfer between companies.
The following types of tape devices are described:

Traditional tapes like 3480 and 3490
IBM Magstar 3590 and 3592
Automated tape library (ATL) 3494
Virtual tape server (VTS).
We also briefly explain the storage area network (SAN) concept.

8.1 Overview of DASD types
Traditional DASD
3380 Models J, E, K
3390 Models 1, 2, 3, 9
DASD based on RAID technology and Seascape

architecture
DS6000 and DS8000 series
Figure 8-1 Overview of DASD types
Traditional DASD
In the era of traditional DASD, the hardware consisted of controllers like 3880 and 3990,
which contained the necessary intelligent functions to operate a storage subsystem. The
controllers were connected to S/390 systems through parallel or ESCON channels. Behind a
controller there were several model groups of the 3390 that contained the disk drives. Based
on the models, these disk drives had various capacities per device. Within each model group,
the various models provide either four, eight, or twelve devices. All A-units come with four
controllers, providing a total of four paths to the 3990 Storage Control. At that time, you were
not able to change the characteristics of a given DASD device.
DASD based on RAID technology

With the introduction of the RAMAC Array in 1994, IBM first introduced storage subsystems
for S/390 systems based on RAID technology. We discuss the various RAID implementations
in Figure 8-2 on page 448.
The more modern IBM DASD products, such as Enterprise Storage Server (ESS), DS6000,
DS8000, and DASD from other vendors, emulate IBM 3380 and 3390 volumes in geometry,
capacity of tracks, and number of tracks per cylinder. This emulation makes all the other
entities think they are dealing with real 3380s or 3390s. Among these entities, we have data
processing people not working directly with storage, JCL, MVS commands, open routines,
access methods, IOS, and channels. One benefit of this emulation is that it allows DASD
manufacturers to implement changes in the real disks, including the geometry of tracks and
cylinders, without affecting the way those components interface with DASD. From an

operating system point of view, device types always will be 3390s, sometimes with much
higher numbers of cylinders, but 3390s nonetheless.
ESS technology
The IBM TotalStorage Enterprise Storage Server (ESS) is the IBM disk storage server,
developed using IBM Seascape architecture. The ESS provides functionality to the family of
e-business servers, and also to non-IBM (that is, Intel-based and UNIX-based) families of
servers. Across all of these environments, the ESS features unique capabilities that allow it to
meet the most demanding requirements of performance, capacity, and data availability that
the computing business requires. See 8.4, Enterprise Storage Server (ESS) on page 453 for
more information about this topic.
Seascape architecture
The Seascape architecture is the key to the development of the IBM storage products.
Seascape allows IBM to take the best of the technologies developed by the many IBM
laboratories and integrate them, thereby producing flexible and upgradeable storage
solutions. This Seascape architecture design has allowed the IBM TotalStorage Enterprise
Storage Server to evolve from the initial E models to the succeeding F models, and to the
later 800 models, each featuring new, more powerful hardware and functional enhancements,
and always integrated under the same successful architecture with which the ESS was
originally conceived. See 8.3, Seascape architecture on page 450 for more information.
Note: In this publication, we use the terms disk or head disk assembly (HDA) for the real
devices, and the terms DASD volumes or DASD devices for the logical 3380/3390s.
Chapter 8. Storage management hardware 447

8.2 Redundant array of independent disks (RAID)
RAID Disks
Primary Alternate
Record X Record X
Raid-1 ABCDEF ABCDEF
Data Data Data Parity
Raid- 3 1/3 Record X 1/3 Record X 1/3 Record X Parity bits

AB EF PP
CD
Data+ Parity Data+ Parity Data+ Parity Data+ Parity

Record X Record Y Record B Parity
Raid-5 ABCDEF #IJKLM PQRSTU PPPPPP
Record W Parity Record V Record T
TRSVAB PPPPPP CDERST QRUBXA
Figure 8-2 Redundant array of independent disks (RAID)
RAID architecture
Redundant array of independent disks (RAID) is a direct access storage architecture where
data is recorded across multiple physical disks with parity separately recorded, so that no loss
of access to data results from the loss of any one disk in the array.
RAID breaks the one-to-one association of volumes with devices. A logical volume is now the
addressable entity presented by the controller to the attached systems. The RAID unit maps
the logical volume across multiple physical devices. Similarly, blocks of storage on a single
physical device may be associated with multiple logical volumes. Because a logical volume is
mapped by the RAID unit across multiple physical devices, it is now possible to overlap
processing for multiple cache misses to the same logical volume because cache misses can
be satisfied by separate physical devices.
The RAID concept involves many small computer system interface (SCSI) disks replacing a
big one. The major RAID advantages are:
Performance (due to parallelism)
Cost (SCSI are commodities)
zSeries compatibility
Environment (space and energy)
However, RAID increased the chances of malfunction due to media and disk failures and the
fact that the logical device is now residing on many physical disks. The solution was

redundancy, which wastes space and causes performance problems as write penalty and
free space reclamation. To address this performance issue, large caches are implemented.
Note: The ESS storage controllers use the RAID architecture that enables multiple logical
volumes to be mapped on a single physical RAID group. If required, you can still separate
data sets on a physical controller boundary for the purpose of availability.
RAID implementations
Except for RAID-1, each manufacturer sets the number of disks in an array. An array is a set
of logically related disks, where a parity applies.
Various implementations certified by the RAID Architecture Board are:

RAID-1 This has simply disk mirroring, like dual copy.
RAID-3 This has an array with one dedicated parity disk and just one I/O request at a
time, with intra-record striping. It means that the written physical block is striped
and each piece (together with the parity) is written in parallel in each disk of the
array. The access arms move together. It has a high data rate and a low I/O rate.
RAID-5 This has an array with one distributed parity (there is no dedicated disk for
parities). It does I/O requests in parallel with extra-record striping, meaning each
physical block is written in each disk. The access arms move independently. It
has strong caching to avoid write penalties; that is, four disk I/Os per write.
RAID-5 has a high I/O rate and a medium data rate. RAID-5 is used by the IBM
2105 controller with 8-disk arrays in the majority of configurations.
RAID-5 does the following:
It reads data from an undamaged disk. This is one single disk I/O operation.
It reads data from a damaged disk, which implies (n-1) disk I/Os, to recreate
the lost data where n is the number of disks in the array.
For every write to an undamaged disk, RAID-5 does four I/O operations to
store a correct parity block; this is called a write penalty. This penalty can be
relieved with strong caching and a slice triggered algorithm (coalescing disks
updates from cache into a single parallel I/O).
For every write to a damaged disk, RAID-5 does n-1 reads and one parity
write.
RAID-6 This has an array with two distributed parity and I/O requests in parallel with
extra-record striping. Its access arms move independently (Reed/Salomon P-Q
parity). The write penalty is greater than RAID-5, with six I/Os per write.
RAID-6+ This is without write penalty (due to log-structured file, or LFS), and has
background free-space reclamation. The access arms all move together for
writes. It is used by the RVA controller.
RAID-10 RAID-10 has a new RAID architecture designed to give performance for striping
and has redundancy for mirroring. RAID-10 is optionally implemented in the
IBM 2105.
Note: Data striping (stripe sequential physical blocks in separate disks) is sometimes
called RAID-0, but it is not a real RAID because there is no redundancy, that is, no parity
bits.

8.3 Seascape architecture
Powerful storage server

Snap-in building blocks
Universal data access
Storage sharing
Data copy sharing
Network
Direct channel
Shared storage transfer
True data sharing
Figure 8-3 Seascape architecture
Seascape architecture
The IBM Enterprise Storage Servers architecture for e-business design is based on the IBM
storage enterprise architecture, Seascape. The Seascape architecture defines
next-generation concepts for storage by integrating modular building block technologies from
IBM, including disk, tape, and optical storage media, powerful processors, and rich software.
Integrated Seascape solutions are highly reliable, scalable, and versatile, and support
specialized applications on servers ranging from PCs to super computers. Virtually all types
of servers can concurrently attach to the ESS, including iSeries and AS/400 systems. As a
result, ESS can be the external disk storage system of choice for AS/400 as well as iSeries
systems in heterogeneous SAN environments.
Seascape has three basic concepts:

DFSMS provides device support for the IBM 2105 Enterprise Storage Server (ESS), a
high-end storage subsystem. The ESS storage subsystem succeeded the 3880, 3990, and
9340 subsystem families. Designed for mid-range and high-end environments, the ESS gives
you large capacity, high performance, continuous availability, and storage expandability. You
can read more about ESS in 8.4, Enterprise Storage Server (ESS) on page 453.

The ESS was the first of the Seascape architecture storage products to attach directly to IBM
System z and open system platforms. The Seascape architecture products come with
integrated storage controllers. These integrated storage controllers allow the attachment of
physical storage devices that emulate 3390 Models 2, 3, and 9, or provide 3380 track
compatibility mode.

The storage system is intelligent and independent, and it can be reached by channels or
through the network. It is powered by a set of fast RISC processors.

Each Seascape product is comprised of building blocks, such as:
Scalable n-way RISC server, PCI-based. This provides the logic of the storage server.
Memory cache from RISC processor memory.
Channel attachments, such as FC-AL, SCSI, ESCON, FICON and SSA.
Network attachments, such as Ethernet, FDDI, TR, and ATM.
These attachments can also implement functions, that is, a mix of network interfaces (to
be used as a remote and independent storage server) and channel interfaces (to be used
as a storage controller interface).
Software building blocks, such as an AIX subset, Java applications, and Tivoli Storage
Manager. High level language (HLL) is more flexible than microcode, and is easier to write
and maintain.
Storage adapters, for mixed storage devices technologies.
Storage device building blocks, such as serial disk (7133), 3590 tape (Magstar), and
optical (3995).
Silos and robots (3494).

Universal data access allows a wide array of connectivity, such as z/OS, UNIX, Linux, and
OS/400, to common data. There are three types of universal access: storage sharing, data
copy sharing, and true data sharing, as explained here.
Storage sharing
Physical storage (DASD or tape) is statically divided into fixed partitions available to a
given processor. It is not a software function. The subsystem controller knows which
processors own which storage partitions. In a sense, only capacity is shared, not data; one
server cannot access the data of the other server. It is required that the manual
reassignment of storage capacity between partitions be simple and nondisruptive.
The benefits are:
Purchase higher quantities with greater discounts
Only one type of storage to manage
Static shifting of capacity as needed
The drawbacks are:
Higher price for SCSI data
Collocation at 20 meters of the SCSI servers
No priority concept between z/OS and UNIX/NT I/O requests

Data copy sharing
Data copy sharing is an interim data replication solution (waiting for a true data sharing)
done by data replication from one volume accessed by a platform to another volume
accessed by another platform. The replication can be done through software or
hardware.There are three ways to implement data copy sharing:
Network: Via network data transfer, such as SNA or TCP/IP. This method has
drawbacks, such as CPU and network overhead; it is still slow and expensive for
massive data transfer.
Direct channel: Direct data transfer between the processors involved using channel or
bus capabilities, referred to as bulk data transfer.
Shared storage transfer: Writing an intermediate flat file by software into the storage
subsystem cache, that is read (and translated) by the receiving processor, so the
storage is shared.
True data sharing
For data sharing between multiple platforms for read/write of a single copy that addresses
the complex issues of mixed data types, file structures, databases, and SCPs, there is no
available solution.

8.4 Enterprise Storage Server (ESS)
Two 6-way RISC processors (668 MHZ)

4.8 GB/sec of aggregate bandwidth
Up to 32 ESCON/SCSI/mixed
Up to 16 FICON and FCP channels
Up to 64 GB of cache
2 GB of NVS cache
18.2/36.4/72.8/145.6 GB capacity disk options
Up to 55.9 TB capacity
8 x 160 MB/sec SSA loops
10,000 rpm and 15,000 rpm disk options
Connects to SAN
RAID-5 or RAID-10
Figure 8-4 Enterprise Storage Server model 800

The IBM Enterprise Storage Server (ESS) is a high performance, high availability capacity
storage subsystem. It contains two six-way RISC processors (668 MHZ) with up to 64 GB
cache and 2 GB of non-volatile storage (NVS) to protect from data loss during power outages.
Connectivity to IBM mainframes is through up to 32 ESCON channels and up to 16 FICON
channels. For other platforms, such as IBM System i, UNIX, or NT, the connectivity is
through up to 32 SCSI interfaces.
Cache
Cache is used to store both read and write data to improve ESS performance to the attached
host systems. There is the choice of 8, 16, 24, 32, or 64 GB of cache. This cache is divided
between the two clusters of the ESS, giving the clusters their own non-shared cache. The
ESS cache uses ECC (error checking and correcting) memory technology to enhance
reliability and error correction of the cache. ECC technology can detect single- and double-bit
errors and correct all single-bit errors. Memory scrubbing, a built-in hardware function, is also
performed and is a continuous background read of data from memory to check for correctable
errors. Correctable errors are corrected and rewritten to cache. To protect against loss of data
on a write operation, the ESS stores two copies of written data, one in cache and the other in
NVS.

NVS cache
NVS is used to store a second copy of write data to ensure data integrity if there is a power
failure or a cluster failure and the cache copy is lost. The NVS of cluster 1 is located in cluster
2 and the NVS of cluster 2 is located in cluster 1. In this way, in the event of a cluster failure,
the write data for the failed cluster will be in the NVS of the surviving cluster. This write data is
then de-staged at high priority to the disk arrays. At the same time, the surviving cluster will
start to use its own NVS for write data, ensuring that two copies of write data are still
maintained. This ensures that no data is lost even in the event of a component failure.
ESS Model 800

The ESS Model 800 has a 2 GB NVS. Each cluster has 1 GB of NVS, made up of four cards.
Each pair of NVS cards has its own battery-powered charger system that protects data even
if power is lost on the entire ESS for up to 72 hours. This model has the following
enhancements:
Model 800 allows 4.8 GBps of aggregate bandwidth.
In the disk interface the ESS has eight Serial Storage Architecture (SSA) loops, each one
with a rate of 160 MBps for accessing the disks. See SSA loops on page 464 for more
information about this topic.
ESS implements RAID-5 or RAID-10 for availability and has eight disks in the majority of
the arrays. See RAID-10 on page 466 for more information about this topic.
Four disks sizes of 18.2, 36.4, 72.8, and 145.6 GB, which can be intermixed. The ESS
maximum capacity is over 55.9 TB with a second frame attached.
ESS Model 750

The ESS 750 is intended for smaller enterprises that need the enterprise-level advanced
functions, reliability, and availability offered by the Model 800, but at a lower entry cost and
size. It is specifically designed to meet the high demands of medium-sized mainframe
environments, and for this reason it is closely tied in with the IBM z890 offering.
The ESS 750 has capabilities similar to the ESS 800. The ESS Model 750 consists of two
clusters, each with a two-way processor and 4 or 8 GB cache. It can have two to six Fibre
Channel/FICON or ESCON host adapters. The storage capacity ranges from a minimum of
1.1 TB up to a maximum of 4 TB. A key feature is that the ESS 750 is upgradeable,
non-disruptively, to the ESS Model 800, which can grow to more than 55 TB of physical
capacity.
Note: Effective April 28, 2006, IBM withdrew from marketing the following products:
IBM TotalStorage Enterprise Storage Server (ESS) Models 750 and 800
IBM Standby Capacity on Demand for ESS offering
For information about replacement products, see 8.16, IBM TotalStorage DS6000 on
page 474 and 8.17, IBM TotalStorage DS8000 on page 477.
SCSI protocol
Although we do not cover other platforms in this publication, we provide here a brief overview
of the SCSI protocol. The SCSI adapter is a card on the host. It connects to a SCSI bus
through a SCSI port. There are two types of SCSI supported by ESS:
SCSI Fast Wide with 20 MBps
Ultra SCSI Wide with 40 MBps

8.5 ESS universal access
Storage consolidation
Storage sharing solution UNIX

IBM System z
with dynamic reallocation
PPRC available for XP and

UNIX - Human control
through Web interface ESS
StorWatch support
IBM System i Windows XP
Figure 8-5 ESS universal access
ESS universal access

ESS is a product designed to implement storage consolidation that puts all of your enterprise
data under the same cover. This consolidation is the first step in achieving server
consolidation, that is, to put all of your enterprise applications under the same z/OS cluster.
PPRC support
IBM includes a Web browser interface called TotalStorage Enterprise Storage Server (ESS)
Copy Services. The interface is part of the ESS subsystem and can be used to perform
FlashCopy and PPRC functions.
Many of the ESS features are now available to non-zSeries platforms, such as PPRC for
Windows XP and UNIX, where the control is through a Web interface.
StorWatch support
On the software side, there is StorWatch, a range of products in UNIX/XP that does what
DFSMS and automation do for System z. The TotalStorage Expert, formerly marketed as
StorWatch Expert, is a member of the IBM and Tivoli Systems family of solutions for
Enterprise Storage Resource Management (ESRM). These are offerings that are designed to
complement one another, and provide a total storage management solution.
TotalStorage Expert is an innovative software tool that gives administrators powerful, yet
flexible storage asset, capacity, and performance management capabilities to centrally
manage Enterprise Storage Servers located anywhere in the enterprise.

8.6 ESS major components
Cages and disk drives
SMP Processors and

Cache
SSA Device Adapters
Host Adapters
Main Power
Supplies
Batteries
Figure 8-6 ESS major components
ESS Model 800 major components

Figure 8-6 shows an IBM TotalStorage Enterprise Storage Server Model 800 and its major
components. As you can see, the ESS base rack consists of two clusters, each with its own
power supplies, batteries, SSA device adapters, processors, cache and NVS, CD drive, hard
disk, floppy disk, and network connections. Both clusters have access to any host adapter
card, even though they are physically spread across the clusters.
At the top of each cluster is an ESS cage. Each cage provides slots for up to 64 disk drives,
32 in front and 32 at the back.
This storage box has two enclosures:

A base enclosure, with:
Two 3-phase power supplies
Up to 128 disk drives in two cages
Feature #2110 for Expansion Enclosure attachment
Host adapters, cluster processors, cache and NVS, and SSA device adapters
An expansion enclosure, with:
Two 3-phase power supplies
Up to 256 disk drives in four cages for additional capacity

8.7 ESS host adapters
Host adapter bays

4 bays
4 host adapters per bay
ESCON host adapters

Up to 32 ESCON links
2 ESCON links per host adapter
2 Gb FICON host adapters

Up to 16 FICON links
1 FICON link per host adapter
Auto speed detection - 1 Gb or 2 Gb
SCSI host adapters

Up to 32 SCSI bus connections
2 SCSI ports per host adapter
2 Gb Fibre Channel host adapters Adapters can be intermixed

Up to 16 Fibre Channel links Any combination of host
adapter cards up to a
1 Fibre Channel port per host adapter maximum of 16
Auto speed detection - 1 Gb or 2 Gb
Figure 8-7 Host adapters
ESS host adapters

The ESS has four host adapter bays, two in each cluster. Each bay supports up to four host
adapter cards. Each of these host adapter cards can be for FICON, ESCON, SCSI, or Fibre
Channel server connection. Figure 8-7 lists the main characteristics of the ESS host
adapters.
Each host adapter can communicate with either cluster. To install a new host adapter card,
the bay must be powered off. For the highest path availability, it is important to spread the host
connections across all the adapter bays. For example, if you have four ESCON links to a host,
each connected to a separate bay, then the loss of a bay for upgrade only impacts one of the
four connections to the server. The same is also valid for a host with FICON connections to
the ESS.
Similar considerations apply for servers connecting to the ESS by means of SCSI or Fibre
Channel links. For open system servers, the Subsystem Device Driver (SDD) program that
comes standard with the ESS can be installed on the connecting host servers to provide
multiple paths or connections to handle errors (path failover) and balance the I/O load to the
ESS.
The ESS connects to a large number of servers, operating systems, host adapters, and SAN
fabrics. A complete and current list is available at the following Web site:
http://www.storage.ibm.com/hardsoft/products/ess/supserver.htm

8.8 FICON host adapters
FICON host adapters

Up to 16 FICON host adapters
One port with an LC connector type
per adapter (2 Gigabit Link)
Long wave or short wave
Up to 200 MB/sec full duplex
Up to 10 km distance with long wave
and 300 m with short wave
Each host adapter communicates
with both clusters
Each FICON channel link can
address all 16 ESS CU images
Logical paths FICON
256 CU logical paths per FICON port zSeries
4096 logical paths per ESS
Addresses
16,384 device addresses per channel
FICON distances
10 km distance (without repeaters)
100 km distance (with extenders)
Figure 8-8 FICON host adapter
FICON host adapters

FICON, or Fiber Connection, is based on the standard Fibre Channel architecture, and
therefore shares the attributes associated with Fibre Channel. This includes the common
FC-0, FC-1, and FC-2 architectural layers, the 100 MBps bidirectional (full-duplex) data
transfer rate, and the point-to-point distance capability of 10 kilometers. The ESCON
protocols have been mapped to the FC-4 layer, the Upper Level Protocol (ULP) layer, of the
Fibre Channel architecture. All this provides a full-compatibility interface with previous S/390
software and puts the zSeries servers in the Fibre Channel industry standard.
FICON versus ESCON

FICON goes beyond ESCON limits:
Addressing limit, from 1024 device addresses per channel to up to 16,384 (maximum of
4096 devices supported within one ESS).
Up to 256 control unit logical paths per port.
FICON channel to ESS allows multiple concurrent I/O connections (the ESCON channel
supports only one I/O connection at one time).
Greater channel and link bandwidth: FICON has up to 10 times the link bandwidth of
ESCON (1 Gbps full-duplex, compared to 200 MBps half duplex). FICON has up to more
than four times the effective channel bandwidth.
FICON path consolidation using switched point-to-point topology.

Greater unrepeated fiber link distances (from 3 km for ESCON to up to 10 km, or 20 km
with an RPQ, for FICON).
These characteristics allow simpler and more powerful configurations. The ESS supports up
to 16 host adapters, which allows for a maximum of 16 Fibre Channel/FICON ports per
machine, as shown in Figure 8-8 on page 458.
Each Fibre Channel/FICON host adapter provides one port with an LC connector type. The
adapter is a 2 Gb card and provides a nominal 200 MBps full-duplex data rate. The adapter
will auto-negotiate between 1 Gb and 2 Gb, depending upon the speed of the connection at
the other end of the link. For example, from the ESS to a switch/director, the FICON adapter
can negotiate to 2 Gb if the switch/director also has 2 Gb support. The switch/director to host
link can then negotiate at 1 Gb.
Host adapter cards

There are two types of host adapter cards you can select: long wave (feature 3024), and short
wave (feature 3025). With long-wave laser, you can connect nodes at distances of up to
10 km (without repeaters). With shortwave laser, you can connect at distances of up to
300 m. These distances can be extended using switches or directors.

8.9 ESS disks
Eight-packs
Set of 8 similar capacity/rpm disk drives packed
together
Installed in the ESS cages
Initial minimum configuration is 4 eight-packs
Upgrades are available increments of 2 eight-packs
Maximum of 48 eight-packs per ESS with expansion
Disk drives
18.2 GB 15,000 rpm or 10,000 rpm
36.4 GB 15,000 rpm or 10,000 rpm
72.8 GB 10,000 rpm
145.6 GB 10,000 rpm
Eight-pack conversions
Capacity and/or RPMs
Figure 8-9 ESS disks
ESS disks
With a number of disk drive sizes and speeds available, including intermix support, the ESS
provides a great number of capacity configuration options.
The maximum number of disk drives supported within the IBM TotalStorage Enterprise
Storage Server Model 800 is 384, with 128 disk drives in the base enclosure and 256 disk
drives in the expansion rack. When configured with 145.6 GB disk drives, this gives a total
physical disk capacity of approximately 55.9 TB (see Table 8-1 on page 461 for more details).
Disk drives
The minimum available configuration of the ESS Model 800 is 582 GB. This capacity can be
configured with 32 disk drives of 18.2 GB contained in four eight-packs, using one ESS cage.
All incremental upgrades are ordered and installed in pairs of eight-packs; thus the minimum
capacity increment is a pair of similar eight-packs of either 18.2 GB, 36.4 GB, 72.8 GB, or
145.6 GB capacity.
The ESS is designed to deliver substantial protection against data corruption, not just relying
on the RAID implementation alone. The disk drives installed in the ESS are the latest
state-of-the-art magneto resistive head technology disk drives that support advanced disk
functions such as disk error correction codes (ECC), Metadata checks, disk scrubbing, and
predictive failure analysis.

Eight-pack conversions
The ESS eight-pack is the basic unit of capacity within the ESS base and expansion racks. As
mentioned before, these eight-packs are ordered and installed in pairs. Each eight-pack can
be configured as a RAID 5 rank (6+P+S or 7+P) or as a RAID 10 rank (3+3+2S or 4+4).
The IBM TotalStorage ESS Specialist will configure the eight-packs on a loop with spare
DDMs as required. Configurations that include drive size intermixing can result in the creation
of additional DDM spares on a loop as compared to non-intermixed configurations. Currently
there is the choice of four new-generation disk drive capacities for use within an eight-pack:
18.2 GB/15,000 rpm disks
Also available is the option to install eight-packs with:

18.2 GB/10,000 rpm disks or
The eight disk drives assembled in each eight-pack are all of the same capacity. Each disk
drive uses the 40 MBps SSA interface on each of the four connections to the loop.
It is possible to mix eight-packs of various capacity disks and speeds (rpm) within an ESS, as
described in the following sections.
Use Table 8-1 as a guide for determining the capacity of a given eight-pack. This table shows
the capacities of the disk eight-packs when configured as RAID ranks. These capacities are
the effective capacities available for user data.
Table 8-1 Disk eight-pack effective capacity chart (gigabytes)

Physical Effective usable capacity (2)
capacity
(raw RAID 10 RAID 5 (3)
Disk capacity)
3 + 3 + 2S 4+4 6+P+S 7+P
size
Array (4) Array (5) Array (6) Array (7)
18.2 145.6 52.50 70.00 105.20 122.74
36.4 291.2 105.12 140.16 210.45 245.53
72.8 582.4 210.39 280.52 420.92 491.08
145.6 1,164.8 420.78 561.04 841.84 982.16

8.10 ESS device adapters
SSA 160 Device Adapters Up to 48 disk drives per loop

4 DA pairs per subsystem Mix of RAID 5 and RAID 10 eight-packs
Each RAID 5 array is 8 disk drives:
4 x 40 MB/sec loop data rate 6+P+S or
2 loops per device adapter pair 7+P
Each RAID 10 array is 8 disk drives:
3+3+2S or
4+4
Mix of different capacity disk drives
2 spares per loop per disk capacity
Same rpm for same capacity disks
S : representation of spare drive
A A A S B B B B C C C C D D D D 1 2 3 S 1 2 3 4
DA A A A A S B B B C C C C D D D D 1' 2' 3' S 1' 2 ' 3' 4'

DA
48 disks
1/2/3/4 1'/2'/3'/4': representation of RAID 10 rank drives
A/B/C/D : representation of RAID 5 rank drives (user data and distributed parity)
Figure 8-10 Device adapters
ESS device adapters

Device adapters (DA) provide the connection between the clusters and the disk drives. The
ESS Model 800 implements faster Serial Storage Architecture (SSA) device adapters than its
predecessor models.
The ESS Storage Server Model 800 uses the latest SSA160 technology in its device adapters
(DAs). With SSA 160, each of the four links operates at 40 MBps, giving a total nominal
bandwidth of 160 MBps for each of the two connections to the loop. This amounts to a total of
320 MBps across each loop. Also, each device adapter card supports two independent SSA
loops, giving a total bandwidth of 320 MBps per adapter card. There are eight adapter cards,
giving a total nominal bandwidth capability of 2,560 MBps. See 8.11, SSA loops on
page 464 for more information about this topic.
SSA loops
One adapter from each pair of adapters is installed in each cluster, as shown in Figure 8-10.
The SSA loops are between adapter pairs, which means that all the disks can be accessed by
both clusters. During the configuration process, each RAID array is configured by the IBM
TotalStorage ESS Specialist to be normally accessed by only one of the clusters. If a cluster
failure occurs, the remaining cluster can take over all the disk drives on the loop.
RAID 5 and RAID 10

RAID 5 and RAID 10 are managed by the SSA device adapters. RAID 10 is explained in
detail in 8.12, RAID-10 on page 466. Each loop supports up to 48 disk drives, and each

adapter pair supports up to 96 disk drives. There are four adapter pairs supporting up to 384
disk drives in total. Figure 8-10 on page 462 shows a logical representation of a single loop
with 48 disk drives (RAID ranks are actually split across two eight-packs for optimum
performance). You can see there are six RAID arrays: four RAID 5 designated A to D, and two
RAID 10 (one 3+3+2 spare and one 4+4).
Disk drives per loop

Each loop supports up to 48 disk drives, and each adapter pair supports up to 96 disk drives.
There are four adapter pairs supporting up to 384 disk drives in total.
Figure 8-10 on page 462 shows a logical representation of a single loop with 48 disk drives
(RAID ranks are actually split across two eight-packs for optimum performance). In the figure
you can see there are six RAID arrays: four RAID 5 designated A to D, and two RAID 10 (one
3+3+2 spare and one 4+4).

8.11 SSA loops
read
SSA operation write
4 links per loop
2 read and 2 write DA
simultaneously in each direction
40 MB/sec on each link
write
read
Loop availability
DA
Loop reconfigures itself
dynamically
write
read
Spatial reuse
Up to 8 simultaneous
DA DA
operations to local group of
disks (domains) per loop
Figure 8-11 SSA loops
SSA operation
SSA is a high performance, serial connection technology for disk drives. SSA is a full-duplex
loop-based architecture, with two physical read paths and two physical write paths to every
disk attached to the loop. Data is sent from the adapter card to the first disk on the loop and
then passed around the loop by the disks until it arrives at the target disk. Unlike bus-based
designs, which reserve the whole bus for data transfer, SSA only uses the part of the loop
between adjacent disks for data transfer. This means that many simultaneous data transfers
can take place on an SSA loop, and it is one of the main reasons that SSA performs so much
better than SCSI. This simultaneous transfer capability is known as spatial release.
Each read or write path on the loop operates at 40 MBps, providing a total loop bandwidth of
160 MBps.
Loop availability
The loop is a self-configuring, self-repairing design that allows genuine hot-plugging. If the
loop breaks for any reason, then the adapter card will automatically reconfigure the loop into
two single loops. In the ESS, the most likely scenario for a broken loop is if the actual disk
drive interface electronics fails. If this happens, the adapter card dynamically reconfigures the
loop into two single loops, effectively isolating the failed disk. If the disk is part of a RAID
array, the adapter card will automatically regenerate the missing disk using the remaining
data and parity disks to the spare disk. After the failed disk is replaced, the loop is
automatically reconfigured into full duplex operation and the replaced disk becomes a new
spare.

Spatial reuse
Spatial reuse allows domains to be set up on the loop. A domain means that one or more
groups of disks belong to one of the two adapter cards, as is the case during normal
operation. The benefit of this is that each adapter card can talk to its domains (or disk groups)
using only part of the loop. The use of domains allows each adapter card to operate at
maximum capability because it is not limited by I/O operations from the other adapter.
Theoretically, each adapter card can drive its domains at 160 MBps, giving 320 MBps
throughput on a single loop. The benefit of domains can reduce slightly over time, due to disk
failures causing the groups to become intermixed, but the main benefits of spatial reuse will
still apply.
If a cluster fails, then the remaining cluster device adapter owns all the domains on the loop,
thus allowing full data access to continue.

8.12 RAID-10
RAID-10 configurations: Eight-pack pair 1

Eight-pack 2
First RAID-10 rank Data Data Data Spare Data Data Data Data
configured in the loop will be: 1 2 3 S 1 2 3 4

3 + 3 + 2S Eight-pack 1
Additional RAID-10 ranks Data Data Data Spare Data Data Data Data
configured in the loop will be 1' 2' 3' S 1' 2' 3' 4'
4+4 Eight-pack pair 2
Eight-pack 4
For a loop with an intermixed
Data Data Data Data Data Data Data Data
capacity, the ESS will assign
1 2 3 4 1 2 3 4
two spares for each capacity.
Eight-pack 3
This means there will be one
Data Data Data Data Data Data Data Data
3+3+2S array per capacity
1' 2' 3' 4' 1' 2' 3' 4'
Figure 8-12 RAID-10
RAID-10
RAID-10 is also known as RAID 0+1 because it is a combination of RAID 0 (striping) and
RAID 1 (mirroring). The striping optimizes the performance by striping volumes across
several disk drives (in the ESS Model 800 implementation, three or four DDMs). RAID 1 is the
protection against a disk failure provided by having a mirrored copy of each disk. By
combining the two, RAID 10 provides data protection and I/O performance.
Array
A disk array is a group of disk drive modules (DDMs) that are arranged in a relationship, for
example, a RAID 5 or a RAID 10 array. For the ESS, the arrays are built upon the disks of the
disk eight-packs.
Disk eight-pack
The physical storage capacity of the ESS is materialized by means of the disk eight-packs.
These are sets of eight DDMs that are installed in pairs in the ESS. Two disk eight-packs
provide for two disk groups, with four DDMs from each disk eight-pack. These disk groups
can be configured as either RAID-5 or RAID-10 ranks.
Spare disks
The ESS requires that a loop have a minimum of two spare disks to enable sparing to occur.
The sparing function of the ESS is automatically initiated whenever a DDM failure is detected
on a loop and enables regeneration of data from the failed DDM onto a hot spare DDM.

A hot DDM spare pool consisting of two drives, created with one 3+3+2S array (RAID 10), is
created for each drive size on an SSA loop. Therefore, if only one drive size is installed on a
loop, only two spares are required. The hot sparing function is managed at the SSA loop
level. SSA will spare to a larger capacity DDM on the loop in the very uncommon situation
that no spares are available on the loop for a given capacity.
Figure 8-12 on page 466 shows the following:

5. In eight-pack pair 1, the array consists of three data drives mirrored to three copy drives.
The remaining two drives are used as spares.
6. In eight-pack pair 2, the array consists of four data drives mirrored to four copy drives.

8.13 Storage balancing with RAID-10
Cluster 1 Cluster 2
1) RAID 10 array 3) RAID 10 array
Loop A 3 + 3 + 2S 4+4 Loop A
4) RAID 10 array 2) RAID 10 array

Loop B 4+4 3 + 3 + 2S Loop B
LSS 0 LSS 1
SSA 01 SSA 11
Figure 8-13 Storage balancing with RAID-10
Logical Storage Subsystem

The Logical Storage Subsystem (LSS), also known as the Logical Subsystem, is a logical
structure that is internal to the ESS. It is a logical construct that groups up to 256 logical
volumes (logical volumes are defined during the logical configuration procedure) of the same
disk format (CKD or FB), and it is identified by the ESS with a unique ID. Although the LSS
relates directly to the logical control unit (LCU) concept of the ESCON and FICON
architectures, it does not directly relate to SCSI and FCP addressing.
ESS storage balancing

For performance reasons, try to allocate storage on the ESS so that it is equally balanced
across both clusters and among the SSA loops. One way to accomplish this is to assign two
arrays (one from loop A and one from loop B) to each Logical Subsystem. To achieve this you
can follow this procedure when configuring RAID-10 ranks:
1. Configure the first array for LSS 0/loop A. This will be a 3 + 3 + 2S array.
2. Configure the first array for LSS 1/loop B. This will also be a 3 + 3 + 2S array.
3. Configure the second array for LSS1/loop A. This will now be a 4 + 4.
4. Configure the second array for LSS 0/loop B. This will also now be a 4 + 4.
Figure 8-13 illustrates the results of this configuration procedure.

8.14 ESS copy services
DATA
MOVER
DATA
MOVER
local point-in-time copy
Concurrent
py
Copy te co
XRC us remo nces
TotalStorage
Sidefile ono
chr dd ista
asyn unlimite
ove r
Km
o 103
C copy up t
TotalStorage
PPR
mote
us re
hrono over
sync -XD copy
PPRC s remote es FlashCopy
ronou al distanc
ynch t
non-s continen
local point-in-time copy
Figure 8-14 ESS copy services
DFSMS copy services

DFSMS provides Advanced Copy Services that include a hardware and software solution to
help you manage and protect your data. These solutions help ensure that your data remains
available 24 hours a day, seven days a week. Advanced Copy Services provide solutions to
disaster recovery, data migration, and data duplication. Many of these functions run on the
IBM TotalStorage Enterprise Storage Server (ESS). With DFSMS, you can perform the
following data management functions:
Use remote copy to prepare for disaster recovery
Move your PPRC data more easily
Remote copy provides two options that enable you to maintain a current copy of your data at
a remote site. These two options are used for disaster recovery and workload migration:
Extended remote copy (XRC)
Peer-to-peer remote copy (PPRC)
There are two types of copy:

An instantaneous copy where all the late updates in the primary are not copied. It is used
for fast backups and data replication in general. The examples in ESS are Concurrent
Copy and Flash Copy.
Mirroring, a never-ending copy where all the updates are mirrored as fast as possible. It is
used for disaster recovery and planned outages. The example in ESS are Enhanced
PPRC service and XRC.

Peer-to-peer remote copy (PPRC)
PPRC is a hardware solution which provides rapid and accurate disaster recovery as well as
a solution to workload movement and device migration. Updates made on the primary DASD
volumes are synchronously shadowed to the secondary DASD volumes. The local storage
subsystem and the remote storage subsystem are connected through a communications link
called a PPRC path. You can use one of the following protocols to copy data using PPRC:
ESCON
Fibre Channel Protocol
Note: Fibre Channel Protocol is supported only on ESS Model 800 with the appropriate
licensed internal code (LIC) level and the PPRC Version 2 feature enabled.
PPRC provides a synchronous volume copy across ESS controllers. The copy is done from
one controller (the one having the primary logical device) to the other (having the secondary
logical device). It is synchronous because the task doing the I/O receives the CPU back with
the guarantee that the copy was executed. There is a performance penalty for distances
longer than 10 km. PPRC is used for disaster recovery, device migration, and workload
migration; for example, it enables you to switch to a recovery system in the event of a disaster
in an application system.
You can issue the CQUERY command to query the status of one volume of a PPRC volume pair
or to collect information about a volume in the simplex state. The CQUERY command is
modified and enabled to report on the status of S/390-attached CKD devices.
See z/OS DFSMS Advanced Copy Services, SC35-0428, for further information about the
PPRC service and the CQUERY command.
Peer-to-peer remote copy extended distance (PPRC-XD)

When you enable the PPRC extended distance feature (PPRC-XD), the primary and recovery
storage control sites can be separated by long distances. Updates made to a PPRC primary
volume are sent to a secondary volume asynchronously, thus requiring less bandwidth.
If you are trying to decide whether to use synchronous or asynchronous PPRC, consider the
differences between the two modes:
When you use synchronous PPRC, no data loss occurs between the last update at the
primary system and the recovery site, but it increases the impact to applications and uses
more resources for copying data.
Asynchronous PPRC using the extended distance feature reduces impact to applications
that write to primary volumes and uses less resources for copying data, but data might be
lost if a disaster occurs. To use PPRC-XD as a disaster recovery solution, customers need
to periodically synchronize the recovery volumes with the primary site and make backups
to other DASD volumes or tapes.
PPRC Extended Distance (PPRC-XD) is a non-synchronous version of PPRC. This means

that host updates to the source volume are not delayed by waiting for the update to be
confirmed in the secondary volume. It also means that the sequence of updates on the
secondary volume is not guaranteed to be the same as on the primary volume.
PPRC-XD is an excellent solution for:

Remote data copy
Remote data migration
Offsite backup
Transmission of inactive database logs

Application disaster recovery solutions based on periodic point-in-time (PiT) copies of the
data, if the application tolerates short interruptions (application quiesce)
PPRC-XD can operate at very long distances (such as continental distances), well beyond
the 103 km supported for PPRC synchronous transmissionsand with minimal impact on the
application. The distance is limited only by the network and channel extender technology
capabilities.
Extended remote copy (XRC)

XRC combines hardware and software to provide continuous data availability in a disaster
recovery or workload movement environment. XRC provides an asynchronous remote copy
solution for both system-managed and non-system-managed data to a second, remote
location.
XRC relies on the IBM TotalStorage Enterprise Storage Server, IBM 3990, RAMAC Storage
Subsystems, and DFSMSdfp. The 9393 RAMAC Virtual Array (RVA) does not support XRC
for source volume capability.
XRC relies on the system data mover, which is part of DFSMSdfp. The system data mover is
a high-speed data movement program that efficiently and reliably moves large amounts of
data between storage devices. XRC is a continuous copy operation, and it is capable of
operating over long distances (with channel extenders). It runs unattended, without
involvement from the application users. If an unrecoverable error occurs at your primary site,
the only data that is lost is data that is in transit between the time when the primary system
fails and the recovery at the recovery site.
You can implement XRC with one or two systems. Let us suppose that you have two systems:
an application system at one location, and a recovery system at another. With these two
systems in place, XRC can automatically update your data on the remote disk storage
subsystem as you make changes to it on your application system. You can use the XRC
suspend/resume service for planned outages. You can still use this standard XRC service on
systems attached to the ESS if these systems are installed with the toleration or transparency
support.
Coupled Extended Remote Copy (CXRC) allows XRC sessions to be coupled together to
guarantee that all volumes are consistent across all coupled XRC sessions. CXRC can
manage thousands of volumes. IBM TotalStorage XRC Performance Monitor provides the
ability to monitor and evaluate the performance of a running XRC configuration.
Concurrent copy
Concurrent copy is an extended function that enables data center operations staff to generate
a copy or a dump of data while applications are updating that data. Concurrent copy delivers
a copy of the data, in a consistent form, as it existed before the updates took place.
FlashCopy service
FlashCopy is a point-in-time copy services function that can quickly copy data from a source
location to a target location. FlashCopy enables you to make copies of a set of tracks, with the
copies immediately available for read or write access. This set of tracks can consist of an
entire volume, a data set, or just a selected set of tracks. The primary objective of FlashCopy
is to create a copy of a source volume on the target volume. This copy is called a
point-in-time copy. Access to the point-in-time copy of the data on the source volume is
through reading the data from the target volume. The actual point-in-time data that is read
from the target volume might or might not be physically stored on the target volume. The ESS
FlashCopy service is compatible with the existing service provided by DFSMSdss. Therefore,
you can invoke the FlashCopy service on the ESS with DFSMSdss.

8.15 ESS performance features
Priority I/O queuing

Custom volumes
Improved caching algorithms
FICON host adapters
Enhanced CCWs
ESS
z/OS
DS6000
DS8000
Figure 8-15 ESS performance features
I/O priority queuing

Prior to ESS, IOS kept the UCB I/O pending requests in a queue named IOSQ. The priority
order of the I/O request in this queue, when the z/OS image is in goal mode, is controlled by
Workload Manager (WLM), depending on the transaction owning the I/O request. There was
no concept of priority queuing within the internal queues of the I/O control units; instead, the
queue regime was FIFO.
With ESS, it is possible to have this queue concept internally; I/O Priority Queueing in ESS
has the following properties:
I/O can be queued with the ESS in priority order.
WLM sets the I/O priority when running in goal mode.
There is I/O priority for systems in a sysplex.
Each system gets a fair share.
Custom volumes
Custom volumes provides the possibility of defining small size 3390 or 3380 volumes. This
causes less contention on a volume. Custom volumes is designed for high activity data sets.
Careful size planning is required.
Improved caching algorithms

With its effective caching algorithms, IBM TotalStorage Enterprise Storage Server Model 800
can minimize wasted cache space and reduce disk drive utilization, thereby reducing its

back-end traffic. The ESS Model 800 has a maximum cache size of 64 GB, and the NVS
standard size is 2 GB.
The ESS manages its cache in 4 KB segments, so for small data blocks (4 KB and 8 KB are
common database block sizes), minimum cache is wasted. In contrast, large cache segments
can exhaust cache capacity when filling up with small random reads. Thus the ESS, having
smaller cache segments, is able to avoid wasting cache space for situations of small record
sizes that are common in interactive applications.
This efficient cache management, together with the ESS Model 800 powerful back-end
implementation that integrates new (optional) 15,000 rpm drives, enhanced SSA device
adapters, and twice the bandwidth (as compared to previous models) to access the larger
NVS (2 GB) and the larger cache option (64 GB), all integrate to give greater throughput while
sustaining cache speed response times.
FICON host adapters

FICON extends the IBM TotalStorage Enterprise Storage Server Model 800s ability to deliver
bandwidth potential to the volumes that need it, when they need it.
Performance enhanced channel command words (CCWs)

For z/OS environments, the ESS supports channel command words (CCWs) that reduce the
characteristic overhead associated to the previous (3990) CCW chains. Basically, with these
CCWs, the ESS can read or write more data with fewer CCWs. CCW chains using the old
CCWs are converted to the new CCWs whenever possible. The cooperation of z/OS software
and the ESS provides the most significant benefits for the applications performance. In other
words, in ESS there is less overhead associated with CCW chains by combining tasks into
fewer CCWs, introducing Read Track Data and Write Track Data CCWs. They allow reading
and writing more data with fewer CCWs. It will be used by z/OS to reduce ESCON protocol for
multiple record transfer chains. Measurements on 4 KB records using an EXCP channel
program showed a 15% reduction in channel overhead for the Read Track Data CCW.

8.16 IBM TotalStorage DS6000
Enterprise Class Storage Solutions

ESS 750 DS6000
75.25
@19.2TB
@4.8TB
5.25
54.5 19
Maximum Configuration Up to 32 HDDs, 5TB Up to 224 HDDs, 67TB

5 TB Configured Weight 2,322 lbs 125 lbs
Maximum Power Consumption 4.83 kVA 0.69 kVA controller
0.48 kVA expansion unit
Figure 8-16 IBM TotalStorage DS6000
IBM TotalStorage DS6000

The IBM TotalStorage DS6000 series is designed to deliver the resiliency, performance, and
many of the key features of the IBM TotalStorage Enterprise Storage Server (ESS) in a small,
modular package.
The DS6000 series offers high scalability and excellent performance. With the DS6800
(Model 1750-511), you can install up to 16 disk drive modules (DDMs). The minimum storage
capability with 8 DDMs is 584 GB. The maximum storage capability with 16 DDMs for the
DS6800 model is 4.8 TB. If you want to connect more than 16 disks, you can use up to 13
DS6000 expansion units (Model 1750-EX1) that allow a maximum of 224 DDMs per storage
system and provide a maximum storage capability of 67 TB.
DS6000 specifications
Table 8-2 summarizes the DS6000 features.
Table 8-2 DS6000 specifications

DS6000
Controllers Dual active
Max cache 4 GB
Max host ports 8-Ports; 2 Gb FC/FICON

DS6000
Max hosts 1024
Max storage/disks 224
Disk types FC 10K: 146 GB, 300 GB

FC 15K: 73 GB
Max expansion mod 13
Max disk lops 4 (2 dual redundant)
Max LUNs 8192 (up to 2 TB LUN size)
RAID levels 5, 10
RAID array sizes 4 or 8 drives
Operating systems z/OS, i5/OS, OS/400, AIX, Sun Solaris, HP UX,

VMWare, Microsoft Windows, Linux
Packaging 3U Controller & Expansion Drawers
Power consumption Controller: 0.69 kVA

Expansion drawer: 0.48 kVA
Modular scalability
The DS6000 is modularly scalable, with optional expansion enclosure, to add capacity to help
meet your growing business needs. The scalability comprises:
Flexible design to accommodate on demand business environments
Ability to make dynamic configuration changes
Add disk drives in increments of 4
Add storage expansion units
Scale capacity to over 67 TB
DS6800 (Model 1750-511)

The DS6800 is a self-contained 3U enclosure that can be mounted in a standard 19-inch
rack. The DS6800 comes with authorization for up to 16 internal FC DDMs, offering up to 4.8
TB of storage capability. The DS6800 allows up to 13 DS6000 expansion enclosures to be
attached. A storage system supports up to 224 disk drives for a total of up to 67.2 TB of
storage.
The DS6800 offers the following features:
Two FC controller cards.
PowerPC 750GX 1 GHz processor.
4 GB of cache.
Two battery backup units (one per each controller card).
Two AC/DC power supplies with imbedded enclosure cooling units.
Eight 2 Gbps device ports.
Connectivity with the availability of two to eight Fibre Channel/FICON host ports. The host
ports auto-negotiate to either 2 Gbps or 1 Gbps link speeds.
Attachment to 13 DS6000 expansion enclosures.

DS6000 expansion enclosure (Model 1750-EX1)
The 3U DS6000 expansion enclosure can be mounted in a standard 19-inch rack. The front of
the enclosure contains the docking sites where you can install up to 16 DDMs.
The DS6000 expansion enclosure contains the following features:
Two expansion controller cards. Each controller card provides the following:
Two 2 Gbps inbound ports
Two 2 Gbps outbound ports
One FC switch per controller card
Controller disk enclosure that holds up to 16 FC DDMs
Two AC/DC power supplies with imbedded enclosure cooling units
Supports attachment to DS6800

8.17 IBM TotalStorage DS8000
ES 800 DS8000
76
75.25
33.25
54.5
Up to 6X ESS base Model 800

Physical capacity from 1.1TB up to 192TB
Machine type 2107
Figure 8-17 IBM TotalStorage DS8000
IBM TotalStorage DS8000

IBM TotalStorage DS8000 is a high-performance, high-capacity series of disk storage that is
designed to support continuous operations. DS8000 series models (machine type 2107) use
the IBM POWER5 server technology that is integrated with the IBM Virtualization Engine
technology. DS8000 series models consist of a storage unit and one or two management
consoles, two being the recommended configuration. The graphical user interface (GUI) or
the command-line interface (CLI) allows you to logically partition storage (create storage
LPARs) and use the built-in Copy Services functions. For high availability, hardware
components are redundant.
The current physical storage capacity of the DS8000 series system can range from 1.1 TB to
192 TB of physical capacity, and it has an architecture designed to scale to over 96 petabytes.
DS8000 models
The DS8000 series offers various choices of base and expansion models, so you can
configure storage units that meet your performance and configuration needs.
DS8100
The DS8100 (Model 921) features a dual two-way processor complex and support for one
expansion frame.
DS8300
The DS8300 (Models 922 and 9A2) features a dual four-way processor complex and

support for one or two expansion frames. The Model 9A2 supports two IBM TotalStorage
System LPARs (Logical Partitions) in one storage unit.
The DS8000 expansion frames (Models 92E and 9AE) expand the capabilities of the base
models. You can attach the Model 92E to either the Model 921 or the Model 922 to expand
their capabilities. You can attach the Model 9AE to expand the Model 9A2.

8.18 DS8000 hardware overview
2-Way (Model 8100)

Two dual processor servers
Up to 128GB Cache
8 to 64 2Gb FC/FICON 4 to 32 ESCON Ports
16 to 384 HDD
Intermixable 73GB 15Krpm, 146/300GB 10Krpm
4-Way (Model 8300)

Two four processor servers
Up to 256GB Cache
8 to 128 2Gb FC/FICON 4 to 64 ESCON Ports
16 to 640 HDD
Intermixable 73GB 15Krpm, 146/300GB 10Krpm
Figure 8-18 DS8000 models
DS8100 (Model 921)

The IBM TotalStorage DS8100, which is Model 921, offers features that include the following:
Dual two-way processor complex
Up to 128 disk drives, for a maximum capacity of 38.4 TB
Up to 128 GB of processor memory (cache)
Up to 16 fibre-channel/FICON or ESCON host adapters
The DS8100 model can support one expansion frame. With one expansion frame, you can
expand the capacity of the Model 921 as follows:
DS8300 (Models 922 and 9A2)

IBM TotalStorage DS8300 models (Model 922 and Model 9A2) offer higher performance and
capacity than the DS8100. The Model 9A2 also enables you to create two storage system
LPARs (or images) within the same storage unit.
Both DS8300 models offer the following features:

Dual four-way processor complex
Up to 256 GB of processor memory (cache)

The DS8300 models can support either one or two expansion frames. With expansion
frames, you can expand the Model 922 and 9A2 as follows:
With one expansion frame, you can support the following expanded capacity and number
of adapters:
With two expansion frames, you can support the following expanded capacity:
Up to 640 disk drives, for a maximum capacity of 192 TB

8.19 Storage systems LPARs
Workload A Workload B
LUN 0 LUN 0
LUN 1 LUN 2
DS8300 Model 9A2 exploits
Logical
LUN 2 Partition LPAR technology, allowing you
Logical
B to run two separate storage
Partition
A
server images
DS8000
Host Host Host Host Host Host Host

Adapters Adapters Adapters Adapters Adapters Adapters Adapters
Higher Bandwidth Fault Tolerant Interconnect
Volatile Volatile N-way

N-way Memory Memory
SMP SMP
LPAR
Persistent Persistent
Memory LPAR LPAR Memory
RAID
RAID RAID
RAID
RAID RAID
Adapters
Adapters Switched Fabric Adapters
Adapters
Adapters Adapters
Figure 8-19 Storage systems LPARs
LPAR overview
A logical partition (LPAR) is a subset of logical resources that is capable of supporting an
operating system. It consists of CPUs, memory, and I/O slots that are a subset of the pool of
available resources within a system. These resources are assigned to the logical partition.
Isolation between LPARs is provided to prevent unauthorized access between partition
boundaries.
Storage systems LPARs

The DS8300 Model 9A2 exploits LPAR technology, allowing you to run two separate storage
server images.
Each Storage System LPAR has access to:

50% of the processors
50% of the processor memory
Up to 16 host adapters
Up to 320 disk drives (up to 96 TB of capacity)
With these separate resources, each Storage System LPAR can run the same or other
versions of microcode, and can be used for completely separate production, test, or other
unique storage environments within this single physical system. This can enable storage

consolidations where separate storage subsystems were previously required, helping to
increase management efficiency and cost effectiveness.
DS8000 addressing capability

Table 8-2 on page 474 shows the DS8000 addressing capability in comparison to an ESS
800.
Table 8-3 Addressing capability comparison

ESS 800 DS8000 DS8000 w/LPAR
Max logical subsystems 32 255 510
Max logical devices 8K 64 K 128 K
Max logical CKD devices 4K 64 K 128 K
Max logical FB devices 4K 64 K 128 K
Max N-Port logins/port 128 509 509
Max N-Port logins 512 8K 16 K
Max logical paths/FC port 256 2K 2K
Max logical paths/CU image 256 512 512
Max path groups/CU image 128 256 256

8.20 IBM TotalStorage Resiliency Family
Copy services
FlashCopy
Mirroring
Metro Mirror (Synchronous PPRC)
Global Mirror (Asynchronous PPRC)
Metro/Global Copy (two or three-site Asynchronous
Cascading PPRC)
Global Copy (PPRC Extended Distance)
Global Mirror for zSeries (XRC) DS6000 can be
configured as an XRC target only
Metro/Global Mirror for zSeries (three-site solution
using Synchronous PPRC and XRC) DS6000 can
be configured as an XRC target only
Figure 8-20 The IBM TotalStorage Resiliency Family
The IBM TotalStorage Resiliency Family

The IBM TotalStorage Resiliency Family is a set of products and features that are designed to
help you implement storage solutions that keep your business running 24 hours a day, 7 days
a week.
These hardware and software features, products, and services are available on the IBM
TotalStorage DS6000 and DS8000 series and IBM TotalStorage ESS Models 750 and 800. In
addition, a number of advanced Copy Services features that are part of the IBM TotalStorage
Resiliency family are available for the DS6000 and DS8000 series. The IBM TotalStorage DS
Family also offers systems to support enterprise-class data backup and disaster recovery
capabilities. As part of the IBM TotalStorage Resiliency Family of software, IBM TotalStorage
FlashCopy point-in-time copy capabilities back up data in the background and allow users
nearly instant access to information about both source and target volumes. Metro and Global
Mirror capabilities create duplicate copies of application data at remote sites. High-speed
data transfers help to back up data for rapid retrieval.
Copy Services
Copy Services is a collection of functions that provides disaster recovery, data migration, and
data duplication functions. Copy Services runs on the DS6000 and DS8000 series and
supports open systems and zSeries environments.
Copy Services functions also are supported on the previous generation of storage systems,
the IBM TotalStorage Enterprise Storage Server.

Copy Services include the following types of functions:
FlashCopy, which is a point-in-time copy function
Remote mirror and copy functions (previously known as Peer-to-Peer Remote Copy or
PPRC), which includes:
IBM TotalStorage Metro Mirror (previously known as Synchronous PPRC)
IBM TotalStorage Global Copy (previously known as PPRC Extended Distance)
IBM TotalStorage Global Mirror (previously known as Asynchronous PPRC)
z/OS Global Mirror (previously known as Extended Remote Copy or XRC)
For information about copy services, see 8.14, ESS copy services on page 469.
Metro/Global Copy function

The Metro/Global Copy function allows you to cascade a PPRC pair with a PPRC-XD pair
such that the PPRC secondary also serves as the PPRC-XD primary. In this configuration, a
primary and secondary pair is established with the secondary located in a nearby site,
protected from primary site disasters. The secondary volume for the PPRC-XD copy can be
located thousands of miles away and continue to be updated if the original primary location
suffered a disaster.
Metro/Global Mirror function

The Metro/Global Mirror function enables a three-site, high availability disaster recovery
solution. It combines the capabilities of both Metro Mirror and Global Mirror functions for
greater protection against planned and unplanned outages.

8.21 TotalStorage Expert product highlights
Netscape or
Internet Explorer
z/OS
TotalStorage Expert
Windows XP or
AIX
AS/400
DS8000
FICON UNIX
ESCON
VTS
Peer-To-Peer
VTS
3494 Library Windows XP
Manager
Figure 8-21 TotalStorage Expert
TotalStorage Expert
TotalStorage Expert is an innovative software tool that gives administrators powerful but
flexible storage asset, capacity, and performance management capabilities to centrally
manage Enterprise Storage Servers located anywhere in the enterprise.
IBM TotalStorage Expert has two available features:

The ESS feature, which supports ESS
The ETL feature, which supports Enterprise Tape Library products
The two features are licensed separately. There are also upgrade features for users of
StorWatch Expert V1 with either the ESS or the ETL feature, or both, who want to migrate to
TotalStorage Expert V2.1.1.
TotalStorage Expert is designed to augment commonly used IBM performance tools such as
Resource Management Facility (RMF), DFSMS Optimizer, AIX Performance Toolkit, and
similar host-based performance monitors. While these tools provide performance statistics
from the host systems perspective, TotalStorage Expert provides statistics from the ESS and
ETL system perspective.
By complementing other performance tools, TotalStorage Expert provides a more

comprehensive view of performance; it gathers and presents information that provides a
complete management solution for storage monitoring and administration.

TotalStorage Expert helps storage administrators by increasing the productivity of storage
resources.
The ESS is ideal for businesses with multiple heterogeneous servers, including zSeries,
UNIX, Windows NT, Windows 2000, Novell NetWare, HP/UX, Sun Solaris, and AS/400
servers.
With Version 2.1.1, the TotalStorage ESS Expert is packaged with the TotalStorage ETL
Expert. The ETL Expert provides performance, asset, and capacity management for the three
IBM ETL solutions:
IBM TotalStorage Enterprise Automated Tape Library, described in IBM TotalStorage
Enterprise Automated Tape Library 3494 on page 495.
IBM TotalStorage Virtual Tape Server, described in Introduction to Virtual Tape Server
(VTS) on page 497.
IBM TotalStorage Peer-to-Peer Virtual Tapeserver, described in IBM TotalStorage
Peer-to-Peer VTS on page 499.
Both tools can run on the same server, share a common database, efficiently monitor storage
resources from any location within the enterprise, and provide a similar look and feel through
a Web browser user interface. Together they provide a complete solution that helps optimize
the potential of IBM disk and tape subsystems.

8.22 Introduction to tape processing
Figure 8-22 Introduction to tape processing
Tape volumes
The term tape refer to volumes that can be physically moved. You can only store sequential
data sets on tape. Tape volumes can be sent to a safe, or to other data processing centers.
Internal labels are used to identify magnetic tape volumes and the data sets on those
volumes. You can process tape volumes with:
IBM standard labels
Labels that follow standards published by:
International Organization for Standardization (ISO)
American National Standards Institute (ANSI)
Federal Information Processing Standard (FIPS)
Nonstandard labels
No labels
Note: Your installation can install a bypass for any type of label processing; however, the
use of labels is recommended as a basis for efficient control of your data.
IBM standard tape labels consist of volume labels and groups of data set labels. The volume
label, identifying the volume and its owner, is the first record on the tape. The data set label,

identifying the data set and describing its contents, precedes and follows each data set on the
volume:
The data set labels that precede the data set are called header labels.
The data set labels that follow the data set are called trailer labels. They are almost
identical to the header labels.
The data set label groups can include standard user labels at your option.
Usually, the formats of ISO and ANSI labels, which are defined by the respective
organizations, are similar to the formats of IBM standard labels.
Nonstandard tape labels can have any format and are processed by routines you provide.
Unlabeled tapes contain only data sets and tape marks.

8.23 SL and NL format
/ /
IBM Standard IBM Standard
IBM IBM Standard
Data Set Data Set
Standard Volume TM Data Set TM TM TM
Header Trailer
Labels Label
Label Label
/ /
/ /
Unlabeled
Tapes Data Set TM TM
/ /
TM= Tapemark
Figure 8-23 SL and NL format
Using tape with JCL

In the job control statements, you must provide a data definition (DD) statement for each data
set to be processed. The LABEL parameter of the DD statement is used to describe the data
set's labels.
Other parameters of the DD statement identify the data set, give volume and unit information
and volume disposition, and describe the data set's physical attributes. You can use a data
class to specify all of your data set's attributes (such as record length and record format), but
not data set name and disposition. Specify the name of the data class using the JCL keyword
DATACLAS. If you do not specify a data class, the automatic class selection (ACS) routines
assign a data class based on the defaults defined by your storage administrator.
An example of allocating a tape data set using DATACLAS in the DD statement of the JCL
statements follows. In this example, TAPE01 is the name of the data class.
//NEW DD DSN=DATASET.NAME,UNIT=TAPE,DISP=(,CATLG,DELETE),DATACLAS=TAPE01,LABEL=(1,SL)
Describing the labels

You specify the type of labels by coding one of the subparameters of the LABEL parameter as
shown in Table 8-4 on page 490.

Table 8-4 Types of labels
Code Meaning
SL IBM Standard Label
AL ISO/ANSI/FIPS labels
SUL Both IBM and user header or trailer labels
AUL Both ISO/ANSI/FIPS and user header or trailer labels
NSL Nonstandard labels
NL No labels, but the existence of a previous label is verified
BLP Bypass label processing. The data is treated in the same manner as though NL had been
specified, except that the system does not check for an existing volume label. The user is
responsible for the positioning.
If your installation does not allow BLP, the data is treated exactly as though NL had been
specified. Your job can use BLP only if the Job Entry Subsystem (JES) through Job class,
RACF through TAPEVOL class, or DFSMSrmm(*) allow it.
LTM Bypass a leading tape mark. If encountered, on unlabeled tapes from VSE.
Note: If you do not specify the label type, the operating system assumes that the data set
has IBM standard labels.

8.24 Tape capacity - tape mount management
3590=10,000 Mb
3490=800 Mb
3480=200 Mb 3592=300,000 Mb
Figure 8-24 Tape capacity
Tape capacity
The capacity of a tape depends on the device type that is recording it. 3480 and 3490 tapes
are physically the same cartridges. The IBM 3590 and 3592 high performance cartridge tape
is not compatible with the 3480, 3490, or 3490E drives. 3490 units can read 3480 cartridges,
but cannot record as a 3480, and 3480 units cannot read or write as a 3490.

Using DFSMS and tape mount management can help you reduce the number of both tape
mounts and tape volumes that your installation requires. The volume mount analyzer reviews
your tape mounts and creates reports that provide you with information you need to effectively
implement the tape mount management methodology recommended by IBM.
Tape mount management allows you to efficiently fill a tape cartridge to its capacity and gain
full benefit from improved data recording capability (IDRC) compaction, 3490E Enhanced
Capability Magnetic Tape Subsystem, 36-track enhanced recording format, and Enhanced
Capacity Cartridge System Tape. By filling your tape cartridges, you reduce your tape mounts
and even the number of tape volumes you need.
With an effective tape cartridge capacity of 2.4 GB using 3490E and the Enhanced Capacity
Cartridge System Tape, DFSMS can intercept all but extremely large data sets and manage
them with tape mount management. By implementing tape mount management with DFSMS,
you might reduce your tape mounts by 60% to 70% with little or no additional hardware

required. Therefore, the resulting tape environment can fully exploit integrated cartridge
loaders (ICL), IDRC, and 3490E.
Tape mount management also improves job throughput because jobs are no longer queued
up on tape drives. Approximately 70% of all tape data sets queued up on drives are less than
10 MB. With tape mount management, these data sets reside on DASD while in use. This
frees up the tape drives for other allocations.
Tape mount management recommends that you use DFSMShsm to do interval migration to
SMS storage groups. You can use ACS routines to redirect your tape data sets to a tape
mount management DASD buffer storage group. DFSMShsm scans this buffer on a regular
basis and migrates the data sets to migration level 1 DASD or migration level 2 tape as soon
as possible, based on the management class and storage group specifications.
Table 8-5 lists all IBM tape capacities supported since 1952.
Table 8-5 Tape capacity of various IBM products

Year Product Capacity (Mb) Transfer Rate (KB/S)
1952 IBM 726 1.4 7.5
1953 IBM 727 5.8 15
1957 IBM 729 23 90
1965 IBM 2401 46 180
1968 IBM 2420 46 320
1973 IBM 3420 180 1,250
1984 IBM 3480 200 3,000
1989 IBM 3490 200 4,500
1991 IBM 3490E 400 9,000
1992 IBM 3490E 800 9,000
1995 IBM 3590 Magstar 10,000 (uncompacted) 9,000 (uncompacted)
1999 IBM 3590E Magstar 20,000 (uncompacted) 14,000
2000 IBM 3590E Magstar XL 20,000/40,000 (dependent on 14,000

Cartridge Model B or E)
2003 IBM 3592 TotalStorage 300,000 (for high capacity 40,000

Enterprise Tape Drive requirements)
or
60,000 (for fast data access
requirements)
For further information about tape processing, see z/OS DFSMS Using Magnetic Tapes,
SC26-7412.
Note: z/OS supports tape devices starting from D/T 3420.

8.25 TotalStorage Enterprise Tape Drive 3592 Model J1A
The IBM 3592 is an "Enterprise Class" tape drive

Data rate 40 MB/s (without compression)
Capacity 300 GB per cartridge (without
compression)
60 GB format to provide fast access to data
Dual ported 2 Gbps Fiber Channel interface

Autonegotiates (1 or 2 Gbps, Fabric or loop
support)
Options may be hard-set at drive
Drive designed for automation solutions

Small form factor
Cartridge similar form factor to 3590 and 3490
Improved environmentals
Figure 8-25 TotalStorage Enterprise Tape Drive 3592 Model J1A
IBM 3592 tape drive

The IBM 3592 tape drive is the fourth generation of high capacity and high performance tape
systems. It was announced in September 2003 and connects to IBM eServer zSeries
systems by way of the TotalStorage Enterprise Tape Controller 3592 Model J70 using
ESCON or FICON links. The 3592 system is the successor of the IBM Magstar 3590 family of
tape drives and controller types.
The IBM 3592 tape drive can be used as a standalone solution or as an automated solution
within a 3494 tape library.
Enterprise class tape drive

The native rate for data transfer increases up to 40 MBps compared to 14 MBps in a 3590
Magstar. The uncompressed amount of data which fits on a single cartridge increases to
300 GB and is used for scenarios where high capacity is needed. The tape drive has a
second option, where you can store a maximum of 60 GB per tape. This option is used
whenever fast access to tape data is needed.
Dual ported 2 Gbps Fiber Channel interface

This tape drive generation connects to the tape controller 3592 Model J70 using Fiber
Channel. SCSI connection, as it used in 3590 configuration, is no longer supported. However,
if you connect a 3590 Magstar tape drive to a 3592 controller, SCSI connection is possible.

Drive designed for automation solutions
The drive has a smaller form factor. Thus, you can integrate more drives into an automated
tape library. The cartridges have a similar form factor to the 3590 and 3490 cartridge, so they
fit into the same slots in a 3494 automated tape library.
Improved environmentals
By using a smaller form factor than 3590 Magstar drives, you can put two 3592 drives in place
of one 3590 drive in the 3494. In a stand-alone solution you can put a maximum of 12 drives
into one 19-inch rack, managed by one controller.

8.26 IBM TotalStorage Enterprise Automated Tape Library 3494
Figure 8-26 3494 tape library
IBM 3494 tape library

Tape storage media can provide low-cost data storage for sequential files, inactive data, and
vital records. Because of the continued growth in tape use, tape automation has been seen
as a way of addressing an increasing number of challenges.
Various solutions providing tape automation, including the following, are available:
The Automatic Cartridge Loader on IBM 3480 and 3490E tape subsystems provides quick
scratch (a volume with no valued data, used for output) mount.
The Automated Cartridge Facility on the Magstar 3590 tape subsystem, working with
application software, can provide a 10-cartridge mini-tape library.
The IBM 3494, an automated tape library dataserver, is a device consisting of robotics
components, cartridge storage areas (or shelves), tape subsystems, and controlling
hardware and software, together with the set of tape volumes that reside in the library and
can be mounted on the library tape drives.
The Magstar Virtual Tape Server (VTS) provides volume stacking capability and exploits
the capacity and bandwidth of Magstar 3590 technology.

3494 models and features
IBM 3494 offers a wide range of models and features, including the following:
Up to 96 tape drives
Support through the Library Control Unit for attachment of up to 15 additional frames,
including the Magstar VTS, for a total of 16 frames, not including the High Availability unit
Cartridge storage capacity of 291 to 6145 tape cartridges
Data storage capacity of up to 1.84 PB (Petabytes) of uncompacted data and 5.52 PB of
compacted data (at a compression rate of 3:1)
Support for the High Availability unit that provides a high level of availability for tape
automation
Support for the IBM Total Storage Virtual Tape Server
Support for the IBM Total Storage Peer-to-Peer VTS
Support for the following tape drives:
IBM 3490E Model F1A tape drive
IBM 3490E Model CxA tape drives
IBM Magstar 3590 Model B1A tape drives
IBM Magstar 3590 Model E1A tape drives
IBM Magstar 3590 Model H1A tape drives
IBM TotalStorage Enterprise Tape Drive 3592 Model J1A
Attachment to and sharing by multiple host systems, such as IBM eServer zSeries,
iSeries, pSeries, S/390, RS/6000, AS/400, HP, and Sun processors
Data paths through FICON, fibre channels, SCSI-2, ESCON, and parallel channels
depending on the tape subsystem installed
Library management commands through RS-232, a local area network (LAN), and
parallel, ESCON, and FICON channels

8.27 Introduction to Virtual Tape Server (VTS)
VTS models:
Model B10 VTS
Model B20 VTS
Peer-to-Peer (PtP) VTS (up to twenty-four 3590
tape drives)
VTS design (single VTS)
32, 64, 128 or 256 3490E virtual devices
Tape volume cache:
Analogous to DASD cache
Data access through the cache
Dynamic space management
Cache hits eliminate tape mounts
Up to twelve 3590 tape drives (the real 3590 volume
contains up to 250,000 virtual volumes per VTS)
Stacked 3590 tape volumes managed by the 3494
Figure 8-27 Introduction to VTS
VTS introduction
The IBM Magstar Virtual Tape Server (VTS), integrated with the IBM Tape Library
Dataservers (3494), delivers an increased level of storage capability beyond the traditional
storage products hierarchy. The host software sees VTS as a 3490 Enhanced Capability
(3490E) Tape Subsystem with associated standard (CST) or Enhanced Capacity Cartridge
System Tapes (ECCST). This virtualization of both the tape devices and the storage media to
the host allows for transparent utilization of the capabilities of the IBM 3590 tape technology.
Along with introduction of the IBM Magstar VTS, IBM introduced new views of volumes and
devices because of the different knowledge about volumes and devices on the host system
and the hardware. Using a VTS subsystem, the host application writes tape data to virtual
devices. The volumes created by the hosts are called virtual volumes and are physically
stored in a tape volume cache that is built from RAID DASD.
VTS models
These are the IBM 3590 drives you can choose:
For the Model B10 VTS, four, five, or six 3590-B1A/E1A/H1A can be associated with VTS.
For the Model B20 VTS, six to twelve 3590-B1A/E1A/H1A can be associated with VTS.
Each ESCON channel in the VTS is capable of supporting 64 logical paths, providing up to
1024 logical paths for Model B20 VTS with sixteen ESCON channels, and 256 logical paths
for Model B10 VTS with four ESCON channels. Each logical path can address any of the 32,
64, 128, or 256 virtual devices in the Model B20 VTS.

Each FICON channel in the VTS can support up to 128 logical paths, providing up to 1024
logical paths for the Model B20 VTS with eight FICON channels. With a Model B10 VTS, 512
logical paths can be provided with four FICON channels. As with ESCON, each logical path
can address any of the 32, 64, 128, or 256 virtual devices in the Model B20 VTS.
Note: Intermixing FICON and SCSI interfaces is not supported.
Tape volume cache

The IBM TotalStorage Peer-to-Peer Virtual Tape Server appears to the host processor as a
single automated tape library with 64, 128, or 256 virtual tape drives and up to 250,000 virtual
volumes. The configuration of this system has up to 3.5 TB of Tape Volume Cache native
(10.4 TB with 3:1 compression), up to 24 IBM TotalStorage 3590 tape drives, and up to 16
host ESCON or FICON channels. Through tape volume cache management policies, the VTS
management software moves host-created volumes from the tape volume cache to a Magstar
cartridge managed by the VTS subsystem. When a virtual volume is moved from the tape
volume cache to tape, it becomes a logical volume.
VTS design
VTS looks like an automatic tape library with thirty-two 3490E drives and 50,000 volumes in
37 square feet. Its major components are:
Magstar 3590 (three or six tape drives) with two ESCON channels
Magstar 3494 Tape Library
Fault-tolerant RAID-1 disks (36 Gb or 72 Gb)
RISC Processor
VTS functions
VTS provides the following functions:
Thirty-two 3490E virtual devices.
Tape volume cache (implemented in a RAID-1 disk) that contains virtual volumes.
The tape volume cache consists of a high performance array of DASD and storage
management software. Virtual volumes are held in the tape volume cache when they are
being used by the host system. Outboard storage management software manages which
virtual volumes are in the tape volume cache and the movement of data between the tape
volume cache and physical devices. The size of the DASD is made large enough so that
more virtual volumes can be retained in it than just the ones currently associated with the
virtual devices.
After an application modifies and closes a virtual volume, the storage management
software in the system makes a copy of it onto a physical tape. The virtual volume remains
available on the DASD until the space it occupies reaches a predetermined threshold.
Leaving the virtual volume in the DASD allows for fast access to it during subsequent
requests. The DASD and the management of the space used to keep closed volumes
available is called tape volume cache. Performance for mounting a volume that is in tape
volume cache is quicker than if a real physical volume is mounted.
Up to six 3590 tape drives; the real 3590 volume contains logical volumes. Installation
sees up to 50,000 volumes.
Stacked 3590 tape volumes managed by the 3494. It fills the tape cartridge up to 100%.
Putting multiple virtual volumes into a stacked volume, VTS uses all of the available space
on the cartridge. VTS uses IBM 3590 cartridges when stacking volumes.
VTS is expected to provide a ratio of 59:1 in volume reduction, with dramatic savings in all
tape hardware items (drives, controllers, and robots).

8.28 IBM TotalStorage Peer-to-Peer VTS
Virtual Tape Controllers Distributed Library
CX1
VTC
Master VTS
I/O VTS
FICON/ESCON
to zSeries
ESCON/FICON
Distributed Library
UI Library
VTC
CX1 I/O VTS

Composite Library
Figure 8-28 IBM TotalStorage Peer-to-Peer VTS
Peer-to-Peer VTS
IBM TotalStorage Peer-to-Peer Virtual Tape Server, an extension of IBM TotalStorage Virtual
Tape Server, is specifically designed to enhance data availability. It accomplishes this by
providing dual volume copy, remote functionality, and automatic recovery and switchover
capabilities. With a design that reduces single points of failure (including the physical media
where logical volumes are stored), IBM TotalStorage Peer-to-Peer Virtual Tape Server
improves system reliability and availability, as well as data access. To help protect current
hardware investments, existing IBM TotalStorage Virtual Tape Servers can be upgraded for
use in this new configuration.
IBM TotalStorage Peer-to-Peer Virtual Tape Server consists of new models and features of
the 3494 Tape Library that are used to join two separate Virtual Tape Servers into a single,
interconnected system. The two virtual tape systems can be located at the same site or at
separate sites that are geographically remote. This provides a remote copy capability for
remote vaulting applications.
IBM TotalStorage Peer-to-Peer Virtual Tape Server appears to the host IBM eServer zSeries
processor as a single automated tape library with 64, 128, or 256 virtual tape drives and up to
500,000 virtual volumes. The configuration of this system has up to 3.5 TB of Tape Volume
Cache native (10.4 TB with 3:1 compression), up to 24 IBM TotalStorage 3590 tape drives,
and up to 16 host ESCON or FICON channels.

In addition to the 3494 VTS components B10, B18, and B20, the Peer-to-Peer VTS consists
of the following components:
The 3494 virtual tape controller model VTC
The VTC in the Virtual Tape Frame 3494 Model CX1 provides interconnection between
two VTSs with the Peer-to-Peer Copy features, and provides two host attachments for the
PtP VTS. There must be four (for the Model B10 or B18) or eight (for the Model B20 or
B18) VTCs in a PtP VTS configuration. Each VTC is an independently operating,
distributed node within the PtP VTS, which continues to operate during scheduled or
unscheduled service of another VTC.
The 3494 auxiliary tape frame model CX1
The Model CX1 provides the housing and power for two or four 3494 virtual tape
controllers. Each Model CX1 can be configured with two or four Model VTCs. There are
two power control compartments, each with its own power cord, to allow connection to two
power sources.
Peer-to-Peer copy features

Special features installed on 3494 Models B10, B18, and B20 in a Peer-to-Peer configuration
provide automatic copies of virtual volumes. These features can be installed on existing VTS
systems to upgrade them to a Peer-to-Peer VTS.
VTS advanced functions

As with a stand-alone VTS, the Peer-to-Peer VTS has the option to install additional features
and enhancements to existing features. These new features are:
Outboard policy management: Outboard policy management enables the storage
administrator to manage SMS data classes, storage classes, management classes, and
storage groups at the library manager or the 3494 specialist.
Physical volume pooling: With outboard policy management enabled, you are able to
assign logical volumes to selected storage groups. Storage groups point to primary
storage pools. These pool assignments are stored in the library manager database. When
a logical volume is copied to tape, it is written to a stacked volume that is assigned to a
storage pool as defined by the storage group constructs at the library manager.
Tape volume dual copy: With advanced policy management, storage administrators have
the facility to selectively create dual copies of logical volumes within a VTS. This function
is also available in the Peer-to-Peer environment. At the site or location where the second
distributed library is located, logical volumes can also be duplexed, in which case you can
have two or four copies of your data.
Peer-to-Peer copy control
There are two types of copy operations:
Immediate, which creates a copy of the logical volume in the companion connected
virtual tape server prior to completion of a rewind/unload command. This mode
provides the highest level of data protection.
Deferred, which creates a copy of the logical volume in the companion connected
virtual tape server as activity permits after receiving a rewind/unload command. This
mode provides protection that is superior to most currently available backup schemes.
Tape volume cache management: Prior to the introduction of these features, there was no
way to influence cache residency. As a result, all data written to the TVC was pre-migrated
using a first-in, first-out (FIFO) method. With the introduction of this function, you now have
the ability to influence the time that virtual volumes reside in the TVC.

8.29 Storage area network (SAN)
LAN
ESCON Fibre Channel

FICON SAN
switches & directors
Any Server to TotalStorage
Any Storage
ESS
Figure 8-29 Storage area network (SAN)
Storage area network

The Storage Network Industry Association (SNIA) defines a SAN as a network whose primary
purpose is the transfer of data between computer systems and storage elements. A SAN
consists of a communication infrastructure, which provides physical connections, and a
management layer, which organizes the connections, storage elements, and computer
systems so that data transfer is secure and robust. The term SAN is usually (but not
necessarily) identified with block I/O services rather than file access services. It can also be a
storage system consisting of storage elements, storage devices, computer systems, and
appliances, plus all control software, communicating over a network.
SANs today are usually built using Fibre Channel technology, but the concept of a SAN is
independent of the underlying type of network.
The major potential benefits of a SAN can be categorized as:

Access
Benefits include longer distances between processors and storage, higher availability, and
improved performance (because I/O traffic is offloaded from a LAN to a dedicated
network, and because Fibre Channel is generally faster than most LAN media). Also, a
larger number of processors can be connected to the same storage device, compared to
typical built-in device attachment facilities.

Consolidation
Another benefit is replacement of multiple independent storage devices by fewer devices
that support capacity sharing; this is also called disk and tape pooling. SANs provide the
ultimate in scalability because software can allow multiple SAN devices to appear as a
single pool of storage accessible to all processors on the SAN. Storage on a SAN can be
managed from a single point of control. Controls over which hosts can see which storage
(called zoning and LUN masking) can be implemented.
Protection
LAN-free backups occur over the SAN rather than the (slower) LAN, and server-free
backups can let disk storage write itself directly to tape without processor overhead.
There are various AN topologies on the base of Fibre Channel networks:

Point-to-Point
With a SAN, a simple link is used to provide high-speed interconnection between two
nodes.
Arbitrated loop
The Fibre Channel arbitrated loop offers relatively high bandwidth and connectivity at a
low cost. For a node to transfer data, it must first arbitrate to win control of the loop. Once
the node has control, it is free to establish a virtual point-to-point connection with another
node on the loop. After this point-to-point (virtual) connection is established, the two nodes
consume all of the loops bandwidth until the data transfer operation is complete. Once the
transfer is complete, any node on the loop can then arbitrate to win control of the loop.
Switched
Fibre channel switches function in a manner similar to traditional network switches to
provide increased bandwidth, scalable performance, an increased number of devices,
and, in certain cases, increased redundancy.
Multiple switches can be connected to form a switch fabric capable of supporting a large
number of host servers and storage subsystems. When switches are connected, each
switchs configuration information has to be copied into all the other participating switches.
This is called cascading.
FICON and SAN

From a zSeries perspective, FICON is the protocol that is used in a SAN environment. A
FICON infrastructure may be point-to-point or switched, using ESCON directors with FICON
bridge cards or FICON directors to provide connections between channels and control units.
FICON uses Fibre Channel transport protocols, and so uses the same physical fiber.
Today zSeries has 2 Gbps link data rate support. The 2 Gbps links are for native FICON,
FICON CTC, cascaded directors and fibre channels (FCP channels) on the FICON Express
cards on z800, z900, and z990 only.

Related publications
The publications listed in this section are considered particularly suitable for a more detailed
discussion of the topics covered in this book.
IBM Redbooks publications

For information about ordering these publications, see How to get IBM Redbooks
publications on page 504. Note that some of the documents referenced here may be
available in softcopy only.
VSAM Demystified, SG24-6105
DFSMStvs Overview and Planning Guide, SG24-6971
DFSMStvs Presentation Guide, SG24-6973
z/OS DFSMS V1R3 and V1R5 Technical Guide, SG24-6979
Other publications
These publications are also relevant as further information sources:
z/OS DFSMStvs Administration Guide, GC26-7483
Device Support Facilities Users Guide and Reference Release 17, GC35-0033
z/OS MVS Programming: Assembler Services Guide, SA22-7605
z/OS MVS System Commands, SA22-7627
z/OS MVS System Messages,Volume 1 (ABA-AOM), SA22-7631
DFSMS Optimizer Users Guide and Reference, SC26-7047
z/OS DFSMStvs Planning and Operating Guide, SC26-7348
z/OS DFSMS Access Method Services for Catalogs, SC26-7394
z/OS DFSMSdfp Storage Administration Reference, SC26-7402
z/OS DFSMSrmm Guide and Reference, SC26-7404
z/OS DFSMSrmm Implementation and Customization Guide, SC26-7405
z/OS DFSMS Implementing System-Managed Storage, SC26-7407
z/OS DFSMS: Managing Catalogs, SC26-7409
z/OS DFSMS: Using Data Sets, SC26-7410
z/OS DFSMS: Using the Interactive Storage Management Facility, SC26-7411
z/OS DFSMS: Using Magnetic Tapes, SC26-7412
z/OS DFSMSdfp Utilities, SC26-7414
z/OS Network File System Guide and Reference, SC26-7417
DFSORT Getting Started with DFSORT R14, SC26-4109
DFSORT Installation and Customization Release 14, SC33-4034
z/OS DFSMShsm Storage Administration Guide, SC35-0421

z/OS DFSMShsm Storage Administration Reference, SC35-0422
z/OS DFSMSdss Storage Administration Guide, SC35-0423
z/OS DFSMSdss Storage Administration Reference, SC35-0424
z/OS DFSMS Object Access Method Application Programmers Reference, SC35-0425
z/OS DFSMS Object Access Method Planning, Installation, and Storage Administration
Guide for Object Support, SC35-0426
Tivoli Decision Support for OS/390 System Performance Feature Reference Volume I,
SH19-6819
Device Support Facilities (ICKDSF) User's Guide and Reference, GC35-0033-35
z/OS DFSORT Application Programming Guide, SC26-7523
z/OS MVS JCL Reference, SA22-7597
Online resources
These Web sites and URLs are also relevant as further information sources:
For articles, online books, news, tips, techniques, examples, and more, visit the z/OS
DFSORT home page:
http://www-1.ibm.com/servers/storage/support/software/sort/mvs
For more information about DFSMSrmm, visit:

http://www-1.ibm.com/servers/storage/software/sms/rmm/
How to get IBM Redbooks publications

You can search for, view, or download books, Redpapers, Hints and Tips, draft publications
and Additional materials, as well as order hardcopy books or CD-ROMs, at this Web site:
ibm.com/redbooks
Help from IBM

IBM Support and downloads
ibm.com/support
IBM Global Services

ibm.com/services

ABCs of z/OS System Programming
Volume 3
Volume 3
(1.0 spine)
0.875<->1.498
460 <-> 788 pages
Volume 3
Volume 3
Back cover
ABCs of z/OS System

Programming
Volume 3
DFSMS, Data set The ABCs of z/OS System Programming is a thirteen-volume collection that
provides an introduction to the z/OS operating system and the hardware INTERNATIONAL
basics, SMS architecture. Whether you are a beginner or an experienced system programmer, TECHNICAL
the ABCs collection provides the information that you need to start your research
SUPPORT
Storage management into z/OS and related subjects. The ABCs collection serves as a powerful technical
tool to help you become more familiar with z/OS in your current environment, or ORGANIZATION
software and
to help you evaluate platforms to consolidate your e-business applications.
hardware This edition is updated to z/OS Version 1 Release 1.
Catalogs, VSAM, The contents of the volumes are:

DFSMStvs Volume 1: Introduction to z/OS and storage concepts, TSO/E, ISPF, JCL, SDSF, and BUILDING TECHNICAL
z/OS delivery and installation INFORMATION BASED ON
Volume 2: z/OS implementation and daily maintenance, defining subsystems, PRACTICAL EXPERIENCE
JES2 and JES3, LPA, LNKLST, authorized libraries, Language Environment, and
SMP/E
Volume 3: Introduction to DFSMS, data set basics, storage management hardware
IBM Redbooks are developed
and software, VSAM, System-Managed Storage, catalogs, and DFSMStvs
by the IBM International
Volume 4: Communication Server, TCP/IP and VTAM Technical Support
Volume 5: Base and Parallel Sysplex, System Logger, Resource Recovery Services Organization. Experts from
(RRS), Global Resource Serialization (GRS), z/OS system operations, Automatic IBM, Customers and Partners
Restart Management (ARM), Geographically Dispersed Parallel Sysplex (GPDS) from around the world create
Volume 6: Introduction to security, RACF, Digital certificates and PKI, Kerberos, timely technical information
cryptography and z990 integrated cryptography, zSeries firewall technologies, based on realistic scenarios.
LDAP, Enterprise Identity Mapping (EIM), and firewall technologies Specific recommendations
Volume 7: Printing in a z/OS environment, Infoprint Server and Infoprint Central are provided to help you
Volume 8: An introduction to z/OS problem diagnosis implement IT solutions more
Volume 9: z/OS UNIX System Services effectively in your
Volume 10: Introduction to z/Architecture, zSeries processor design, zSeries environment.
connectivity, LPAR concepts, HCD, and HMC
Volume 11: Capacity planning, performance management, RMF, and SMF
Volume 12: WLM
Volume 13: JES3 For more information:
ibm.com/redbooks
SG24-6983-03 ISBN 0738434094

SG 246983

Uploaded by

Copyright:

Available Formats

SG 246983

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

SG 246983

Uploaded by

Copyright:

Available Formats

Front cover

ABCs of z/OS System

Storage management software

Catalogs, VSAM, DFSMStvs

ABCs of z/OS System Programming Volume 3

Fourth Edition (March 2010)

Copyright International Business Machines Corporation 2010. All rights reserved.

Chapter 1. DFSMS introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Chapter 2. Data set basics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Chapter 3. Extended access volumes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

Copyright IBM Corp. 2010. All rights reserved. iii

Chapter 4. Storage management software . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

iv ABCs of z/OS System Programming Volume 3

Chapter 5. System-managed storage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239

Chapter 6. Catalogs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325

vi ABCs of z/OS System Programming Volume 3

Chapter 7. DFSMS Transactional VSAM Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . 377

Chapter 8. Storage management hardware . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 445

Related publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 503

viii ABCs of z/OS System Programming Volume 3

Copyright IBM Corp. 2010. All rights reserved. ix

The following terms are trademarks of other companies:

x ABCs of z/OS System Programming Volume 3

The ABCs of z/OS System Programming is a thirteen-volume collection that provides an

This edition is updated to z/OS Version 1 Release 1.

The contents of the volumes are:

Volume 4: Communication Server, TCP/IP and VTAM

Volume 6: Introduction to security, RACF, Digital certificates and PKI, Kerberos,

Volume 7: Printing in a z/OS environment, Infoprint Server and Infoprint Central

Volume 8: An introduction to z/OS problem diagnosis

Volume 9: z/OS UNIX System Services

Volume 10: Introduction to z/Architecture, zSeries processor design, zSeries connectivity,

Volume 11: Capacity planning, performance management, RMF, and SMF

Volume 12: WLM

Volume 13: JES3

The team who wrote this book

Copyright IBM Corp. 2010. All rights reserved. xi

The fourth edition was updated by Paul Rogers.

Now you can become a published author, too!

xii ABCs of z/OS System Programming Volume 3

Stay connected to IBM Redbooks

Chapter 1. DFSMS introduction

DFSMSdfp is a base element of z/OS. DFSMSdfp is automatically included with z/OS.

The other elements of DFSMSDFSMSdss, DFSMShsm, DFSMSrmm, and

Copyright IBM Corp. 2010. All rights reserved. 1

Tape Library dsname.f.data Tape

Figure 1-1 Introduction to data management

2 ABCs of z/OS System Programming Volume 3

Figure 1-2 Data Facility Storage Management Subsystem

Chapter 1. DFSMS introduction 3

Importance of DFSMS elements

The other elements of DFSMSDFSMSdss, DFSMShsm, DFSMSrmm, and

4 ABCs of z/OS System Programming Volume 3

DFSMSdfp provides the following functions:

Figure 1-3 DFSMSdfp functions

Chapter 1. DFSMS introduction 5

Managing devices with DFSMSdfp

Tape mount management

Distributed data access with DFSMSdfp