You are on page 1of 110

Front cover

A Practical Guide to ICF
Catalogs
Keith Winnard
Jose Gilberto Biondo Jr
Gert Laumann
Norbert Schlumberger
Jeanne Vangsness
Chao Xing CX Ma

Redbooks

International Technical Support Organization
A Practical Guide to ICF Catalogs
June 2015

SG24-8262-00

Note: Before using this information and the product it supports, read the information in “Notices” on
page vii.

First Edition (June 2015)
This edition applies to Version 2, Release 1, of DFSMS (product number 5650-ZOS).

© Copyright International Business Machines Corporation 2015. 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 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
Trademarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii
IBM Redbooks promotions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi
Authors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi
Now you can become a published author, too . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xii
Comments welcome. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii
Stay connected to IBM Redbooks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii
Chapter 1. Technical overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Introduction to the ICF catalog environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Catalogs and data sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 ICF catalog-related components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3.1 Catalog address space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.3.2 Master catalog . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.3.3 User catalogs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.3.4 VTOC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.3.5 VTOC index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.3.6 VVDS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.3.7 In Storage Cache . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.3.8 Catalog data space cache . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3.9 Enhanced Catalog Sharing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3.10 SMSVSAM and Record Level Sharing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3.11 Catalog parmlib . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3.12 ICF catalogs and tapes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.4 Catalog scenarios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.4.1 Scenario 1: Single master catalog with multiple user catalogs . . . . . . . . . . . . . . . . 7
1.4.2 Scenario 2: Start new LPAR sharing the same master catalog . . . . . . . . . . . . . . . 8
1.4.3 Scenario 3: Split master catalog into two systems . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.4.4 Scenario 4: System consolidation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.5 Reporting. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.5.1 What types of reporting are available? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.6 The need for reviews. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Chapter 2. Planning integrated catalog facility catalog environments . . . . . . . . . . . .
2.1 The plan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.1.1 Why plans are necessary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.1.2 What to expect in a technical plan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.1.3 Effect on other areas. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2 Understanding catalog search . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.3 Specifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.3.1 Technical restrictions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.3.2 Estimating catalog size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.3.3 Setting the control interval and control area size of a catalog. . . . . . . . . . . . . . . .

© Copyright IBM Corp. 2015. All rights reserved.

13
14
14
14
15
15
16
16
16
17

iii

iv

2.4 Planning considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.4.1 Integrity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.4.2 Availability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.4.3 Security . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.4.4 Operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.4.5 Documentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.4.6 ICF catalog performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.4.7 Storage management subsystem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.4.8 Features including parmlibs and caching considerations . . . . . . . . . . . . . . . . . . .
2.4.9 System software considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.4.10 Testing environments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.5 Housekeeping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.6 Reporting. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.7 Monitoring and automation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.7.1 Catalog contention detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.7.2 CA RECLAIM of VSAM KSDS data sets. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.7.3 Notify extents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.7.4 Health checks done by IBM Health Checker for z/OS . . . . . . . . . . . . . . . . . . . . .
2.8 Suggested practices summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

17
17
17
18
18
18
19
23
24
29
29
30
31
31
31
33
34
34
35

Chapter 3. Deploying ICF catalogs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.1 About the deployment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2 Scenario 1: Simple ICF catalog environment structure . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.1 Scenario 1 Target Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.2 Scenario 1: Task list . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.3 Scenario 1: Deployment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3 Scenario 2: Adding a second LPAR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3.1 Scenario 2: Target environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3.2 Scenario 2: Task list . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3.3 Scenario 2: Deployment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4 Scenario 3: Multiple MCATs and UCATs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4.1 Scenario 3: Target environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4.2 Scenario 3: Task list . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4.3 Scenario 3: Deployment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4.4 Scenario 4: Acquisition and integration. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4.5 Scenario 4: Acquisition and integration environments . . . . . . . . . . . . . . . . . . . . .
3.4.6 Scenario 4 Task list. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4.7 Scenario 4 Deployment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

37
38
38
38
39
40
42
42
43
44
45
46
46
47
51
51
52
53

Chapter 4. Maintaining ICF catalogs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.1 Maintenance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.1.1 Backup of ICF catalogs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.1.2 Reorganizing ICF catalogs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.1.3 REPRO MERGECAT as maintenance tool. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.1.4 Other types of ICF catalog maintenance. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.1.5 Recovery of ICF Catalogs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.1.6 ICF catalog diagnostics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.1.7 Modify CATALOG commands. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

57
58
58
58
59
59
60
69
69

A Practical Guide to ICF Catalogs

. . . . . . .2. . . . . . . . . . . . . . . . . . . . . . . 5. . 5. . .1. . . .2 Documentation . . . . . . 5. . . . .1. . . . . 5.1 Preparing to migrate to an RLS enabled ICF Catalog. . . . . . . . . . .3. . . . . . . . . . . . . . . 5. . . .2. . . . . . . . . . . . . . . . . . . . . 5. . . . . . . . 5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. . . . . . . . . . . . . . . . . . . . . . .3 ICF catalog possible RLS states . . . . . . . . . . . . . . 5. .Chapter 5. .1 Benchmarks for ICF catalogs using RLS mode . . . . . . . 5. 5. . . . . . . . . . . . . . . . . . . . . . .1. . . 5. . . . . . . . . . . . . . . . . . . . . . . . . . 5.2 Considerations for implementing RLS. . . . .1 Requirements for RLS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4 Fallback from using RLS mode on ICF Catalogs . . . . . . . . . . . . . . . . .5 Catalog maintenance considerations in RLS environment . . . . . . . . . . . . . . .4 Comparison with other enhanced catalog features. . .2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. . .2 Sizing the RLS catalog environment for RLS buffers and CF cache . . . . . . . . . . . . .1. . . . . . . . . . . . . . . . . . . . . . .2. . .3 Migrating to RLS enabled ICF catalog . . . . . . . . .2. . .5 New parameters on the DEFINE USERCATALOG command . Record-level sharing support for ICF catalogs . . . . . . . . .6 Using REPRO MERGECAT on an RLS-managed catalog .1. . . . . .2. . . . . . . 5. . . . 5.1 RLS for catalog overview . . . . . . . . . . . . . . .3. . . . . . . . . . . Contents 73 74 74 75 75 76 76 78 78 78 81 83 83 84 89 89 89 90 v . . . . . . . .2 Support for ICF catalogs using RLS . . . . . .7 Migration considerations . . . . . . . . . . . . . .2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3 Basic test results . . . . . . . 5. . . . . . . . . . . . . . . . . . . . . . . . . .

vi A Practical Guide to ICF Catalogs .

These examples have not been thoroughly tested under all conditions. it is the user's responsibility to evaluate and verify the operation of any non-IBM product. or service may be used. program.S. or service that does not infringe any IBM intellectual property right may be used instead. program. which illustrate programming techniques on various operating platforms. Questions on the capabilities of non-IBM products should be addressed to the suppliers of those products. serviceability. BUT NOT LIMITED TO. Furthermore. This information could include technical inaccuracies or typographical errors. program. cannot guarantee or imply reliability. vii . 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. Changes are periodically made to the information herein.Notices This information was developed for products and services offered in the U.A. marketing or distributing application programs conforming to the application programming interface for the operating platform for which the sample programs are written. © Copyright IBM Corp. compatibility or any other claims related to non-IBM products. their published announcements or other publicly available sources. and products. NY 10504-1785 U. Any reference to an IBM product. IBM may have patents or pending patent applications covering subject matter described in this document. Users of this document should verify the applicable data for their specific environment. 2015. Information concerning non-IBM products was obtained from the suppliers of those products. IBM may not offer the products. EITHER EXPRESS OR IMPLIED. this statement may not apply to you. for the purposes of developing. The materials at those websites are not part of the materials for this IBM product and use of those websites is at your own risk. IBM has not tested those products and cannot confirm the accuracy of performance. IBM may use or distribute any of the information you supply in any way it believes appropriate without incurring any obligation to you. Some states do not allow disclaimer of express or implied warranties in certain transactions. You may copy.A. All rights reserved.S. The furnishing of this document does not grant you any license to these patents. 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. Some measurements may have been made on development-level systems and there is no guarantee that these measurements will be the same on generally available systems. therefore. therefore. IBM. using. Armonk. the examples include the names of individuals. MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. or features discussed in this document in other countries. brands. Therefore. or function of these programs. and distribute these sample programs in any form without payment to IBM. some measurements may have been estimated through extrapolation. You can send license inquiries. Any functionally equivalent product. Any performance data contained herein was determined in a controlled environment. Any references in this information to non-IBM websites are provided for convenience only and do not in any manner serve as an endorsement of those websites. COPYRIGHT LICENSE: This information contains sample application programs in source language. these changes will be incorporated in new editions of the publication. in writing. program. or service. services. To illustrate them as completely as possible. North Castle Drive. IBM Corporation. modify. However. Consult your local IBM representative for information about the products and services currently available in your area. or service is not intended to state or imply that only that IBM product. companies. the results obtained in other operating environments may vary significantly. Actual results may vary. This information contains examples of data and reports used in daily business operations. THE IMPLIED WARRANTIES OF NON-INFRINGEMENT. to: IBM Director of Licensing. INCLUDING. All of these names are fictitious and any similarity to the names and addresses used by an actual business enterprise is entirely coincidental.

Trademarks IBM. product. A current list of IBM trademarks is available on the web at http://www.com are trademarks or registered trademarks of International Business Machines Corporation in the United States. and ibm.shtml The following terms are trademarks of the International Business Machines Corporation in the United States. or both: Easy Tier® IBM® Parallel Sysplex® RACF® Redbooks® Redbooks (logo) ® Resource Measurement Facility™ RMF™ z Systems™ z/OS® The following terms are trademarks of other companies: Other company. other countries.com/legal/copytrade. indicating US registered or common law trademarks owned by IBM at the time this information was published. the IBM logo. viii A Practical Guide to ICF Catalogs . or both. or service names may be trademarks or service marks of others. Such trademarks may also be registered or common law trademarks in other countries. These and other IBM trademarked terms are marked on their first occurrence in this information with the appropriate symbol (® or ™). other countries.ibm.

® Qualified IBM Business Partners may place a full page promotion in the most popular Redbooks publications.IBM REDBOOKS PROMOTIONS IBM Redbooks promotions Find & read thousands of IBM Redbooks publications Search. Imagine the power of being seen by users who download millions of Redbooks publications each year! ibm. featuring your business or solution with a link to your web site. save & organize favorites Get up-to-the-minute Redbooks news & announcements Link to the latest Redbooks blogs & videos Download Now Android iOS Get the latest version of the Redbooks Mobile App Promote your business in an IBM Redbooks publication ® Place a Sponsorship Promotion in an IBM Redbooks publication. bookmark.com/Redbooks About Redbooks Business Partner Programs .

THIS PAGE INTENTIONALLY LEFT BLANK .

Data Set Services (DFSMSdss). delivered. In addition. Jose’s areas of expertise include installing and maintaining storage products. Norbert provides marketing support for DFSMSrmm. DFSMSdss. However. DFSMShsm. His experience is mostly with IBM products. and deployment activities to help you create and maintain a balanced and efficient catalog environment.Preface This IBM® Redbooks® publication gives a broad understanding of integrated catalog facility (ICF) catalog environments. such as the Data Facility Storage Management Subsystem (DFSMS) components Data Facility Product (DFSMSdfp). but he also works with OEM software products. including 24 years of experience with DFSMSrmm. Jose Gilberto Biondo Jr is an IT Specialist in Integrated Technology Delivery. Norbert Schlumberger is an IT Architect with IBM Germany. He is currently helping achieve this efficiency across Nordic regions by also having hardware consolidated in fewer data centers. He has over 35 years of experience in storage software and storage management for IBM and client systems. to support the DFSMSrmm business. © Copyright IBM Corp. He has seven years of experience in z/OS. He has many tools available. he has also worked with original equipment manufacturer (OEM) software and hardware products. He has developed. Server Systems Operations/Storage Management in IBM Denmark. Gert Laumann is an IT Specialist in Integrated Technology Delivery. and DFSMSrmm. planning. working with storage management since 2007. such as DFSMSdfp. He is the team leader for the Scandinavian mainframe storage team. and IBM TS7700 Virtual Engines. Poughkeepsie Center. power. It includes suggestions for design. Keith Winnard is the ITSO Redbooks Publications Project Leader for IBM z/OS® and related topics at the ITSO. 2015. Hierarchical Storage Manager (DFSMShsm). and Removable Media Manager (DFSMSrmm). he sets up shared direct access storage device (DASD) and tape environments to mitigate data center constraints on floor space. working with storage management since 1989. and educates clients around the world about DFSMSrmm. Poughkeepsie Center. ServerSystems Operations/Storage Management in IBM Brazil. and cooling. including IBM TotalStorage 3494 and IBM TotalStorage 3495 Automated Tape Libraries (ATLs). Virtual Tape Servers (VTSs). He is experienced in blending traditional z/OS environments and applications with web middleware and applications. Authors This book was produced by a team of specialists from around the world working at the IBM International Technical Support Organization (ITSO). Four scenarios are provided to illustrate sample implementations of typical activities associated with an organization’s requirements. He has presented on many mainframe-related topics. and process automation. His focus has mainly been automation and standardization across mainframe client platforms. such as a Tape Copy Tool. He joined information technology (IT) in 1977 and has worked for various clients and IBM Business Partners. xi . All rights reserved. He has 30 years of experience in z/OS. His areas of expertise include performing conversions from vendor tape management products to DFSMSrmm and new DFSMSrmm implementations. Jose works mainly with IBM storage products.

Discover more about the residency program. SG24-6116. He has supported clients in his areas of expertise.12 DFSMS Technical Update. grow your career. including Implementing an IBM b-type SAN with 8 Gbps Directors and Switches. He has 5 years of experience in the area of Mainframe Technical Support and Customer Service Delivery. all at the same time. and z/OS V1. and become a published author. IBM System Modification Program/Extended (IBM SMP/E) for z/OS. browse the residency index. Your efforts will help to increase product acceptance and customer satisfaction. and has spent 19 of those years working with customers and developers on the DFSMS product set. including z/OS. Jeannie has assisted with other IBM Redbooks publications. She has been with IBM for 26 years. Chao Xing CX Ma is an Advisory IT Specialist in IBM China. Jeannie has co-authored several Redbooks publications. Join an ITSO residency project and help write a book in your area of expertise. Poughkeepsie Center Now you can become a published author.html xii A Practical Guide to ICF Catalogs . as you expand your network of technical contacts and relationships. and apply online at: ibm. IBM Parallel Sysplex®. while honing your experience using leading-edge technologies. IBM z Systems™ Software. IBM San Jose Bob Haimowitz Development Support Team (DST). IBM San Jose Barbara McDonald DFSMS Product Strategist. and disaster recovery solutions. and you can participate either in person or as a remote resident working from your home base. Residencies run from two to six weeks in length.Jeanne Vangsness is a DFSMS Software Technical Support Engineer specializing in the DFSMShsm product area. too Here’s an opportunity to spotlight your skills. DFSMS.com/redbooks/residencies. She provides feedback to the DFSMShsm development in several areas to help make improvements to the functionality of DFSMShsm. Thanks to the following people for their contributions to this project: Patricia Driscoll Choi Tony Hui Xu Rich Lungerhausen DFSMS Catalog Development. SG24-7895.

redbooks.ibm.Comments welcome Your comments are important to us.com 򐂰 Mail your comments: IBM Corporation.html Preface xiii .com/redbooks 򐂰 Send your comments in an email: redbooks@us.ibm. HYTD Mail Station P099 2455 South Road Poughkeepsie.facebook. NY 12601-5400 Stay connected to IBM Redbooks 򐂰 Find us on Facebook: http://www. We want our books to be as helpful as possible.ibm.com/rss. and workshops with the IBM Redbooks weekly newsletter: https://www.linkedin.nsf/subscribe?OpenForm 򐂰 Stay current on recent Redbooks publications with RSS Feeds: http://www.com/ibmredbooks 򐂰 Look for us on LinkedIn: http://www. Send us your comments about this book or other IBM Redbooks publications in one of the following ways: 򐂰 Use the online Contact us review Redbooks form: ibm.com/IBMRedbooks 򐂰 Follow us on Twitter: http://twitter.com/Redbooks.com/groups?home=&gid=2130806 򐂰 Explore new Redbooks publications. International Technical Support Organization Dept.redbooks. residencies.

xiv A Practical Guide to ICF Catalogs .

1 . All rights reserved. deployment.1 Chapter 1. The chapter contains information about the following topics: 򐂰 A brief introduction 򐂰 A summary of ICF catalog-related components 򐂰 Sample scenarios of ICF catalog environments Four scenarios are used throughout this book to provide a consistent point of reference for you. © Copyright IBM Corp. and maintenance of the ICF catalog environments of each scenario. or want to broaden your existing knowledge and appreciation of ICF catalog environments. this chapter aims to present you with a high-level understanding of why there is a need for catalogs and the roles they play. Subsequent chapters deal with the planning. Technical overview Welcome to the Technical Overview. 2015. If you are relatively new to the integrated catalog facility (ICF) catalogs.

1. partners. Therefore. testing. the areas of marketing. Some data sets might only be used by a single application. safe. demonstrate the deployment of those plans. These areas need processes. 򐂰 The second source is internal to the organization: – Data is generated from processes as part of the transaction fulfillment. and processes need data. has a direct effect on the number of data sets required to store it. This book describes the basic components required to provide a solution to the previous four questions. Well-structured multiple ICF catalog environments are essential to accommodate an organization’s data sets.2 Catalogs and data sets Organizations typically consist of multiple applications. – Data is generated for analytical and reporting activities to gain a better understanding of the organization’s well-being and opportunities. access. and show how to maintain control of your ICF catalog environment. 2 A Practical Guide to ICF Catalogs .1. and backup purposes. suggest four linked scenarios and their relevant plans. including how it is categorized. there can be several versions or iterations of the data sets used. The increasing rate of data growth presents challenges for the definition. and usually has many data sets associated with it. availability. management. human resources. suppliers. consider that two general sources are responsible for the creation of data: 򐂰 The first source is external to the organization. logistics. An organization’s data can be categorized and stored in separate data sets for use by different applications. We explain considerations for planning environments to accommodate the solution. where others might be accessed by multiple applications. In addition. and anyone who interacts with the organization. so the scope of this book is focused on z/OS. Without it. For perspective. this number can range from tens of thousands to millions. An ICF catalog is a key component in determining a data set’s location. Data is captured from transactions with clients. production. Where does data come from? For the purposes of this IBM Redbooks publication. and so on cannot exist. sales. Each application has a discrete function in the organization. and efficient ICF catalog environment? ICF catalogs are part of the IBM z/OS environment. Questions abound relating to data. but the scope of this book aims to answer four fundamental questions: 򐂰 򐂰 򐂰 򐂰 How do I create an ICF catalog environment? What are the considerations for planning an ICF catalog environment? How might I deploy ICF catalogs? How do I maintain a balanced. and maintenance of ICF catalogs. several components and tools are required to control the creation and subsequent management of all of these data sets. order processing. The voluminous nature of data in the modern world. such as for development. ICF catalogs are also a key cornerstone to data storage and management.1 Introduction to the ICF catalog environment The most unique feature of any organization is its data: The data defines the organization.

any change to that environment might have an effect on other components in or related to that environment. or if you are resolving unexpected situations. Consider this as a sort of checklist for which components might be available. and to assist in minimizing the risk of data loss or outage. too.Careful planning is essential to successfully create and maintain an ICF catalog environment. This helps you to sustain optimum value to the organization. The master catalog acts as the user catalog inventory and contains details of all of the UCATs running on a z/OS system. You will find several UCATs on z/OS systems.3 ICF catalog-related components Figure 1-1 on page 4 shows which z/OS-related components can play a part in catalog processing. depending on your setup and which features you use. The actual number depends on how the organization has decided to categorize its data and data sets. 1. which is a term that can represent or mean other things. The terms buffers and caches are sometimes interchangeable. Buffers do not always represent cache. but that is only in certain contexts. Technical overview 3 . The important point here is to understand that an ICF catalog is part of an integrated environment. As part of the planning process. Chapter 1. ICF catalogs are divided into two categories: 򐂰 The master catalog (MCAT) 򐂰 The user catalog (UCAT) The MCAT is the catalog of catalogs. The next section provides an overview of the related components that are necessary to build an ICF catalog environment. and other key data sets and objects. the ongoing monitoring and refinement of the ICF catalog environment must be regularly reported and reviewed. Caching is an integral part of the ICF catalog environment if acceptable performance is to be achieved. Therefore. perhaps the figure can serve as a guide for where to look if you want to make changes to your ICF catalog environment. In addition to holding information about the UCATs. Alternatively. it also contains entries for critical data sets relating to the operating system (OS) software.

The following sections list the related components with a short description of what these are and how they play a role in the catalog environment. where all searches for data sets start. This is described further throughout this book. CAS holds control block information for open catalogs.1 Catalog address space Core catalog management is done in the catalog address space (CAS). which is started at the time of initial program load (IPL). it was moved to a new volume). it remains open until CAS determines that something fundamental about the catalog has changed (for example. the searches are directed to the user catalog assigned for the individual search (based on the alias).Figure 1-1 show components related to z/OS. A user catalog is opened by CAS when the first request is directed to it. 4 A Practical Guide to ICF Catalogs . and the catalog control blocks need to be rebuilt. CAS can be displayed and managed by commands. which are listed and explained in more detail later in this book. and it remains open until closed by command. virtual lookaside facility (VLF). The areas where catalog information can be cached are represented in the In Storage Cache (ISC). Communication between cache and user address spaces happens using cross-memory services. From there. or storage management subsystem (SMS) Virtual Storage Access Method (VSAM) components. Figure 1-1 Catalog component overview Figure 1-1 shows the basic catalog structure and environment from the master catalog. 1. The user catalog points to the volume where the data set is for this search. and for ALIAS objects and volume serial numbers (VOLSERs).3. Alternatively.

3. The SYSCATxx member of SYS1. entries are removed from the cache based on a least recently used (LRU) basis.3. 1. There is one VTOC per volume. then the information in the LOADxx member will be used instead of that contained in the SYSCATxx member. the location of every data set and the number of extents allocated to it. 1. User catalogs hold information about user and application data sets.2 Master catalog The MCAT is the highest level in the catalog structure. However.NUCLEUS contains information about the MCAT. The purpose of the VTOC index is to speed up locating data sets on a volume by using an index rather than reading serially through the VTOC.PARMLIB can also be used to identify the MCAT and. The LOADxx member of SYS1.3. 1. The amount of memory used is fixed.3. When specifying the data set name. VLF is not specified.6 VVDS The VSAM volume data set (VVDS) holds additional information about the data sets on the volume. and catalogs are cached in ISC.3. The MCAT must be identified at IPL time. but can be built based on the VTOC with the BUILD IX function in IBM Device Support Facilities (ICKDSF) at a later date.3 User catalogs A user catalog’s data sets are pointed to with an alias in the master catalog. Technical overview 5 . When located. among other details.1. the MCAT contains an object called the ALIAS (there can be many aliases). if this is the case. 1. and when the cache fills. and VSAM data set extent information. For these. It includes. It is the ALIAS that identifies and points to the UCAT containing the entry for the requested data set.7 In Storage Cache The ISC is the memory in CAS reserved for caching. This information includes catalog information. The VTOC index is normally created at volume initialization. Therefore. such as the data set name and the corresponding VOLSERs. Data Facility Storage Management Subsystem (DFSMS) information. Chapter 1. ISC is the default caching location. 1. there is no limit to the number of entries. a search is made in the master catalog for an alias.4 VTOC The volume table of contents (VTOC) holds information about the data sets on the volume.5 VTOC index The VTOC index is a key-sequenced register of all data sets on the volume. this does not apply to master catalogs.3. In addition. for example. the alias entry points to the user catalog that contains the data set. where information about UCATs and system-critical data sets are to be found (SYS1 data sets are cataloged in the master catalog).

Catalog parmlib can replace settings for the catalog. 1.3.3. Tape data sets are defined as non-VSAM data sets when cataloged in an ICF catalog. SMSVSAM can switch RLS-managed catalogs in and out of RLS mode. RLS access has removed the previous approach of enqueuing on the entire catalog (SYSIGGV2). or VLF caching.1. Catalogs are not limited by storage at a catalog level. is based in a data space. Many settings would previously require IPL. It is defined through the COFVLFxx member in SYS1. These scenarios are intended to relate to requirements during an organization’s system and data growth.10 SMSVSAM and Record Level Sharing SMSVSAM is the started task handling catalog VSAM record-level sharing (RLS) requests.11 Catalog parmlib Catalog parmlib was introduced in z/OS V1. previously done in the LOADxx or SYSCATxx member. such as DFSMS Removable Media Manager (DFSMSrmm).13. 6 A Practical Guide to ICF Catalogs . This is intended to improve performance and availability.3. the common operations used for them. At the point when the storage becomes saturated.3. and also the restriction or limitation for catalog processing. therefore. Tape data sets do not need to be cataloged like DFSMS-managed disk data sets. and instead enqueues at the record level. but now can be dynamically modified by restarting CAS.PARMLIB. Tape data sets are also cataloged in an ICF catalog. but use whatever storage is assigned to the data space in total.4 Catalog scenarios This book uses four sample ICF catalog environment scenarios to help illustrate different aspects of ICF catalog environments. Each scenario introduces another layer of complexity to the previous one. In this way. For tape data sets. an LRU algorithm starts removing entries from the VLF cache. RLS access is now available to ICF user catalogs. 1.3. they are normally registered in a Tape Management system.12 ICF catalogs and tapes Tape data sets do not as such differ from disk data sets in terms of cataloging. The four scenarios show how MCAT and UCATs work. There are currently more than 20 keywords that can decide the ICF catalog settings. such as deleting and redefining these data sets. It is advised to have a CF structure for RLS-managed catalogs to ensure optimum buffering in RLS mode. 1. 1. CAS connects to this during initialization.8 Catalog data space cache Catalog data space cache (CDSC). though.9 Enhanced Catalog Sharing The Enhanced Catalog Sharing (ECS) gives far better performance for shared catalogs. 1. All non-VSAM catalog handling. applies to tape. ECS uses a coupling facility resource management (CFRM) policy in the coupling facility (SYSIGGCAS_ECS). they are protected and easily accessible without being registered in an ICF catalog.

Table 1-1 ICF catalog relationship in Scenario 1 User catalog Connected Master catalog Available for LPARs UCAT A1 MCAT A LPAR A UCAT A2 MCAT A LPAR A UCAT B1 MCAT A LPAR A UCAT B2 MCAT A LPAR A Chapter 1. we repeat the scenario diagrams. LPAR A MCAT A UCAT A1 UCAT A2 UCAT B1 UCAT B2 Figure 1-2 ICF catalog design in Scenario 1 You can see the detailed relationship between the catalogs in Table 1-1. In the deployment chapter. their respective plans. and the output associated with those commands. the job control language (JCL) commands. and all four UCATs connect to MCAT A. LPAR A uses MCAT A as the MCAT to IPL. It shows one single logical partition (LPAR) with one MCAT and four UCATs.1 Scenario 1: Single master catalog with multiple user catalogs This scenario is the basic configuration of a simplified z/OS system.Subsequent chapters describe planning considerations for each of these scenarios. 1. as shown in Figure 1-2.4. Technical overview 7 .

All user catalogs are available for both LPARs. All other system operations are not included. LPAR A could be the production environment. Figure 1-3 shows a clear picture of the catalog structure. The new LPAR (LPAR B) could be used as an application development and testing environment. LPAR B UCAT A2 MCAT A LPAR A.2 Scenario 2: Start new LPAR sharing the same master catalog As the organization grows. LPAR B UCAT B2 MCAT A LPAR A. This scenario introduces a new LPAR using the same MCAT. LPAR B UCAT B1 MCAT A LPAR A. LPAR B A Practical Guide to ICF Catalogs . Table 1-2 ICF catalog relationship in Scenario 2 8 User catalog Connected Master catalog Available for LPARs UCAT A1 MCAT A LPAR A.1. Important: This part only focuses on the catalog perspective. it might be advisable to expand from the single LPAR to a sysplex or multi-LPARs with data sharing using a CF.4. LPAR A LPAR B MCAT A UCAT A1 UCAT A2 UCAT B1 UCAT B2 Figure 1-3 ICF catalog design in Scenario 2 You can see the target structure of each catalog in Table 1-2.

3 Scenario 3: Split master catalog into two systems In certain circumstances the organization might want to split the sysplex into two independent systems.1. though others will be dedicated to a specific LPAR. MCAT B LPAR A. The new system has several requirements: 򐂰 A new master catalog 򐂰 A move of a subset of user catalogs to the new separate catalog structures 򐂰 A connection of those user catalogs to the new master catalog There might be still a requirement to share some but not all data sets. LPAR B UCAT B1 MCAT A. you must pay attention to non-shared user catalogs UCAT A1 and UCAT B2. Technical overview 9 . LPAR B uses MCAT B then UCAT B2 dedicated connect to MCAT B. LPAR A LPAR B MCAT Split MCAT A UCAT A1 MCAT B UCAT A2 UCAT B1 Shared Shared UCAT B2 Figure 1-4 ICF catalog design in Scenario 3 Table 1-3 shows the final status of each catalog after splitting.4. Table 1-3 ICF catalog relationship after split in Scenario 3 User catalog Connected Master catalog Available for LPARs UCAT A1 MCAT A LPAR A UCAT A2 MCAT A. but UCAT A2 and UCAT B1 are shared to LPAR A and LPAR B. they can’t be used by the other LPAR. You can see in Figure 1-4 that MCAT B is the new Master catalog split from MCAT A. LPAR B UCAT B2 MCAT B LPAR B Chapter 1. MCAT B LPAR A. and therefore some of the catalogs will have to be shared.

1. Table 1-4 ICF catalog relationship before consolidation in Scenario 4 User catalog Connected Master catalog Available for LPARs UCAT A1 MCAT A LPAR A UCAT A2 MCAT A. LPAR B UCAT B2 MCAT B LPAR B UCAT C1 MCAT C LPAR C UCAT C2 MCAT C LPAR C After the new system has been examined and has settled in. the newly acquired would be an independent system with its own MCAT and UCATs. but initially keep the data separate.4. Suppose our sample organization acquires another organization. As shown in Figure 1-5. In this case. MCAT B LPAR A. LPAR B UCAT B1 MCAT A. MCAT B LPAR A. LPAR A LPAR C MCAT C UCAT C1 LPAR B MCAT A UCAT C2 UCAT A1 MCAT B UCAT A2 Shared UCAT B1 UCAT B2 Shared Figure 1-5 ICF catalog design before consolidation in Scenario 4 See the detailed relationship between catalogs before consolidation in Table 1-4. So. 10 A Practical Guide to ICF Catalogs .4 Scenario 4: System consolidation This scenario deals with a consolidation and integration of ICF catalog environments. the organization might want to restructure the catalog further for more efficient running. we call this LPAR C. and two user catalogs (UCAT C1 and UCAT C2). The organization might want to consolidate all of the LPARs onto its own mainframe. LPAR C is an independent system with its own master catalog (MCAT C). and therefore gains another LPAR.

LPAR C UCAT B2 MCAT B LPAR B UCAT C1 MCAT A LPAR A. Figure 1-6 describes the final relationship of catalogs after MCAT A and MCAT C consolidation.The aim might be to merge MCAT C to MCAT A. LPAR C UCAT A2 MCAT A. MCAT B LPAR A. LPAR C UCAT C2 MCAT A LPAR A. then load LPAR C with MCAT A. it is important to generate reports that can show you the information you need in a simple and efficient format. and business grow. Unlike alerts that are issued when an error is identified on the system. where one team can predict and prevent outages. and the other identify business opportunities and application trends. applications. Reports are valuable to both technical and management teams. LPAR B. LPAR B. MCAT C will not be used. After this. Technical overview 11 . it is necessary to identify what types of reports are available. and enable the readers to forecast issues. LPAR C 1. and help guide projects. and to what groups this information is relevant. Chapter 1. identify opportunities. MCAT B LPAR A. connect UCAT C1 and UCAT C2 to MCAT A. LPAR C UCAT B1 MCAT A. what information they deliver.5 Reporting To understand how your data. LPAR C LPAR A MCAT A MCAT C UCAT C1 LPAR B UCAT C2 MCAT B UCAT A1 UCAT A2 Shared UCAT B1 UCAT B2 Shared Figure 1-6 ICF catalog design after consolidation in Scenario 4 You can see the final status of each catalog after system consolidation in Table 1-5. Table 1-5 ICF catalog relationship after consolidation in Scenario 4 User catalog Connected Master catalog Available for LPARs UCAT A1 MCAT A LPAR A. reports should consist of relevant data grouped into a format that is easy to understand. If reports can be used by different areas.

and delete requests. update. circumstances. report. The most effective environments are planned and based on understanding variables. All ICF catalogs and related components must be available to support the creation and retrieval of data. The OS software must provide the functions and tools to perform access to data sets. Today. The environment should be configured in a manner that successfully contributes to meeting SLAs. or a slow and steady creep in several areas. The content must be valid. so it is possible to identify the catalogs. The environment’s structure should facilitate both permanent and temporary changes to the catalog’s configuration as required. data creation and access. A low cache hit can lead to higher processor usage or performance degradation. 12 A Practical Guide to ICF Catalogs . data creation and retrieval are on an unprecedented scale as social media expands. The environment must be fit for the purpose of. and modify the environment. It should not be perceived as a hindrance. from the catalog: Catalog Provide information about the number of accesses to the catalog for insert. Performance Show performance statistics for all the types of catalog events. A defined and quantified tolerance ought to be built into the plan. 򐂰 Functionality. result in the destabilization of the environment in one or more of the pillars described previously. faster and deeper analytics are needed. collaborative initiatives grow. This might be due to a single instance (such as a company acquisition). analyze. measure. this is unlikely to re-establish the environment’s effectiveness.1. and affect service level agreements (SLAs). This information can be used to identify possible catalog bottlenecks. 1. In the short term. which are defined by default. 򐂰 Availability.1 What types of reporting are available? You can request the following statistic reports. and therefore the applications in the catalog that have a high data set activity. such as the requirements. 򐂰 Performance.5.6 The need for reviews An ICF catalog environment stands on six pillars: 򐂰 Integrity. The purpose of the reviews is to not only to prevent destabilization. if exceeded. but also to try to provide an ongoing improvement and contribution to the organization. and goals of the business. including cache hit percentage. and contribute to all aspects of. The review process is described in more detail in later sections. and pressures are put on an organization to perform on demand. 򐂰 Flexibility. read. These tolerance values represent the points that. Exceeding the thresholds unexpectedly causes reactive tactical behavior to minimize the effect of the issue. and to monitor. which can affect application performance. Note: There are other reports that can be created from System Management Facilities (SMF) records by the technical team to help troubleshoot issues and monitor the system. Cache Display the catalog cache information. 򐂰 Effectiveness.

© Copyright IBM Corp. Considerations and suggestions are given to help improve the chances of producing a realistic plan that meets your organization’s needs. 13 . All rights reserved. 2015. we describe how integrated catalog facility (ICF) catalog environments might be planned.2 Chapter 2. Planning integrated catalog facility catalog environments In this chapter.

1 The plan The planning of ICF catalog environments considers various aspects related to business needs. The planning phase is where the thinking and designing takes place. depending on the activity. In addition. Discuss the plan thoroughly with all interested parties to set common expectation levels. The goal is to create a plan for the design and processes that both builds a balanced ICF catalog environment. defined. security. and maintained. and the best blend achieved. and explains the approach taken. Previous plans are useful documentary evidence of why and how configurations have been designed.2. environment.1.1 Why plans are necessary A structured ICF catalog environment enables ever-changing data to be accessed both effectively and efficiently. it provides a record of how a particular need was met.2 What to expect in a technical plan Concise and clear plans are preferable. too. and must be used in a cohesive manner to ensure that the environment’s goals are met. and ensure that its effectiveness is maintained. and ensure that everyone understands both their own roles and the roles of others. Each of the related components plays an essential part. This does not happen by luck alone. 2. stopping and starting other components) Who is responsible for delivering the task Review points Backout instructions if the tasks fail or cause post-implementation issues A Practical Guide to ICF Catalogs . and other local standards and guidelines: 򐂰 򐂰 򐂰 򐂰 򐂰 򐂰 򐂰 򐂰 14 Objective of the plan Deliverables Task identifier and description Components to be changed Activities related to the tasks (for example. Other factors might need to be considered. legal requirements. Other tasks can also appear. The plan contains a series of tasks.1. The objective must be clear and unambiguous. Each related component’s functionality and configuration must be considered. Expect to see some or all of the following points in the plan. The following examples show why a plan is needed: 򐂰 򐂰 򐂰 򐂰 򐂰 򐂰 򐂰 The first implementation of a new system or logical partition (LPAR) Addition of a secondary LPAR to the original implementation and data sharing Separation of data sharing Acquisition of another business and merger of different systems As a result of a regular review A temporary configuration to ease the business through a transitional phase Planning for disaster recovery and business continuity The plan provides the path to be followed to set objectives. and provide the appropriate outcomes to satisfy the needs of the initiative. 2. and technical capabilities. meet the requirements.

A catalog locate request comes from a batch job or a user. together with the volume serial numbers (VOLSERs) of the data set being searched for.3 Effect on other areas ICF catalogs are a key component of the operational configuration. there can be an effect on other areas. The request proceeds to the volume table of contents (VTOC) and Virtual Storage Access Method (VSAM) volume data set (VVDS) where the data set is located. when changes are made to the ICF catalog environment or a related component. 2. The following immediate areas might be affected: 򐂰 򐂰 򐂰 򐂰 򐂰 Operational housekeeping Security administration Operational automation Performance management Reporting The following other areas are not so immediate. The data set is located in the UCAT. but should still be considered: 򐂰 Disaster recovery 򐂰 Capacity planning 򐂰 Hardware configuration 2. Chapter 2. Therefore. where a key-sequenced search for the data set is performed. and the request is directed to that UCAT.2 Understanding catalog search Catalog standard search order is a process that occurs when you attempt to locate a data set. and are used by all of the online and batch workloads. The request can be completed. The ALIAS also contains information relating to where the data set is located (disk or tape). 4. such as using the CAT statement to point to a nonstandard search order catalog. attempts to locate the data set follows the standard search order. Planning integrated catalog facility catalog environments 15 . The ALIAS points to a UCAT. and the ALIAS leads it to the user catalog (UCAT). Search enters the master catalog (MCAT). The search might use the ALIAS object to go through the ICF catalog hierarchy to find a data set entry.1. 3. The following steps represent a typical catalog search: 1. Unless other parameters are set.2. The search starts in the MCAT based on the ALIAS of the data set in scope.

2. SC23-6853. 򐂰 The BCSs cannot be compressed or striped.3. and VVDS. Therefore. excessive secondary extents can decrease catalog performance. For IBM z/OS V1R13 or later. 򐂰 A BCS data set can have up to 123 extents. 򐂰 To exploit extended addressability. 16 A Practical Guide to ICF Catalogs . and be defined as extended format (DSNTYPE=EXT).3 Specifications If performance and scalability are key requirements. although secondary extents of BCS do not cause any functional problem. 򐂰 By default.2. because the only extended format option available for a BCS is extended addressable. A practical maximum number is 3000 aliases per catalog.1 Technical restrictions Consider the following technical restrictions: 򐂰 The basic catalog structure (BCS) must be a VSAM key-sequenced data set (KSDS). you can increase the number of possible aliases to a theoretical limit of over 500. 2. the number of aliases that a catalog can have is limited by the maximum record size for the MCAT. a catalog should be ideally defined with sufficient space to avoid excessive or unnecessary secondary extents.000 aliases (depending on the length of alias names).3. tape volume catalog. but can only occupy space on a single volume. If the extended addressability format is used. the BCS must be storage management subsystem (SMS)-managed. The required space of a BCS depends on the type and number of data sets and objects that are cataloged in the BCS. consider the following topics when planning your ICF catalog structure. They may vary for future z/OS versions. if the EXTENDEDALIAS option is enabled.1. see the topic about determining catalog size in z/OS DFSMS Managing Catalogs. Therefore. 򐂰 A BCS is limited to 4 gigabytes (GB) unless it is defined as an extended format BCS. Note: The previous technical limitations are based on z/OS V2. the size limit of the BCS is determined by the control interval size multiplied by 4 GB. a BCS with a control interval size of 4 kilobytes (KB) has a maximum data set size of 16 TB.2 Estimating catalog size Based on operational experience. The required amount of BCS space for each type of data set or object can vary according to the following factors: 򐂰 򐂰 򐂰 򐂰 򐂰 򐂰 򐂰 Length of the data set or component names Number of volumes per data set Number of relationships between components Number of alternative indexes for a data set Number of paths Presence of security information Presence of SMS information For details about space estimation and calculation of BCS.

backing up. you need to prove it. consider the implications of the intended changes you plan to make. Chapter 2. which requires a degree of flexibility. Are there other areas that will need to take actions? The following topics are areas to consider during the planning. Be more granular in the design. This might be listing.4 Planning considerations Whether planning a full ICF catalog environment or modifying an existing one. and provide the opportunity for a quicker and simpler recovery. 2.2 Availability Planning for availability is the high priority. In addition. the performance of BCS is also related to the setting of control interval (CI) and control area (CA). but before the components are made available for general use. Several catalogs servicing multiple applications can reduce the chance of an outage. ICF catalog availability planning includes designing your environment with the minimum risk of experiencing a partial or full outage. assume that integrity already exists at the start of the exercise. Beware of the implications of enabling one ICF catalog to have too many applications accessing data through it. 򐂰 Run regular utilities to check that the catalog is accessible and has integrity. Data sets access is severely limited or even impossible without the catalogs. assessing record counts. or any other action to prove that the entity is intact and operable. 򐂰 Set up test catalogs to develop catalog recovery skills. and post-implementation phases. depending on the cause and size of the outage. 2. minimize the effect of that outage. Planning integrated catalog facility catalog environments 17 .3 Setting the control interval and control area size of a catalog As a VSAM KSDS.3.4. However. implementation. Also devise a method to test the integrity of affected components when the implementation activities are complete. For the practical setting guideline of catalog CI and CA. SC23-6853. 򐂰 Back up the catalogs regularly. and practice them appropriately. Use whatever means or utilities to prove that the ICF catalogs or related components are intact. 2. or enable it to take the default value.1 Integrity How can a plan include integrity? Unless the objective of the plan is to specifically restore integrity to a broken component. The following considerations are typical: 򐂰 Choose where to allocate your catalogs to minimize the exposure to unnecessary risk of deletion or corruption. see the topic about setting the catalog control interval and area size in z/OS DFSMS Managing Catalogs. you can specify the CI size in the job control language (JCL). ICF catalogs and their related components are essential for data access in a z/OS environment. When defining a new BCS with the access method services (IDCAMS) multifunction utility.2.4. the design should include an error-free recovery in the shortest time possible.

and the complexity levels of the applications.3 Security Discuss the proposed environment or new entity with security administration to avoid inadvertently creating exposures. to help increase availability levels. When either party plans a change. because their areas are so closely connected.4 Operations The storage management team must work closely with the operational support teams. and the processes for maintaining those environments. ICF catalog environment portability enables a catalog to be moved from one system to another. Pay particular attention to documenting the environments and their structure. presents better recovery options. or might prolong the recovery period to a point beyond acceptable business continuity thresholds. The documentation must be self-contained and portable. Reduce the risk of a failing catalog from spreading to other catalogs or affecting multiple applications. and addresses relevant topics concerning that subject. Should production.4. with minimum risk and a structured approach.4. and test data be in the same catalog? Is it wiser to separate different applications? The answers depend on the policies and guidelines of the organization. development. Recovering a system with no documentation might not be possible. Be aware that some documentation is stored off-site. This is not acceptable if the business is to continue. carefully plan how catalogs can be repaired from another system. How will a partial implementation of a change be handled in terms of documentation? Perhaps circumstances arise whereby an implementation is only partially achieved. The plan must include a section to handle the change control of the documentation. 2. It must also have a simple structure. 2. and be easy to follow. Sharing ICF catalogs across more than one system.5 Documentation Documentation is most effective when it accurately reflects the subject. Availability levels can also be improved by considering what kinds of data sets are cataloged into each ICF catalog. This can lead to documentation becoming inaccurate. or is planned to take place in several steps over several weeks. explore the following areas: 򐂰 򐂰 򐂰 򐂰 򐂰 Effect on the operational schedule Operational jobs access to data sets Changes to the automation requirements Operational procedures and supporting documentation Potential effect on service level agreements (SLAs) 2. the other party must be included to help assess any potential risks and resolve them.Simple structures and environments are easier to understand than overly complex configurations. and approve the new environment status. In particular. and the situation can become more critical if one of the implementation steps is backed out. At a simpler system level. Disaster recovery documentation should be kept in line with all systems concerned. either through temporary or permanent sharing. test changes. This must also be updated as appropriate. Include tasks for security administration to set up. or to a disaster recovery site. or as part of a highly functional sysplex.4. 18 A Practical Guide to ICF Catalogs .

򐂰 Review the documentation regularly to ensure that it accurately describes the ICF environments and configurations. you might later choose to change the values that you initially specified. because splits might happen immediately.4. The default is STRNO + 2. Rather than use the BUFFERSPACE option. BUFND could be specified as STRNO + 1. it is suggested that you follow a few guidelines as to parameters in your DEFINE USERCATALOG command. 򐂰 Keep multiple copies of the documentation in several locations. The STRNO (string number) option defines the number of parallel reads to the ICF catalog. To ensure the best possible performance on your ICF catalogs. The setup. sharing. and monitoring. Use STRNO=3 as a good start (the default is 2). For SHAREOPTIONS and RECORDSIZE. If it adds value. Limit access to only those who need it. Without this separation. and perform selected tuning activities based on any performance anomalies. if the system fails. especially away from the system it is describing. It provides an indication of how well the ICF catalog environment meets workload demands. 򐂰 Align the documentation with the respective systems. use BUFNI and BUFND instead. ICF catalogs grow. because you cannot control where the access goes to the ICF catalog in terms of the key. 4). The previous values can be changed with the IDCAMS ALTER command. then use it. BUFNI (number of buffers on the index) should be specified. consider the use of video or other suitable media. and applications change in response to day-to-day activities. 򐂰 Ensure that the documentation is secure. 򐂰 Use whatever media suits your purpose. if they suffer from lack of buffers or you want to have a standardized setup. After you have defined the ICF catalog. and improve it if they are not already in place. Chapter 2.6 ICF catalog performance This section focuses on the setup and features that can affect ICF catalog performance.The following suggestions are to help you plan your documentation: 򐂰 Create the documentation in a modular format. 2. Planning integrated catalog facility catalog environments 19 . This command also enables you to change your existing catalogs easily. from allocation of a new ICF catalog to caching. Good reporting is essential. you might need to review your setup. Never define more than 6-7 unless you have a specific reason to. Based on trends and preset thresholds. Specify this as the number of index levels +1 (for example. in this case 4. use defaults. focus on the setup of your environment. is covered. from allocation of a catalog to using features that can affect performance. FREESPACE(0 0) is suggested. data volumes increase. What worked well at creation time might not work as effectively now. Definition of ICF catalogs When defining a new ICF catalog. In addition to the traditional approach. 򐂰 Treat the recovery job repository as you would the documentation. the documentation might be unavailable.

this design works well. However. When a catalog is in RLS mode. but performance might be affected if you need to reload catalog records frequently. 20 A Practical Guide to ICF Catalogs . direct reads are more likely than sequential reads.Example 2-1 illustrates how the DEFINE USERCATALOG might be coded. it should add capacity for some time. this is even better. We have three ways of using caching: 򐂰 In Storage Cache (ISC) 򐂰 Catalog data space cache (CDSC) 򐂰 Record-level sharing (RLS) buffering based on SMSVSAM You can only use one cache type at any point in time. You can find more information about CA reclaim in DFSMS Managing Catalogs. which can only be used by this UCAT. cache records are flushed based on least recently used (LRU). ISC and virtual lookaside facility (VLF) are not being used. When this is fully used. because creating extents takes time. buffers for a given catalog are flushed if updates or accesses are made from another LPAR. If your ICF catalog is facing direct reads. Every UCAT has its fixed storage. only the key-sequenced accessed records are cached. For ICF catalogs. You can allocate catalogs on the same volume. ISC is caching in the catalog address space (CAS). In a shared environment. CISZ should be larger than 4 KB. In an environment where the ICF catalog is only accessed from one system. Example 2-1 Example of defining a user catalog with tuning parameters DEFINE USERCATALOG (NAME(YOUR. Allocate a secondary extent in more than one cylinder. a small CISZ might be a better option. and balance your catalog input/output (I/O) on more volumes. CISZ (CISIZE) should be kept at 4 KB. because hot extents might be moved to SSD if needed. limit this to a few catalogs. For the MCAT. In Storage Cache and catalog data space cache caching ICF catalog caching is covered elsewhere in this document in more detail. If you have frequent sequential reads on your catalogs. In this topic. we focus on when you should prioritize using one cache option rather than other options. and to reduce contention. If this is a disk controller with solid-state drive (SSD) and IBM Easy Tier®. This flushing ensures integrity. it depends.CATALOG) FREESPACE(0 0) CYLINDER(10 10) VOLUME(vvvvvv) EATTR(OPT) STRNO(3)) DATA(CISZ(4096) BUFND(4)) INDEX(CISZ(4096) BUFNI(4)) For indexes. Rather. RLF uses SMSVSAM buffering. CA reclaim should also be used to avoid some extents by reusing CAs that were freed up. all accessed records are cached (both sequential and key sequenced). and prevent each CA split from requiring new extents. When it happens. Consider allocating your ICF catalogs on dedicated volumes to minimize the effect of other workload on your catalog performance. For UCATs. SC23-6853. For data. ISC is the default cache.

we need to mention this here as a tuning option. it is probably a good idea to take the catalog trashing the cache out of CDSC to increase the hit ratio for the remaining catalogs. because RLS uses SMSVSAM buffers. ICF catalog sharing and Enhanced Catalog Sharing For better integrity. where enqueue is happening on the VVDS. RLS is an alternative to using ISC or CDSC for buffering and performance. However. the resource use is simply too high compared to the benefit of caching. Chapter 2. The effect on performance is also a consideration. a counter. A VSAM volume record (VVR) update in a shared environment needs an enqueue and read of the VVDS before doing the update. We suggest that all three of these RESERVEs (including SYSIGGV2) be converted. Consider that a break-even point for cache hit ratio is around 20%. Example 2-2 Example of how to disable an ICF catalog from default caching F CATALOG. Only then does flushing of cache records start. ICF catalog updates in a shared environment need extra steps and checks for integrity reasons.Having a situation as described previously for ISC. and performance in a shared ICF catalog environment. ECS mode is a technique used by the ECS feature. where the individual user catalogs are able to use storage until the limit (=MAXVIRT as specified in COFVLFxx member) is reached. can cause a performance delay in a busy environment with many ICF catalog updates. Using ECS improves ICF catalog processing significantly. This cache is one big buffer. and identification of the change. ECS uses a structure in the CF for caching (SYSIGGCAS_ECS). availability.1 ICF catalogs can use RLS mode.NOISC RLS caching Because Data Facility Storage Management Subsystem (DFSMS) V2. CDSC caching happens in a separate data space. SMSVSAM buffers are either stored in a 31-bit SMSVSAM data space. “Record-level sharing support for ICF catalogs” on page 73. called VVDS mode. Otherwise. and is another tuning option if it is not already implemented. Therefore. the time stamp VVR (also called integrity VVR) is stored in the CF. again based on an LRU algorithm. Rather than using the traditional VVR mode. This technique. or in the coupling facility (CF). Example 2-2 shows how an ICF catalog is taken out of default caching. Planning integrated catalog facility catalog environments 21 . Benchmark tests show significant improvement in ICF catalog tasks. Any catalog update from any system updates this VVR with a time stamp. Any action done to SYSZVVDS must also be done to SYSVTOC to avoid deadlocks. the SYSIGGV2 resource must be converted to a SYSTEMS enqueue for catalogs using Enhanced Catalog Sharing (ECS). But you can only do that if the catalogs are shared only within a sysplex. you might want to move this shared catalog to CDSC (or VLF). RLS is covered in more detail in Chapter 5. ISC and CDSC are not used for RLS-managed ICF catalogs. a 64-bit SMSVSAM address space. If the hit ratio goes below this.

or for moving to another caching option.ZOSR1A (ISC) * * 93% 1 15 14 0 0 0 * * UCAT. A good overall report is the performance report.CACHE command *CAS****************************************************************** * HIT% -RECORDS. the extract and report jobs should be submitted regularly.--PURGE-.REPORT. After the weekly extract. Capture the output from these commands. and it is suggested to be 20% and above. This must be investigated per ICF catalog. This report summarizes all catalog events. consider using an automation package to issue these commands on a regularly defined interval. and trends and deviations can be calculated based on this.-DELETES. A high purge might also affect performance.REPORT. See Example 2-4 for an extract of the output. which reports by ICF catalog.ZOSR1A * * 0 0 16 0 0 10 16 8 * * UCAT. nullify (reset by ICF catalog command) the ICF catalog counters.REPORT. See Example 2-3 for an extract of this command output. 22 A Practical Guide to ICF Catalogs . and STRNO). and also gives you an idea over time about how much your environment grows. BUFNI. captured.Monitoring To follow up on ICF catalog performance.CATSTATX command. which reports for your entire environment and not by ICF catalog. issue the F CATALOG.PERFORMANCE command. and then format and record it into a data set for subsequent retrieval. If ICF catalogs continually stay below 20%.CATSTAT *CAS**************************************************************** * ADDS UPDATES GETS GETUPD DELETES BUFNI BUFND STRNO * * * * UCAT. By doing this. A useful report to produce weekly is the cache report. Example 2-3 Extract of output from F CATALOG. use the F CATALOG. For command-issued reports. Once a week.-SHR UPD.-SEARCHES --FOUND-. and of the buffering (BUFND. If the solution is to put one or more ICF catalogs in CDSC cache. then perhaps disable the ICF catalog from ISC cache. and then stored for further processing to create trend and exception reports. It now also reports on CPU TIME and ELAPSE TIME (these two items were added in z/OS V2. Example 2-4 Extract of output from F CATALOG. To create this report. A catalog statistics report issued using the F CATALOG. only this report reports on control area (CA) reclaim and CA reuse.VTFMTAPE * * 57 181 749 181 0 10 16 8 * A similar report is created with the F CATALOG.REPORT.* * * * UCAT. they are candidates for removal from the cache.1). and can identify tuning candidates by ICF catalog name. your reports are always based on a comparable period (one week).CACHE command. which gives the most valuable information based on your systems’ behaviors and profile.CATSTATS command gives an overview of the activity on the individual catalogs.REPORT. a display of the counters could be shown.VTFMTAPE (ISC) * * 66% 1 3 2 0 0 5 * The cache report shows cache hit ratio. and should be investigated and responded too.REPORT. To produce this report.

Example 2-5 RMF3 reports as displayed in option 10.4. it’s suggested for catalogs containing entries of SMS-managed data also to be SMS-managed. it’s important for you to work with Storage Management Team to define the appropriate SMS policy for catalogs. and 12 in SYSPLEX option 01 02 03 04 05 06 07 08 09 10 11 12 SYSSUM SYSRTD SYSWKM SYSENQ CFOVER CFSYS CFACT CACHSUM CACHDET RLSSC RLSDS RLSLRU Sysplex performance summary Response time distribution Work Manager delays Sysplex-wide Enqueue delays Coupling Facility overview Coupling Facility systems Coupling Facility activity Cache summary Cache detail VSAM RLS activity by storage class VSAM RLS activity by data set VSAM LRU overview (SUM) (RTD) (WKM) (ES) (CO) (CS) (CA) (CAS) (CAD) (RLS) (RLD) (RLL) 2. consider how to use SMS settings to manage your catalogs.System Management Facilities reports System Management Facilities (SMF) data can also be used for catalog reporting. which includes the Data Class. SC23-6853. You can create your own reports based on Restructured Extended Executor (REXX) code or similar. From the primary panel. In Sysplex Reports on the next panel. see the topic about using catalogs with the SMS in z/OS DFSMS Managing Catalogs. 򐂰 SMS can just be used for ICF-type catalogs. Therefore. However. select S for SYSPLEX. 2. Storage Class. 11.7 Storage management subsystem When planning catalog structure. There is a close relationship between catalog and SMS: 򐂰 All SMS-managed permanent data sets must be cataloged. Management Class. Chapter 2. and Storage Group for SMS constructs and ACS routines. Look into the following SMF record types to investigate ICF catalogs: 򐂰 򐂰 򐂰 򐂰 򐂰 򐂰 Type 60: VVR updated Type 61: ICF define Type 62: VSAM Open Type 64: VSAM Close (new for ICF catalogs from z/OS V 1. there are also RLS reports in IBM Resource Measurement Facility™ (RMF3) that you can run: 1.13) Type 65: ICF delete Type 66: ICF alter Resource Measurement Facility reports For the RLS-managed catalogs. you will see these 3 RLS reports at the bottom of the panel. 򐂰 The extended format catalogs using extended addressability must be SMS-managed. as shown in Example 2-5. 򐂰 Both SMS-managed and non-SMS-managed data sets can be cataloged to the same catalog. 򐂰 The catalog data set itself doesn’t need to be SMS-managed. considering the performance and availability requirements of different catalogs. Planning integrated catalog facility catalog environments 23 . For more guidance.

or by caching in CDSC instead. which can later be used to end the request or cause an abnormal end of task (abend) if. If you do not want the catalog in ISC for some reason. diagnosed. This task is assigned a CAS ID. the request is not satisfied.8 Features including parmlibs and caching considerations Catalog setup offers granularity in settings that can improve the overall performance. Parmlib parameter settings determine how CAS operates. The CAS address space can also be restarted. For the MCAT. CDSC caching CDSC caching happens in a data space. a service task is assigned for that request. For UCATs. such as how many concurrent tasks can run in CAS and how many user catalogs can be open at the same time. ISC is the default catalog caching if no other caching is requested. and is designed to have tasks timed out or restarted if errors occur. Caching in ISC is caching happening within the CAS itself.PARMLIB. records start being removed from the ISC cache based on an LRU algorithm. Important ICF catalogs tasks The following areas will help to keep your ICF catalog environment effective and efficient. and holds all user catalog and alias names in control blocks within the address space. This address space is started at initial program load (IPL). CAS has a specialty task called the Analysis Task that periodically checks CAS for errors. automation. again based on an LRU algorithm. Implementing the appropriate settings can help you build a more robust and stable environment for the systems software and business applications running in your organization. and only accessed records are cached. all records referred are cached in ISC except for alias records. Catalog address space The CAS is the main address space that services catalog requests. each catalog is dedicated a certain amount of storage in ISC. Caching catalogs in VLF is not a constraint at a catalog level as in ISC. and monitoring of your catalogs. Adding catalogs to VLF requires a stop and start of VLF to have the added catalogs start using VLF. When this limit is reached for a user catalog. CAS can be modified.4. Catalog records are cached until the data space limit is reached (set by the MAXVIRT parameter in the COFVLFxx member). which are stored in a separate table in storage. for any reason. Only then do catalog records start being removed. 24 A Practical Guide to ICF Catalogs . you need to remove this from cache by command. so this must be started to have the caching in effect. CDSC uses VLF.2. This is defined in the COFVLFxx member in SYS1. In Storage Cache caching ISC is one way of caching catalog records. and displayed by modify commands. When a user requests a catalog function.

15 blocks per ICF catalog is the suggested number. 2. ECS uses a structure in the CF for buffers. Choose Supporting Products and click Coupling Facility Structure Sizer Tool.BCS1) EMAJ(UCAT.See Example 2-6 for a sample of the COFVLFxx member. 5 was chosen).com/systems/z/advantages/pso/index. Select this option and you reach the window shown in Figure 2-1. found on the following website: http://www. because VVDS enqueues are reduced by ECS. nnnn is the number of 4 KB blocks that you can specify. and the default value is 4096. On this window. this value should be (15*50) = 750 blocks. you find a selection for Enhanced Catalog Sharing to the left. so if you have 50 ICF catalogs. Figure 2-1 CF sizer pane for estimating ECS structure size 3. The first step in implementing ECS is to size and allocate this structure. Chapter 2. Planning integrated catalog facility catalog environments 25 . type in the number of ICF catalogs that you expect to have under ECS control (in the example. Example 2-6 Example of COFVLFxx member CLASS NAME(IGGCAS) EMAJ(UCAT. Here you specify by name the catalogs that you want added to CDSC caching. The minimum value is 256 megabytes (MB). Next. and it is an important performance feature.ibm. Sizing is done most easily by using the CFSizer. and the potential of having ICF catalog hangs or lockouts can also be reduced by using ECS. perform the following steps: 1. Next. Enhanced Catalog Sharing ECS is also covered in some detail in the performance chapter of this book. and how to enable the feature.html On the website. It is also an important sharing feature. focus on the planning considerations for ECS.BCS2) … MAXVIRT(4096|nnnn) In this case.

Define the ECS structure in the coupling facility resource management (CFRM) policy and activate the policy. In addition to the ECSHARING attribute on the ICF catalog.If you want to stop ECS activity on one system. and beneath that the JCL to create the structure. CFSizer also lists the JCL that can be used for allocating the structure. All other systems stay connected to the CF structure. Activate AUTOADD automatically at IPL time by setting the indicator in the SYSCATxx in SYS1. making it possible to assign any existing ICF catalogs for ECS sharing. you must use the IDCAMS keyword ECSHARING. Calculations done by CFSizer advise 2 MB in this case. the ECS mode must be enabled by issuing the command shown in Example 2-7. 9. 26 A Practical Guide to ICF Catalogs . 6. ECS mode and VVDS mode are mutually exclusive. but you can take an ICF catalog out of ECS mode on all systems by issuing the following command: F CATALOG ECSHR(REMOVE.ECSHR(DISCONNECT) This command disconnects ECS from the coupling facility structure on the system. When you submit this window. Example 2-7 Example of enabling ECS mode F CATALOG. This action should connect all ECS-eligible systems to the ECS structure. Figure 2-2 CF sizer result window with calculated structure size 5. where the command is issued. Alternatively.The ECSHARING attribute is still on the catalog. Figure 2-2 shows an example of the sizing information in the window. but the ICF catalog reverts to VVDS mode. This keyword is also supported on the IDCAMS ALTER command.catname) 11. issue the following command: F CATALOG. activate AUTOADD from there. the results come back showing you the size calculated at the top. In addition to doing the calculation. AUTOADD will connect the ECS eligible ICF catalogs to the structure at the next access to the ICF catalog. too.ECSHR(AUTOADD) 8.4. To allocate a new ICF catalog as an ECS candidate.NUCLEUS. 7.catname) 10. You can return the ICF catalog to ECS mode by using the following command: F CATALOG ECSHR(ENABLE. but quiesce their ECS activity. though. if you are using the IGGCATxx member.

ACTIVE command from the operator console. RLS obtains buffers dynamically. when enabled. that can contain most ICF catalog settings. Planning integrated catalog facility catalog environments 27 . Most of these can be changed dynamically by command. 򐂰 All systems sharing a catalog in ECS mode must be using the same CF and be in the same global resource serialization (GRS) complex. define two sharing control data sets (SHCDSs) and one spare.PARMLIB. for recovery and restart. If SMSVSAM is not started. because they might be needed using system-managed buffering (SMB). a new IGGCATxx member in SYS1.Note that Enhanced Catalog Sharing does not support the following scenarios: 򐂰 An ICF catalog cannot use ECS mode and VVDS mode protocols at the same time.13 introduced the catalog parmlib. SC23-6853. Table 2-1 on page 28 shows keywords that can be specified in the catalog parmlib. and which keywords can be updated by command. and do not require an IPL to be updated. plus a data class with the buffering requirements. The role of the SMSVSAM task RLS is supported on ICF catalogs from DFSMS V2. The table shows the individual keywords. Chapter 2. a storage class is required. For sharing and buffering. features. Having the keywords specified in catalog parmlib enables you to dynamically change most of these if you face an issue where a changed value would be needed. you can start the address space after IPL by issuing the V SMS. Buffering For catalogs in RLS mode. These data sets hold critical restart information for SMSVSAM. add all of the keywords with the defaults.SMSVSAM.0 DFSMS Managing Catalogs. Using this parmlib member. SMSVSAM and record-level sharing This section describes aspects of the SMSVSAM task. the catalog records are placed in RLS local buffer pools or CF cache structures. See more detail about RLS for catalogs. Catalog parmlib z/OS V1.PARMLIB. The SMSVSAM address space automatically starts at IPL if the RLSINIT (YES) keyword is specified in the IGDSMSxx member of SYS1. it is possible to gain an understanding of the catalog settings just by looking into the parmlib member’s specifications. see Chapter 5. “Record-level sharing support for ICF catalogs” on page 73. See more on Enhanced Catalog Sharing in the chapter about ICF catalog tasks. unless you want a different value on some of these. 򐂰 A maximum of 1024 catalogs can be shared using ECS from a single system. SMSVSAM is the server that handles RLS requests in general. Only the master catalog name and master catalog volume need to go into the SYSCATxx member in SYS1. and parmlib in z/OS V2R1.PARMLIB. and for the ICF catalogs with RLS enabled. Additionally. The storage class is assigned with a cacheset and lockset identifying cache and lock structures in the CF.1. their defaults.NUCLEUS or the LOADxx member in SYS1. It is advised to use catalog parmlib and.

Table 2-1 lists the catalog parmlib keywords. CNT 1 (CNT) YES EXTENDEDALIAS Extended alias can be created NO ENABLE | DISABLE GDGFIFOENABLE First-in first-out (FIFO) ordering on the generation data group (GDG) NO Not applicable (N/A) NOTIFYEXTENT ICFCAT extents exceeded (IEC361I) 80% YES SYMREC Enabling of SYMREC creation YES ENABLE | DISABLE SYS% ON OFF SYS% to SYS1 conversion enabled OFF YES SYSIGGV2 SYSIGGV2 contention setting 10 minutes per notify CONTENTION SYSZTIOT SYSZTIOT contention setting 10 minutes per notify CONTENTION SYSZVVDS SYSZVVDS contention setting 10 minutes per notify CONTENTION TAPEHLQ High-level qualifier (HLQ) Tape Volume Catalog SYS1 N/A TASKMAX Maximum number of user service tasks in CAS 180 YES TASKMIN Minimum number of user service tasks in CAS 60 YES TASKTABLESIZE Maximum number of catalog tasks 200 N/A UPDTFAIL IEC390I warning catalog update error YES ENABLE | DISABLE VVDSSPACE Default allocation size for VVDS Tracks(10. MOD. RSN. Table 2-1 Catalog parmlib overview 28 Keyword Description Default Modify command Aliaslevel Multi-level alias number (1 .9999) 1024 YES DELFORCEWNG Warning MSG IDC1997I or IDC1998I YES ENABLE | DISABLE DELRECOVWNG Warning MSG IDC1999I NO ENABLE | DISABLE DSNCHECK Syntax check on names added YES ENABLE | DISABLE DUMP Dynamic dumps available? OFF YES DUMPON RC.4) 1 YES ALLOCLCK ALLOCLCK contention setting 10 minutes per notify CONTENTION AUTOADD ECS AUTOADD (ON/OFF) OFF ECSHR CATMAX Maximum number of open catalogs (1 .10) YES VVRCHECK Enhanced VVR checking on VVDS? NO ENABLE | DISABLE A Practical Guide to ICF Catalogs .

Each new release of z/OS and DFSMS continues to add functionality to this area. plus the chance of installing products and fixes on these volumes. and do not affect the production systems. Also take the opportunity to discuss and evaluate new functionality introduced by the system software upgrade or maintenance implementation that would add value to business applications or environments. The inclusion of storage managers into the system software upgrade and maintenance process help prove that the existing ICF catalog environments have not been negatively affected by the upgrade or maintenance application.9 System software considerations The need for ever-improving ways to manage storage. you can test upgrades on your main products.2. to an entire LPAR or SYSPLEX environment. merge catalogs. Chapter 2. incompatibility. consider creating one of the following environments to enable your support team to perform the minimum necessary tasks before implementing projects on a production system: 򐂰 A set of DASD space Provide DASD space to help support teams test allocation of data sets and catalogs. You are highly advised to implement a test environment for your business. and others. customization. 򐂰 A set of DASD space with bootable volumes This environment has all of the advantages of the previous option. Establish connections with the parties responsible for the installation. and other tasks. and back out the changes almost instantaneously when there is a failure. The test system can be started for testing on availability windows. 2. gaps on implementation plan. fix a catalog in error. and used to assist recover production if there is an outage. These resources can vary from a simple set of direct access storage device (DASD) and tape volumes. If sufficient resources are not available to implement an LPAR or SYSPLEX test system.4. and implementation of system software to assess any added or modified functionality. testing. Planning integrated catalog facility catalog environments 29 . The scenarios are isolated or protected from other environments. 򐂰 A copy of the production system with the main products installed In addition to the other advantages already mentioned.4. drives system software innovation to provide enhanced tools to meet modern demands to store and retrieve data. Other changes can be introduced by software maintenance.10 Testing environments A test system can be created and used by several resources that are available for support teams to create test scenarios. A test environment helps reduce the number of outages and performance degradation on production systems related to program code errors. test SMS policies. both at the physical and logical level.

usage trends. and the output saved on data sets for reporting purposes. These tasks can be configured to run through automation or a job scheduler. IDCAMS checks if the data or index structure contains errors. Because this function is time-consuming and uses space. Weekly Catalog performance The CAS provides plenty of reports on catalogs and cache performance.5 Housekeeping Whether you are planning to go through a system maintenance. depending on your organization’s needs. Define plans to perform the necessary action. it is suggested to plan and run housekeeping tasks to maintain the system’s performance and prevent errors. It is advised to run EXAMINE before catalogs move or reorganize. and tries to fix it if requested. Monthly VVDS errors Run IDCAMS DIAGNOSE against VVDSs to detect integrity errors. Define jobs to extract and process the data to create catalog performance reports. consolidation. 30 A Practical Guide to ICF Catalogs . or only perform routine tasks. Weekly Index and data integrity errors IDCAMS also provides DIAGNOSE to identify and fix index and data integrity errors. Table 2-2 Housekeeping in your environment Threat Prevention plan Frequency Catalog loss You should always make sure that you back up a valid catalog. Also run EXAMINE before catalogs move or reorganize to ensure catalog integrity. Monthly Catalog growth Set up a job or automation to collect catalog allocation information. so you should run the DIAGNOSE and EXAMINE commands before each backup. to ensure catalog integrity. plan volume dumps based on volume importance and availability windows. and other information. You can run backups using different solutions to extend your recovery options.2. At least once a day Index and data structure errors The IDCAMS EXAMINE option can be used to identify and fix index and data structure errors. This can include allocation extents and data set usage. As required Remember: All of the prevention plans and frequencies are suggestions. split. and can vary from site to site. Weekly VVDS loss Full-volume memory dumps are considered the best way to reduce the effects from a VVDS loss. as shown in Table 2-2.

or a redrive action can be initiated. have a purpose. to enable identification of usage peaks. identify system trends. Monthly Catalog performance This report contains information for each catalog event activity. 򐂰 ALLOCLCK. and so on. The need for automation goes beyond the following options. Planning integrated catalog facility catalog environments 31 . You can use this report to identify events that are taking too long to run. identify bottlenecks. and threats. 2. only the SYSZTIOT (input/output table) resource could be notified on. based on customer needs.1 introduced the following three extra resources that can be alerted on (besides SYSZTIOT): 򐂰 SYSZVVDS. Use this report to analyze the number of requests processed for a catalog. gets. Plan creating specific jobs to extract and handle the information for use. list application changes. Monthly Catalog statistics Shows number of adds. Plan to create reports that are meaningful to the teams and the business.7 Monitoring and automation This section focuses on built-in functions in the standard ICF catalog software. Chapter 2. Used to serialize access to associated VVDS records. Reports can analyze your data growth. be easy to follow. and delete for a specified catalog. Monthly VLF statistics You can create jobs to extract VLF statistics from SMF records to define the best VLF cache values for catalogs. and plan corrective actions. A CAS internal lock that ensures serialization of access to most catalog allocation events. Each report must have a specific function.6 Reporting A well-structured reporting process is vital to the good management of a system. On demand SMF data There are other information available on SMF data to report on catalogs behavior. Table 2-3 shows some reports that can be defined on your systems. updates. Used to serialize access to associated catalog resources. trends.7. and store multiple versions of these reports. Some of the reports can be fed into a process (such as REXX routines) and subsequent actions automatically deployed. You can also create your own reports. Table 2-3 Types of reports and most common usage Report type Report description Frequency Catalog cache Cache reports provide information related to cache searches and hits. 򐂰 SYSIGGV2. and be effective in maintaining high-quality ICF environments. On demand 2. A low cache hit can result in increased CPU usage for catalog task and low performance. But DFSMS V2. Catalog cache report is not applicable for RLS catalogs. but you should understand and consider them.2. be informative.1 Catalog contention detection Catalog contention detection is a useful feature that monitors core catalog resources for timeouts. including the count of events and average time. Other reports can be issued for review purposes. At timeout notification can be issued. Initially.

3. The following actions are valid: 򐂰 N for Notification Only 򐂰 R for Redrive The catalog detection function monitors resources looking into wait time.CONTENTION F CATALOG. providing information about the task. If the same task is still pending after another 5 minutes. and the request is resubmitted to the catalog for processing. Catalog Contention time-out and redrive settings can be displayed or changed by command. 5. 2. 4.Besides monitoring. If this is set to 0. Example 2-9 Display and modify command for catalog contention F CATALOG. At end the status will be updated in the next system check (within 30 seconds). Issuing message IEC393I next continues at 15-minute intervals until the wait has ended.PARMLIB catalog member. Following is the action sequence that starts based on the extended wait time: 1. Wait time and actions by resource must be set in the IGGCATxx member of the SYS1. and issues alerts on those exceeding wait time.CONTENTION(SYSZTIOT. The second command changes SYZTIOT to have the redrive option. as shown in Example 2-8. A symptom record is written to LOGREC.R) The top command is the display command.R) ALLOCLCK(0. the function is disabled for this resource. an abnormal end of task (abend) of the service task occurs. catalog contention detection now also has the possibility of triggering an action per resource monitored. See Example 2-10 for the output from the display command following.R) SYSIGGV2(13.N) The first parameter on the individual resource in Example 2-8 is the timeout value. The second parameter is the action. Example 2-8 Catalog contention detection: Sample settings SYSZTIOT(11. message IEC393I is issued again. as shown in Example 2-9. Message IEC393I is issued. Example 2-10 F CATALOG CONTENTION output display IEC351I CATALOG ADDRESS SPACE MODIFY COMMAND ACTIVE IEC359I CATALOG CONTENTION REPORT *CAS***************************************************** * RESOURCE THRESHOLD(MIN) ACTION(S) * *CAS***************************************************** * SYSZTIOT 11 N * * SYSZVVDS 12 NR * * SYSIGGV2 13 NR * * ALLOCLCK INACTIVE N * *CAS***************************************************** * ACTION KEY * * N = NOTIFY OPERATOR R = REDRIVE REQUEST * *CAS***************************************************** IEC352I CATALOG ADDRESS SPACE MODIFY COMMAND COMPLETED 32 A Practical Guide to ICF Catalogs .. When redrive is active the first time a service task with an active resource passes the contention threshold.N) SYSZVVDS(12.

in the IGDSMSxx member in SYS1. Before the CA RECLAIM function.816 * * ERASES (DELETES) = 29.EXAMPLE * * INSERTS (ADDS) = 37. The default value is NONE. CAs in KSDSs with the IMBED option active. In this case. CAs must be freed up before reuse is possible. ICF catalog maintenance can be particularly disruptive. directly issue the activation command: SETSMS CA_RECLAIM(DATACLAS).469 * * UPDATES = 73. Example 2-11 CA-reclaim s and CA-reuses *CAS***************************************************************** * CATALOG NAME = MY. Alternatively.REPORT. Planning integrated catalog facility catalog environments 33 .706 * * RETRIEVES = 6.PARMLIB. and is enabled by adding a new keyword.7. Therefore. Note that CA RECLAIM does not work under the following circumstances: 򐂰 򐂰 򐂰 򐂰 򐂰 Partially empty CAs. Activate the CA RECLAIM function by refreshing your IGDSMSxx member by issuing the SET SMS=xx command. See Example 2-11.2.870K * * RETRIEVES FOR UP = 133. CA with RBA 0 and the CA with the highest key of the KSDS. where the CA RECLAIM option is set to Y.CATSTATX(catalog_name) command displays reclaim activity and other ICF catalog information.2 CA RECLAIM of VSAM KSDS data sets The CA RECLAIM function (only supported on VSAM KSDSs) enables reuse of empty CAs by adding freed-up CAs to the free CA list. CAs already empty before CA RECLAIM is enabled. Assign a DATACLAS to the data sets in scope. The F CATALOG. because many applications can use the same ICF catalog. which can improve the availability figures for key application data sets. the empty CA would need to wait until the next REORG of the VSAM KSDS.CATALOG. This maximizes the usage of your VSAM KSDS. and makes those more efficient by reducing the need for performing a REORG. CA RECLAIM also works for ICF catalogs. CA_RECLAIM(DATACLAS). This makes what was previously unusable now eligible for immediate reuse. issue an IDCAMS ALTER RECLAIMCA command. For existing data sets. the CA RECLAIM is useful.661 * * CA-RECLAIMS = 1 * * CA-REUSES = 80 * * BUFNI SETTING = 7 * * BUFND SETTING = 11 * * STRNO SETTING = 3 * *CAS***************************************************************** IEC352I CATALOG ADDRESS SPACE MODIFY COMMAND COMPLETE Chapter 2. Application opening a data set with global shared resources (GSR). xx is the suffix of your active IGDSMS member.

2.7. Note that for RLS catalogs. you receive a notification in SYSLOG when the threshold is reached. Example 2-12 Notification in SYSLOG. From the IBM Health Checker panel in IBM z/OS System Display and Search Facility (SDSF). Example 2-13 IBM Health Checker: Current checks related to ICF catalog and RLS NAME CATALOG_IMBED_REPLICATE CATALOG_RNLS VSAMRLS_CFCACHE_MINIMUM_SIZE VSAMRLS_CFLS_FALSE_CONTENTION VSAMRLS_DIAG_CONTENTION VSAMRLS_QUIESCE_STATUS VSAMRLS_SHCDS_CONSISTENCY VSAMRLS_SHCDS_MINIMUM_SIZE VSAMRLS_SINGLE_POINT_FAILURE CheckOwner IBMCATALOG IBMCATALOG IBMVSAMRLS IBMVSAMRLS IBMVSAMRLS IBMVSAMRLS IBMVSAMRLS IBMVSAMRLS IBMVSAMRLS State ACTIVE(ENABLED) ACTIVE(ENABLED) ACTIVE(DISABLED) ACTIVE(DISABLED) ACTIVE(ENABLED) ACTIVE(ENABLED) ACTIVE(DISABLED) ACTIVE(DISABLED) ACTIVE(DISABLED) Statu EXCEPTION-LOW SUCCESSFULL GLOBAL-lpar GLOBAL-lpar SUCCESSFUL SUCCESSFUL GLOBAL-lpar GLOBAL-lpar GLOBAL-lpar Health check limits are based on IBM suggested values. when extent limit is reached on your ICF catalog IEC361I CATALOG catalogname (data or index) HAS REACHED xx% OF THE MAXIMUM EXTENTS You can use the warning message to trigger a notification from your automation package. but can be changed by setting a policy in the IBM Health Checker HZSPRMxx parmlib. as shown in Example 2-12. where CAS was restarted. With this parameter. Example 2-13 shows an extract of the catalog-related and RLS-related health checks. type CF.The previous figures show activity since the last reset of the catalog figures.3 Notify extents NOTIFYEXTENTS is one of the IGGCATxx options that you might already have considered and enabled.7.4 Health checks done by IBM Health Checker for z/OS The IBM Health Checker for z/OS also has warnings on ICF catalog-related setup. you can monitor which health checks of all possible checks are enabled or disabled. Note that for RLS catalogs. To access SDSF. See the description of the following CA reclaim fields: CA-RECLAIMS The number of CA reclaims performed by VSAM on this catalog over the reported interval. and in this way be able to plan and prepare a REORG (or maybe a reallocation with a larger primary and secondary size) of your ICF catalog before the limit of 123 extents is reached. this information shows up as N/A because the values are not currently available. this information shows up as N/A because the values are not currently available. but 80 CAs became reusable. If this is set to 80 (the default). 2. Only one reclaim was done. or by a MODIFY command. 34 A Practical Guide to ICF Catalogs . you set the value for the maximum allowed number of extents for your ICF catalogs. CA-REUSES The number of CA reuses performed by VSAM on this catalog over the reported interval. or since the most recent IPL.

set BUFNI=TI .(HURBA/CASZ) +1. 򐂰 Use cylinder allocation and secondary allocation of more than one cylinder. 򐂰 Remove Imbed and Replicate. but general guidelines can be set as a rule to assist you in your considerations on how to manage the ICF catalogs environment.VSAMRLS_SINGLE_POINT_FAILURE) Read more in the documentation for IBM Health Checker: IBM Health Checker for z/OS User’s Guide. we enabled a health check of VSAMRLS. 򐂰 Only reorganize your ICF catalogs if the ICF catalog grows in too many extents. 2.CHECK=(IBMVSAMRLS. and in the background. Planning integrated catalog facility catalog environments 35 . an ACTIVATE command is issued to the IBM Health Checker task. 򐂰 Set BUFND=2. 򐂰 Given the following definitions. 򐂰 Monitor your ICF catalogs and keep them free of errors. we have already made quite a few suggestions. In this IBM Redbooks publication. To summarize this advice. keep your control interval (CI) size (CISIZE or CISZ) on the data and index portion at 4 KB (unless you have many sequential reads). 򐂰 For a shared ICF catalog use share option (3 4). In Example 2-14. Chapter 2. 򐂰 Always apply current maintenance.8 Suggested practices summary ICF catalog best practices depend on your setup and your organization’s requirements. 򐂰 In general. 򐂰 Do not code BUFFERSPACE. but also consider ECS or RLS on your ICF catalogs. Extra index buffers improve direct access: – – – – BUFNI is the number of buffers on the index.ACTIVATE. 򐂰 Set STRNO=3. TI is the total number of index records. 򐂰 CDSC cache is suggested. Example 2-14 IBM Health Checker ACTIVATE command F HZSPROC. SC23-6843. 򐂰 Also use CA Reclaim on your ICF catalogs. 򐂰 Have your backup and recovery procedures in place. you can activate or deactivate checks through line commands. the following list describes them: 򐂰 Use FREESPACE(0 0) at allocation time and let splits happen. 򐂰 Involve other teams in your plans and gain their approval.From the IBM Health Checker panel. and to give an overview of leading practices. HURBA is the high usage relative byte address number. CASZ = CISZ x CI/CA.

36 A Practical Guide to ICF Catalogs .

Deploying ICF catalogs Chapter 1. depicted integrated catalog facility (ICF) environments scenarios. However. they are illustrative. and for continuity purposes are used for the sample plan extracts and deployments covered in this chapter. All rights reserved. “Technical overview” on page 1. © Copyright IBM Corp. 37 . 2015.3 Chapter 3. These configurations and environments are much simpler than the typical organization.

and four user catalogs (UCATs).3. commands. 3. which consists of three subtopics: 򐂰 The scenario 򐂰 A task list related to the catalog environment 򐂰 Sample job control language (JCL).1 About the deployment We take a common approach to each scenario. the following methods: 򐂰 IBM z/OS Data Facility Storage Management Subsystem (DFSMS) Access Method Services utility (IDCAMS) export 򐂰 Logical data set dump 򐂰 Full volume memory dump 3. and output When implementing a change on a system. but are not limited to.1 Scenario 1 Target Environment Figure 3-1 shows the target environment. single master catalog (MCAT). so you have a second recovery plan in case the first plan fails. We strongly suggest that you back up your catalogs before implementing changes. perform all of the actions necessary to make sure that you can recover to a valid point in time if the changes fail for any reason. such as the ones described in this chapter.2.2 Scenario 1: Simple ICF catalog environment structure This section describes how to create the ICF catalog environment for a single logical partition (LPAR). LPAR A MCAT A UCAT A1 UCAT A2 Figure 3-1 Single LPAR environment 38 A Practical Guide to ICF Catalogs UCAT B1 UCAT B2 . Additional considerations and steps might be required depending on your environment. Some backup techniques include. The activities shown are related to the ICF catalogs. It is also advised to use different backup tools for performing the backup.

and report procedures. Table 3-1 Scenario 1 activities Step Activity Define a master catalog. None of the user catalogs in this plan are shared. include the following information: 򐂰 Recovery point objective (RPO) 򐂰 Backup method 򐂰 Backup and recovery procedures Create monitoring. Consider the following points when you define the system MCAT: 򐂰 The catalog size 򐂰 The volume to allocate the catalog 򐂰 The shareoptions 򐂰 The definitions for the sys1 data sets necessary to start the system from a powered off state Define user catalogs. The monitoring and report planning include defining the following tasks: 򐂰 What reports to extract 򐂰 The frequency the reports run 򐂰 How to store reports 򐂰 Error messages and alerts tracked by automation 򐂰 How to respond to alerts and errors Create a maintenance plan. Before updating IGGCATxx or LOADxx members.3. Include the following tasks on your maintenance plan: 򐂰 What frequency the maintenance should run 򐂰 Negotiated availability windows for performing REORGs 򐂰 What reports to use to decide if maintenance is required Chapter 3.2. To create a DR plan. define the following information: 򐂰 Where to place catalog parameters 򐂰 The default values used by the system Create a disaster recover (DR) plan. When defining the ALIASes. automation. Deploying ICF catalogs 39 . Consider the following points when you define the system UCATs: 򐂰 The catalog size 򐂰 The volume to allocate the catalog 򐂰 The shareoptions 򐂰 Volume catalog (VOLCAT) planning considerations to include Define the ALIASes and connect them to catalogs. the system has one master catalog and four user catalogs. Upon completion of this plan. plan for the following aspects: 򐂰 ALIAS growth 򐂰 Grouping application 򐂰 Making sure that only required data sets are cataloged on the MCAT Update the IGGCATxx and LOADxx members. single environment.2 Scenario 1: Task list Table 3-1 shows the steps for creating a new.

The name of the new catalog to be used later by LPAR A is SYS1. Example 3-1 Define the MCAT DEFINE USERCATALOG (NAME(SYS1. In addition. Example 3-2 Define user catalog DEFINE USERCATALOG (NAME(SYS1. It is suggested to dedicate a volume to the MCAT to avoid contention between the catalog and other data sets on the volume. and data growth. but have the appropriate name. and the output from the command.UCATB2) ICFCAT STORAGECLASS(NONSMS) VOLUME(VOLUC2) CYLINDERS(25 10) 40 A Practical Guide to ICF Catalogs .UCATB2 It is suggested that UCATs are placed on separate volumes to provide a more balanced environment in terms of availability.3 Scenario 1: Deployment The first step to implement scenario one consists of creating the MCAT to be used by the system to be run in LPAR A: 1.UCATA2 SYS1. The new catalog that you are defining is. The other three catalogs use the same parameters.2. Example 3-2 shows the IDCAMS parameters to build the SYS1.UCATB2 user catalog. MAXIMUM CONDITION CODE WAS 0 2.MCATA. a user catalog on the driving system but is the master catalog on LPAR A. HIGHEST CONDITION CODE WAS 0 IDC0002I IDCAMS PROCESSING COMPLETE.MCATA) ICFCAT STORAGECLASS(NONSMS) VOLUME (VOLMC1) CYLINDERS (25 10) FREESPACE(0 0) SHR(3 3) STRNO(3)) DATA (CISZ(4096) BUFND(4)) INDEX(BUFNI(4)) IDC0510I CATALOG ALLOCATION STATUS FOR VOLUME VOLMC1 IS 0 IDC0512I NAME GENERATED-(I) SYS1. possible contention issues. define the four UCATs to be used in the LPAR A environment: – – – – SYS1. and performance.MCATA. clear structure.3. consider application usage. Next. Example 3-1 shows a sample user catalog DEFINE command. in effect.UCATB1 SYS1. It is assumed that you already have a system running that you can use to define the catalog for the new LPAR A.CATINDEX IDC0001I FUNCTION COMPLETED. Important: Do not be confused by the DEFINE USERCATALOG statement.UCATA1 SYS1. See Example 3-1 for the IDCAMS parameters defining the master catalog.

MCATA) IDC0510I CATALOG ALLOCATION STATUS FOR VOLUME VOLUC2 IS 0 IDC0512I NAME GENERATED-(I) SYS1. Deploying ICF catalogs 41 . HIGHEST CONDITION CODE WAS 0 DEFINE ALIAS(NAME(TIVOLI) RELATE(SYS1. 4.UCATB2)) CAT(SYS1. Use a DEFINE ALIAS command to create all of the necessary ALIASes. See Example 3-3.UCATB2 DEVICETYPE(3390) VOLUMES(VOLUC2)) CATALOG(SYS1. HIGHEST CONDITION CODE WAS 0 IDC0002I IDCAMS PROCESSING COMPLETE.MCATA) IDC0001I FUNCTION COMPLETED.UCATA1)) CAT(SYS1. HIGHEST CONDITION CODE WAS 0 The next step adds ALIAS entries to the new master catalog to direct searches and actions to the appropriate user catalog. you might want to EXPORT DISCONNECT the catalogs from the current MCAT after catalogs are defined.MCATA) IDC0001I FUNCTION COMPLETED.MCATA) IDC0001I FUNCTION COMPLETED.MCATA) IDC0001I FUNCTION COMPLETED. MAXIMUM CONDITION CODE WAS 0 3. HIGHEST CONDITION CODE WAS 0 Chapter 3. to prevent users from inadvertently updating other system’s catalogs. HIGHEST CONDITION CODE WAS 0 DEFINE ALIAS(NAME(GDPS) RELATE(SYS1. Example 3-4 Define ALIAS DEFINE ALIAS(NAME(DB2) RELATE(SYS1.UCATB2. See Example 3-4.CATINDEX IDC0001I FUNCTION COMPLETED. 5. use the IMPORT CONNECT command to connect the UCATs into MCATA. HIGHEST CONDITION CODE WAS 0 DEFINE ALIAS(NAME(CICS) RELATE(SYS1.UCATA2)) CAT(SYS1. When you have created the new UCAT from the driving system. Consider also running DEFINE ALIAS commands on other LPARs if they should have access to the data. Example 3-3 IMPORT CONNECT catalogs IMPORT CONNECT OBJECTS( SYS1.MCATA) - IDC0603I CONNECT FOR USER CATALOG SYS1.UCATB2 SUCCESSFUL IDC0001I FUNCTION COMPLETED. This establishes a relationship and structure between the MCAT and the UCATs.UCATB1)) CAT(SYS1. and point to the correct catalogs. After creating the catalogs.FREESPACE(0 0) SHR(3 3) STRNO(3)) DATA (CISZ(4096) BUFND(4)) INDEX(BUFNI(4)) CAT(SYS1.

but which appear in Table 3-1 on page 39. LPAR A could be the production environment.MCATA) IDC0001I FUNCTION COMPLETED. policies. 42 A Practical Guide to ICF Catalogs . it is also necessary to catalog the minimum system data sets to start the system.3.3 Scenario 2: Adding a second LPAR As the organization grows. Example 3-5 Define system data sets before system start DEFINE NONVSAM (NAME('SYS1. 3. The new LPAR B can be used as an application development and testing environment.PARMLIB') DEVT(3390) VOLUMES(SYSV01)) CAT(SYS1.LOADLIB') DEVT(3390) VOLUMES(SYSV01)) CAT(SYS1. Example 3-5 shows the DEFINE command for three system data sets on SYSV01.PROCLIB') DEVT(3390) VOLUMES(SYSV01)) CAT(SYS1. Before starting the system LPAR A using the new MCAT. and customize them to your needs. HIGHEST CONDITION CODE WAS 0 DEFINE NONVSAM (NAME('SYS1.6.MCATA) IDC0001I FUNCTION COMPLETED. MAXIMUM CONDITION CODE WAS 0 You have now completed the initial set up of your ICF catalog environment. 3. HIGHEST CONDITION CODE WAS 0 IDC0002I IDCAMS PROCESSING COMPLETE. and standards. it can be advisable to expand from the single LPAR to a SYSPLEX or multi-LPARs with data sharing using a coupling facility (CF). Consider the other tasks not shown here. HIGHEST CONDITION CODE WAS 0 DEFINE NONVSAM (NAME('SYS1.MCATA) IDC0001I FUNCTION COMPLETED.1 Scenario 2: Target environment This scenario introduces a new LPAR using the same master catalog. See Figure 3-2 on page 43 for a clear picture of the catalog structure.

PAGE=. Consider the effects that this will have on the environments. Add a new LPAR to the DR plan. SWAP=. Input/output (I/O) generation (IOGEN) is planned to be the same as for the existing LPAR. You should also use &SYSNAME for the SMFPRMxx parameters DSNAME= and SID=. Deploying ICF catalogs 43 . Table 3-2 Scenario 2 activities Step Activity Prepare the existing environment (LPAR A) for sharing. An initial program load (IPL) is planned to be done from the same IPL volume with existing parameters. Use the symbolic name &SYSNAME wherever possible. User catalogs that are going to be shared should have these changes: 򐂰 Change the share option to SHR(3 4) if not already in place.3. and report procedures. and NONVIO=. You may have to make changes to your automation routines to accommodate the shared environments. Make changes to the catalog maintenance plan. DUPLEX=. The documentation for recovery procedures should be updated. 򐂰 RLS is a strong option to consider. Make changes to the existing MCAT.LPAR A LPAR B MCAT A UCAT A1 UCAT A2 UCAT B1 UCAT B2 Figure 3-2 Introducing LPAR B 3.LOGREC.2 Scenario 2: Task list LPAR B is to be added to an existing LPAR A environment connected to the existing master catalog MCAT A. Prepare existing user catalogs for sharing. “Record-level sharing support for ICF catalogs” on page 73 is dedicated to RLS. 򐂰 Move shared catalogs to the virtual lookaside facility (VLF) for most efficient access. Add any alias that might be unique for the new LPAR. Chapter 3. See Table 3-2. You should use &SYSNAME for the IEASYSxx parameters VIODSN=. 򐂰 򐂰 򐂰 Make changes to catalog monitoring. LOGREC=SYS1. Recovery can be done from the other LPAR if one is down. automation. 򐂰 Ensure that catalogs are on a device defined as a shared unit control block (UCB). Make changes to the DR plan. 򐂰 Consider using Enhanced Catalog Sharing (ECS). because the LPARs can now be used as backup for each other. Chapter 5. and determine what areas you want to monitor and report on to ensure smooth operational running. Define RPO and recovery time objective (RTO) for the new LPAR. Document recovery demands for the new LPAR.&SYSNAME. for example.

3 Scenario 2: Deployment The first step is to check if the current catalogs have the correct shareoptions values set: 1.4) IDC0531I ENTRY SYS1.4.UCATA2 ALTERED IDC0001I FUNCTION COMPLETED. HIGHEST CONDITION CODE WAS 0 ALTER SYS1.3) SPEED UNIQUE NOERASE INDEXED 2.UCATA1 ALTERED IDC0001I FUNCTION COMPLETED. as shown on Example 3-7.4. You can do this by issuing a LISTC ENT(/) CAT(/) ALL command next to the catalog name on TSO ISPF option 3. HIGHEST CONDITION CODE WAS 0 ALTER SYS1. You can alter the shareoptions value with the IDCAMS ALTER command. Example 3-6 Displaying catalog shareoptions DATA ------. HIGHEST CONDITION CODE WAS 0 ALTER SYS1.UCATA1 SHAREOPTIONS(3. The necessary value to share the catalog is 3.MCATA SHAREOPTIONS(3.UCATB2 SHAREOPTIONS(3. Example 3-7 Alter catalog shareoptions ALTER SYS1.MCATB HISTORY DATASET-OWNER-----(NULL) CREATION--------2014. If you already have this value set up.4) IDC0531I ENTRY SYS1. HIGHEST CONDITION CODE WAS 0 ALTER SYS1.3. you can skip to the next step.3.4) IDC0531I ENTRY SYS1.MCATA ALTERED IDC0001I FUNCTION COMPLETED.321 RELEASE----------------2 EXPIRATION------0000.UCATB1 ALTERED IDC0001I FUNCTION COMPLETED.SYS1.UCATB2 ALTERED IDC0001I FUNCTION COMPLETED. See Example 3-6 for a sample of the output.4) IDC0531I ENTRY SYS1.UCATB1 SHAREOPTIONS(3.000 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.UCATA2 SHAREOPTIONS(3. HIGHEST CONDITION CODE WAS 0 44 A Practical Guide to ICF Catalogs .4) IDC0531I ENTRY SYS1.

CLOSE(SYS1.UCATA2 ECSHR IDC0531I ENTRY SYS1.com/abstracts/sg245594.UCATB2)' ICH70001I GILBION LAST ACCESS AT 16:05:39 ON MONDAY.CLOSE(SYS1.UCATA1 ALTERED IDC0001I FUNCTION COMPLETED. Alter the catalog attribute to use ECS on your environment if possible. HIGHEST CONDITION CODE WAS 0 ALTER SYS1.COND CODE 0000 5.CLOSE(SYS1. changing backup polices.4 Scenario 3: Multiple MCATs and UCATs In our third scenario.UCATA1 ECSHR IDC0531I ENTRY SYS1. For more information about ECS. See Example 3-9 for a sample of how to do this. HIGHEST CONDITION CODE WAS 0 ALTER SYS1. and the data will be reloaded on the next catalog request. Because the CLOSE command will only remove the data from CATALOG memory. Some of this data is shared by both LPARs. Chapter 3.MCATA)' 5 // COMMAND 'F CATALOG.UCATB1)' 8 // COMMAND 'F CATALOG. we create a new MCAT to be used by LPAR B. It is also advised to use ECS to improve catalog performance. HIGHEST CONDITION CODE WAS 0 4.UCATA2 ALTERED IDC0001I FUNCTION COMPLETED. See Example 3-8.UCATB1 ALTERED IDC0001I FUNCTION COMPLETED.UCATA1)' 6 // COMMAND 'F CATALOG. It is also required to define a cache structure and update the COFVLFxx member.html?Open Example 3-8 Define Enhanced Catalog Sharing ALTER SYS1.ibm. In order for new changes to take effect.CLOSE(SYS1. HIGHEST CONDITION CODE WAS 0 ALTER SYS1. HIGHEST CONDITION CODE WAS 0 ALTER SYS1. Deploying ICF catalogs 45 . and so on. Before starting the systems. see Enhanced Catalog Sharing and Management. SG24-5594 on the following website: http://www. there should be no affect to system or applications.3.redbooks.MCATA ALTERED IDC0001I FUNCTION COMPLETED. it is necessary to close the catalog on CATALOG address space. and move application data between the LPARs. NOVEMBER 17.UCATA2)' 7 // COMMAND 'F CATALOG.UCATB2 ECSHR IDC0531I ENTRY SYS1.UCATB1 ECSHR IDC0531I ENTRY SYS1. although other data sets are exclusive to each LPAR.UCATB2 ALTERED IDC0001I FUNCTION COMPLETED. Example 3-9 Close catalogs to activate new settings 3 //COMMAND EXEC PGM=IEFBR14 4 // COMMAND 'F CATALOG. including updating parmlibs. extra steps might be necessary. copying data sets.CLOSE(SYS1.MCATA ECSHR IDC0531I ENTRY SYS1. 3.STEP WAS EXECUTED . 2014 IEF142I REFCAT COMMAND .

򐂰 򐂰 򐂰 Make changes to catalog monitoring. . Define RPO and RTO for the new LPAR. Table 3-3 Scenario 3 activities Step Activity Examine and Diagnose the old MCAT and UCATs. LPAR A LPAR B MCAT Split MCAT A MCAT B UCAT A1 UCAT A2 UCAT B1 Shared Shared UCAT B2 Figure 3-3 Create a new master catalog for LPAR B and introduce sharing 3. automation. and report procedures. Make changes to the DR plan. Table 3-3 illustrates the suggested plan for this scenario. 򐂰 Ensure that catalogs are on a device defined as shared UCB. some UCATs can be shared by two MCATs. Recovery procedures (documentation) should be updated. 򐂰 Diagnose the integrity of the current MCAT. Add any alias that might be unique for the new LPAR. At the same time.3. and move several UCATs connecting to the new MCAT. Make changes to the catalog maintenance plan. 46 A Practical Guide to ICF Catalogs Add a new LPAR to the DR plan.2 Scenario 3: Task list This scenario describes splitting one MCAT into two separate MCATs. because the LPARs can now be used as backup for each other (recovery can occur from the other LPAR if one is down). Document recovery demands for the new LPAR. Make changes to the existing MCAT. 򐂰 Move shared catalogs to VLF for most efficient access.4. and share UCATS A2 and B1 between both LPARs (Figure 3-3). consider using ECS.4.1 Scenario 3: Target environment We create MCAT B on LPAR B. UCATs that are going to be shared should have these changes made: 򐂰 Change the share option to SHR(3 4) if it is not already in place. Make sure that the current MCAT and UCATs are error-free: 򐂰 Run IDCAMS VERIFY and EXAMINE commands on the current MCAT. Prepare existing UCATs for sharing. No changes. 򐂰 For most optimum sharing.

depending on your needs. EXAMINE.NO ERRORS DETECTED IDC0001I FUNCTION COMPLETED. Deploying ICF catalogs 47 . We also ran a LISTC job to list catalog entries. Example 3-11 Output from the Listcat catalogs IDCAMS SYSTEM SERVICES LISTING FROM CATALOG -. Chapter 3. HIGHEST CONDITION CODE WAS 0 DIAGNOSE ICFCATALOG INDATASET(SYS1. AVERAGE KE IDC11774I CURRENT INDEX CISIZE IS 3584. RECOMMENDED MINIMUM INDEX CISI IDC01724I INDEXTEST COMPLETE . to reduce the exposure to disk failures and contention.UCATA1 THE NUMBER OF ENTRIES PROCESSED WAS: AIX -------------------0 ALIAS -----------------0 CLUSTER ---------------2 DATA ------------------2 GDG -------------------0 INDEX -----------------1 NONVSAM -----------15968 PAGESPACE -------------0 PATH ------------------0 SPACE -----------------0 USERCATALOG -----------0 TAPELIBRARY -----------0 TAPEVOLUME ------------0 TOTAL -------------15973 3. See Example 3-11. You can use your current MCAT as a model. and the output directed to data sets. or use specific allocation attributes. HIGHEST CONDITION CODE WAS 0 EXAMINE NAME(SYS1. This is necessary to create a valid recovery point if there is a failure.UCATA1) IDC0001I FUNCTION COMPLETED.3 Scenario 3: Deployment Make sure that all catalogs are error-free before creating the catalogs’ backup and starting the split project: 1.4.UCATA1) INDEXTEST NODATATEST ERRORLIMIT(1000) IDC01700I INDEXTEST BEGINS IDC11773I 731 KEYS PROCESSED ON INDEX LEVEL 1.3. Example 3-10 VERIFY. AVERAGE KE IDC11773I 7 KEYS PROCESSED ON INDEX LEVEL 2. It is not recommended to place the new MCAT on the same volume as the current MCAT.SYS1. to compare the data set and ALIAS definition after the system split. and DIAGNOSE catalogs VERIFY DATASET(SYS1. HIGHEST CONDITION CODE WAS 0 2. EXAMINE.UCATA1) IDC0001I FUNCTION COMPLETED. and DIAGNOSE commands against a user catalog. Define a new master catalog for LPAR B. Example 3-10 shows the job output from VERIFY.

(001)-DTDSC(01).UCATB2 CATALOG NAME MCAT.321 16:49:58 DATA SET FILTERING IS COMPLE AND 0 FAILED FOR OTHER REASONS ADR454I (001)-DTDSC(01). Check the output to confirm that the backup job ran with no errors. THE FOLLOWING DATA SETS WERE SUCCESSFULLY PROC CLUSTER NAME SYS1. using different tools. TASKID 001 HAS BEEN ASSIGNED TO COMMAND 'DUMP ADR109I ADR016I ADR006I ADR801I (R/I)-RI01 (01). an ADRDSSU logical or physical memory dump.Z1C.UCATB2 OUTFILE(EXPDD) TEMPORARY IDC0005I NUMBER OF RECORDS PROCESSED WAS 5971 IDC0594I PORTABLE DATA SET CREATED SUCCESSFULLY ON 11/17/14 AT 16:49:59 IDC1147I IT IS RECOMMENDED THAT DIAGNOSE AND EXAMINE BE RUN BEFORE IDC1147I IMPORT OF CATALOG IDC0001I FUNCTION COMPLETED. or other tools available on your environment.Example 3-12 shows the IDCAMS parameters.SANDBOX.SBOX00 COMPONENT NAME SYS1. See Example 3-13.321 16:49:57 EXECUTION BEGINS 2014. create a backup copy of your catalogs.UCATB2.CATINDEX IDC0001I FUNCTION COMPLETED. to provide an alternative restore option. (001)-PRIME(01). Example 3-13 Back up your catalogs DUMP DS(INCL( SYS1.UCATB2 COMPONENT NAME SYS1. (001)-STEND(01).UCATB2 )) OUTDD(OUT1) COMPRESS OPT(4) ALLDATA(*) ALLEXCP ADR101I (R/I)-RI01 (01). Next. This can be done by using the IDCAMS EXPORT command. HIGHEST CONDITION CODE WAS 0 4.321 16:49:58 EXECUTION ENDS EXPORT SYS1. Example 3-12 Define a new MCAT for LPAR B DEFINE USERCATALOG (NAME(SYS1.CATINDEX ADR006I (001)-STEND(02).MCATB) ICFCAT STORAGECLASS(NONSMS) VOLUME (VOLMC2) CYLINDERS (25 10) FREESPACE(10 10) SHR(3 3) STRNO(3)) DATA (CISZ(4096) BUFND(4)) INDEX(BUFNI(4)) IDC0510I CATALOG ALLOCATION STATUS FOR VOLUME VOLMC2 IS 0 IDC0512I NAME GENERATED-(I) SYS1.MCATB. HIGHEST CONDITION CODE WAS 0 48 A Practical Guide to ICF Catalogs .321 16:49:57 INITIAL SCAN OF USER CONTROL RACF LOGGING OPTION IN EFFECT FOR THIS TASK 2014. Create at least two backup copies of your catalogs. 2014. 2014.

If you disconnect the ALIAS from the catalog. HIGHEST CONDITION CODE WAS 0 IDC0002I IDCAMS PROCESSING COMPLETE. Remember that NOMERGECAT requires the target catalog to be empty to complete the copy.Attention: Avoid using a high-level qualifier that might be disconnected from your system. Chapter 3. use REPRO NOMERGECAT to copy the necessary information from MCAT A to MCAT B. Deploying ICF catalogs 49 .UCATB2' CAT('SYS1. IDC0005I NUMBER OF RECORDS PROCESSED WAS 110 IDC0001I FUNCTION COMPLETED. the data sets being moved will be deleted from the source catalog. Attention: If you use the MERGECAT option.UCATB2 DURING EXPORT DISCONNECT WAS IDC0144I VOLUC2 IDC0001I FUNCTION COMPLETED.1 or later.MCATA') DISCONNECT IDC0144I VOLUME SERIAL FOR SYS1. Then. 5.1 or later. Example 3-14 shows the parameters to do this. This can lead to system failures and outages to your system if system data sets are removed from the MCAT. the REPRO NOMERGECAT step should have also imported and connected the ALIAS to the new master catalog. You might need to EXPORT DISCONNECT the UCATs that will not be accessed by the new MCAT.MCATB) NOMERGECAT IDC11468I NVR/VVR NOW POINTS TO TARGET CATALOG. Make sure that all of the ALIASes on the catalogs being exported will not be used on LPAR A. all catalog and ALIAS information is also copied. Example 3-15 Export disconnect catalog EXPORT 'SYS1. you might not be able to access your data set unless you recatalog it. 6. Next.MCATA) OUTDATASET(SYS1. HIGHEST CONDITION CODE WAS 0 If you are running z/OS V2. Example 3-14 REPRO NOMERGECAT REPRO INDATASET(SYS1. you can run REPRO commands to move these data set to other catalogs. use the EXPORT DISCONNECT command to remove the catalog that will not be accessed from LPAR A. See Example 3-15. If necessary. MAXIMUM CONDITION CODE WAS 0 Tip: If you are running z/OS 2.

UCATB2) IDC0001I FUNCTION COMPLETED.UCATB2) INDEXTEST NODATATEST ERRORLIMIT(1000) IDC01700I INDEXTEST BEGINS IDC11773I 275 KEYS PROCESSED ON INDEX LEVEL 1.UCATB2)) CAT(SYS1. HIGHEST CONDITION CODE WAS 0 Tip: If the user catalog was already imported. Analyze the output to identify and correct any errors encountered. If you are running earlier versions of z/OS or are not sure if the ALIASes were also copied. you receive a duplicate data set name for the existing ALIAS.MCATB) IDC0001I FUNCTION COMPLETED. RECOMMENDED MINIMUM INDEX CISI IDC01724I INDEXTEST COMPLETE .UCATB2) NOINDEXTEST DATATEST ERRORLIMIT(1000) IDC01701I DATATEST BEGINS IDC01709I DATATEST COMPLETE . Example 3-17 Define ALIAS to new master catalog DEFINE ALIAS(NAME(CICS) RELATE(SYS1.MCATB) IDC0001I FUNCTION COMPLETED. HIGHEST CONDITION CODE WAS 0 DEFINE ALIAS(NAME(GDPS) RELATE(SYS1.UCATB1)) CAT(SYS1.UCATA2)) CAT(SYS1. AVERAGE KE IDC11774I CURRENT INDEX CISIZE IS 3584.MCATB) IDC0603I CONNECT FOR USER CATALOG SYS1. HIGHEST CONDITION CODE WAS 0 EXAMINE NAME(SYS1. and DIAGNOSE catalogs VERIFY DATASET(SYS1.MCATB) IDC0001I FUNCTION COMPLETED. HIGHEST CONDITION CODE WAS 0 EXAMINE NAME(SYS1. Define the ALIAS used on LPAR B. Example 3-18 VERIFY. Example 3-16 Connect catalogs to new master catalog IMPORT CONNECT OBJECTS( SYS1. See Example 3-17. EXAMINE. 8. If the ALIAS information was already copied during the REPRO NOMERGECAT step.7.NO ERRORS DETECTED 50 A Practical Guide to ICF Catalogs . you receive a duplicate data set name error message. HIGHEST CONDITION CODE WAS 0 9. as shown in Example 3-16.UCATB2 DEVICETYPE(3390) VOLUMES(VOLUC2)) CATALOG(SYS1. you can run the IMPORT CONNECT command to connect user catalogs. See Example 3-18.NO ERRORS DETECTED IDC0001I FUNCTION COMPLETED. EXAMINE. and DIAGNOSE against the catalogs. Check the integrity of your catalogs after the split by running a new VERIFY. HIGHEST CONDITION CODE WAS 0 DEFINE ALIAS(NAME(TIVOLI) RELATE(SYS1.UCATB2 SUCCESSFUL IDC0001I FUNCTION COMPLETED. AVERAGE KE IDC11773I 3 KEYS PROCESSED ON INDEX LEVEL 2.

Suppose that our sample organization acquires another organization. So. as shown in Figure 3-4. Deploying ICF catalogs 51 . 3. HIGHEST CONDITION CODE WAS 0 10. In this case. and therefore gains another LPAR.Also run a new backup of your catalogs.4.IDC01708I 275 CONTROL INTERVALS ENCOUNTERED IDC01710I DATA COMPONENT CONTAINS 5971 RECORDS IDC01711I DATA COMPONENT CONTAINS 0 DELETED CONTROL INTERVALS IDC01712I MAXIMUM LENGTH DATA RECORD CONTAINS 221 BYTES IDC01722I 96 PERCENT FREE SPACE IDC0001I FUNCTION COMPLETED.4 Scenario 4: Acquisition and integration This scenario deals with a consolidation and integration of ICF catalog environments.UCATB2) IDC0001I FUNCTION COMPLETED. we call this LPAR C. 3. LPAR A LPAR C MCAT C UCAT C1 LPAR B MCAT A UCAT C2 UCAT A1 MCAT B UCAT A2 Shared UCAT B1 UCAT B2 Shared Figure 3-4 Acquisition of LPAR C Chapter 3. but initially keep the data separate.4. update any documentation and processes that are changed by the catalog split. The organization might want to consolidate all of the LPARs onto its own mainframe.5 Scenario 4: Acquisition and integration environments The ICF catalog environment for LPAR C would appear as a separate system initially. the newly acquired would be an independent system with its own MCAT and UCATs. HIGHEST CONDITION CODE WAS 0 DIAGNOSE ICFCATALOG INDATASET(SYS1.

LPAR C LPAR A MCAT A MCAT C UCAT C1 LPAR B UCAT C2 MCAT B UCAT A1 UCAT A2 Shared UCAT B1 UCAT B2 Shared Figure 3-5 Integration of LPAR C into the full ICF catalog environment 3. it is common that users. 򐂰 For most optimum sharing. Back up all of the catalogs involved in the merge process. Table 3-4 Scenario 4 activities: 52 Step Activity Examine and diagnose catalogs. Back up the catalogs.6 Scenario 4 Task list This task list in Table 3-4 describes merging MCAT C into MCAT A.4. In addition. Confirm that all necessary resources (direct access storage device (DASD) and tape) are configured and accessible on the receiving system. MCAT C is no longer required. 򐂰 Move shared catalogs to VLF for more efficient access. During a system merge. A Practical Guide to ICF Catalogs . consider catalog availability and sizing. it also includes consideration steps for identifying data set names conflicts. to protect system resiliency. 򐂰 Diagnose the integrity of the catalogs. Check if user catalogs have the necessary settings for sharing: 򐂰 Change shareoptions to SHR(3 4) if not already in place. Analyze data set conflicts. consider the following points: 򐂰 Analyze your systems to identify all data set and ALIAS conflicts. applications. 򐂰 If user catalogs will also be merged. and even user catalog names match on both systems. Prepare user catalogs for sharing. Check your resources. consider using ECS.The goal to integrate the LPAR C ICF catalog environment into the LPAR B structure might look something like Figure 3-5. 򐂰 Define if any ALIASes or data sets should be renamed. Make sure that all catalogs are error-free before performing backups and system merge: 򐂰 Run IDCAMS VERIFY and EXAMINE commands on all of the current catalogs. To prevent errors related to duplicate data set names during your merge. UCATs C1 and C2 are now connected to MCAT A.

򐂰 Check the job output to identify and solve any remaining data set names conflicts. and DIAGNOSE commands against the catalogs.4. check the integrity of your catalogs. 3. AVERAGE KEY LENGTH IDC11774I CURRENT INDEX CISIZE IS 3584.7 Scenario 4 Deployment This sections describes the steps to consolidate two MCATs into a single MCAT used by both LPARs. Example 3-20 Sample error detected by diagnose IDC21364I ERROR DETECTED BY DIAGNOSE: ICFCAT ENTRY: SYS1.MCATA (D) RECORD: 00000000000000000000000000000000000000000000 OFFSET: X'004A' REASON: 11 . EXAMINE.UNCOMPLETED DELETE DETECTED IDC21365I ICFCAT RECORD DISPLAY: RECORD: 00000000000000000000000000000000000000000000 000000 00DB0034 C3004800 2D000000 00000000 00000000 000020 00000000 00000000 00000000 00000000 00000000 000040 FF481432 1F00000F 01000012 C4004480 0BE2E8E2 000060 FFFFFFFF FFFFFF08 14321F00 000F0100 001E0400 /00 /00 00000000 00000014 F14BD4C3 E5D6D3D4 00000000 01FFFFFF C1E3C100 C3F13010 000000 FFFFFF 001401 200F82 Chapter 3. shut down the sending system.UCATB1) IDC0001I FUNCTION COMPLETED. We advise you to run at least the VERIFY. Examine and diagnose catalogs. EXAMINE. 򐂰 Define the ALIASes on the receiving MCAT. Deploying ICF catalogs 53 . You can also use more tools to check catalog integrity. Analyze and fix any errors encountered before proceeding to next steps. During the merge window. Before performing the catalog merge. HIGHEST CONDITION CODE WAS 0 2. Create a new backup of your catalogs to protect the system from errors and outages. See Example 3-19.NO ERRORS DETECTED IDC0001I FUNCTION COMPLETED. It is assumed that LPAR A and LPAR C have access to each other’s storage devices: 1. AVERAGE KEY LENGTH IDC11773I 3 KEYS PROCESSED ON INDEX LEVEL 2. and DIAGNOSE catalogs VERIFY DATASET(SYS1. and perform the following tasks: 򐂰 Merge catalogs with the MERGECAT option.UCATB1) 0016500 INDEXTEST NODATATEST ERRORLIMIT(1000) 0016600 IDC01700I INDEXTEST BEGINS IDC11773I 275 KEYS PROCESSED ON INDEX LEVEL 1.Step Activity Merge master catalog and define ALIASes. HIGHEST CONDITION CODE WAS 0 EXAMINE NAME(SYS1.UCATB1) IDC0001I FUNCTION COMPLETED. and DIAGNOSE commands against the catalogs to make sure that they have no errors after system merge. Back up your new configuration. EXAMINE. Example 3-20 shows a sample error detected during DIAGNOSE processing. Run new VERIFY. Example 3-19 VERIFY. HIGHEST CONDITION CODE WAS 0 DIAGNOSE ICFCATALOG INDATASET(SYS1. RECOMMENDED MINIMUM INDEX CISIZE IS 3 IDCAMS SYSTEM SERVICES TIME: 09:44:2 IDC01724I INDEXTEST COMPLETE .

Example 3-22 Sample catalog backup EXPORT SYS1. Example 3-21 Sample LISTCAT output IDCAMS SYSTEM SERVICES 09:50:46 TIME: LISTING FROM CATALOG -. Next.UCATC1 THE NUMBER OF ENTRIES PROCESSED WAS: AIX -------------------0 ALIAS -----------------0 CLUSTER ---------------2 DATA ------------------2 GDG -------------------0 INDEX -----------------1 NONVSAM ------------1000 PAGESPACE -------------0 PATH ------------------0 SPACE -----------------0 USERCATALOG -----------0 TAPELIBRARY -----------0 TAPEVOLUME ------------0 TOTAL --------------1005 THE NUMBER OF PROTECTED ENTRIES SUPPRESSED WAS 0 IDC0001I FUNCTION COMPLETED.REASON CODE: 11 IDC0001I FUNCTION COMPLETED. create backup copies of your catalogs. It is suggested to create at least two backup copies using different techniques for extra recovery protection.MCATA (D) . to compare the data prior to and after the merge. HIGHEST CONDITION CODE WAS 0 EXPORT SYS1. Also create a LISTCAT copy of your catalogs. HIGHEST CONDITION CODE WAS 0 4. HIGHEST CONDITION CODE WAS 8 3. HIGHEST CONDITION CODE WAS 0 54 A Practical Guide to ICF Catalogs . See Example 3-22.MCATA OUTFILE(EXPDD) TEMPORARY IDC0005I NUMBER OF RECORDS PROCESSED WAS 12 IDC0594I PORTABLE DATA SET CREATED SUCCESSFULLY ON 11/17/19 AT 16:50:43 IDC1147I IT IS RECOMMENDED THAT DIAGNOSE AND EXAMINE BE RUN BEFORE IDC1147I IMPORT OF CATALOG IDC0001I FUNCTION COMPLETED.MCATC OUTFILE(EXPDD) TEMPORARY IDC0005I NUMBER OF RECORDS PROCESSED WAS 12 IDC0594I PORTABLE DATA SET CREATED SUCCESSFULLY ON 11/17/19 AT 16:49:58 IDC1147I IT IS RECOMMENDED THAT DIAGNOSE AND EXAMINE BE RUN BEFORE IDC1147I IMPORT OF CATALOG IDC0001I FUNCTION COMPLETED.SYS1.000080 00003000 00000000 00000000 0000001B C9004D00 14E2E8E2 F14BD4C3 C1E3C1 0000A0 C3C1E3C9 D5C4C5E7 00001401 FFFFFFFF FFFFFFFF 0814321F 00000F01 00001E 0000C0 00E5D6D3 D4C3F130 10200F82 00000020 00000000 00000000 000000 IDC21363I THE FOLLOWING ENTRIES HAD ERRORS: SYS1. See Example 3-21.

4. see 3. Make sure that all other required changes are in place before performing the REPRO MERGECAT. and define the ALIASes on the correct MCAT. “Scenario 3: Multiple MCATs and UCATs” on page 45. For detailed information about the IMPORT CONNECT and DEFINE ALIAS commands. shut down LPAR C and perform the REPRO MERGECAT to move the necessary entries to the LPAR A MCAT.5.VVDS. and updating system parmlibs. These changes might include renaming data sets to avoid data set name duplication.TEST.V98 IDC3013I DUPLICATE DATA SET NAME IDC3009I ** VSAM CATALOG RETURN CODE IS 8 .V180 IDC0639I SPHERE CONVERSION STARTED FOR SYS1.TEST. Chapter 3.REASON CODE IS IGG0CLEH-38 Tip: If you are running earlier than z/OS version 2. Deploying ICF catalogs 55 .UCATC2 IDC0639I SPHERE CONVERSION STARTED FOR SYS1.UCATC2 IDC01402I SPHERE CONVERSION COMPLETED FOR SYS1. using system variables for data set and member names.TEST. Check the output and correct any errors identified.V97 IDC0639I SPHERE CONVERSION STARTED FOR SYS1.VVOLMC3 IDC21409I CONVERSION FAILED FOR SPHERE SYS1.TEST. Example 3-23 REPRO MERGECAT IDC0639I SPHERE CONVERSION STARTED FOR SYS1. During your change window.V180 IDC01402I SPHERE CONVERSION COMPLETED FOR SYS1. Example 3-23 shows some data sets moved. you should also IMPORT CONNECT the UCATs. and errors related to data set name conflict. 6.1.

UCATC2) IDC0001I FUNCTION COMPLETED. create a new backup of your catalog environment and keep it for an extended period. After starting the systems.NO ERRORS DETECTED IDC01708I 47 CONTROL INTERVALS ENCOUNTERED IDC01710I DATA COMPONENT CONTAINS 1004 RECORDS IDC01711I DATA COMPONENT CONTAINS 0 DELETED CONTROL INTERVALS IDC01712I MAXIMUM LENGTH DATA RECORD CONTAINS 221 BYTES IDC01722I 99 PERCENT FREE SPACE IDC0001I FUNCTION COMPLETED. HIGHEST CONDITION CODE WAS 0 8. monitoring.UCATC2) 00291001 NOINDEXTEST DATATEST ERRORLIMIT(1000) 00292001 IDC01701I DATATEST BEGINS IDC01709I DATATEST COMPLETE . HIGHEST CONDITION CODE WAS 0 EXAMINE NAME(SYS1. update any documentation. RECOMMENDED MINIMUM INDEX CISIZE IS 35 IDC01724I INDEXTEST COMPLETE . and automations currently running on your systems. and DIAGNOSE after MERGECAT VERIFY DATASET(SYS1. Finally. EXAMINE. Example 3-24 VERIFY. HIGHEST CONDITION CODE WAS 0 DIAGNOSE ICFCATALOG INDATASET(SYS1. AVERAGE KEY LENGTH: IDC11774I CURRENT INDEX CISIZE IS 3584. Also.NO ERRORS DETECTED IDC0001I FUNCTION COMPLETED. 56 A Practical Guide to ICF Catalogs . run a new catalog integrity check and analyze any possible errors.7.UCATC2) 00169001 INDEXTEST NODATATEST ERRORLIMIT(1000) 00169101 IDC01700I INDEXTEST BEGINS IDC11773I 47 KEYS PROCESSED ON INDEX LEVEL 1. reporting.UCATC2) IDC0001I FUNCTION COMPLETED. See Example 3-24. HIGHEST CONDITION CODE WAS 0 EXAMINE NAME(SYS1.

4 Chapter 4. 2015. Maintaining ICF catalogs This chapter provides information about the following areas of maintenance for integrated catalog facility (ICF) catalogs: 򐂰 Backup 򐂰 Reorganization 򐂰 Recovery 򐂰 Recovery using IBM z/OS Data Facility Storage Management Subsystem (DFSMS) V2. All rights reserved. 57 .1 features 򐂰 Forward Recovery using the Integrated Catalog Forward Recovery Utility (ICFRU) 򐂰 Modify Catalog commands © Copyright IBM Corp.

An ICF catalog backup is at least a recovery point that you can go back to rather than having to re-create all catalog pointers manually. cross-reference the ICF catalogs and the corresponding backup to ensure that they are always in sync. 4. Check your procedures for implementing ICF catalog backup to make sure that you have all of the ICF catalogs backed up.1 Backup of ICF catalogs The backup of ICF catalogs is essential to being able to recover an ICF catalog environment from a failure.4. 4. or DFSMS Hierarchical Storage Manager (DFSMShsm). and also due to the risk you always have when working with ICF catalogs. as described later in this chapter. DFSMS Data Set Services (DFSMSdss). but also consider if you should back up more often. Consider taking more than one backup and maybe use different tools. and an individual backup cycle for each catalog. but because many applications today run continuously.1 Maintenance This section describes various aspects of ICF maintenance. for example. such as the IDCAMS EXPORT command. especially if you have hundreds of thousands of data sets on thousands of disk volumes. Whenever you create a new ICF catalog. only reorganize when your ICF catalog has reached a high number of extents. Reorganization of ICF catalogs should be minimized due to its disruptive nature. such as the following items: 򐂰 򐂰 򐂰 򐂰 Criticality of the individual ICF catalog Frequency of updates to the ICF catalog Size (you do not want to lose a big catalog with many data set pointers) Customer service level agreement (SLA) to meet recovery time objective (RTO) and recovery point objective (RPO) Back up once a day is typical. because you must consider which ICF catalog is in scope. a backup should be scheduled. 58 A Practical Guide to ICF Catalogs . your recovery procedure will of course be more complex. or if you should also differentiate your backup based on criticality or other criteria. Backup of your ICF catalogs might depend on more criteria. As a guideline. also consider extending the primary and secondary allocation of the ICF catalog to prevent future requirements for reorganization. leaving you time to schedule a reorganization in time before the maximum limit of 123 extents is reached. In addition. If you choose to differentiate. A forward recovery procedure should also be prepared in addition to having your backup.1. Before the reorganization. Often. this means that this activity must happen in maintenance windows where the application can be closed down. Daily backup is advised unless you have ICF catalogs with extremely low activity. A limit could be 80 extents.2 Reorganizing ICF catalogs The reorganization of ICF catalogs is disruptive. This can increase your recovery potential if you have a severe ICF catalog issue. this is becoming more and more difficult. and can mean outage for business-critical applications while the reorganization goes on.1. based on ICFRU for example. This manual task can be unrealistic.

Using CA Reclaim can reduce the need for reorganization. Example 4-1 shows how multilevel alias aaaa.bbbb is moved to a different ICF catalog.1. You might also do this if the number of entries in the ICF catalog brings it to a size where having that many entries for one application in one user catalog increases your risk. If one high-level qualifier thrashes one of your ICF catalogs due to high activity. Another reason for moving selected catalog entries to a different ICF catalog can also be a very high-activity application that increases the number of ICF catalog’s extents.Do not do reorganization to prevent control interval (CI) and control area (CA) splits. features change. and should also be implemented on your ICF catalogs.1. you of course have to create this with an IDCAMS DEFINE ALIAS statement. Let these happen.4 Other types of ICF catalog maintenance You might need to work with your ICF catalogs for other reasons than already mentioned. because they will reoccur anyway after you have done the reorganization.bbbb ) MERGECAT /* If this is a new multilevel alias. perhaps through using a multilevel alias on your target ICF catalog. Example 4-1 Example of moving data sets for a multilevel alias to a different ICF catalog REPRO INDATASET(your source catalog ) OUTDATASET(your target catalog LEVEL(aaaa. Alternatively. Moving the qualifier isolates this problem and avoids affecting other high-level qualifiers and applications due to reorganization. IMBED and REPLICATE are removed. which can be implemented using an IDCAMS ALTER command. consider moving that qualifier to a dedicated ICF catalog. You could then offload the ICF catalog. and might need to be extended. either by command or using IDCAMS ALTER statements. When you redefine the ICF catalog. 4. Do this by backing up the ICF catalog and import it again. the REPRO MERGECAT function can also be used to balance ICF catalogs across the entire ICF catalog environment. Chapter 4. Many changes to the ICF catalog itself are dynamic. Catalogs can grow in size over the years.3 REPRO MERGECAT as maintenance tool The IDCAMS REPRO MERGECAT function is primarily a migration tool used for moving data to a different ICF catalog at one of the following levels: 򐂰 Full ICF catalog level 򐂰 Data set level 򐂰 Alias level From a maintenance perspective. 4. Relate this to the new ICF catalog to be able to access data there. Also. causing the frequent need for reorganization. where new disk technology has made parameters obsolete. only move maybe one critical application to its own ICF catalog. It is advised to remove the IMBED and REPLICATE parameters for performance reasons. as for IMBED and REPLICATE. An example of this is CA Reclaim. Maintaining ICF catalogs 59 . This is appropriate if the input/output (I/O) rate on one ICF catalog grows too high.

and to do so on a regular schedule. to ensure integrity. After the procedures have been outlined. Example 4-2 Example of locking an ICF catalog using the RECOVER command F CATALOG. This command works for both Virtual Storage Access Method (VSAM) record-level sharing (RLS)-managed and non-RLS-managed catalogs. A LOCK is generally used when the catalog is being reorganized or recovered. a SUSPEND and RESUME option is also available on the RECOVER command. Restriction: Only UCATs can be locked. you should build the jobs and prepare the commands needed to perform the recovery.1. You can release the LOCK status again by issuing the UNLOCK command shown in Example 4-3. with the message UNABLE TO ACCESS LOCKED CATALOG.RECOVER. Suspending catalog access while performing maintenance and recovery A significant enhancement that was added in DFSMS V2.1 for ICF catalogs is the MODIFY RECOVER command to support ICF catalog recovery procedures.RECOVER.5 Recovery of ICF Catalogs ICF catalog recovery procedures should be well-documented and prepared. it is not accessible until an UNLOCK is issued. The final step is to test the recovery procedure. LOCK fails any user request while the LOCK is active. both of which are explained in this chapter. The intent of the following topics is to describe the current features that can be used in catalog recovery. The catalog appears as LOCKED in a catalog display. 60 A Practical Guide to ICF Catalogs . It is not possible to lock an MCAT.SUSPEND(ucat name) This command performs a serialized close of the ICF catalog across the sysplex. As an alternative to the RECOVER LOCK and UNLOCK command pair. In a catalog display. See the RECOVER LOCK command issued in Example 4-2. Also note that if a catalog is locked during initial program load (IPL). but also has a corresponding SUSPEND option. The RECOVER LOCK command is equivalent to the IDCAMS LOCK function.4.LOCK(ucat name) This command locks and closes the ICF catalog sysplex-wide. to make the best possible and least disruptive recovery. or by an IDCAMS ALTER LOCK command. The SUSPEND command prevents unauthorized requests to the catalog. as shown in Example 4-4.RECOVER. You perform a SUSPEND on an ICF catalog.UNLOCK(ucat name) This command releases a LOCK set by a previous RECOVER LOCK command. Trying to access the catalog. the catalog appears as in a suspended state. you get return code 186. Example 4-3 Example of unlocking an ICF catalog using RECOVER command F CATALOG. Example 4-4 Example of suspending an ICF catalog using the RECOVER command F CATALOG. It can be used for locking and unlocking the ICF catalog while maintenance or recovery goes on.

RLSTST) IEC351I CATALOG ADDRESS SPACE MODIFY COMMAND ACTIVE IEC352I MODIFY CATALOG UCAT.RECOVER. commands.RLSTST) IEC351I CATALOG ADDRESS SPACE MODIFY COMMAND ACTIVE IEC352I MODIFY CATALOG UCAT.SUSPEND(UCAT. and the SUSPEND and RESUME. Maintaining ICF catalogs 61 .RECOVER. Example 4-7 Messages returned when doing a RECOVER UNLOCK command F CATALOG. Example 4-5 Example of resuming an ICF catalog using the RECOVER command F CATALOG. or as a fully qualified catalog name. ensuring full integrity. Example 4-9 Messages received when issuing a RECOVER RESUME command F CATALOG.RLSTST TO STATE UNLOCK SUCCESSFUL IEC352I CATALOG ADDRESS SPACE MODIFY COMMAND COMPLETED Example 4-8 and Example 4-9 show a comparison of the messages that you receive when you issue corresponding RECOVER SUSPEND and RECOVER RESUME commands. Chapter 4. but they are failed using the LOCK and UNLOCK option. as shown in Example 4-5. It is also important to mention that the ALTER LOCK and SUSPEND command now works sysplex-wide.RECOVER. Example 4-8 Messages received when issuing a RECOVER SUSPEND command F CATALOG.RLSTST) IEC351I CATALOG ADDRESS SPACE MODIFY COMMAND ACTIVE IEC352I MODIFY CATALOG UCAT. The SUSPEND and RESUME option might be preferable to the LOCK and UNLOCK option. you can release catalog activity by issuing the RECOVER command with a RESUME option.RECOVER.RESUME(UCAT.UNLOCK(UCAT.RESUME(ucat name) The user catalog name can be specified generically.RLSTST TO STATE RESUME SUCCESSFUL IEC352I CATALOG ADDRESS SPACE MODIFY COMMAND COMPLETED Important: You need to ensure appropriate access to the IGG.LOCK(UCAT.RLSTST TO STATE SUSPEND SUCCESSFUL IEC352I CATALOG ADDRESS SPACE MODIFY COMMAND COMPLETED The messages from the RECOVER RESUME look like those shown in Example 4-9. Also. because catalog requests during maintenance and recovery are queued and restarted using the SUSPEND and RESUME option.CATLOCK resource to be able to perform the LOCK and UNLOCK.RLSTST) IEC351I CATALOG ADDRESS SPACE MODIFY COMMAND ACTIVE IEC352I MODIFY CATALOG UCAT.RECOVER.RLSTST TO STATE LOCK SUCCESSFUL IEC352I CATALOG ADDRESS SPACE MODIFY COMMAND COMPLETED Example 4-7 shows messages returned from a RECOVER UNLOCK command. Example 4-6 shows some of the messages returned when you LOCK an ICF catalog using RECOVER. The messages related to the SUSPEND are shown in Example 4-8.To resume catalog activity after maintenance or recovery. Example 4-6 Messages returned when performing a RECOVER LOCK command F CATALOG.

At this point. you no longer have to maintain the ALIAS pointers yourself. as shown in Example 4-11. even if the ICF catalog is defined on another volume or device type.1 introduced new options for the DELETE USERCATALOG and DEFINE USERCATALOG statements. the ALIAS was visible. NODISCONNECT and RECONNECT subparameters scenario For visualization of how the new NODISCONNECT and RECONNECT subparameters work.RLSTST USERCATALOG RECOVERY NODISCONNECT /* Next. 62 A Practical Guide to ICF Catalogs . to keep and maintain alias information about the deletion and creation of an ICF catalog. the user catalog was defined again with the RECONNECT keyword to connect to the existing ALIAS pointers. The ICF catalog is defined again with the RECONNECT parameter. RECONNECT also supports rebuilding alias information. with one data set created.4 after defining the UCAT. there are no data sets.DELETE USERCATALOG and DEFINE USERCATALOG using NODISCONNECT and RECONNECT DFSMS V2. RECONNECT can be used on the DEFINE USERCATALOG command to have the newly defined catalog connected to the existing alias information. a DELETE of an ICF catalog is done using the keyword NODISCONNECT. The following figure shows an example of deleting an ICF catalog using the NODISCONNECT keyword. NODISCONNECT retains alias information at the deletion of a user catalog. Example 4-11 Example of defining an ICF catalog using the RECONNECT parameter //IDCAMS EXEC PGM=IDCAMS //SYSPRINT DD SYSOUT=* //SYSIN DD * DEFINE USERCATALOG (NAME(UCAT. so that you do not have to rebuild this information manually. With DELETE USERCATALOG NODISCONNECT and DEFINE USERCATALOG RECONNECT parameters. When we looked up the ALIAS using IBM z/OS Interactive System Productivity Facility (ISPF) 3. Example 4-10 Example of deleting an ICF catalog using the NODISCONNECT parameter //IDCAMS EXEC PGM=IDCAMS //SYSPRINT DD SYSOUT=* //SYSIN DD * DELETE UCAT.RLSTST ) ICFCATALOG VOLUME(MHL1A0) TRK(5 1) STORCLAS(SCRLS) DATACLAS(WELCHRLS) LOG(NONE) RECONNECT FREESPACE(0 0) NOIMBED NOREPLICATE) DATA (CISZ(4096)) /* The only ALIAS in this test catalog was RLSTST. See Example 4-10. If you use this parameter.

RLSTST PREALLOCATED. THE FOLLOWING DATA SETS WERE NOT PROCESSED FROM THE LOG RLSTST ADR454I (001)-TDLOG(01). The restore was resubmitted (using the BCSRECOVER and SUSPEND keywords).CATLOCK facility class was provided.DFSMSdss awareness of LOCK and SUSPEND keywords DFSMSdss also has the awareness of LOCK and SUSPEND when an ICF catalog is restored.. Maintaining ICF catalogs 63 .RLSTST INDDNAME(TAPE) OUTDDNAME(DASD) BCSRECOVER(LOCK) )) - The RESTORE job was submitted and failed as expected.. you can specify either LOCK or SUSPEND. Example 4-14 Restore of ICF catalog using DFSMSdss and BCSRECOVER keyword ADR442I (001)-FRLBO(01).RLSTST. Example 4-13 DFSMSdss fails on BSCRECOVER(LOCK) due to missing access ICH408I USER(MHLRES1 ) GROUP(SYS1 ) NAME(MHLRES1/ name ) IGG.RLSTST COMPONENT UCAT. We did this so that we could demonstrate that DFSMSdss tries to do a LOCK during recovery. THE FOLLOWING DATA SETS WERE SUCCESSFULLY PROCESSED Chapter 4. PROCESSING SUSPENDED USER CATALOG UCAT. IN CATALOG MCAT.SANDBOX.CATINDEX ADR372W (001)-TDNVS(03). whichever is selected. When the ICF catalog has been successfully recovered.RLSTST. Example 4-12 Recovering an ICF catalog using DFSMSdss with keyword BCSRECOVER(LOCK) RESTORE DATASET( INCLUDE( UCAT. and DFSMSdss changes the status of the ICF catalog being restored to either of these. we tried to restore an ICF catalog without having the required access to the IGG. DFSMSdss performs the corresponding UNLOCK or RESUME to reestablish normal ICF catalog access.RLSTST WAS RESTORED CATALOG MCAT. CLUSTER UCAT.. CLUSTER UCAT. READ access to the IGG. as shown in Example 4-14.SBOX00 COMPONENT UCAT.CATLOCK resource. To correct the error. On the BCSRECOVER keyword.8 ADR489I (001)-TDLOG(02).RLSTST WAS DUMPED USING RECORD LEVEL SHARI GMT TIMESTAMP IS: 2014. The job ended successfully this time.Z1C.RLSTST ADR963I (001)-TDLOG(02). ADR480W (001)-TDLOG(01).275 00:04:50. In the following scenario..CATLOCK CL(FACILITY) INSUFFICIENT ACCESS AUTHORITY ACCESS INTENT(READ ) ACCESS ALLOWED(NONE ) Example 4-13 shows that DFSMSdss does an implicit LOCK (or SUSPEND) of an ICF catalog before restoring the ICF catalog from a backup. We used these DFSMSdss parameters to perform the recover. as shown in Example 4-13. See Example 4-12. ALIAS RLSTST NOT RESTORED FOR USER CATALOG UCAT. DATA SET UCAT. ADR360I (001)-TDVSM(02).

and System Management Facilities (SMF) data. Start by issuing the F CATALOG. This ensures the integrity of the backups. DFSMSdss performs a SUSPEND of the catalog. we redefine the catalog using the RECONNECT parameter (counterpart to NODISCONNECT). the SUSPEND also closes the catalog across the sysplex. 6. in addition to RLSQUIESCE or RLSENABLE. You still need backups as your recovery checkpoint. DFSMSdss detects the suspended status and restores the empty catalog without using additional serialization. we assume that we need to delete the current ICF catalog and allocate it on a different volume. 5. the only outstanding action is to release the SUSPEND of the catalog. The catalog is then protected against concurrent updates. to perform forward recovery from this checkpoint to a current point in time. When the restore has been successfully completed. Only data set control blocks (DSCBs) and VSAM volume records (VVRs) for the ICF catalog are being deleted. because this information is already in place. In this scenario. For the non-RLS-managed ICF catalogs. DFSMSdss finds the user catalog preallocates. DFSMSdss enqueues on SYSIGGV2. Most importantly. DFSMSdss also has the awareness that the memory dump was performed using RLS access (implicit QUIESCE before doing the memory dump. Restore can now happen using DFSMSdss.1: 1.Before processing the restore. ICFRU performs the forward recovery from this point in time to the current point in time. After the ICFRU forward recovery of the ICF catalog has been completed. you need to do forward recovery from the time that the ICF catalog backup was taken until the current point in time. except from the SUSPEND. The SUSPEND issued by DFSMSdss is resumed again after a successful RESTORE. and SUSPEND parameters. while the SUSPEND parameter ensures integrity throughout the recover. Catalog recovery using the current features Recovery of a catalog might be changed to use the newest DFSMS V2. the RLS-managed ICF catalogs are quiesced. See the following section. Using the VOLSER parameter updates across the sysplex. for an RLS-managed ICF catalog. as done in DFSMSdss before V2. Next. We are now ready to delete the catalog from its current location (having verified the we have a valid backup). Use the VOLSER(vvvvvv).RESUME(ucat name) command.RECOVER. This queues catalog requests against this catalog without failing them. Now. ensuring integrity. 64 A Practical Guide to ICF Catalogs . 2. Perform this action with the F CATALOG.RECOVER. and RECONNECT connects existing alias information. LOG(NONE).SUSPEND(ucat name) command.1 enhancements. Delete should occur with the NODISCONNECT parameter to maintain alias pointers. and consequently does not restore the alias. 4. Backup of ICF catalogs can be done in multiple ways: 򐂰 Using IDCAMS EXPORT 򐂰 Using DFSMShsm 򐂰 Using DFSMSdss Using DFSMSdss as the backup tool.1. look at the steps for a catalog recovery based on DFSMSdss and the features introduced in DFSMS V2. “ICF catalog forward recovery using ICFRU” on page 65. You can accomplish this by using the ICFRU tool that uses selected SMF records for the forward recovery. and RESUME after the DUMP had finished). 3.

Next.0 DFSMS Managing Catalogs. it can be used as input in an IDCAMS IMPORT. and RENAMES). You must also provide the following information as input parameters: – – – – – ICF catalog name Start date and time for the forward recovery End date and time for the forward recovery Maximum gap time in minutes in your SMF records. Even though it was not created by an IDCAMS EXPORT command.ICF catalog forward recovery using ICFRU This section describes the ICFRU functionality. Introduction to ICFRU ICFRU is the recovery tool that can perform forward recovery of your ICF catalogs based on a point-in-time IDCAMS EXPORT backup and SMF records. the analysis and processing step uses the IDCAMS EXPORT copy as input. “Catalog recovery using the current features” on page 64. Additionally. you need the ICF catalog-related SMF records in the time span from the last successful IDCAMS EXPORT of the ICF catalog to the current point in time. You must use ICF catalog-related SMF records from all logical partitions (LPARs) having access to the ICF catalog in scope for recovery. to bring the ICF catalog to the current point in time. and in descending date and time order. to be able to apply later changes to the ICF catalog (DELETES. If a data set occurs in SMF as a deletion. and only looks at the most recent processing of the individual data set as seen in the SMF data. the input is the SMF records for the time period since the most recent IDCAMS EXPORT up to the current point in time. Beneath the job control language (JCL) for the selection and validation step (the CRURRSV program). it is not written to the output file. If a data set is seen on the IDCAMS EXPORT. but not in the SMF records. Before being processed in the analysis and processing step. and has all of the ICF catalog updates up to the current point in time. this section shows the processing steps to give you an understanding of how ICFRU works: 1. DEFINES. The CRURRAP program ends up creating an IDCAMS EXPORT file. SMF records are sorted in data set name order. SC23-6853. Note that ICFRU can only do forward recovery based on this. Read this section to see how ICFRU can complement the recovery scenario described in the previous section. Maximum clock difference in seconds (between your LPARs) Chapter 4. ICFRU processing steps ICFRU is explained in much more detail in z/OS V2R1. ICFRU performs the forward recover based on two programs in ICFRU: 򐂰 Integrated Catalog Forward Recovery Record Selection and Validation (CRURRSV) 򐂰 Integrated Catalog Forward Recovery Record Analysis and Processing (CRURRAP) The record selection and validation step validates the SMF records and issues alerts on time gaps that might indicate missing SMF data. However. Of course. it is written to the output file. Maintaining ICF catalogs 65 . How ICFRU works ICFRU needs a successful backup of your ICF catalog in order to be recovered based on IDCAMS EXPORT. any creates are added to the output file.

REGION=1024K.RECS. //* ’STARTDATE’.smfdump(0). //* ’GAPTIME’. //* ’STOPDATE’.DCB=BUFNO=60 // DD DSN=system1.name’. // UNIT=SYSDA.smfdump(-1). Example 4-15 JCL example for the ICFRU selection and validation step //******************************************************** //* INTEGRATED CATALOG FORWARD RECOVERY UTILITY * //* JCL EXAMPLE .LOAD.RLSE).smfdump(0).(10.’hh:mm:ss’.DISP=SHR.’STARTTIME’.smfdump(-n). // ’mmmm’. // ’mm/dd/yy’.CATLG). //* ’CLOCKDIFFERENCE’) //STEPLIB DD DSN=USER.DISP=(NEW.DISP=SHR.NAME’. //******************************************************** //RRSV EXEC PGM=CRURRSV.DCB=BUFNO=60 66 A Practical Guide to ICF Catalogs . * //* MODIFY THE STEPLIB DD STATEMENTS TO POINT TO THE * //* INSTALLATION LOAD MODULE LIBRARY CONTAINING CRURRSV. * //* SPECIFY PARAMETERS APPROPRIATE TO THIS RECOVERY.’STOPTIME’.The JCL for the selection and validation step is shown in Example 4-15.DCB=BUFNO=60 //SMFOUT DD DSN=SMF. * //******************************************************** //* RUN CRURRSV TO EXTRACT APPROPRIATE DATA.DISP=SHR.* //* THIS JCL EXECUTES RECORD SELECTION AND VALIDATION. * //* MODIFY THE SMFIN DD STATEMENT TO POINT TO THE * //* SMF INPUT DATA SETS.DCB=BUFNO=60 // DD DSN=system2.CAT.smfdump(-n).DISP=SHR //SYSUDUMP DD SYSOUT=* //SYSPRINT DD SYSOUT=* //SYSLOG DD SYSOUT=* //SMFIN DD DSN=system1. // PARM=(’catalog.’hh:mm:ss’.DCB=BUFNO=60 // DD DSN=system2.SPACE=(CYL.DCB=BUFNO=60 // DD DSN=system2.DISP=SHR.DISP=SHR. // ’ssss’) //*PARM=(’CATALOG.DISP=SHR.DCB=BUFNO=60 // DD DSN=system1. * //* SMFOUT MAY BE PASSED TO THE SORT STEP.1).smfdump(-1). // ’mm/dd/yy’. * //* MODIFY THE SMFOUT DD STATEMENT TO CONFORM TO YOUR * //* INSTALLATION CONVENTIONS.

EQUALS SORT FIELDS=(218. * //* THE SORTIN MAY BE PASSED FROM THE CRURRSV STEP.RECS.CAT. The next step is to sort the SMF records extracted by the CRURRSV program in data set order. as shown in Example 4-16.44.DISP=(NEW.CAT.11.* //* THIS JCL EXECUTES THE NECESSARY SORT STEP.BI.2.FILSZ=E10000. // UNIT=SYSDA. * //******************************************************** //* SORT THE OUTPUT FROM CRURRSV.CATLG). Maintaining ICF catalogs 67 .PD.SORTED. * //* THE SORTOUT MAY BE PASSED TO THE CRURRAP STEP. * //* MODIFY THE SORTOUT DD STATEMENT TO CONFORM TO YOUR * //* INSTALLATION CONVENTIONS.262.A. and in descending date and time order.D) /* Chapter 4.RECS. //******************************************************** //SORT EXEC PGM=ICEMAN //SYSOUT DD SYSOUT=* //SORTIN DD DSN=SMF. * //* MODIFY THE SORT STEP AS APPROPRIATE.4. The input for this sort is your SMF data from the time span between your most recent IDCAMS EXPORT backup and the current point in time.4.D.SPACE=(CYL. TO INVOKE * //* THE SORT PRODUCT AVAILABLE IN YOUR INSTALLATION * //* (THIS EXAMPLE INVOKES DFSORT.DISP=SHR //* DD DSN=concatenations if necessary //SORTOUT DD DSN=SMF.RLSE) //SYSIN DD * OPTION DYNALLOC=SYSDA. Example 4-16 Sort of SMF records before using these in ICFRU analysis and processing step //******************************************************** //* INTEGRATED CATALOG FORWARD RECOVERY UTILITY * //* JCL EXAMPLE .A.(10.BI.1).7.1.CH.

NEW.’STOPTIME’.CATALOG. It is suggested that you run daily jobs to extract these record types into “ICFRU-ready” SMF files. //* ’STOPDATE’. The final step is to build an up-to-date IDCAMS EXPORT file for recovery of the catalog. 68 A Practical Guide to ICF Catalogs . * //* MODIFY THE EXPIN DD STATEMENT TO POINT TO * //* THE EXPORT DATA SET TO BE USED AS THE RECOVERY BASE. * //* THE SMFIN DD STATEMENT MAY SPECIFY A DATA SET * //* PASSED FROM THE SORT STEP. Example 4-17 shows a sample of the analysis and processing step. You must provide the same input parameters as for the selection and validation step.CAT. During the extract of ICF catalog SMF data from a large volume of SMF data. // PARM=(’catalog. you should have jobs prepared that can extract from current SMF data sets. //******************************************************** //RRAP EXEC PGM=CRURRAP. // ’mm/dd/yy’.EXPORT.DCB=BUFNO=60 //EXPIN DD DSN=Baplicat. * //* SPECIFY PARAMETERS APPROPRIATE TO THIS RECOVERY: * //* THEY SHOULD BE THE SAME AS THE ONES USED IN CRURRSV. * //* MODIFY THE STEPLIB DD STATEMENTS TO POINT TO THE * //* INSTALLATION LOAD MODULE LIBRARY CONTAINING CRURRAP. Time is critical when you encounter ICF catalog errors. struggling with building JCL and potential JCL errors can cost you valuable time and complicate the recovery. In addition to having the ICFRU jobs ready with good documentation.CATALOG. // ’mm/dd/yy’.SORTED.(10. * //* MODIFY THE EXPOUT DD STATEMENT TO CONFORM TO YOUR * //* INSTALLATION CONVENTIONS.* //* THIS JCL EXECUTES RECORD SELECTION AND PROCESSING. Example 4-17 JCL example of ICFRU analysis and processing step (CRURRAP) //******************************************************** //* INTEGRATED CATALOG FORWARD RECOVERY UTILITY * //* JCL EXAMPLE .DISP=SHR.DISP=(NEW. The input for this file is your sorted SMF records and the IDCAMS EXPORT of your ICF catalog. //* ’CLOCKDIFFERENCE’) //STEPLIB DD DSN=USER.DISP=SHR.3. //* ’STARTDATE’.65. // ’ssss’) //*PARM=(’CATALOG.’hh:mm:ss’. // UNIT=SYSDA. so that you need less processing time when an ICF catalog failure occurs. //* ’GAPTIME’.LOAD.REGION=1536K. Also.DCB=BUFNO=60 Preparations for ICFRU forward recovery ICFRU will use SMF record types 61.66.1).name’.BACKUP(0).’STARTTIME’.NAME’.DISP=SHR //SYSUDUMP DD SYSOUT=* //SYSPRINT DD SYSOUT=* //SYSLOG DD SYSOUT=* //SMFIN DD DSN=SMF.RLSE).’hh:mm:ss’. your procedures should also have been tested in a controlled test system or environment.CATLG).RECS. * //******************************************************** //* RUN CRURRAP USING OUTPUT FROM SORT. // DCB=BUFNO=60 //EXPOUT DD DSN=Baplicat.SPACE=(CYL. // ’mmmm’.

some commands might be repeated. F CATALOG.VUNALLOCATE Deallocate a VVDS from CAS. 4. so that you can handle them promptly before they escalate. It is imperative to always understand any ICF catalog problem and root cause fully. Use an automation tool to pick up the most possible ICF catalog error messages. Of course.7 Modify CATALOG commands Catalog address space (CAS) can be operated for display and actions. any ICF catalog issue requires other considerations related to your environment and to the individual case. F CATALOG. because ICF catalog errors that grow in extent can have a large effect on both the information technology (IT) operations.SYS% Apply maintenance from one system. we grouped the commands into types of actions to outline which commands could be useful for you in these scenarios. Table 4-1 ICF catalog MODIFY commands grouped by type Type of command Command Description Operator command F CATALOG. In addition. Due to this grouping. such as ICF catalog full messages. In the following table. 4. Performing ICF catalog diagnosis and recovery requires good skills and experience in ICF catalog handling.VCLOSE Close a VSAM volume data set (VVDS). Chapter 4. The intent of Table 4-1 is to help you decide which command you might use for a certain task. because they provide warnings through nonzero return codes back to your job scheduling package.1.1. Frequent IDCAMS DIAGNOSE jobs should be scheduled.UNALLOCATE Deallocate a catalog from CAS. We split the commands into these groups: 򐂰 򐂰 򐂰 򐂰 򐂰 򐂰 򐂰 򐂰 Operator commands Information about CAS and CAS activity Fixing temporary catalog problems Temporary modifications to ICF catalogs Catalog recovery RLS-related catalog commands Catalog parmlib commands Enhanced catalog sharing commands The commands related to these main actions are listed according to this grouping.CLOSE Close an ICF catalog. It is therefore strongly advised to detect any early warnings on any errors that your ICF catalogs might have. Maintaining ICF catalogs 69 . F CATALOG. These warnings alert you to potential problems.This short introduction focuses on the ICFRU procedure. and on the business. because they can appear in more than one context. however. and to have a good backup of the current status before starting the recovery.6 ICF catalog diagnostics Diagnosing ICF catalog errors is an important discipline. F CATALOG. it requires a well-prepared toolbox (JCL library with prepared jobs for any kind of ICF catalog operation or recovery). at the earliest stage possible.

F CATALOG.ENTRY Display catalog entry points.CATALOG. elapsed times.END End a catalog task. MODIFY.REPORT Create a catalog report.ALLOCATE Open a catalog. F CATALOG. F CATALOG.CATMAX Change the maximum number of allowed catalogs. F CATALOG.ALIASLEVEL Change an aliaslevel.LIST Lists currently active CAS service tasks.CATSTATX Create a statistics report (extended). F CATALOG.UNALLOCATE Deallocate a catalog. F CATALOG. F CATALOG. MODIFY.CLOSE Close a catalog.CATALOG. F CATALOG.VUNALLOCATE Deallocate a VVDS.REPORT. Fixing temporary catalog problems Temporary modifications to ICF catalog 70 A Practical Guide to ICF Catalogs .DETAIL List a job in queue in a catalog request.ABEND Terminate a catalog task. F CATALOG. F CATALOG.TASKMAX(nnn) Change the maximum allowed tasks. F CATALOG.CATSTATS Create a statistics report. F CATALOG. F CATALOG. F CATALOG.ECSHR(STATUS) Display Enhanced Catalog Sharing (ECS). their related job names.TAKEDUMP Have a catalog take an SVC memory dump.RESTART Restart CAS. F CATALOG.REPORT.{ISC|NOISC} Set the usage of In Storage Cache (ISC). F CATALOG.CONTENTION Display the catalog contention setting.LISTJ(jobname).PERFORMANCE Create a performance report. and unique identifications. F CATALOG. F CATALOG.VCLOSE Close a VVDS.ALLOCATED Display allocated ICF catalogs.REPORT.Type of command Command Description Information about CAS and CAS activity F CATALOG. F CATALOG. F CATALOG. F CATALOG. F CATALOG.UNALLOCATE Deallocate a catalog. F CATALOG.{VLF|NOVLF} Add or remove a catalog to or from the virtual lookaside facility (VLF).CLOSE Close a catalog.{SYS%ON|SYS%OFF Apply maintenance from one system. F CATALOG.

F CATALOG.RECOVER.ENABLE(DELFORCEWNG) Enables the display of message IDC1997I and IDC1998I when a DELETE VVDS RECOVERY or DELETE USERCATALOG FORCE is performed. F CATALOG. F CATALOG.DELRECOVWNG(ENABLE | DISABLE) Turn warning message IDC1999I on or off.UNLOCK(ucat name) Unlock a command after recovery.nn. D SMS. F CATALOG.a.DUMPON Set dynamic DUMP on CAS.) Display or modify the contention setting on ALLOCLCK. F CATALOG.n. F CATALOG.RLSENABLE(ucatname) Enable the ICF catalog for RLS usage. Catalog recovery RLS-related catalog commands Catalog parmlib command Chapter 4. F CATALOG. F CATALOG.RESUME(ucat name) Resume a catalog after a suspend. F CATALOG.Type of command Command Description Temporary modifications to ICF catalog (continued) F CATALOG.CFLS Display the coupling facility (CF) lock structure.DISABLE(DSNCHECK) Perform a syntax check on added names.nn. F CATALOG. F CATALOG.RLSENABLE.DISABLE(DELFORCEWNG) Disables the warning message IDC1997I or IDC1998I when attempting to use the DELETE VVDS RECOVERY or DELETE USERCATALOG FORCE command.ENABLE Turn warning message IDC1997I on or off.SUSPEND(ucat name) Suspend a catalog for recovery. F CATALOG.LOCK(ucat name) Lock a catalog for recovery.CONTENTION. F CATALOG. F CATALOG.RECOVER. F CATALOG.CATMAX(1024) Display the maximum number of open catalogs.SYSTEM Quiesce a whole system from RLS.EXTENDALIAS(ENABLE | DISABLE) Enable or disable Extended Alias.DSNCHECK(ENABLE | DISABLE) Perform a syntax check on added names. F CATALOG. F CATALOG.(SYSIGGV2.DISABLE(DELFORCEWNG).RLSQUIESCE.RECOVER.RECOVER. F CATALOG. Maintaining ICF catalogs 71 .RLSQUIESCE(ucatname) Quiesce an RLS-enabled ICF catalog.SYMREC(ENABLE | DISABLE) Enable SYMREC creation.CONTENTION.a) Display or modify the SYSIGGV2 contention setting. F CATALOG.ECSHR(AUTOADD) Activate ECS on ECS-eligible catalogs.SYSTEM Enable a whole system for RLS.NOTIFYEXENT(nn) Notify if the ICFCAT maximum number of allowed extents was exceeded. D SMS.CFLF(IGWLOCK00) Display the IGWLOCK00 structure. F CATALOG.ALIASLEVEL(n) Display the number of alias levels.(ALLOCLOCK. F CATALOG. F CATALOG.

CONTENTION(SYSZTIOT. F CATALOG.STATUS(ucatname) Display ECS status.ECSHR(DISCONNECT) Disconnect ECS from a structure. F CATALOG.nn.TASKMIN(nnn) Set the minimum number of user service tasks in CAS. F CATALOG.Type of command Command Description Catalog parmlib command (continued) F CATALOG.a) Display or modify the SYSZTIOT contention setting. F CATALOG.ECSHR(CONNECT) Connect ECS to a structure.sec) Change the VVDS allocation default.a) Display or modify the SYSZVVDS contention setting.VVRCHECK(ENABLE | DISABLE) Enable or disable a VVR check. F CATALOG.REMOVE(catname) Remove a catalog from ECS.UPDTFAIL(ENABLE | DISABLE) Activate the IEC390I catalog update error. F CATALOG. F CATALOG. F CATALOG.nn. Enhanced Catalog Sharing commands 72 A Practical Guide to ICF Catalogs .VVDSSPACE.ECSHR. F CATALOG.TASKMAX(nnn) Set the maximum number of user service tasks in CAS.ECSHR.CONTENTION(SYSZVVDS.(prim. F CATALOG.

possible states. The function requires data sets using RLS control to be SMS-managed. VSAM RLS uses a CF-based Lock Manager and CF Cache Manager in the implementation of record-level sharing. sizing. RLS provides multisystem sharing at a record level across a sysplex using the coupling facility (CF).5 Chapter 5. support for catalogs. and consists of the following areas: 򐂰 RLS for catalog overview. and new parameters 򐂰 Considerations for implementing RLS. RLS is one of many access modes to VSAM. This can be done by using the job control language (JCL) RLS parameter. fallback. © Copyright IBM Corp. including preparing to migrate. or by specifying MACRF=RLS in the access method control block (ACB). Record-level sharing support for ICF catalogs This chapter describes the Virtual Storage Access Method (VSAM) record-level sharing (RLS) for integrated catalog facility (ICF) catalogs. 73 . The following resources are other types of access: 򐂰 Nonshared resources (NSRs) 򐂰 Local shared resources (LSRs) 򐂰 Global shared resources (GSRs) These access types can be referred to as non-RLS access to distinguish RLS from other modes of access. All rights reserved. including requirements. and maintenance 򐂰 Basic tests results. migrating. and is set at OPEN time. including benchmarks and documentation RLS access for VSAM is widely used today for both IBM z/OS Data Facility Storage Management Subsystem (DFSMS) Transactional VSAM Services (DFSMStvs) and non-transactional applications. enhanced features. 2015.

if you do not regularly have to do disruptive maintenance on your ICF catalogs due to split or catalog reorganization. Moving to 64-bit buffering using RLS reduces the input/output (I/O) activity. improving performance by not having to serialize on the SYSIGGV2 bcsname. or to LOG(UNDO) or LOG(ALL) if the data set is recoverable. If SMSVSAM is not started.5. which includes an IBM Parallel Sysplex. SMSVSAM restart depends on the SHCDSs being available.1. 74 A Practical Guide to ICF Catalogs . As always required with VSAM RLS data sets. Using RLS for ICF catalogs provides locking at a record level.1 Requirements for RLS RLS use for ICF catalogs requires the base RLS setup. SC23-6860 for information about how to set up the initial RLS environment. SMSVSAM holds SYSIGGV2 bcsname SHARE while a catalog is opened for RLS access. See z/OS DFSMSdfp Storage Administration. The SMSVSAM started task must be up and running. Also.1 introduces RLS usage for ICF catalogs and for volume catalogs (VOLCATs). Alter the attribute to LOG(NONE) if unrecoverable. The improvements described here are also expected to solve performance bottlenecks. the SMSVSAM address space must be active. Define and activate two of these data sets for normal use. availability is improved. 5.ACTIVE command from the operator console. The purpose of the enhancement is to improve performance and availability on ICF catalogs. which are key for RLS recovery because they hold vital information for it. This ensures catalog data integrity from programs relying on SYSIGGV2 to serialize the catalog. The data sets in scope for RLS use must have their attributes changed as related to unrecoverable and recoverable options. In addition. You need to define the sharing control data sets (SHCDSs). and also one for spare purposes. and to reduce the need for customers to split ICF catalogs to meet availability requirements. add the storage class created for this RLS application to have the appropriate cache set and CF cache assigned. The SMSVSAM address space automatically starts at initial program load (IPL) if you specify the RLSINIT (YES) keyword in the IGDSMSxx member of SYS1.PARMLIB.1 RLS for catalog overview DFSMS V2. you need a storage class assigned with a Cacheset and Lockset identifying cache and lock structures in the CF. you can start the address space after IPL by issuing the V SMS. The following information is part of the contents of the SHCDSs: 򐂰 򐂰 򐂰 򐂰 򐂰 The name of the CF lock structure in use A list of subsystems and their status A list of open data sets using the CF A list of data sets with unbound locks A list of data sets in permit non-RLS state SHCDSs are basically VSAM linear data sets. For example. In addition.SMSVSAM.

The only exception is if you encounter an issue that forces you to revert to non-RLS mode. or one that was last closed in RLS mode.5. The catalog has been defined or altered to have the LOG attribute. The LOG attribute must be defined. 򐂰 RLS enabled. This is mandatory to be able to access the catalog in RLS mode (as for all other VSAM data sets). OPEN is only possible for non-RLS access. However. in this implementation. These commands work sysplex wide. These commands are not intended for daily usage. A catalog in RLS mode can be defined as a catalog currently open in RLS mode. the ICF catalogs can be in one of the following states: 򐂰 Eligible. For pre-DFSMS V2. Chapter 5. only unrecoverable catalogs are supported. ICF catalogs to support RLS access mode.1. Catalogs can be defined or altered in one of four states: 򐂰 򐂰 򐂰 򐂰 Eligible RLS quiesced RLS enabled RLS mode These states are described in detail in 5. In this case. it requires that all instances of SMSVSAM be shut down. 򐂰 RLS mode. ready for being accessed in RLS mode. Record-level sharing support for ICF catalogs 75 . and have the catalogs perform a serialized close across the sysplex. they are suspended while the QUIESCE or ENABLE process happens.3. or that you use the IDCAMS SHCDS CFRESETDS(user_cat_name) command to reset the RLS-IN-USE indicator. New operator commands and define parameters have also been provided to support the transition of ICF catalogs between RLS and non-RLS access.1.2 Support for ICF catalogs using RLS New and existing parameters are needed to define new. even when submitted from only one logical partition (LPAR). or to alter existing. this either requires that SMSVSAM be up and running. 5. As RLS enabled. The RLS-IN-USE indicator can also be set in the catalog’s VVR.3 ICF catalog possible RLS states In an environment using RLS. The RLS quiesced indicator is set to YES in the catalog’s VSAM volume record (VVR) entry. The OPEN in this state still occurs as non-RLS. it should remain in RLS mode. 򐂰 RLS quiesced. so the attribute is LOG(NONE). Important: The ability to move to RLS-accessed ICF catalogs from non-RLS-accessed catalogs is only supported for user catalogs (UCATs). Catalog requests issued during transition are not failed. RLS enable indicator is set to YES in the catalog’s VVR entry. The status ends up being RLS enabled or RLS quiesced. The catalog can be taken out of RLS mode by command. the catalog can only be opened in RLS mode. However.1 systems. “ICF catalog possible RLS states”. It is not supported for master catalogs (MCATs).1. When an ICF catalog has been migrated to RLS mode. You also need parameters to set the type of initial access (non-RLS or RLS) at the first access of the catalog. Regard these commands as migration or fallback tools.

76 A Practical Guide to ICF Catalogs . the two parameters directly related to RLS catalogs. RLS obtains buffers dynamically as they are needed.4 Comparison with other enhanced catalog features RLS for catalog can coexist with other advanced catalog features.1. it still makes sense to use ECS. Table 5-1 on page 77 is an overview of the parameters added in DFSMS V2. The DATACLAS controls the buffer pool (64 bit) and caching options through the RLSOVETHEBAR and RLSCFCACHE settings. Catalogs using RLS do not need to access the VVR in cache. BUFND. new parameters were added to the DEFINE USERCATALOG command. Storage is obtained in the catalog address space (CAS). The STORAGECLASS cacheset defines which cache structure to use. and is intended for shared catalogs.5. because RLS mode is not supported for MCATs at this point. 5. these settings are ignored in an RLS environment. Virtual storage is possible within a range of 1 MB (megabyte) . using system-managed buffering (SMB).5 New parameters on the DEFINE USERCATALOG command To support the transition to RLS. such as STRNO. Cached records are stored in a data space. we compare enhanced catalog features with the RLS for catalog feature: Enhanced Catalog Sharing Enhanced Catalog Sharing (ECS) places the ICF catalogs self-describing VVR in the CF (SYSIGGCAC_ECS) to reduce I/O. For other tuning parameters. For catalogs in RLS mode. This is also the case for RLSQUIESCE and RLSENABLE. Catalog Data Space Cache Catalog Data Space Cache (CDSC) uses virtual lookaside facility (VLF). For SUSPEND and RESUME. Locking happens at a record level to serialize access to individual catalog records rather than enqueuing on the SYSIGGV2 bcsname resource (entire catalog). divided into 4 KB (kilobyte) blocks.1. The parameters define the availability level for readiness to use RLS mode on the ICF catalogs. the catalog is either in the SUSPEND state or the RESUME state.1. In this section. the catalog records are placed in RLS local buffer pools or CF cache structures. The enhanced catalog features can be replaced by RLS. It is intended for non-shared catalogs.2 GB (gigabytes). The catalog status is either QUIESCED or ENABLED. For the MCAT. and BUFNI on the DEFINE USERCATALOG command. In Storage Cache In Storage Cache (ISC) has a maximum of 3041 basic catalog structure (BCS) records.

that LOG(NONE) is used to prepare for RLS. The following new parameters support RLS: RLSQUIESCE The catalog is accessed in non-RLS mode following the define.Table 5-1 RLS added parameters New DEFINE user catalog parameters Description Default SUSPEND Requests are suspended until a RESUME has been issued. or RESUME a catalog. using the new RLSENABLE parameter. or if the RACF FACILITY class is not active. RLSQUIESCE RLSENABLE The catalog will be accessed in RLS mode after the allocation. See the OA45623 description for further information. RLSENABLE The catalog is accessed in RLS mode after the allocation. The SUSPEND parameter was mutually exclusive with the LOCK parameter but since OA45623 this is no longer the case. Chapter 5. If the IGG. Record-level sharing support for ICF catalogs 77 . The command releases a SUSPEND state. N/A LOG(NONE) Catalog is eligible for access in RLS mode. an authorized user with READ access to the IBM Resource Access Control Facility (IBM RACF®) STGADMIN profile can still access and repair a locked catalog. UNLOCK. RESUME RLSQUEISCE The catalog is accessed in non-RLS mode following the define. Example 5-1 Example of defining an ICF catalog using RECONNECT and RLSENABLE DEFINE USERCATALOG (NAME(UCAT. you cannot LOCK. RLSQUIESCE RECONNECT Request the new catalog to use existing ALIAS information.BNC) ICFCATALOG VOLUME(MHL1A0) TRK(50 50) STORCLAS(SCRLS) DATACLAS(WELCHRLS) LOG(NONE) RECONNECT RLSENABLE FREESPACE(20 20) NOIMBED NOREPLICATE) DATA (CISZ(32768)) Note. Value appears as NULL If the catalog is in LOCKED or SUSPENDED mode. Example 5-1 shows how to define a new ICF catalog. At the same time.CATLOCK profile is not defined. RESUME RESUME The request for the catalog is executed immediately. other operations against the catalog are failed or queued.RLSTST. and RLSENABLE enables OPEN in RLS mode. SUSPEND.

This is a total of 67.578 Total in bytes and MB 511.000 bytes.108. and maintenance. Important: It is suggested that all systems in the sysplex have migrated to DFSMS V2. An average size chosen could be 2000 bytes (total is then 608. Add up the buffer sizes based on the number of catalogs you have.2. sizing. and add more as needed to meet performance requirements. and monitored as the conversion of the ICF catalogs to RLS access mode progresses. This is because shared catalogs cannot be accessed in RLS mode if they are opened on a pre-DFSMS V2.456 bytes.864 bytes for the VLF-managed catalogs.722 The current buffer need totals over 500 MB.2 Sizing the RLS catalog environment for RLS buffers and CF cache To ensure sufficient RLS local buffering and CF cache size for your RLS-managed ICF catalogs.435.280. Look into the following values: 򐂰 Current catalog data space cache (CDSC) using VLF. For four LPARs. For four systems. BCS records vary in size. including preparation. this is 268. An RLS Lock Set Name must have been defined.715. this adds up to 243. migrating.1 before starting the migration.144 ISC 243.5.456 499.000 bytes. Capture the current MAXVIRT value (in 4 kilobyte (KB) blocks). fallback. The following list summarizes the basic requirements for RLS: 򐂰 򐂰 򐂰 򐂰 򐂰 SMSVSAM must be active. Regard this as a minimum.1 system. Table 5-2 Buffer requirement calculation Component Current usage in bytes Current usage in MB COFVLF 268.2. 78 A Practical Guide to ICF Catalogs . Again. The 10 ICF catalogs using ISC cache currently use (10 x 3041 x 2000) = 60. A CF cache size of 500 MB could then be allocated as a start. you can start migrating your first ICF catalog to use RLS mode. An RLS Cache Set must have been defined.820. Sum up the total for the entire sysplex.384 for the MAXVIRT value (times 4 KB).200 bytes per catalog).456 262.000 237. An RLS cache structure must have been defined in the CF. Calculate 3041 records per catalog. look at the buffer sizes in your current environment. Your COFVLFxx member specifies 16. The catalogs in scope must be SMS-managed. You have a sysplex with four LPARs sharing 50 ICF catalogs (10 ISC-managed and 40 in CDSC).1 Preparing to migrate to an RLS enabled ICF Catalog When you have the base RLS environment in place. calculate the total for the sysplex. Buffer calculation scenario Calculate a sample scenario.280.2 Considerations for implementing RLS The following sections describe RLS implementation considerations. 򐂰 ISC is the default. The requirement for buffering all existing catalogs then totals as shown in Table 5-2. 5.435. 5. and the type of caching.

set RlsFixedPoolsize in IGDSMSxx to 500 MB. use the RlsAboveTheBarMaxPoolSize parameter. follow these steps: 1.RlsFixedPoolsize and RlsAboveTheBarMaxPoolSize parameters To meet the requirement in the previous scenario regarding RLS local buffers. Set this parameter to limit use for the RLS buffer (the default is 0). To limit use of storage above the bar. The CFSizer can be found on the following site: http://www. as shown in Figure 5-1.com/systems/support/z/cfsizer To size a structure.ibm. This value specifies the maximum number of buffers under and above the bar. Coupling Facility Structure Sizer tool The Coupling Facility Structure Sizer Tool (CFSizer) is a web-based application that returns structure sizes based on the latest CFLEVEL for the IBM products that use the CF. Record-level sharing support for ICF catalogs 79 . CFSizer can be used to generate the exact settings and the JCL for creating the Cache Structure. Select VSAM RLS from the product links on the left side navigation bar. Figure 5-1 CFSizer Tool Chapter 5. Open the CFSizer web page.

In the VSAM RLS page. Figure 5-3 CF sizer results for allocating VSAM RLS cache structure for RLS It is suggested that you create a dedicated structure for ICF catalogs. At least one structure must be selected to submit a request for a structure size. This is to ensure that if catalog buffers are flushed from the cache. it is based on the catalog activity. 80 A Practical Guide to ICF Catalogs . and not based on other activity in the CF. We selected the VSAM RLS CACHE structure and used the estimated value of 500 MB.2. Calculated results based on your input are shown in Figure 5-3. You can paste these statements into your CFRM policy and modify them as necessary. The CFSizer presents you with sample coupling facility resource management (CFRM) statements based on the input that you provided. Figure 5-2 VSAM RLS CACHE structure sizing using CFSizer tool 3. Enter the required information and click Submit to continue. as shown in Figure 5-2. select one or more structures to size by clicking the check box for the cache structure that you want to size.

If you want a system-wide scope to change all RLS enabled catalogs in one command across the sysplex. Keywords such as STRNO. The same command syntax is valid for RLSQUIESCE. non-RLS catalogs have the value NULL in the LOG parameter. issue the F CATALOG. you can start using RLS mode on ICF catalogs. Record-level sharing support for ICF catalogs 81 .RLSTST TO STATE RLSENABLE SUCCESSFUL IEC352I CATALOG ADDRESS SPACE MODIFY COMMAND COMPLETED The F CATALOG command in Example 5-4 is directed toward a specific ICF catalog. Example 5-2 DEFINE USERCATALOG example of RLS ready ICF catalog DEFINE USERCATALOG (NAME(UCAT.1. issue the F CATALOG. Example 5-4 F CATALOG RLSENABLE messages returned in SYSLOG F CATALOG. By default.2. No RLS QUIESCE command is needed at this point.SYSTEM command.3 Migrating to RLS enabled ICF catalog After upgrading all of the systems in the sysplex to DFSMS V2.RLSENABLE command.RLSTST) ICFCATALOG VOLUME(MHL1A0) TRK(5 1) STORCLAS(SCRLS) DATACLAS(WELCHRLS) LOG(NONE) FREESPACE(20 20) NOIMBED NOREPLICATE) DATA (CISZ(4096)) A LISTCAT following immediately after the DEFINE shows that the user catalog has the required LOG(NONE) attribute and also is VSAM QUIESCED already. Example 5-4 shows SYSLOG messages issued in response to the command. An example of a DEFINE is shown in Example 5-2 (note the LOG(NONE) option).5.RLSENABLE(UCAT. and BUFNI have been removed because they are not used in an RLS environment. The LOG(NONE) attribute puts the catalog in RLS quiesced mode. BUFND. See Example 5-3.RLSTST) IEC351I CATALOG ADDRESS SPACE MODIFY COMMAND ACTIVE IEC352I MODIFY CATALOG UCAT. This is done by either defining a completely new ICF catalog for test purposes. or by altering an existing ICF catalog to have the LOG(NONE) attribute. Chapter 5.RLSENABLE. Example 5-3 Extract from LISTCAT of newly defined ICF catalog with LOG(NONE) attribute LOG(NONE) RLSDATA LOG ------------------NONE RECOVERY REQUIRED --(NO) VSAM QUIESCED ------(YES) RLS IN USE ---------(NO) To make the ICF catalog that we defined fully RLS-enabled.

and a shared RLS candidate catalog is open in this system. RLS enablement is also indicated by RLS IN USE field. This means that the catalog is no longer quiesced. E-ECS SHARED. C-CLOSED. K-LOCKED. This command lists all open catalogs. D-DELETED.RLSTST) IEC351I CATALOG ADDRESS SPACE MODIFY COMMAND ACTIVE IEC353I SERVICE 'CLRC' RETURNED RC-236-030. as shown in an extract from a LISTCAT of a user catalog in an RLS enabled state. and displays settings for the open catalogs in this environment. Example 5-6 F CATALOG. V-VLF.A LISTCAT of the user catalog now shows the VSAM QUIESCED field changed from YES to NO. * R-SHARED. Example 5-7 shows the error messages displayed when trying to RLS enable a catalog shared with a lower-level system. S-SMS. I-ISC. W-SUSPENDED *CAS*************************************************************** IEC352I CATALOG ADDRESS SPACE MODIFY COMMAND COMPLETED Note the U indicator for an RLS-enabled catalog. you get an error message when trying to enable this catalog.RLSTST TO STATE RLSENABLE UNSUCCESSFUL IEC353I CATALOG ADDRESS SPACE MODIFY UNSUCCESSFUL You get a similar error message when trying to do the RLS enable using an IDCAMS ALTER command. UCAT. Example 5-5 Listcat example of ICF catalog in RLS enabled state. 82 A Practical Guide to ICF Catalogs .OPEN command.RLSTST ******************************************************************* * Y/N-ALLOCATED TO CAS.SBXHS8 0001 UCAT. See Example 5-5.RLSENABLE(UCAT. Example 5-6 shows a display of an RLS enabled ICF Catalog using the F CATALOG.OPEN command output listing RLS state -F CATALOG.OPEN IEC351I CATALOG ADDRESS SPACE MODIFY COMMAND ACTIVE IEC348I ALLOCATED CATALOGS *CAS*************************************************************** * FLAGS -VOLSER-USER-CATALOG NAME * YSU-R. which has changed to YES. but enabled for RLS. U-RLS. A-ATL. QUIESOFF. If you have a lower-level system in your sysplex. Example 5-7 Error message trying to RLSENABLE an ICF catalog shared with lower-level system F CATALOG.OPEN command. RLSDATA LOG ------------------NONE RECOVERY REQUIRED --(NO) VSAM QUIESCED -------(NO) RLS IN USE --------(YES) Another way of displaying the RLS readiness of an ICF catalog is to use the F CATALOG.RLSTST IEC353I MODIFY CATALOG UCAT.

MAXIMUM CONDITION CODE WAS 8 5.0 DFSMSdfp Storage Administration. it is also suggested that you have a fallback plan in place.REASON CODE IS IGG0CLRC-30 IDC0532I **ENTRY UCAT. but earlier level systems need SMSVSAM active to switch catalogs out of RLS mode. the process to disable the use of RLS. 5. For more information. Chapter 5. and the data set will have consistent read integrity. and of course for other users. 򐂰 YES indicates that the source and target data sets will be opened using RLS. There are new IDCAMS keywords to manage RLS source or target catalogs: 򐂰 RLSSOURCE(NO|YES|QUIESCE) 򐂰 RLSTARGET(NO|YES|QUIESCE) Using these parameters influences processing as follows: 򐂰 NO indicates that the source and target data sets will be opened using NSR. and catalog management reverts to normal catalog and VSAM I/O. Record-level sharing support for ICF catalogs 83 . Backing up your RLS managed ICF catalogs Backing up your RLS-managed user catalogs with IDCAMS can be done as in prior releases.Example 5-8 shows error messages displayed with the IDCAMS ALTER RLSENABLE on a shared catalog. we look at catalog maintenance in an RLS environment.RLSTST RLSENABLE IDC3014I CATALOG ERROR IDC3009I ** VSAM CATALOG RETURN CODE IS 236 . 򐂰 QUIESCE indicates that the source and target data sets will be opened using RLS. Use the F CATALOG. SMSVSAM is the server for all RLS processing.RLSTST NOT ALTERED IDC0001I FUNCTION COMPLETED. is straightforward.2.5 Catalog maintenance considerations in RLS environment New keywords have been introduced in support of RLS enablement for catalogs. SC23-6860. and the data set will be quiesced before processing any entries. and RLS processing fails if SMSVSAM for some reason is not available.RLSQUIESCE(ucat name) command or the F CATALOG. see z/OS V2R1.RLSQUIESCE. Example 5-8 IDCAMS RLSENABLE command failure due to sharing with lower-level system IDCAMS SYSTEM SERVICES TIME: 17:37:27 ALTER UCAT.SYSTEM command. New catalog commands support fallback.4 Fallback from using RLS mode on ICF Catalogs SMSVSAM recovery is essential for ICF catalogs using RLS. For ICF catalogs using RLS. Before enabling RLS use on ICF catalogs. which has a detailed description about recovery of SMSVSAM and the RLS environment. In this section. if SMSVSAM and the RLS environment encounter problems.2. HIGHEST CONDITION CODE WAS 8 IDC0002I IDCAMS PROCESSING COMPLETE.

DFSMShsm backs up aliases.RLSTST1.. you can choose whether to access the catalog using RLS... to control the integrity of that catalog.. and is able to back up catalogs... RLS access is QUIESCED during backup. Remember to LOCK or SUSPEND the catalog before performing the restore.RLSTST )) OUTDDNAME (TAPE) /* DFSMS Hierarchical Storage Manager backup DFSMS Hierarchical Storage Manager (DFSMShsm) is also able to back up catalogs.BACK2...CATLOCK in order to be able to work with a locked catalog. and enables RLS when the backup is finished. Example 5-10 Example of using DFSMSdss as the backup tool for ICF catalog //MHLRES1D JOB .SPACE=(TRK..RLSTST OUTFILE(DD1) TEMPORARY RLSSOURCE(QUIESCE) DFSMS Data Set Services backup When using DFSMS Data Set Services (DFSMSdss) for backing up RLS-managed catalogs.RLSE) //SYSOUT DD SYSOUT=* //SYSIN DD * EXPORT UCAT.CATLG)..REGION=4096K //SYSPRINT DD SYSOUT=* //TAPE DD DISP=(NEW.....2.DCB=(BLKSIZE=27920) //SYSIN DD * DUMP DS(INC(UCAT. DFSMShsm is a privileged user.IDCAMS backup Using IDCAMS for backing up a catalog.. but also suspends availability during backup..(2... Example 5-9 shows a sample IDCAMS backup of an ICF catalog using RLSSOURCE(QUIESCE)..... Also see an example of DFSMSdss backing up an ICF catalog in Example 5-10. and will be resumed when IDCAMS finishes backup... // SPACE=(TRK. // UNIT=3390. and also restores them when restoring the catalog.MHLRES1A.. // DSN=MHLRES1.1))... DFSMSdss performs the QUIESCE implicitly during backup.6 Using REPRO MERGECAT on an RLS-managed catalog Running the REPRO MERGECAT function in a shared environment from a non-RLS-enabled ICF catalog to an RLS-enabled ICF catalog requires the QUIESCE done on the target catalog before you are allowed to issue a REPRO MERGECAT.(2. // DSN=MHLRES1.. 5.1)..DSS.CATLG). Using this parameter ensures integrity... You need to ensure that DFSMShsm has access to IGG..BACKUP1. Otherwise. there is no specific keyword for QUIESCE. //STEPT006 EXEC PGM=ADRDSSU. When QUIESCE is specified... 84 A Practical Guide to ICF Catalogs . Example 5-9 IDCAMS backup of ICF catalog using the RLSSOURCE parameter //CAT004 EXEC PGM=IDCAMS //SYSPRINT DD SYSOUT=* //DD1 DD DISP=(.

096 MSEC * * SPHERE DEQ 5. or through standard z/OS tools.627 MSEC * * BCS ENQ Shr Sys 889.020 MSEC * * VVDS RESERVE Shr 1.015 MSEC * * CAXWA DEQ 47.365 MSEC * * BCS DEQ 960.505 0.VSBOX01 ) OUTDATASET(UCAT.035 MSEC * * VVDS DEQ CI 222. or since the last reset of these statistics through the F CATALOG. Example 5-12 Extract of ICF catalog performance report -F CATALOG.947 0.187 1..REPORT.947 0.79 on 11/01/2014 * * -----CATALOG EVENT-----COUNT-.368 0.015 0.PERFORMANCE(RESET) command. which is extensive..505 0. In the following section.001 MSEC * * VDSPM DEQ 865. It is just a snapshot from the beginning of the report. Reporting on both non-RLS-managed and RLS-managed catalogs is done through F CATALOG commands.002 MSEC * * VDSPM ENQ 865.457K 0.---AVERAGE--.MHLRES1A IDC3351I ** VSAM OPEN RETURN CODE IS 168 IDC3003I FUNCTION TERMINATED. This is not the entire report. Note the Statistics since time stamp that shows the starting time for this summary report.MHLRES1A ) LEVEL(MHLRES1A) MERGECAT IDC3300I ERROR OPENING UCAT.538 0. The REPORT PERFORMANCE command should be issued regularly to identify potential performance bottlenecks.014 MSEC * * CAXWA ENQ Shr 47..001 MSEC * . such as IBM z/OS Resource Measurement Facility (IBM RMF™).951 0.Attempting to issue the REPRO MERGECAT without having the target catalog quiesced for RLS access ends up giving the error message shown in Example 5-11. there are several reporting options. Record-level sharing support for ICF catalogs 85 .443K 0.* * Entries to Catalog 598. An example of the output from this report is shown in Example 5-12. we look into the reporting options.. Using the F CATALOG.REPORT.301 MSEC * * VVDS DEQ 1.PERFORMANCE command returns the I/O statistics of the catalog address space since the last IPL.PERFORMANCE IEC351I CATALOG ADDRESS SPACE MODIFY COMMAND ACTIVE IEC359I CATALOG PERFORMANCE REPORT *CAS*************************************************** * Statistics since 11:05:51. CONDITION CODE IS 12 Reporting on catalogs To track your catalog efficiency.312 0.106 MSEC * * BCS ENQ Excl Sys 8.RLSTST.951 0. Chapter 5.016 MSEC * * VVDS RESERVE CI 222. Example 5-11 Example of REPRO mergecat from non RLS managed ICF catalog to RLS managed REPRO INDATASET(UCAT. because your catalogs are a key component in I/O activity.312 0.REPORT.257 0.036 MSEC * * SPHERE ENQ Excl Sys 5.continues .048 MSEC * * VVDS RESERVE Excl 17. This is the case for both non-RLS-managed and RLS-managed catalogs.RLSTST.

Example 5-14 Example of report from F CATALOG.. and should be at least 20% to reach a reasonable value..ZOSR1A (ISC) * * 94% 1 17 16 0 0 0 * * UCAT. Measuring for an hour might not represent your normal catalog access. continues .REPORT. and STRNO information.REPORT.VTFM (ISC) * * 99% 6 335. You can reset the statistics by issuing the F CATALOG.-SHR UPD.. The F CATALOG. about a specific user catalog.--PURGE-.. BUFNI.CLOSE command.REPORT. Through this.. This command can report on one or more catalogs with the use of asterisks (*) or wildcards in the CATNAME specification.VSTEST1 * * 0 0 18 0 0 4 4 2 * . all counters have been reset except for the purge values.CACHE command -F CATALOG.REPORT.CACHE IEC351I CATALOG ADDRESS SPACE MODIFY COMMAND ACTIVE IEC359I CATALOG CACHE REPORT *CAS****************************************************************** * HIT% -RECORDS.-DELETES.Cache data space efficiency The F CATALOG.CATSTATS(catname) gives you I/O statistics.REPORT.CATSTATS..CATSTATS command -F CATALOG.CATSTATS IEC351I CATALOG ADDRESS SPACE MODIFY COMMAND ACTIVE IEC359I CATALOG I/O STATS REPORT *CAS**************************************************************** * ADDS UPDATES GETS GETUPD DELETES BUFNI BUFND STRNO * * * * UCAT. The new CATSTATX parameter on the F CATALOG.699 0 0 0 * * UCAT.VTFM * * 0 0 23 0 0 10 16 8 * * UCAT.VSTEST1 (ISC) * * 94% 1 17 16 0 0 0 * * UCAT.VSBOX11 (ISC) * * 98% 3 220 217 0 0 0 * .CACHE command reports on the cache data space efficiency. Example 5-13 Sample cache report based on F CATALOG. a new cache structure is built.-SEARCHES --FOUND-.* * * * UCAT.ZOSR1A * * 0 0 18 0 0 10 16 8 * * UCAT. The hit rate is reported. See Example 5-14 for an extract of this report.REPORT.VTFMTAPE * * 57 181 752 181 0 10 16 8 * * UCAT. including BUFND.. The statistics must be measured over time to be valid.705 335. as specified in the CATNAME parameter. 86 A Practical Guide to ICF Catalogs .REPORT. When the catalog is reopened after this. A sample extract of the cache report is shown in Example 5-13.CATSTATX(catname) command reports the same information as F CATALOG. continues . but it includes CA-RECLAIMS and CA-REUSES information as well.

This report is reporting one ICF catalog at a time, as shown in Example 5-15.
Example 5-15 Example of F CATALOG,REPORT,CATSTATX report output

-F CATALOG,REPORT,CATSTATX(UCAT.RLSTST)
IEC351I CATALOG ADDRESS SPACE MODIFY COMMAND ACTIVE
IEC359I EXTENDED CATALOG STATS
*CAS*****************************************************************
* CATALOG NAME
= UCAT.RLSTST
*
* INSERTS (ADDS)
=
0
*
* UPDATES
=
0
*
* RETRIEVES
=
18
*
* RETRIEVES FOR UP =
0
*
* ERASES (DELETES) =
0
*
* CA-RECLAIMS
=
0
*
* CA-REUSES
=
0
*
* BUFNI SETTING
=
4
*
* BUFND SETTING
=
10
*
* STRNO SETTING
=
2
*
* AVG ELAPSED TIME =
5.000 MSEC
*
* AVG CPU TIME
=
842.000 USEC
*
*CAS*****************************************************************
The information about CA-RECLAIMS, AVG ELAPSED TIME, and AVG CPU TIME has been added in
z/OS V2.1.
RMFMON III has RLS activity reports. Choose the SYSPLEX option, and there are several
reports on RLS. The RMFMON III Sysplex panel with these reports is displayed in
Example 5-16.

RMF3 reports on RLF activity
Choose RMF Sysplex Report. Three RLS reports are available in this menu.
Example 5-16 Available VSAM RLS reports in RMF3

1
2
3
4
5
6
7
8
9
10
11
12

SYSSUM
SYSRTD
SYSWKM
SYSENQ
CFOVER
CFSYS
CFACT
CACHSUM
CACHDET
RLSSC
RLSDS
RLSLRU

Sysplex performance summary
Response time distribution
Work Manager delays
Sysplex-wide Enqueue delays
Coupling Facility overview
Coupling Facility systems
Coupling Facility activity
Cache summary
Cache detail
VSAM RLS activity by storage class
VSAM RLS activity by data set
VSAM LRU overview

(SUM)
(RTD)
(WKM)
(ES)
(CO)
(CS)
(CA)
(CAS)
(CAD)
(RLS)
(RLD)
(RLL)

Coupling facility lock structure report
The D SMS,CFLS command displays information about the CF lock structure. This information
includes the lock rate, lock contention rate, false contention rate, average number of requests
waiting for locks, the lock structure size, and primary structure information.
If the lock structure is in duplex mode, secondary structure information is also displayed.

Chapter 5. Record-level sharing support for ICF catalogs

87

Output from the D SMS,CFLS command is shown in Example 5-17.
Example 5-17 Coupling facility lock structure report.

D SMS,CFLS
IEE932I 853
IGW320I 07:42:24 Display SMS,CFLS(IGWLOCK00 )
PRIMARY STRUCTURE:IGWLOCK00 VERSION:CBE13E285CE22FD3 SIZE:14336K
RECORD TABLE ENTRIES:27365 USED:6
LOCK STRUCTURE MODE: SIMPLEX STATUS: ENABLE
System Interval LockRate ContRate FContRate WaitQLen
SC64 1 Minute 0.0 0.000 0.000 0.00
SC64 1 Hour 0.0 0.000 0.000 0.00
SC64 8 Hour 0.0 0.000 0.000 0.00
SC64 1 Day 0.0 0.000 0.000 0.00
(04) 1 Minute 0.0 0.000 0.000 0.00
(04) 1 Hour 0.0 1.923 0.000 0.00
(04) 8 Hour 0.0 5.048 0.000 0.00
(04) 1 Day 0.0 4.968 0.000 0.00
***************** LEGEND ******************
LockRate = number of lock requests per second
CONTRATE = % of lock requests globally managed
FCONTRATE = % of lock requests falsely globally managed
WaitQLen = Average number of requests waiting for locks

SMF reports
SMF is another source of information. You need a reporting tool that can post-process the
SMF data. Choose the option that you have in your organization.
SMF type 42 subtypes 15 through 19 can be used for summary reporting on RLS use. The
following subtypes are available: 򐂰 򐂰 򐂰 򐂰 򐂰

Subtype 15. VSAM RLS Storage Class Response Time Summary.
Subtype 16. VSAM RLS Data Set Response Time Summary.
Subtype 17. VSAM RLS Coupling Facility Lock Structure Usage.
Subtype 18. VSAM RLS CF Cache Partition Usage.
Subtype 19. VSAM RLS Local Buffer Manager least recently used (LRU) statistics
summary.

Use the V SMS,MONDS(spherename),ON command to collect subtype 16 statistics. Additionally,
data set collection for SMF type 42 must be turned on when using RMF. For catalogs and
VSAM in general, all of the catalog-specific SMF records can be used: 򐂰 򐂰 򐂰 򐂰 򐂰 򐂰

88

Type 60. VVR Updated.
Type 61. ICF Define.
Type 62. VSAM OPEN.
Type 64. VSAM CLOSE (new for catalogs in z/OS 1.13 and later).
Type 65. ICF Delete.
Type 66. ICF Alter.

A Practical Guide to ICF Catalogs

5.2.7 Migration considerations
It is necessary for SMSVSAM to be enabled on all systems in the sysplex using RLS mode for
catalogs. This is valid for DFSMS V2.1 systems, as for lower-level systems participating in the
sysplex. It is also required that all systems in the sysplex have migrated to DFSMS V2.1
before implementing RLS mode for shared ICF catalogs.
Even if you turn on toleration support on the earlier systems, a candidate catalog to move to
RLS is only usable in a system earlier than V2.1 in RLS Quiesced mode. It is not usable when
the catalog enters RLS Enabled mode.

5.3 Basic test results
The following sections describe benchmarks and documentation.

5.3.1 Benchmarks for ICF catalogs using RLS mode
Benchmarks comparing VLF-managed and RLS-managed catalogs shows significant
improvements in performance.
One particular benchmark on massive DELETES showed that elapsed time was reduced
90%, and processor usage was reduced over 75%.
On CREATES, similar numbers were 56% and 81% reduction. For GENERIC READS,
however, the savings were 32% in elapse time, but processor usage was over 15% higher.
For DIRECT READS, the corresponding figures are a 24% reduction in elapse time and a
processor consumption that is 14% higher.
We did a small study creating and deleting 300,000 data sets in a non-RLS-managed
environment compared to an RLS-managed environment. An RLS quiesced ICF catalog was
created, and the 300,000 data set creates and deletes were done in eight batch jobs
sequentially. Elapse time and processor time was captured on those.
Next, the ICF catalog was RLS enabled, and the same 300,000 data sets were created and
deleted in the same way, again capturing elapse time and processor time. In the graphical
presentation in Figure 5-4, the improvement on elapse time using RLS-managed catalogs is
obvious in our case. In average, elapsed time on data set creates was reduced by 57.2%.

Chapter 5. Record-level sharing support for ICF catalogs

89

5.8% reduction in elapsed time moving from a non-RLS-managed ICF catalog to an RLS-managed ICF catalog. the comparison between non-RLS-managed and RLS-managed deletes showed a 31.000 data set creates and data set deletes in a non-RLS-managed environment with creates and deletes in an RLS-managed environment only showed a little improvement. see the following documentation: 򐂰 򐂰 򐂰 90 z/OS DFSMS Managing Catalogs.000 data sets. Figure 5-5 Non-RLS-managed deletes versus RLS-managed deletes Comparing the processor consumption on the 300. SC23-6860 A Practical Guide to ICF Catalogs .See Figure 5-4 and Figure 5-5 for the results of our controlled tests.2 Documentation When planning for implementation of RLS for catalogs.3. SC23-6853 z/OS DFSMS Access Method Services Commands. SC23-6846 z/OS DFSMSdfp Storage Administration. as shown in Figure 5-5. Figure 5-4 Non-RLS-managed creates versus RLS-managed creates Deleting the same 300. This was moving from a non-RLS-managed ICF catalog to an RLS-managed ICF catalog in our scenario.

473” 90<->249 pages .A Practical Guide to ICF Catalogs (0.2”spine) 0.17”<->0.

.

.

com/redbooks . ibm.Back cover SG24-8262-00 ISBN 073844068X Printed in U.S.A.