IBM InfoSphere Information Server
Installation and Configuration Guide
March 2011
International Technical Support Organization
REDP-4596-00
© Copyright International Business Machines Corporation 2011. All rights reserved.
Note to U.S. Government Users Restricted Rights -- Use, duplication or disclosure restricted by GSA ADP
Schedule Contract with IBM Corp.
First Edition (March 2011)
This edition applies to Version 8.1 of IBM Information Server.
Note: Before using this information and the product it supports, read the information in
“Notices” on page ix.
© Copyright IBM Corp. 2011. All rights reserved. iii
Contents
Notices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
Trademarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . x
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xi
The team who wrote this paper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xi
Now you can become a published author, too! . . . . . . . . . . . . . . . . . . . . . . . . xii
Comments welcome. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xii
Stay connected to IBM Redbooks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii
Chapter 1. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 IBM services offerings. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Platform specification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2.1 Client: Windows platform specification . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2.2 Server: Windows platform specification . . . . . . . . . . . . . . . . . . . . . . . 3
1.2.3 Server: UNIX and Linux platform specification . . . . . . . . . . . . . . . . . . 5
1.2.4 Server: z/OS platform specification. . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Chapter 2. Platform specifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.1 Client: Windows platform specification . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.1.1 Suggested 32-bit version operating systems . . . . . . . . . . . . . . . . . . 12
2.1.2 Suggested web browsers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.1.3 Required assets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.1.4 Suggested memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.2 Server: Windows platform specification . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.2.1 Suggested memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.2.2 Required disk space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.2.3 C++ compiler. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.2.4 Embedded MKS OEM. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.3 Server: UNIX/Linux Platform specification . . . . . . . . . . . . . . . . . . . . . . . . 14
2.3.1 Suggested memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.3.2 Required disk space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.3.3 C++ compiler. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.4 Server: z/OS platform specification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.4.1 Required memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.4.2 Required disk space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.4.3 C++ compiler. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
Chapter 3. Capacity planning. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.1 Minimums . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
iv IBM InfoSphere Information Server Installation and Configuration Guide
3.2 Processor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.3 Memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.4 Swap space. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.5 Disk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.5.1 Staging disk for input and output files . . . . . . . . . . . . . . . . . . . . . . . . 23
3.5.2 Scratch/sort work areas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.5.3 Resource areas for parallel data sets . . . . . . . . . . . . . . . . . . . . . . . . 24
Chapter 4. Installation and configuration . . . . . . . . . . . . . . . . . . . . . . . . . . 25
4.1 Pre-installation overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
4.2 Pre-installation checklist . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
4.3 Reviewing release notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
4.3.1 IBM InfoSphere Information Server release notes . . . . . . . . . . . . . . 28
4.3.2 WebSphere Application Server release notes . . . . . . . . . . . . . . . . . 28
4.4 Planning, installation, and configuration . . . . . . . . . . . . . . . . . . . . . . . . . . 29
4.5 Reviewing migrating to InfoSphere Information Server. . . . . . . . . . . . . . . 29
4.6 Choosing and validating the architecture or topology . . . . . . . . . . . . . . . . 30
4.6.1 Two-tier deployment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
4.6.2 Three-tier deployment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
4.6.3 Four-tier deployment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
4.6.4 Cluster and grid deployments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
4.6.5 Wide area network deployments. . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
4.7 Validating system requirements for all tiers. . . . . . . . . . . . . . . . . . . . . . . . 34
4.8 Verifying domain requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
4.9 Verifying database requirements for metadata repository. . . . . . . . . . . . . 35
4.10 Verifying database requirements for Information Analyzer analysis . . . . 36
4.11 Verifying and configure disks, volume groups, file systems . . . . . . . . . . 37
4.11.1 RAID or SAN configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
4.11.2 InfoSphere Information Server file systems . . . . . . . . . . . . . . . . . . 38
4.11.3 Software installation directories . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
4.11.4 Database storage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
4.11.5 InfoSphere Information Server Project directories . . . . . . . . . . . . . 41
4.11.6 Dataset and Scratch directories . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
4.11.7 Extending the DataStage project for external entities . . . . . . . . . . . 43
4.11.8 File staging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
4.12 Verifying and configuring OS and resource limits . . . . . . . . . . . . . . . . . . 44
4.12.1 UNIX kernel parameters for all platforms . . . . . . . . . . . . . . . . . . . . 45
4.12.2 UNIX user (shell) parameters for all platforms . . . . . . . . . . . . . . . . 45
4.12.3 AIX system configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
4.12.4 HP-UX system configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4.12.5 RedHat and SUSE Linux system configuration. . . . . . . . . . . . . . . . 50
4.12.6 Solaris 9 system configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
4.12.7 Solaris 10 system configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
Contents v
4.13 Verifying connectivity and network configuration . . . . . . . . . . . . . . . . . . 54
4.14 Configuring operating system users, groups, and permissions. . . . . . . . 56
4.14.1 Privileged installation user . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
4.14.2 Required operating system users . . . . . . . . . . . . . . . . . . . . . . . . . . 56
4.14.3 Domain (WebSphere Application Server) user registry . . . . . . . . . 57
4.14.4 Engine (DataStage) user setup. . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
4.14.5 Engine (DataStage) user setup on Windows . . . . . . . . . . . . . . . . . 59
4.15 Verifying and installing C++ compiler and runtime libraries . . . . . . . . . . 59
4.16 Verifying InfoSphere Information Server connector requirements . . . . . 60
4.17 Downloading and installing InfoSphere Information Server . . . . . . . . . . 60
4.18 Performing complete system backup . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
4.19 Identifying and configuring file systems. . . . . . . . . . . . . . . . . . . . . . . . . . 61
4.19.1 Software installation directory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
4.19.2 DataStage Projects (repository) directory . . . . . . . . . . . . . . . . . . . . 63
4.19.3 Data set and sort directories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
4.19.4 Extending the DataStage project for external entities . . . . . . . . . . . 69
4.19.5 File staging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
4.19.6 File system sizing example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
4.20 Connectivity and network configuration. . . . . . . . . . . . . . . . . . . . . . . . . . 77
4.20.1 Network port usage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
4.20.2 UNIX NIS configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
4.20.3 Windows network configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
4.21 Configuring OS users, groups, and associated permissions. . . . . . . . . . 79
4.21.1 UNIX user configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
4.21.2 Windows user configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
4.22 C++ compiler and runtime library requirements . . . . . . . . . . . . . . . . . . . 82
4.22.1 Development systems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
4.22.2 Deployment systems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
4.23 Checking product release notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
4.24 Installing DataStage/Parallel Framework . . . . . . . . . . . . . . . . . . . . . . . . 85
4.24.1 Installing multiple DataStage Servers on UNIX. . . . . . . . . . . . . . . . 85
4.24.2 Installing plug-ins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
4.24.3 UNIX install requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
4.24.4 Windows installation requirements . . . . . . . . . . . . . . . . . . . . . . . . . 87
4.25 Verifying the installation log file . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
4.26 Installing DataStage patches. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
4.27 Installing and configuring optional components . . . . . . . . . . . . . . . . . . . 89
4.28 Configuring post-installation operating system settings . . . . . . . . . . . . . 89
4.28.1 Securing JobMon ports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
4.28.2 Post-installation configuration of Windows 2003 Server . . . . . . . . . 89
4.28.3 UNIX cluster configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
4.28.4 Windows cluster configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
4.29 Configuring the DataStage environment and default settings . . . . . . . . . 99
vi IBM InfoSphere Information Server Installation and Configuration Guide
4.29.1 Setting the DataStage environment . . . . . . . . . . . . . . . . . . . . . . . . 99
4.29.2 Altering the DataStage dsenv on UNIX. . . . . . . . . . . . . . . . . . . . . 100
4.29.3 Suggested default settings for all projects . . . . . . . . . . . . . . . . . . 101
4.30 Configuring the DataStage administrator environment . . . . . . . . . . . . . 102
4.30.1 Setting the UNIX and LINUX administrator environments . . . . . . 102
4.30.2 Setting the Windows 2003 environment . . . . . . . . . . . . . . . . . . . . 102
4.31 Configuring and verifying database connectivity . . . . . . . . . . . . . . . . . . 103
4.31.1 DB2 configuration for Enterprise stage . . . . . . . . . . . . . . . . . . . . . 103
4.31.2 Informix configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
4.31.3 Oracle configuration for Enterprise stage or connector. . . . . . . . . 106
4.31.4 Sybase configuration. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
4.31.5 Teradata configuration for Enterprise Stage . . . . . . . . . . . . . . . . . 110
4.31.6 Netezza connectivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
4.32 Configuring and verifying ODBC connectivity . . . . . . . . . . . . . . . . . . . . 114
4.32.1 Configuring ODBC access on UNIX . . . . . . . . . . . . . . . . . . . . . . . 115
4.32.2 Setting up DSNs on UNIX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
4.32.3 Configuring ODBC access on Windows 2003 Server . . . . . . . . . . 116
4.32.4 ODBC readme notes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
4.33 Creating and verifying project location . . . . . . . . . . . . . . . . . . . . . . . . . 116
4.34 Verifying project security settings and roles . . . . . . . . . . . . . . . . . . . . . 118
4.35 Configuring and verifying client installations . . . . . . . . . . . . . . . . . . . . . 118
4.35.1 DataStage Multi-Client Manager . . . . . . . . . . . . . . . . . . . . . . . . . . 119
4.35.2 WAN development considerations . . . . . . . . . . . . . . . . . . . . . . . . 120
4.35.3 Secure client installation considerations . . . . . . . . . . . . . . . . . . . . 120
4.35.4 Enterprise Application PACKs. . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
Chapter 5. Parallel configuration files. . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
Appendix A. Configurations and checklists . . . . . . . . . . . . . . . . . . . . . . . 127
Installation and configuration checklist. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
DataStage administrator UNIX environment . . . . . . . . . . . . . . . . . . . . . . . . . 129
Installing and configuring multiple server instances . . . . . . . . . . . . . . . . . . . . 130
Configuring remote DB2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
Setting up DB2 connectivity for remote servers . . . . . . . . . . . . . . . . . . . . 134
Configuring multiple DB2 instances in one job . . . . . . . . . . . . . . . . . . . . . 140
Troubleshooting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
Performance notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
Summary of settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
Increasing DataStage Server Edition memory on AIX . . . . . . . . . . . . . . . . . . 143
Using HP-UX 11 memory on Windows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
Estimating the size of a parallel data set . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
Windows XP Service Pack 2 firewall configuration . . . . . . . . . . . . . . . . . . . . 147
DataStage ports used in Windows platforms . . . . . . . . . . . . . . . . . . . . . . . . . 152
Contents vii
Pre-installation checklist. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
Installation and configuration checklist. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
InfoSphere Information Server installation settings . . . . . . . . . . . . . . . . . . . . 158
Online documentation and link summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
Network ports used by InfoSphere Information Server . . . . . . . . . . . . . . . . . 162
Glossary of terminology and abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . 164
Example user setup for UNIX environments . . . . . . . . . . . . . . . . . . . . . . . . . 164
viii IBM InfoSphere Information Server Installation and Configuration Guide
© Copyright IBM Corp. 2011. All rights reserved. ix
Notices
This information was developed for products and services offered in the U.S.A.
IBM may not offer the products, services, or features discussed in this document in other countries. Consult
your local IBM representative for information about the products and services currently available in your
area. Any reference to an IBM product, program, or service is not intended to state or imply that only that
IBM product, program, or service may be used. Any functionally equivalent product, program, or service that
does not infringe any IBM intellectual property right may be used instead. However, it is the user's
responsibility to evaluate and verify the operation of any non-IBM product, program, or service.
IBM may have patents or pending patent applications covering subject matter described in this document.
The furnishing of this document does not give you any license to these patents. You can send license
inquiries, in writing, to:
IBM Director of Licensing, IBM Corporation, North Castle Drive, Armonk, NY 10504-1785 U.S.A.
The following paragraph does not apply to the United Kingdom or any other country where such
provisions are inconsistent with local law: INTERNATIONAL BUSINESS MACHINES CORPORATION
PROVIDES THIS PUBLICATION "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESS OR
IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF NON-INFRINGEMENT,
MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Some states do not allow disclaimer
of express or implied warranties in certain transactions, therefore, this statement may not apply to you.
This information could include technical inaccuracies or typographical errors. Changes are periodically made
to the information herein; these changes will be incorporated in new editions of the publication. IBM may
make improvements and/or changes in the product(s) and/or the program(s) described in this publication at
any time without notice.
Any references in this information to non-IBM Web sites are provided for convenience only and do not in any
manner serve as an endorsement of those Web sites. The materials at those Web sites are not part of the
materials for this IBM product and use of those Web sites is at your own risk.
IBM may use or distribute any of the information you supply in any way it believes appropriate without
incurring any obligation to you.
Information concerning non-IBM products was obtained from the suppliers of those products, their published
announcements or other publicly available sources. IBM has not tested those products and cannot confirm
the accuracy of performance, compatibility or any other claims related to non-IBM products. Questions on
the capabilities of non-IBM products should be addressed to the suppliers of those products.
This information contains examples of data and reports used in daily business operations. To illustrate them
as completely as possible, the examples include the names of individuals, companies, brands, and products.
All of these names are fictitious and any similarity to the names and addresses used by an actual business
enterprise is entirely coincidental.
COPYRIGHT LICENSE:
This information contains sample application programs in source language, which illustrate programming
techniques on various operating platforms. You may copy, modify, and distribute these sample programs in
any form without payment to IBM, for the purposes of developing, using, marketing or distributing application
programs conforming to the application programming interface for the operating platform for which the
sample programs are written. These examples have not been thoroughly tested under all conditions. IBM,
therefore, cannot guarantee or imply reliability, serviceability, or function of these programs.
x IBM InfoSphere Information Server Installation and Configuration Guide
Trademarks
IBM, the IBM logo, and ibm.com are trademarks or registered trademarks of International Business
Machines Corporation in the United States, other countries, or both. These and other IBM trademarked
terms are marked on their first occurrence in this information with the appropriate symbol (® or ™),
indicating US registered or common law trademarks owned by IBM at the time this information was
published. Such trademarks may also be registered or common law trademarks in other countries. A current
list of IBM trademarks is available on the Web at http://www.ibm.com/legal/copytrade.shtml
The following terms are trademarks of the International Business Machines Corporation in the United States,
other countries, or both:
AFS™
AIX®
DataStage®
DB2®
IBM®
Informix®
InfoSphere®
MVS™
Orchestrate®
Passport Advantage®
QualityStage®
RACF®
Redbooks®
Redpaper™
Redbooks (logo) ®
System z®
VisualAge®
WebSphere®
z/OS®
The following terms are trademarks of other companies:
Netezza Performance Server, Netezza, NPS, and N logo are trademarks or registered trademarks of IBM
International Group B.V., an IBM Company.
Intel, Itanium, Intel logo, Intel Inside logo, and Intel Centrino logo are trademarks or registered trademarks of
Intel Corporation or its subsidiaries in the United States and other countries.
Linux is a trademark of Linus Torvalds in the United States, other countries, or both.
Microsoft, Windows, and the Windows logo are trademarks of Microsoft Corporation in the United States,
other countries, or both.
Java, and all Java-based trademarks and logos are trademarks or registered trademarks of Oracle and/or its
affiliates.
UNIX is a registered trademark of The Open Group in the United States and other countries.
Other company, product, or service names may be trademarks or service marks of others.
© Copyright IBM Corp. 2011. All rights reserved. xi
Preface
This IBM® Redpaper™ publication provides suggestions, hints and tips,
directions, installation steps, checklists of prerequisites, and configuration
information collected from several IBM InfoSphere® Information Server experts.
It is intended to minimize the time required to successfully install and configure
the InfoSphere Information Server.
The information in this document is based on field experiences of experts who
have implemented InfoSphere Information Server. In certain cases, the
suggestions documented here might differ from the product documentation.
However, except where noted, this document is intended to supplement, and not
replace, the product documentation and readme files.
The primary audience for this document is administrators who have been trained
on InfoSphere Information Server. The information in some sections might also
be relevant for technical architects, system administrators, and developers.
The team who wrote this paper
This paper was produced by the following author, along with contributions from
several of his colleagues.
Patrick (Danny) Owen has been a Field Engineer with the Center Of Excellence
for Information Server since 2003. He specializes in complex installs, grid, high
availability, and performance for complex and advanced needs customers. Danny
has published in the field of computer science on topics such as optical character
recognition and algorithms for mapping water vapor on the moon. He graduated
from the University of Arkansas at Little Rock with a Bachelor of Science degree
in computer science.
Thanks to the following people who contributed to the development and
publication of this paper:
Chuck Ballard, Project Manager
Mary Comianos, Publications Management
Emma Jacobs, Graphics
IBM San Jose, CA
xii IBM InfoSphere Information Server Installation and Configuration Guide
Now you can become a published author, too!
Here’s an opportunity to spotlight your skills, grow your career, and become a
published author—all at the same time! Join an ITSO residency project and help
write a book in your area of expertise, while honing your experience using
leading-edge technologies. Your efforts will help to increase product acceptance
and customer satisfaction, as you expand your network of technical contacts and
relationships. Residencies run from two to six weeks in length, and you can
participate either in person or as a remote resident working from your home
base.
Find out more about the residency program, browse the residency index, and
apply online at:
ibm.com/redbooks/residencies.html
Comments welcome
Your comments are important to us!
We want our papers to be as helpful as possible. Send us your comments about
this paper or other IBM Redbooks® publications in one of the following ways:
Use the online Contact us review Redbooks form found at:
ibm.com/redbooks
Send your comments in an email to:
Mail your comments to:
IBM Corporation, International Technical Support Organization
Dept. HYTD Mail Station P099
2455 South Road
Poughkeepsie, NY 12601-5400
Preface xiii
Stay connected to IBM Redbooks
Find us on Facebook:
http://www.facebook.com/IBMRedbooks
Follow us on Twitter:
http://twitter.com/ibmredbooks
Look for us on LinkedIn:
http://www.linkedin.com/groups?home=&gid=2130806
Explore new Redbooks publications, residencies, and workshops with the
IBM Redbooks weekly newsletter:
https://www.redbooks.ibm.com/Redbooks.nsf/subscribe?OpenForm
Stay current on recent Redbooks publications with RSS Feeds:
http://www.redbooks.ibm.com/rss.html
xiv IBM InfoSphere Information Server Installation and Configuration Guide
© Copyright IBM Corp. 2011. All rights reserved. 1
Chapter 1. Introduction
This IBM Redbooks publication highlights various IBM InfoSphere Information
Server installation topologies, platform-specific requirements, prerequisites,
checklists, hardware resource suggestions, hardware configurations, I/O
configuration suggestions, user and group management, InfoSphere Information
Server Architecture, and post-install configurations. Our purpose is to enable an
administrator, with some experience with InfoSphere Information Server, to
choose a deployment topology and IS configuration to meet the needs of their
enterprise, and to be successful in deploying InfoSphere Information Server.
Towards that end, we strongly suggest that you read this document in its entirety
before making final choices. This will give the best foundation for ensuring that
the choices are made with as broad an understanding and informed
consideration as possible.
The primary audience for this document is administrators who have been trained
on InfoSphere Information Server. The information in some sections might also
be relevant for technical architects, system administrators, and developers.
The following sections of this chapter start by describing the platforms,
environments, and specifications for the IBM InfoSphere Information Server.
1
2 IBM InfoSphere Information Server Installation and Configuration Guide
1.1 IBM services offerings
IBM Information Management provides a broad set of services offerings
designed to maximize success with the Information Management product suite
through standard practices that have been developed across numerous
successful deployments. Intended to establish a foundation of product
knowledge and guidelines, these strategic workshops are tailored to the
customer's existing environments, standards, and methodologies.
Within the overall project life cycle, a comprehensive set of services offerings is
available using the entire Information Management suite. Complete details on
each offering are available through IBM Information Management Services:
http://www.ibm.com/software/data/
1.2 Platform specification
The information in this section provides guidelines based on field experiences. In
certain cases the suggestions provided here are not the same as those in the
installation documentation. The installation documentation typically deals with
minimum requirements, whereas the numbers contained below are based on
best practice experience and will typically yield more satisfactory performance
than configurations based on the minimum requirements. All platforms have
additional considerations that are documented in the DataStage Install and
Upgrade Guide and readme files.
1.2.1 Client: Windows platform specification
The IBM InfoSphere DataStage® client is tightly coupled to the DataStage
server. Despite what might be indicated in the release notes, the DataStage
client and server versions should always match unless you have been directed
otherwise by support.
By installing the DataStage Multi-Client Manager on the client workstation, you
can maintain multiple DataStage client versions on a single machine.
Chapter 1. Introduction 3
Suggested 32-bit operating system versions
The following 32-bit operating system versions are suggested:
Windows XP Professional Service Pack 2
Windows Vista Business, Windows Vista Ultimate, and Windows
Vista Enterprise
Windows Server 2003 Service Pack 2
Suggested web browsers
The following web browsers are suggested:
Microsoft Internet Explorer 6 Service Pack 2
Microsoft Internet Explorer 7
Mozilla Firefox 2
Required assets
The following assets are required:
A screen resolution of 1024x768 or better with True Color (24-bit)
.NET framework v1.1 (included in the DataStage Client Install CD if not
already installed)
Suggested memory
A minimum of 2 GB memory is suggested. Additional memory is beneficial as the
size of the flows or the number of columns being processed increases.
Required disk space
The following disk space for InfoSphere Information Server Client Products is
required:
10 MB in \Windows\System32
840 MB \IBM\Information Server
1.2.2 Server: Windows platform specification
DataStage for Windows release 8.1 requires a PC with an Intel processor (or
equivalent) running 32-bit Windows Server 2003 Service Pack 2.
4 IBM InfoSphere Information Server Installation and Configuration Guide
Suggested memory
The following memory is suggested:
The minimum amount of memory for installing the InfoSphere Information
Server client tier is 2 GB.
The minimum amount of memory for installing the InfoSphere Information
Server services and engine tiers on the same computer, or on separate
computers, is 4 GB.
Memory requirements depend on the type of processing, the volume of parallel
processing, the size and number of simultaneously running InfoSphere
DataStage and IBM InfoSphere QualityStage® jobs, and the memory
requirements of the operating system and other applications (such as relational
databases).
Evaluate the following factors to determine the memory requirements for
your system:
The number of InfoSphere Information Server product components on the
same server
Other software on the same server
Performance requirements
Size and complexity of your configuration
Extent of activity and the number of concurrent clients that access
your system
Required disk space
The following disk space is required:
2.6 GB for InfoSphere Application Server
1.0 GB for IBM DB2®
1.4 GB for the InfoSphere Information Server components
2.5 GB for the metadata repository database
1.5 GB for the InfoSphere Information Analyzer analysis database
2 GB of temporary space during the installation
C++ compiler
On development systems, a C++ compiler is required to compile jobs with
parallel transformers:
Microsoft Visual C++ .NET 2003
Microsoft Visual Studio 2005 Professional Edition C++
Microsoft Visual Studio .NET 2005 Express Edition C++
Chapter 1. Introduction 5
Embedded MKS OEM
DataStage for Windows installations (and requires) a special OEM version of the
MKS Framework that provides UNIX-style compatibility through runtime libraries,
scripts, and utilities.
1.2.3 Server: UNIX and Linux platform specification
The DataStage server component runs on the following platforms that are
supported by release 8.1:
IBM AIX® 5.3, 6.1
HP-UX (PA-RISC) 11i v2 (11.23), 11iv3
HP-UX (Itanium) 11i v2 (11.23), 11iv3
Linux: Red Hat Enterprise Linux Advanced Server 4 on AMD or Intel
Linux: Red Hat Enterprise Linux Advanced Platform 5 on AMD or Intel
Linux: SUSE Linux Enterprise Server 10 on AMD or Intel
Linux: SUSE Linux Enterprise Server 10 on IBM System z®
Sun Solaris 9 and 10
Suggested memory
As with any configuration, actual memory requirements depend on the type of
processing, degree of parallelism, size and number of simultaneously running
DataStage jobs, and memory requirements by the operating system and other
applications (such as relational databases).
The following memory suggestions are for DataStage only:
For 4 - 16 processors: Two GB of memory per processor is generally
adequate, but 3 GB is preferred for high-performance environments.
For 16 or more processors: Less than 2 GB of memory per processor is
needed except for instances with DataStage jobs that use very large lookups
or hash aggregators, or when running large, complex DataStage jobs
simultaneously.
Important: The OEM version of MKS that is installed with DataStage for
Windows includes make, as well as the header files and libraries necessary to
build custom components, BuildOps, and transformers. As such, the
DataStage MKS OEM distribution is different from OEM MKS installations
included with other products, such as InfoSphere QualityStage.
6 IBM InfoSphere Information Server Installation and Configuration Guide
Required disk space
The following disk space is required:
2.6 GB for InfoSphere Application Server
1.0 GB for DB2
1.4 GB for the InfoSphere Information Server components
2.5 GB for the metadata repository database
1.5 GB for the InfoSphere Information Analyzer analysis database
2 GB of temporary space during the installation
100 MB per project for design metadata and logs (might grow significantly
over time)
25 MB of free space in /var
Sufficient storage space for any data that is to be held in DataStage tables
or files
Additional space to allow for temporary data storage while a DataStage job
is running
C++ compiler
On development systems, a C++ compiler is required to compile jobs with
parallel transformers. When installing the C++ compiler for your machine, ensure
that all packages are installed.
Important: Only the compilers and versions in Table 1-1 are compatible with
DataStage. IBM certifies DataStage for specific compiler releases for a
given platform.
Chapter 1. Introduction 7
Table 1-1 lists the supported compilers.
1
Table 1-1 Supported compilers
1
IBM InfoSphere Information Server at:
http://www.ibm.com/software/data/infosphere/info-server/overview/requirements.html
Operating system C++ compilers Runtime components and
additional requirements
64-bit AIX 5.3
64-bit AIX 6.1
May 2008 XL C/C++ Enterprise Edition
V8.0 for AIX PTF, Reference #4019338
August 2008 XL C/C++ Enterprise
Edition V9.0 for AIX, Reference
#4020144
October 2008 XL C/C++ Enterprise
Edition V10.1 for AIX, Reference
#4021392
The runtime libraries are installed
with the operating system.
HP-UX on PA-RISC aCC: HP ANSI C++ B3910B A.03.85 The runtime libraries are installed
with the operating system.
64-bit HP-UX 11i v2 on
Intel Itanium
64-bit HP-UX 11i v3 on
Intel Itanium
6.16 aCC: HP C/C++ B3910B A.06.14
6.16 aCC: HP C/C++ B3910B A.06.20
The runtime libraries are installed
with the operating system.
32-bit Red Hat Enterprise
Linux Advanced Server 4
acc 3.4 The runtime libraries are installed
with the operating system.
64-bit Red Hat Enterprise
Linux Advanced Server 4
acc 3.4.6 Available on the Red Hat Linux
installation media:
glibc-devel-2.3.4-2.25.i386.rpm.
32-bit and 64-bit Red Hat
Enterprise Linux 5
Advanced Platform
acc 4.1.2 The runtime libraries are installed
with the operating system.
64-bit Solaris 9
64-bit Solaris 10
Sun Studio 10, 11, or 12 The runtime libraries are installed
with the operating system.
32-bit and 64-bit SUSE
Linux Enterprise Server
10
acc 4.1.2 The runtime libraries are installed
with the operating system.
8 IBM InfoSphere Information Server Installation and Configuration Guide
1.2.4 Server: z/OS platform specification
To install the parallel engine, your IBM z/OS® system should meet the following
hardware and software requirements:
Red Hat Enterprise Linux 5 Advanced Platform on IBM System z.
SUSE Linux Enterprise Server 10 on IBM System z.
IBM z/800, 2-way processor (2066-0X2), or an LPAR that is equivalent or
bigger than that.
z/OS Version 1.3 and later.
To determine this, use the DISPLAY IPLINFO command and note the
RELEASE value.
IBM C/C++ compiler 1.3.
Java 1.4 or greater.
Review APARs OA06361 and OA07784 for applicability to your system.
z/OS 1.3 requires UQ77835, z/OS 1.4 requires UQ77836, and DB2 v8.x
requires UQ89056.
Required memory
The required memory is 2 GB or more in the LPAR. To determine this, use the
DISPLAY M command and look for the line that says HIGH REAL STORAGE ADDRESS
IS nnnnM. nnnn should be 2048 or more.
The following disk space is required:
250 MB of free disk space for product installation.
100 MB per project for design metadata and logs (might grow significantly
over time)
At least 500 MB of scratch disk space per processing node
Sufficient storage space for any data that is to be held in DataStage tables
or files
C++ compiler
A C++ compiler is required to compile jobs with parallel transformers. When
installing the C++ compiler for your machine, ensure that all packages are
installed. Note the directory where the compiler is installed, because you will
need it for system configuration: IBM C/C++ compiler Version 1 Release 2 or
later.
Chapter 1. Introduction 9
For more information about detailed requirements, go to the IBM InfoSphere
Information Server, Version 8.1, system requirements page at:
http://www-01.ibm.com/support/docview.wss?rs=14&uid=swg21315971
The installer must verify requirements, because this page is updated whenever
omissions are discovered.
Note: You can install anywhere on your UNIX System Services machine, but
do not install at a mount point because the installation attempts to rename the
installation directory to support subsequent maintenance and upgrades. If you
do attempt to install at a mount point, the installation will still work, but you will
receive warning messages that might be confusing. The top-level directory is
subsequently identified by the environment variable $APT_ORCHHOME.
10 IBM InfoSphere Information Server Installation and Configuration Guide
© Copyright IBM Corp. 2011. All rights reserved. 11
Chapter 2. Platform specifications
The information in this section provides guidelines based on field experiences. In
certain cases the suggestions documented here are not the same as those in the
installation documentation. The installation documentation typically deals with
minimum requirements, whereas the numbers contained below are based on
best practice experience and will typically yield more satisfactory performance
than configurations based on the minimum requirements. All platforms have
additional considerations that are documented in the DataStage Install and
Upgrade Guide and readme files that are available with the product
documentation.
2
12 IBM InfoSphere Information Server Installation and Configuration Guide
2.1 Client: Windows platform specification
In this section we describe the Windows platform specification.
2.1.1 Suggested 32-bit version operating systems
The suggested 32-bit version operating systems are:
Windows XP Professional Service Pack 2
Windows Vista Business, Windows Vista Ultimate, and Windows
Vista Enterprise
Windows Server 2003 Service Pack 2
2.1.2 Suggested web browsers
The suggested web browsers are:
Microsoft Internet Explorer 6 Service Pack 2
Microsoft Internet Explorer 7
Mozilla Firefox 2
2.1.3 Required assets
The required assets are:
A screen resolution of 1024x768 or better is suggested with True Color
(24-bit).
.NET framework v1.1 (included in the DataStage Client Install CD if not
already installed).
Note: The DataStage client is tightly coupled to the DataStage server. Despite
what might be indicated in the release notes, the DataStage client and server
versions should always match unless you have been directed otherwise
by support.
Installing the DataStage Multi-Client Manager on the client workstation allows
you to maintain multiple DataStage client versions on a single machine.
Chapter 2. Platform specifications 13
2.1.4 Suggested memory
A minimum of 2 GB memory is suggested. Additional memory is beneficial as the
size of the flows or the number of columns being processed increases.
Required disk space
The required disk space is:
InfoSphere Information Server Client Products: 10 MB in \Windows\System32
About 1 GB
InfoSphere Information Server Business Glossary Anywhere: 3 MB
2.2 Server: Windows platform specification
DataStage for Windows release 8.1 requires a PC with an Intel processor (or
equivalent) running 32-bit Windows Server 2003 Service Pack 2.
2.2.1 Suggested memory
The suggested memory is:
The minimum amount of memory for installing the IBM InfoSphere Information
Server client tier is 2 GB.
The minimum amount of memory for installing the InfoSphere Information
Server services and engine tiers on the same computer, or on separate
computers, is 4 GB.
Memory requirements depend on the type of processing, the volume of parallel
processing, the size and number of simultaneously running InfoSphere
DataStage and InfoSphere QualityStage jobs, and the memory requirements of
the operating system and other applications (such as relational databases).
Evaluate the following factors to determine the memory requirements for
your system:
The number of InfoSphere Information Server product components on the
same server
Other software on the same server
Performance requirements
Size and complexity of your configuration
Extent of activity and the number of concurrent clients that access
your system
14 IBM InfoSphere Information Server Installation and Configuration Guide
2.2.2 Required disk space
The required disk space is:
2.6 GB for InfoSphere Application Server
1.0 GB for DB2
1.4 GB for the InfoSphere Information Server components
2.5 GB for the metadata repository database
1.5 GB for the InfoSphere Information Analyzer analysis database
2 GB of temporary space during the installation
2.2.3 C++ compiler
On development systems, a C++ compiler is required to compile jobs with
parallel transformers:
Microsoft Visual C++ .NET 2003
Microsoft Visual Studio 2005 Professional Edition C++
Microsoft Visual Studio .NET 2005 Express Edition C++
2.2.4 Embedded MKS OEM
DataStage for Windows installs (and requires) a special OEM version of the MKS
Framework that provides UNIX-style compatibility through runtime libraries,
scripts, and utilities.
2.3 Server: UNIX/Linux Platform specification
The DataStage server component runs on the following platforms supported by
release 8.1:
IBM AIX 5.3, 6.1
HP-UX (PA-RISC) 11i v2 (11.23), 11iv3
HP-UX (Itanium) 11i v2 (11.23), 11iv3
Linux: Red Hat Enterprise Linux Advanced Server 4 on AMD or Intel
Linux: Red Hat Enterprise Linux Advanced Platform 5 on AMD or Intel
Important: The OEM version of MKS that is installed with DataStage (for
Windows) includes make, as well as the header files and libraries necessary
to build custom components, BuildOps, and transformers. As such, the
DataStage MKS OEM distribution is different from OEM MKS installations
included with other products, such as InfoSphere QualityStage.
Chapter 2. Platform specifications 15
Linux: SUSE Linux Enterprise Server 10 on AMD or Intel
Linux: SUSE Linux Enterprise Server 10 on IBM System z
Sun Solaris 9 and 10
2.3.1 Suggested memory
As with any configuration, actual memory requirements depend on the type of
processing, degree of parallelism, size and number of simultaneously running
DataStage jobs, and memory requirements by the operating system and other
applications (such as relational databases).
The following memory suggestions are for DataStage only:
For 4 - 16 processors: Two GB of memory per processor is generally
adequate, but 3 GB is preferred for high-performance environments.
For 16 or more processors: Less than 2 GB of memory per processor is
needed except for instances with DataStage jobs that use very large lookups
or hash aggregators, or when running large, complex DataStage jobs
simultaneously.
2.3.2 Required disk space
The required disk space is:
2.6 GB for InfoSphere Application Server
1.0 GB for DB2
1.4 GB for the InfoSphere Information Server components
2.5 GB for the metadata repository database
1.5 GB for the InfoSphere Information Analyzer analysis database
2 GB of temporary space during the installation
100 MB per project for design metadata and logs (might grow significantly
over time)
25 MB of free space in /var
Sufficient storage space for any data that is to be held in DataStage tables
or files
Additional space to allow for temporary data storage while a DataStage job
is running
16 IBM InfoSphere Information Server Installation and Configuration Guide
2.3.3 C++ compiler
On development systems, a C++ compiler is required to compile jobs with
parallel transformers. When installing the C++ compiler for your machine, ensure
that all packages are installed.
Table 2-1 shows the list of supported compilers, which was compiled from:
http://www.ibm.com/software/data/infosphere/info-server/overview/requir
ements.html
Table 2-1 Supported compilers
Important: Only the following compilers and versions are compatible with
DataStage. IBM certifies DataStage for specific compiler releases for a
platform.
Operating system C++ compilers Runtime components and
additional requirements
64-bit AIX 5.3
64-bit AIX 6.1
May 2008 XL C/C++ Enterprise
Edition V8.0 for AIX PTF,
Reference #4019338
August 2008 XL C/C++
Enterprise Edition V9.0 for AIX,
Reference #4020144
October 2008 XL C/C++
Enterprise Edition V10.1 for AIX,
Reference #4021392
The runtime libraries are installed
with the operating system.
HP-UX on PA-RISC aCC: HP ANSI C++ B3910B
A.03.85
The runtime libraries are installed
with the operating system.
64-bit HP-UX 11i v2 on Intel
Itanium
64-bit HP-UX 11i v3 on Intel
Itanium
6.16 aCC: HP C/C++ B3910B
A.06.14
6.16 aCC: HP C/C++ B3910B
A.06.20
The runtime libraries are installed
with the operating system.
32-bit Red Hat Enterprise Linux
Advanced Server 4
gcc 3.4 The runtime libraries are installed
with the operating system.
64-bit Red Hat Enterprise Linux
Advanced Server 4
gcc 3.4.6 Available on the Red Hat Linux
installation media:
glibc-devel-2.3.4-2.25.i386.rpm.
Chapter 2. Platform specifications 17
2.4 Server: z/OS platform specification
To install the parallel engine, your z/OS system should meet the following
hardware and software requirements:
Red Hat Enterprise Linux 5 Advanced Platform on IBM System z.
SUSE Linux Enterprise Server 10 on IBM System z.
IBM z/800, two-way processor (2066-0X2), or an LPAR that is equivalent to or
larger than that.
z/OS Version 1.3 and later.
To determine this, use the DISPLAY IPLINFO command and note the
RELEASE value.
IBM C/C++ Compiler 1.3.
Java 1.4 or later.
Review APARs OA06361 and OA07784 for applicability to your system.
z/OS 1.3 requires UQ77835, z/OS 1.4 requires UQ77836, and DB2 v8.x
requires UQ89056.
2.4.1 Required memory
The required memory is 2 GB of memory or more in the LPAR. To determine this,
use the DISPLAY M command and look for the line that says HIGH REAL STORAGE
ADDRESS IS nnnnM. nnnn should be 2048 or more.
32-bit and 64-bit Red Hat
Enterprise Linux 5 Advanced
Platform
gcc 4.1.2 The runtime libraries are installed
with the operating system.
64-bit Solaris 9
64-bit Solaris 10
Sun Studio 10, 11 or 12 The runtime libraries are installed
with the operating system.
32-bit and 64-bit SUSE Linux
Enterprise Server 10
gcc 4.1.2 The runtime libraries are installed
with the operating system.
Operating system C++ compilers Runtime components and
additional requirements
18 IBM InfoSphere Information Server Installation and Configuration Guide
2.4.2 Required disk space
The required disk space is:
250 MB of free disk space for product installation
100 MB per project for design metadata and logs (might grow significantly
over time)
At least 500 MB of scratch disk space per processing node
Sufficient storage space for any data that is to be held in DataStage tables
or files
2.4.3 C++ compiler
A C++ compiler is required to compile jobs with parallel transformers. When
installing the C++ compiler for your machine, ensure that all packages are
installed. Note the directory where the compiler is installed, as it will be needed
for system configuration: IBM C/C++ Compiler Version 1 Release 2 or later.
For more information about detailed requirements, see IBM InfoSphere
Information Server, Version 8.1, system requirements at:
http://www.ibm.com/support/docview.wss?rs=14&uid=swg21315971
The installer must verify the requirements, because this page is updated
whenever omissions are discovered.
Note: You can install anywhere on your UNIX System Services machine, but
do not install at a mount point because the installation attempts to rename the
installation directory to support subsequent maintenance and upgrades. If you
do attempt to install at a mount point, the installation still works, but you
receive warning messages that might be confusing. The top-level directory is
subsequently identified by the environment variable $APT_ORCHHOME.
© Copyright IBM Corp. 2011. All rights reserved. 19
Chapter 3. Capacity planning
With any application, the most accurate sizing is determined from real-world
measurements of the actual developed job flows running on the target platform.
Unfortunately, this luxury is not always available, so educated estimates must be
used based on prior experience and available information, with certain elements
(such as disk usage) more determinant than others. This section is intended to
provide rough estimates and guidelines for sizing a DataStage environment.
3
20 IBM InfoSphere Information Server Installation and Configuration Guide
3.1 Minimums
For all but the smallest of applications, a minimum production configuration
consists of:
Four processors.
Two GB of real memory per processor, although 3 GB is suggested. With
more memory-intensive applications, 4 GB is better.
Swap space should be set to at least twice the real memory.
Disk space (see 3.5, “Disk” on page 23).
3.2 Processor
Consider a 4-processor system as a minimum for a production InfoSphere
DataStage environment. Two processors are sufficient for development and
certain testing environments, including quality assurance. But if you are not
prepared to put four processors to work on your production data, you will likely
never realize the benefits from a parallel processing engine.
Next consider the type of processing that you will be doing. Aggregations,
decimal calculations, and complex data parsing are generally considered more
CPU intensive than sorts and string manipulations. So taking into account what
type of work the data flows are going to be doing can be helpful in deciding
whether to recommend more memory versus spending the same dollars on disk
or processor.
For all but the smallest systems, add processors in groups of four. For example,
8-way or 12-way systems are suggested, but for a 6-processor system (with
budgetary constraints), the money is better spent on additional memory or
upgrading the I/O subsystem.
3.3 Memory
For all but the smallest deployments, consider 3 GB per processor as a
minimum. One of the main benefits of parallelism is to avoid landing the data to
the disk subsystem, thereby avoiding the time delay of an I/O transaction. For
data flows that are going to require heavy use of sorts and partitioning (such as
RDBMS sources or targets, SCD type data flows) consider 4 GB of real memory
per processor.
Chapter 3. Capacity planning 21
On systems with a large number of processors (16 or more), actual memory
requirements might be less than these guidelines. Consider that DataStage is a
32-bit application, so that most processes (with the exception of certain
operators such as lookup and hash aggregator) are limited to 2 GB of memory
per process. In these large systems, less than 2 GB of memory per processor
should be adequate, unless job processing includes very large lookups or hash
aggregators, or if running large complex DataStage jobs simultaneously.
As an example, let us consider the impact of sorting on a particular data flow and
how memory could affect the outcome. Consider system
alpha: four processors,
8 GB memory, 100 GB disk working with file sizes of approximately 10 GB for
processing. During a relatively simple data flow we are able to alter the amount
of real memory consumed by the sort stage. We are able to allocate
approximately 6 GB of real system memory for sorting. So at any given moment
we are swapping out 4 GB of the data file to disk to perform the sorting activity.
Now consider system
beta: four processors, 12 GB memory, 100 GB disk
working with the same files sizes. During a relatively simple data flow we are able
to allocate 10 GB of real memory to the sorting operation. This means that
effectively, we do not land any data to disk during the sort operations. The
difference in performance of these two data systems running these data flows
will be tremendous.
Remember, however, that sort operations are performed in parallel, and
partitioned. Because sort memory is allocated per partition, the amount of
memory required for an in-memory sort is dependent on partition skew (how data
is distributed across partitions). If data is not evenly distributed across partitions,
it might be necessary to allocate memory for the largest partition. Total sort
memory requirements depend on the degree of parallelism multiplied by the
configured sort memory usage.
Lookups are another area of consideration for memory. For non-database
normal lookups, each reference link is loaded into shared memory, which
cannot be swapped on most operating systems. Also, because normal
lookups allocate memory in a single shared memory block, a contiguous free
block of the required size must be available in shared memory. If the server is
shared with other systems (for example, databases), then shared memory might
become segmented.
22 IBM InfoSphere Information Server Installation and Configuration Guide
As with any configuration, actual memory requirements depend on the type of
processing, degree of parallelism, size and number of simultaneously running
DataStage jobs, and memory requirements by the operating system and other
applications (such as relational databases). In general, total memory
requirements depend on the number of:
Processes generated by the job at run time, and the:
– Size of the job (number of stages)
– Degree of operator combination
– Degree of parallelism (config file, node pool assignments)
Buffer operators and buffer size (default is 3 MB/partition/buffer)
Simultaneously active sorts (default is 20 MB/sort/partition)
Lookups (depends on combined size of all in-memory reference tables)
Hash aggregators (2 K per unique key column value/calculation)
Jobs running simultaneously (maximum requirement based on job schedule)
3.4 Swap space
In most UNIX environments, swap is allocated as needed, and is generally set to
1.5x to 2x physical memory. Certain UNIX environments calculate swap space
based on the incremental not complete size. Your system administrator will have
this information to assist in sizing.
Solaris allocates swap space differently from other UNIX platforms. Swap space
is preallocated for every child process using the memory requirements of the
parent process (and correspondingly each shared library used). Because each
parallel engine job executes as a hierarchical group of processes at run time,
swap space requirements on Solaris will depend on the number of simultaneous
DataStage jobs and their degree of complexity. On Solaris platforms only, start
with 8 GB per processor for swap allocation. Following this guideline, 128 GB of
swap space would be suggested for a 16 processor system. However, when
running a large number of jobs simultaneously, or when running very large jobs,
the swap requirements might be greater than this starting point.
Guidelines for minimizing the number of processes generated at run time are
given in the Dataflow Design Standard Practice.
Chapter 3. Capacity planning 23
3.5 Disk
One possible suggestion is a disk subsystem that has four main areas. The four
main areas form a traditional batch model for data processing wherein data
moved in from a flat file and out to a flat file is processed more efficiently. When
reading to or writing from some other type of source/target, this design would
need to be modified. Consider:
Staging for input files
Staging for output files
Scratch/sort work areas
Resource areas for the
safe storage of parallel data sets
Often, we can use a tried-and-true method of estimating the space required for
each class of storage developed from experience with databases: raw-data
volume times 2.5. To use this method, you must have at least a partial inventory
of the data that you intend to store, its frequency of use, and its retention period.
More detailed internal DataSet sizing information is available in the DataStage
Parallel Job Advanced Developer's Guide. Keep in mind that experience with
your data and implementation will cause adjustments to the formulas.
3.5.1 Staging disk for input and output files
The staging areas for input and output files should be able to contain at least two
of the largest expected input/output files, plus an additional 25 - 35% for growth
over time. This suggestion stems from experience in such environments where it
is sometimes necessary to rerun a previous file and still have room to contain
today's processing data. To guard against disk failure, redundant storage (for
example, RAID 5) is suggested.
3.5.2 Scratch/sort work areas
The scratch/sort areas have several unique characteristics. First, it is optimum if
they are local to the system where the engine is being run. However, it is
suggested that you not secure these disk areas, as the RAID penalty is often
more of a hindrance to good performance than the benefit of securing temporary
files that only exist during the job run. The following formulas are rough estimates
Note: Storage considerations for the last two items (scratch/sort and
resource) is dependent on the file systems identified in the parallel
configuration file used at run time. For more information about building and
tuning parallel configuration files, see Chapter 5, “Parallel configuration files”
on page 123.
24 IBM InfoSphere Information Server Installation and Configuration Guide
only. There are more exact formulas, but they usually require information that is
not available during the install phase of an implementation.
The size of each scratch/sort area should be as follows. For each simultaneous
sort use (X x 1.35)/N, where:
X is the size of the largest data file.
N is the degree of parallelism used by most jobs.
For example, consider a 4 GB input file on an 8-way system. The calculation
would be (4 GB x 1.35)/8 processor = 675 MB for each scratch/sort space, or
(4 x 1.35)/8 per partition, and there are eight in this case.
When there is more than one simultaneous sort (within multiple branches of one
job or across separate, simultaneously running jobs), the total sort requirements
will be the sum of each sort.
3.5.3 Resource areas for parallel data sets
Parallel data sets are most often used for intermediate storage of data between
jobs (end-to-end parallelism) and for checkpoint/restart. This area's size can be
calculated in roughly the same manner as the scratch/sort areas with the
exception that it might be necessary to store more than one version of any given
data set. For example, it might be necessary to have the previous night's data set
for delta processing, or some number of prior runs for recovery purposes.
Remember that since data sets are persistent, there needs to be allocated
storage space for all data sets, not just those needed for the currently running
job. The formulas that follow are rough estimates only. Exact formulas are
available, but usually require information that is not available during the install
phase of an implementation.
Therefore, the calculation for each resource disk area should be (X x 1.35)/N,
where:
X is the total size of all data to be stored concurrently.
N is the number of processors expected for most jobs that run in parallel.
Using our example job only and storing a week’s worth of data for a single job
only, there will be (4 GB x 5 copies x 1.35) / 8 processor = ~3.38 GB per partition.
So, in this case, there will be eight partitions.
For a more detailed estimation for the size of a parallel data set, see “Estimating
the size of a parallel data set” on page 146.
© Copyright IBM Corp. 2011. All rights reserved. 25
Chapter 4. Installation and
configuration
This chapter provides information about installing and configuring IBM
InfoSphere Information Server. It includes a pre-installation checklist that
describes the known factors for the five layers of an InfoSphere Information
Server 8.x installation.
4
26 IBM InfoSphere Information Server Installation and Configuration Guide
4.1 Pre-installation overview
IBM InfoSphere Information Server provides a unified foundation for enterprise
information architectures, combining the capabilities of DataStage, QualityStage,
Information Analyzer, Business Glossary, InfoSphere Information Services
Director, and Metadata Workbench.
InfoSphere Information Server is installed in the following five layers:
Client
Metadata repository
Domain (platform services)
Engine
Documentation
Product components are installed in each tier, depending on the install selections
that are made.
To ensure a successful installation that meets functional and performance
requirements, it is critical that overall planning and prerequisites be met for all
tiers. This document provides a detailed methodology for planning an InfoSphere
Information Server installation.
Where possible, links are provided for additional details and reference
documentation is mentioned. The information in this document is based on field
experiences. In certain cases, the suggestions documented here might differ
from the product documentation. Except where noted, this document is intended
to supplement, not replace, the product documentation and readme files.
Complete product documentation for IBM InfoSphere Information Server is only
available by installing the documentation tier of the product install. Additional and
updated documentation is available through the IBM InfoSphere Information
Server Information Center at:
http://publib.boulder.ibm.com/infocenter/iisinfsv/v8r0/index.jsp
Important: Establishing an InfoSphere Information Server environment that
meets performance expectations requires a capacity planning exercise:
reviewing deployment architecture, server, disk, and network configuration,
data sources, targets, data volumes, processing requirements, and service
level agreements.
Although minimum system requirements are referenced in this checklist,
capacity planning is outside the scope of this document.
Chapter 4. Installation and configuration 27
4.2 Pre-installation checklist
The checklist in Table 4-1 outlines the areas that must be reviewed and the steps
that must be completed prior to installing InfoSphere Information Server. A copy
of this checklist is also included in “Pre-installation checklist” on page 155.
Table 4-1 Pre-installation checklist
Specific details on each step can be found in subsequent sections of this
document, the release notes, and the IBM Information Server Planning,
Installation, and Configuration Guide, GC19-1048-07.
Complete Task
Review release notes (InfoSphere Information Server, IBM WebSphere®
Application Server, DB2).
Review the IBM Information Server Planning, Installation and
Configuration Guide, GC19-1048-07.
If migrating from previous versions of DataStage or QualityStage, review
Migrating to IBM Information Server Version 8.
Choose and validate installation architecture/topology.
Validate system requirements for all tiers (engine, domain, repository,
client, documentation).
Verify domain (WebSphere Application Server) requirements.
Verify database requirements for the metadata repository.
If applicable: Verify database requirements for Information Analyzer
analysis database.
Verify and configure disks, volume groups, and file systems.
Verify and configure operating system and resource limits.
Verify connectivity and network configuration.
Configure operating system users, groups, and associated permissions.
Verify and install C++ compiler or runtime libraries, or both.
Verify Information Server connector requirements.
Download and install fix pack packages (InfoSvr, WebSphere, DB2).
Perform complete system backup.
28 IBM InfoSphere Information Server Installation and Configuration Guide
4.3 Reviewing release notes
Release notes contain the latest information about a particular release of IBM
InfoSphere Information Server including issues resolved, known issues, and
workarounds. There are separate release notes that should be consulted for IBM
WebSphere Application Server and (if applicable) DB2.
4.3.1 IBM InfoSphere Information Server release notes
Although a copy of the InfoSphere Information Server release notes is included
with a product installation, a later version might be available online in the IBM
Information Center and should be consulted before any installation. For more
information, see “IBM Information Server release notes” in the IBM InfoSphere
Information Server Information Center at:
http://publib.boulder.ibm.com/infocenter/iisinfsv/v8r1/topic/com.ibm.swg
.im.iis.productization.iisinfsv.nav.doc/containers/cont_iisinfsrv_rnote.
html
When reviewing release notes, be sure to choose the version that matches the
base installation version of InfoSphere Information Server, for example, 8.1.
All InfoSphere Information Server fix packs and patches that are applied after a
base installation have separate, corresponding release notes that should also be
reviewed. For example, an 8.1 fix pack is applied to an existing 8.1 base
installation.
4.3.2 WebSphere Application Server release notes
The domain tier of IBM InfoSphere Information Server 8.1 requires WebSphere
Application Server standalone release 6.0.2.27. No other versions are supported
with V8.1. On most platform configurations, WebSphere Application Server is
included with the InfoSphere Information Server installer, but there are some
exceptions (namely, 64-bit platforms).
Release notes for IBM WebSphere Application Server v6.0.2 are available at:
http://publib.boulder.ibm.com/infocenter/wasinfo/v6r0/topic/com.ibm.
websphere.base.doc/info/aes/ae/v6rn.html
Complete documentation for WebSphere Application Server v6.0.2 is available
through the Info Center at the following location:
http://publib.boulder.ibm.com/infocenter/wasinfo/v6r0/topic/com.ibm.
websphere.base.doc/info/welcome_base.html
Chapter 4. Installation and configuration 29
DB2 Enterprise Server Edition release notes
The InfoSphere Information Server metadata repository can be installed in DB2
Enterprise Server Edition v9, Oracle 10g R2, or Microsoft SQL Server 2005.
When installing InfoSphere Information Server, if you choose to install the
metadata server option, a new copy of DB2 v9.1 (32 bit) or DB2 v9.5 (64 bit) is
installed. You can find the release notes for DB2 v9.1 at:
http://publib.boulder.ibm.com/infocenter/db2luw/v9/topic/com.ibm.db2.udb
.doc/doc/c0023859.htm
4.4 Planning, installation, and configuration
This pre-installation checklist provides a methodology and supplemental
information about preparing for an installation of IBM InfoSphere Information
Server:
Planning to install IBM InfoSphere Information Server
Preparing your system for installation
Installing IBM InfoSphere Information Server
Configuring IBM InfoSphere Information Server
Installing additional components
Troubleshooting installations
Removing IBM InfoSphere Information Server
For details about these steps, see IBM Information Server Planning, Installation,
and Configuration Guide, GC19-1048-07. This guide is part of the IBM
InfoSphere Information Server documentation, which is included with the product
installation package. The documentation tier can be installed separately, without
installing the rest of the InfoSphere Information Server.
4.5 Reviewing migrating to InfoSphere Information
Server
If migrating from earlier versions of DataStage or QualityStage, review the
content in the Migrating to IBM Information Server Version 8 Guide, which is
included with the product installation package. The documentation tier can be
installed separately, without installing the rest of the InfoSphere Information
Server.
30 IBM InfoSphere Information Server Installation and Configuration Guide
4.6 Choosing and validating the architecture or
topology
This section explains how to choose and validate the installation architecture
topology. The architecture of IBM InfoSphere Information Server is organized into
four major layers, which are listed in Table 4-2.
Table 4-2 Architecture layers
Layer Description Architectural notes
Client Administration, analysis,
and development user
interfaces and optional
MetaBrokers and bridges.
Multiple clients can access a single InfoSphere Information
Server. Client and server version must match.
Using v8 MultiClientManager, multiple client versions can
be installed on a single workstation.
Requires 32-bit versions of Windows XP, Windows Vista,
or Windows 2003.
Metadata
repository
Database that stores
InfoSphere Information
Server settings,
configuration, design, and
runtime metadata.
A single metadata repository database is defined for each
InfoSphere Information Server installation.
DB2 (included or customer supplied) v9.1 for 32 bit, v9.5
for 64 bit.
Oracle 10g R2 (customer supplied).
SQL Server 2005 (customer supplied).
Domain InfoSphere Information
Server common and
product-specific services.
A single domain is defined for each InfoSphere Information
Server installation.
Requires WebSphere Application Server (included or
customer supplied) Release 6.0.2 Fix Pack 27 or later fix
packs only.
Standalone (non-network) profile only.
Engine Runtime engine that
executes all InfoSphere
Information Server tasks.
Includes engine, connector,
PACK, and service agents
(logging, ASB, JobMon,
PerfMon).
Multiple engines on separate server environments might
be registered in a single InfoSphere Information Server
domain (cluster or grid deployment).
Only one v8 engine can be installed on a single server.
Can co-exist with multiple v7 DataStage engines.
Chapter 4. Installation and configuration 31
In any InfoSphere Information Server installation, the release level (version + fix
pack + patches) must match on all layers.
4.6.1 Two-tier deployment
Figure 4-1 illustrates a classic two-tier deployment:
Clients: Client applications on Windows system
Server 1: All other components (metadata repository, domain, engine) on the
same Linux, UNIX, or Windows server
Figure 4-1 Two-tier deployment
In this configuration, the single server housing the metadata repository, domain,
and engine should have a minimum of 8 GB of RAM, in addition to meeting the
combined disk requirements outlined in the InfoSphere Information Server
system requirements.
The InfoSphere Information Server versions of all components must match.
Because the complexity of tracking the state of different layers is simplified, the
two-tier deployment is optimal for high-availability (failover) scenarios.
Important: Although the architecture of InfoSphere Information Server offers
many theoretical deployment architectures, only the following subset of
configuration tiers is suggested and supported.
Any deviation from these deployments must be reviewed by IBM Support and
Engineering.
Metadata
Repository
Domain
Client
Engine
Ser ver 1
32 IBM InfoSphere Information Server Installation and Configuration Guide
4.6.2 Three-tier deployment
Figure 4-2 illustrates a classic three-tier deployment:
Clients: Client applications on Windows system
Server 1: Metadata server (metadata repository and domain) on the same
Linux, UNIX, or Windows server
Server 2: Engine on Linux, UNIX, or Windows server
Figure 4-2 Three-tier deployment
In the three-tier configuration, both backend servers (metadata server and
engine) must be located in the same physical data center, and should be
connected to the same network subnet.
The operating system of the metadata server and the engine servers should be
the same.
4.6.3 Four-tier deployment
The four-tier deployment, which next segments the metadata repository from the
domain, has the following configuration:
Clients: Client applications on Windows system
Server 1: Metadata repository on Linux, UNIX, or Windows server
Server 2: Domain on Linux, UNIX, or Windows server
Server 3: Engine on Linux, UNIX, or Windows server
For performance reasons, this configuration is not suggested unless the
metadata repository and domain servers are connected by a dedicated, private,
high-speed network connection.
Metadata
Repository
Domain
Client
Engine
Server 1
Server 2
Chapter 4. Installation and configuration 33
In the four-tier configuration, all backend servers must be located in the same
physical data center, and should be connected to the same network subnet. The
connection between these two machines must be low latency to avoid negatively
impacting the UI environment.
4.6.4 Cluster and grid deployments
In all configurations, multiple InfoSphere Information Server (v8) engines can be
installed or cross-mounted on physically separate servers. By including these
servers in a parallel configuration file at run time, InfoSphere Information Server
processing can span across a single server boundary.
Figure 4-3 illustrates a cluster and grid deployment.
Figure 4-3 Cluster and grid deployment
In a cluster or grid deployment, all engine servers must be the same operating
system, must be located in the same physical data center, and should be
connected by a dedicated, private high-speed network connection.
4.6.5 Wide area network deployments
Due to the data exchanges that occur between the InfoSphere Information
Server layers, it is strongly suggested that all tiers be located in the same local
area network (LAN). When deploying in a wide area network (WAN)
configuration, a network-hosting tool (for example, Citrix or Windows Remote
Desktop) must be used to host the InfoSphere Information Server clients.
Metadata
Repository
Domain
Client
Engine
Engine
Engine
34 IBM InfoSphere Information Server Installation and Configuration Guide
4.7 Validating system requirements for all tiers
After selecting the installation topology, IBM InfoSphere Information Server
system requirements should be verified for the client, metadata repository,
domain, and engine.
For more information about the IBM InfoSphere Information Server system
requirements, see the IBM InfoSphere Information Server, Version 8.1, system
requirements at:
http://www.ibm.com/support/docview.wss?rs=14&uid=swg21315971#dqx1
windowsdiskspace
4.8 Verifying domain requirements
The domain layer requires WebSphere Application Server Release 6.0.2 Fix
Pack 27 or later fix packs only (WebSphere Application Server should return
6.0.2.27 as its version). Specific hardware and software requirements for
WebSphere Application Server Release 6 can be found online:
For WebSphere Application Server 6.0.2 hardware requirements, see:
http://www-1.ibm.com/support/docview.wss?rs=180&uid=swg27007250
For WebSphere Application Server 6.0.2 software requirements, see:
http://www-1.ibm.com/support/docview.wss?rs=180&uid=swg27007256
On most platforms, when you install InfoSphere Information Server, you can
install a new copy of WebSphere Application Server or use an existing
installation.
On certain 64-bit platforms (64-bit RedHat Linux, 64-bit SUSE Linux, HP-UX
Itanium) it is necessary to download the WebSphere Application Server 6.0.2
installer and fix packs from IBM Passport Advantage®. For specific instructions
that are included in the IBM InfoSphere Information Server system requirements
see:
http://www-01.ibm.com/support/docview.wss?rs=14&uid=swg21315971#dqx1
windowsdiskspace
To install the domain layer in an existing installation of WebSphere Application
Server it must meet the following requirements:
Be at Release 6.0.2 Fix Pack 27 or later fix pack (WebSphere Application
Server 6.0.2.27).
Be a standalone (not network) deployment.
Chapter 4. Installation and configuration 35
Use a new, empty profile.
The profile must be named server1.
4.9 Verifying database requirements for metadata
repository
The metadata repository layer can be installed on DB2 v9.1, DB2 v9.5, Oracle
10g R2, or Microsoft SQL Server 2005. By default, the metadata repository
database is named
xmeta.
When installing InfoSphere Information Server, you can install a new copy of
DB2 Version 9.1 (or 9.5, depending on the target platform) or use an existing
installation. If you want to use Microsoft SQL Server or Oracle, you must install
and configure them before you install InfoSphere Information Server.
Specific hardware and software requirements for DB2 can be found online. For
DB2 v9 requirements see the following website:
http://publib.boulder.ibm.com/infocenter/db2luw/v9/topic/com.ibm.db2.udb
.uprun.doc/doc/r0025127.htm
To install the metadata repository on an existing installation of DB2, it must meet
the following requirements:
DB2 Enterprise Server Edition Release 9.1 (or 9.5)
Database cannot be partitioned
On certain 64-bit platforms (64-bit RedHat Linux, 64-bit SUSE Linux, HP-UX
Itanium) it is necessary to download the DB2 9.5 installer and fix packs from IBM
Passport Advantage. Specific instructions are included in the IBM InfoSphere
Information Server system requirements at:
http://www.ibm.com/support/docview.wss?rs=14&uid=swg21315971#dqx1
windowsdiskspace
During the installation of InfoSphere Information Server, if you choose to install
the metadata repository on a new copy of DB2, the metadata repository
database will be created as part of the installation.
Important: To minimize risk and to ensure that the domain requirements are
satisfied, use the version of WebSphere bundled with the IBM InfoSphere
Information Server installation, except in specific configurations (64-bit
installations of RedHat Linux, SUSE Linux, or HP-UX Itanium).
36 IBM InfoSphere Information Server Installation and Configuration Guide
If you choose to install the metadata repository on an existing installation of DB2,
Oracle 10g R2, or SQL Server 2005:
The metadata repository (xmeta) database must be created before the
InfoSphere Information Server installation.
Database connectivity must be configured from the domain tier to the
metadata repository database.
Database creation scripts are included in the Database Support subdirectory of
the InfoSphere Information Server installation. For information about creating the
metadata repository database, see IBM Information Server Planning, Installation,
and Configuration Guide, GC19-1048-07.
4.10 Verifying database requirements for Information
Analyzer analysis
This section explains how to verify database requirements for the Information
Analyzer analysis database. If deploying Information Analyzer, the Information
Analyzer analysis database can be installed on DB2 v9.1 and v9.5, Oracle 10g
R2, or Microsoft SQL Server 2005. By default, the Information Analyzer Analysis
database is named
IADB.
When installing InfoSphere Information Server, you can install a new copy of
DB2 Version 9.1 (or 9.5) or use an existing installation. If you want to use
Microsoft SQL Server or Oracle, you must install and configure them before
you install InfoSphere Information Server.
If you choose to install the Information Analyzer Analysis database on an existing
installation of DB2, Oracle 10g R2, or SQL Server 2005:
The Information Analyzer Analysis database (IADB) must be created before
the InfoSphere Information Server installation.
Database connectivity must be configured from the domain tier to the
Information Analyzer Analysis database.
Important: The InfoSphere Information Server requirements recommend a
minimum of 3 GB for the metadata repository. However, you must closely
monitor the database as it grows to ensure that sufficient space is available in
the tablespaces and underlying file systems.
As table definitions, job designs, reports, and runtime metadata are created,
the size of the metadata repository can grow significantly.
Chapter 4. Installation and configuration 37
Database creation scripts are included in the Database Support subdirectory of
the InfoSphere Information Server installation. For information about creating the
Information Analyzer Analysis database, see the IBM Information Server
Planning, Installation, and Configuration Guide, GC19-1048-07.
4.11 Verifying and configure disks, volume groups, file
systems
In this section we describe how to verify and configure disks, volume groups, and
file systems.
4.11.1 RAID or SAN configuration
IBM InfoSphere Information Server uses file system mount points for its
installation, libraries, temporary storage, and data set storage.
Ultimate performance of an InfoSphere Information Server job (DataStage,
QualityStage, and Information Analyzer) depends on all components being
optimized. When discussing disk (including RAID and SAN) configuration,
maximum performance is a combination of maximum bandwidth (controllers,
disk) and minimized contention.
The following guidelines can assist in the configuration of RAID or SAN
technologies:
Minimize contention between temporary (scratch, buffer, sort) and data
file systems.
Minimize contention between the disks and controllers associated with
InfoSphere Information Server file systems and other applications or servers.
Consider isolating multiple mount points to separate high-speed disk
interconnects and controllers.
Consider the trade-off between the granularity of file systems and underlying
configuration versus available, unused storage.
Important: Although the InfoSphere Information Server System requirements
recommend a minimum of 3 GB for the Information Analyzer Analysis
database, the actual size depends on the size of the sources to be analyzed.
Unless using sampled analysis, the Information Analyzer Analysis database
can be larger than the combined size of all analyzed data sources.
38 IBM InfoSphere Information Server Installation and Configuration Guide
Do not create overly complex device configurations. These can be difficult to
administer and use, and might not offer a corresponding performance
improvement.
If possible, test your configuration with expertise from hardware, storage,
operating system, and application (InfoSphere Information Server) resources.
Remember that previous experience in designing I/O systems for non-parallel
engines might lead to non-optimal configurations for InfoSphere Information
Server.
The optimal disk configuration will strike a balance between cost, complexity,
ease of administration, and ultimate performance.
4.11.2 InfoSphere Information Server file systems
IBM InfoSphere Information Server requires file systems for:
Software install directories
– IBM InfoSphere Information Server
– Domain (WebSphere Application Server)
– Database server
Database storage
– Metadata repository (xmeta)
– Information Analyzer analysis database
InfoSphere Information Server project (runtime shadow repository) directories
Used to capture runtime metadata (such as error messages and logging
information) for the InfoSphere Information Server parallel engine.
Synchronized to the metadata repository.
Data file storage
– InfoSphere Information Server engine temporary storage (scratch, temp,
buffer)
– InfoSphere Information Server artifacts (such as surrogate key files)
– InfoSphere Information Server parallel data set segment files
– Staging and archival storage for any source/target files
File system requirements:
File systems should be expandable without requiring destruction and
recreation.
Local file systems should be reserved for temporary and scratch storage.
Chapter 4. Installation and configuration 39
Data and install directories should be created on high-performance shared file
systems.
Always use NFS
hard mounts. (Soft mounts can lead to corruption of
InfoSphere Information Server requiring a reinstall or restore from backup.)
Some components and plug-ins might require additional file system permissions.
For example, the DB2 Enterprise Stage requires that the db2instance ID has
read access to a subdirectory under $TMPDIR for db2load.
4.11.3 Software installation directories
Table 4-3 describes the software installation directories.
Table 4-3 Software installationdirectories
Installing file system requirements
Activities for installing file system requirements are:
Do not install InfoSphere Information Server components (InfoSphere
Information Server, WebSphere Application Server, and DB2) on a top-level
mount point. Always install InfoSphere Information Server components to a
subdirectory within a mount point. The installer components change
ownership and permissions of directories that are installed, and not all mount
points allow these changes, causing the install to fail.
Important: Each storage class should be isolated in separate file systems to
accommodate their various performance and capacity characteristics and
backup requirements.
The default installation directories are best suited for small prototype
environments.
Installation directory Contents Default UNIX path
Information Server InfoSphere Information Server engine, libraries,
communication agents, Job Monitor, Performance
Monitor, Java JRE, and uninstall files
/opt/IBM/Information
Server/
Domain WebSphere Application Server executables, libraries:
InfoSphere Information Server shared services
InfoSphere Information Server product-specific
services
/opt/IBM/WebSphere/
AppServer/
Database server DB2 v9.1 database server executables, libraries /opt/IBM/db2/V9/
40 IBM InfoSphere Information Server Installation and Configuration Guide
On Windows platforms, InfoSphere Information Server cannot be installed on
virtual drives (mapped or SUBST drives).
The installation directories for WebSphere Application Server and (if
applicable) DB2 must be empty. If the target directory is not empty, the
installer will attempt to create another directory, which will not be properly
referenced by the InfoSphere Information Server paths.
All file systems used by InfoSphere Information Server for installation must
have read and write permissions for the primary DataStage users and groups.
Ensure that each level of the target install directories is set to 755
permissions.
InfoSphere Information Server, WebSphere Application Server, and (if
applicable) DB2 must be installed into separate directories. They cannot be
installed in the same directory level.
On UNIX platforms, the DB2 installer must have write access to several file
systems, including /var and /usr/local/bin. This requirement is not always
satisfied when these directories are mounted from network file systems such
as IBM AFS™, or if these directories are configured as read-only file systems.
For cluster or grid implementations, share the installation file systems across
all servers (with the same fully qualified paths).
4.11.4 Database storage
The two types of data storage are:
Metadata repository (xmeta)
Information Analyzer analysis database
On UNIX platforms, using the supplied DB2 database engine, databases are
created in the home directory (/home/db2inst1/) of the DB2 instance owner, by
default.
When using an existing installation of DB2 9.1, 9.5, Oracle 10g R2, or Microsoft
SQL Server 2005, table spaces are created by the database administrator using
existing file systems or raw disk devices.
Important: Closely monitor the metadata repository and (if applicable)
Information Analyzer Analysis databases to ensure that sufficient space is
available in the tablespaces and underlying file systems as the databases
grow. Perform regular, scheduled backups on these databases.
Chapter 4. Installation and configuration 41
4.11.5 InfoSphere Information Server Project directories
The InfoSphere Information Server Project (runtime shadow repository) Directory
is used to capture runtime metadata (such as error messages and logging
information) for the InfoSphere Information Server parallel engine. This data is
synchronized into the metadata repository on periodic intervals for later
reporting. The InfoSphere Information Server Project Directory can also store
legacy Server Edition hash files. Project directories can grow to contain
thousands of files and subdirectories depending on the number of projects, the
number of jobs, and the volume of logging information retained about each job.
During the installation process, the projects subdirectory is created in the
InfoSphere Information Server installation directory. By default, the DataStage
administrator client creates projects in the /opt/IBM/InformationServer/
Server/Projects/ subdirectory.
During installation, Information Analyzer also creates its own project within
this directory.
The following are InfoSphere Information Server Project Directory guidelines:
Do not create DataStage projects in the default directory within the installation
file system, because disk space is typically limited. Create projects in their
own file system.
On most operating systems, it is possible to create file systems at non-root
levels. Create a separate file system for the projects directory within the
InfoSphere Information Server installation. Back up any existing projects
(retaining ownership and permissions) before mounting a separate file system
over the existing /opt/IBM/InformationServer/Projects directory.
For cluster or grid implementations, it is generally best to share the projects
file system across servers (with the same fully qualified path).
Project naming considerations
The name of an InfoSphere Information Server Project is limited to a maximum of
54 characters. The project name can contain alphanumeric characters and it can
contain underscores.
Project names cannot use the following reserved words:
ds
DS
uv
UV
42 IBM InfoSphere Information Server Installation and Configuration Guide
Project names should be maintained in unison with source code control. As
projects are promoted through source control, the name of the phase and the
project name should reflect the version, in the form:
<Phase>_<ProjectName>_<version>
Where Phase corresponds to the phase in the application development life cycle
(Table 4-4).
Table 4-4 Application development lifecycle
Project directory monitoring and maintenance
Effective management of space is important to the health and performance of a
project. As jobs are added to a project, new directories are created in this file
tree, and as jobs are run, their log entries multiply. These activities cause file
system stress (for example, more time to insert or delete DataStage
components, longer update times for logs). Failure to perform routine project
maintenance (for example, remove obsolete jobs and manage log entries) can
cause project obesity and performance issues.
4.11.6 Dataset and Scratch directories
During installation, two directories are created within the InfoSphere Information
Server installation directory tree for storage of temporary and intermediate data
files used by the parallel engine:
/opt/IBM/InformationServer/Server/Datasets (persistent storage
between jobs)
/opt/IBM/InformationServer/Server/Scratch (temp storage for sort and
buffer overflow)
Phase name Phase description
dev Development
it Integration test
uat User acceptance test
prod Production
Important: The project file system should be monitored to ensure that
adequate free space remains. If the Project file system runs out of free space
during runtime activity, the repository may become corrupted, requiring a
restore from backup.
Chapter 4. Installation and configuration 43
The InfoSphere Information Server installer creates a default parallel
configuration file (/opt/IBM/InformationServer/Configurations/default.apt)
that references the default datasets and scratch directories.
Parallel configuration files are used to assign resources (such as processing
nodes, disk, and scratch file systems) at run time when a job is run by the
InfoSphere Information Server engine. For more information about parallel
configuration files, see the IBM Information Server Parallel Job Developer Guide,
LC18-9891-02.
Dataset and scratch directory guidelines
The following are dataset and scratch directory guidelines:
Dataset and scratch file systems should be created outside of the InfoSphere
Information Server installation directory.
Scratch file systems should be created on local (internal) storage.
Dataset and install directories should be created on high-performance,
shared file systems with the same fully qualified path on all engine servers.
Using a shared file system also facilitates high-availability failover scenarios.
The default.apt configuration file should be edited to reference the newly
created dataset and scratch file systems, and to ensure that these directories
are used by any other parallel configuration files.
For optimal performance, file systems should be created in high-performance,
low-contention storage.
For best performance, and to minimize storage impact on development
activities, separate file systems should be created for each data and scratch
resource partition.
On systems where multiple phases are shared on the same server, consider
separating data and scratch storage to different file systems for each
deployment phase to completely isolate each environment. This might be
required for security compliance in some situations.
4.11.7 Extending the DataStage project for external entities
For DataStage and QualityStage environments, it is suggested that another
directory structure be created to integrate all aspects of a DataStage application
that are managed outside of the DataStage Projects repository. This hierarchy
should include directories for secured parameter files, data set header files,
custom components, IBM Orchestrate® schemas, SQL, and shell scripts. It
might also be useful to store custom job logs and reports.
44 IBM InfoSphere Information Server Installation and Configuration Guide
4.11.8 File staging
Use a separate staging file system and directory structure to store, manage, and
archive various source data files.
4.12 Verifying and configuring OS and resource limits
On most platforms, InfoSphere Information Server, WebSphere Application
Server, and the database server have specific operating system (kernel) and
user (shell) resource requirements. For the latest versions of these requirements,
go to the following web pages:
IBM InfoSphere Information Server system requirements:
http://www.ibm.com/support/docview.wss?rs=14&uid=swg21315971#dqx1win
dowsdiskspace
WebSphere Application Server 6.0.2.27 software requirements:
http://www.ibm.com/support/docview.wss?rs=180&uid=swg27007256
DB2 v9.1, 9.5 requirements:
http://publib.boulder.ibm.com/infocenter/db2luw/v9/topic/com.ibm.db2
.udb.uprun.doc/doc/r0025127.htm
If using an existing Oracle10gR2 or SQL Server 2005 database server for the
metadata repository or Information Analyzer analysis database, reference the
database vendor's system requirements.
Notes on operating system requirements
Keep in mind the following notes about operating system requirements:
On UNIX installations of IBM InfoSphere Information Server, the tables in the
following sections list the minimum requirements for UNIX kernel parameters.
These settings give the generic names for the kernel parameters. The actual
name and the case vary for each UNIX platform.
Always make a backup of the kernel settings before making these
changes. On many platforms, this can be accomplished by backing up the
/etc/system file.
On certain platforms, it might be necessary to rebuild the kernel with changes
to these parameters. All changes should only be made by a trained UNIX
system administrator.
For readability, the table values in this section include commas. Omit them
when setting the actual parameters.
Chapter 4. Installation and configuration 45
4.12.1 UNIX kernel parameters for all platforms
Table 4-5 lists the minimum requirements for kernel parameters on all UNIX
platforms. These settings give the generic names for the kernel parameters. The
actual name and the capitalization vary for each platform.
The minimum values depend on the architecture as well. For example, DB2
might require a minimum value of 1024, but the engine might require 2048. If
they are both installed on the same server, you must take the higher of the two.
Table 4-5 lists the UNIX Kernel parameters.
Table 4-5 UNIX Kernel parameters
4.12.2 UNIX user (shell) parameters for all platforms
On all UNIX platforms, adjust the per-user (shell) parameters to ensure that the
InfoSphere Information Server parallel engine has sufficient resources. Table 4-6
lists these parameters.
Table 4-6 UNIX user parameters
Note: For readability, the table values in this section include commas. Omit
them when setting parameters.
Kernel
parameter
Description InfoSphere
Information
Server
minimum
DB2
minimum
Notes
MAXUPROC Maximum
number of
processes
100 per
processor
Set to at least
100 processes
per node.
NOFILES Number of
open files per
process
1,000
User parameter Description Suggested setting
umask Default file permissions 022
ulimit Maximum number of user processes 8192
46 IBM InfoSphere Information Server Installation and Configuration Guide
4.12.3 AIX system configuration
On AIX systems, if you intend to use the DataStage Job Scheduler, you must the
permissions on the /usr/spool/cron/atjobs directory from 770 to 775
(rwxrwxr-x).
AIX kernel parameters
Table 4-7 lists the minimum requirements for UNIX kernel parameters for AIX
installations of IBM InfoSphere Information Server.
Table 4-7 AIX kernel parameters
AIX user (shell) parameters for parallel jobs
Table 4-8 lists the additional per-user (shell) parameters that should be adjusted
on AIX to ensure that the InfoSphere Information Server parallel engine has
sufficient resources.
Table 4-8 AIX user parameters for parallel jobs
Attention: For readability, the table values in this section include commas.
Omit them when setting parameters.
Kernel
parameter
Description InfoSphere
Information
Server minimum
Notes
SHMMAX Maximum shared
memory segment size
536870912 If disk caching is
turned on and
DISKCACHE is larger
than 512, this must be
set higher.
SHMMNI Shared memory
identifiers
2000
SHMSEG Maximum number of
shared memory
segments per process
200
User parameter Description Suggested setting
fsize Largest file that a user
can create
2 GB - 512 bytes
(4,194,303 512-byte blocks)
data Largest data segment (heap)
that a program can have
128 MB minimum
(262,144 512-byte blocks)
or -1 for no limit
Chapter 4. Installation and configuration 47
AIX User (shell) parameters
Table 4-9 lists the additional per-user (shell) parameters on AIX.
Table 4-9 AIX user (shell) parameters
In addition, on AIX platforms, tuning the additional kernel parameters listed in
Table 4-10 might improve performance for the InfoSphere Information Server
parallel engine. Work with your AIX system administrator to determine optimal
values.
Table 4-10 Additional kernel parameters
stack Largest stack size that a
program can have
32 MB minimum
(65,536 512-byte blocks)
rss Maximum amount of physical
memory a that user’s process
can use
64 MB minimum
(131,072 512-byte blocks)
or more
User parameter Engine tier Service tier
MAXUPROC 200 1,000 or unlimited
NOFILES 1,000 10,000
User parameter Description Suggestion
maxperm Maximum number
of permanent
buffer pages for file
I/O
Tune to limit the amount of physical memory
used for file system I/O so that more memory
is dedicated for Enterprise Edition (EE)
processes.
somaxconn Maximum number
of socket
connections
Might need to increase this value for large
EE jobs with lots of processes. When
increasing maxuproc is not enough,
increase this kernel parameter.
rbr Release-behind-w
hen-reading
Consider setting this file system mount
option for permanent data set storage.
Because large permanent data sets are
always read sequentially and never in
reverse or random order, this might improve
performance.
User parameter Description Suggested setting
48 IBM InfoSphere Information Server Installation and Configuration Guide
4.12.4 HP-UX system configuration
This section provides the HP-UX kernel parameters.
HP-UX kernel parameters
Table 4-11 lists the minimum requirements for UNIX kernel parameters on
HP-UX installations of IBM InfoSphere Information Server.
After installing DB2 on HP-UX, run the db2osconf command to verify the
suggested kernel settings.
Table 4-11 HP-UX kernel parameters
CIO Concurrent I/O Investigate the use of CIO for both disk and
scratch disk resources. CIO results in a
performance similar to raw devices, which
might benefit access to both permanent data
sets and temporary sort files. CIO is good for
I/O with large block sizes.
User parameter Description Suggestion
Attention: For readability, the table values in this section include commas.
Omit them when setting parameters.
Kernel
parameter
Description InfoSphere
Information
Server
minimum
WebSphere
Application
Server
minimum
Notes
MAXFILES Maximum number of open
files per process
1448
MSGMAP Number of entries in
SystemV IPC message
space resource map
2319
MSGMAX Maximum message size in
bytes
32,768 65,535
MSGMNB Maximum bytes per
message queue
32,768 65,535
MSGMNI Maximum number of
system-wide SystemV IPC
message queues
2317
Chapter 4. Installation and configuration 49
On 32-bit platforms, HP-UX limits the maximum size of a shared memory
segment to 1.75 GB. When processing extremely large in-memory reference
tables in a DataStage Lookup Stage, you might need to configure HP-UX
memory windowing. For details about this process, see “Example user setup for
UNIX environments” on page 164.
MSGSEG Number of system VIPC
message segments
7,168 32,767
MSGSZ Message size 32
MSGTQL Maximum number of
SystemV IPC messages
2317
NFILE Maximum number of files
open simultaneously
16,219
NFLOCKS Maximum number of file
locks
5793
NIDNOE Maximum number of HFS
file system open inodes
4055
NPROC Number of simultaneous
processes
2912
SEMMNI Number of semaphore
identifiers (system wide)
2896
SEMMNS Total number of semaphores 5794
SEMMNU Maximum number of undo
structures
2896
SHMMAX Maximum shared memory
segment size
307,200,000 2,897,215,488 If disk caching is turned
on and DISKCACHE is
larger than 512, this
must be set higher.
SHMSEG Maximum number of shared
memory segments per
process
200
Kernel
parameter
Description InfoSphere
Information
Server
minimum
WebSphere
Application
Server
minimum
Notes
50 IBM InfoSphere Information Server Installation and Configuration Guide
HP-UX user (shell) parameters
Table 4-12 lists the additional per-user (shell) parameters that should be adjusted
on HP-UX to ensure that the InfoSphere Information Server parallel engine has
sufficient resources.
Table 4-12 HP-UX user (shell) parameters
4.12.5 RedHat and SUSE Linux system configuration
You cannot install InfoSphere Information Server on a version of RedHat Linux 4
that has been upgraded from RedHat v3 (upgrade configurations are missing
required libraries).
RedHat and SUSE Linux kernel parameters
Table 4-13 lists the minimum requirements for UNIX kernel parameters on
RedHat Linux installations of IBM InfoSphere Information Server.
Table 4-13 RedHat and SUSE Linux kernel parameters
User parameter Description Suggested setting
maxdsiz Maximum size of data segment 2,039,480,320
(1945 MB)
maxssiz Maximum size of stack 82,837,504 (79 MB)
maxtsiz Maximum size of text segment 1,073,741,824
rss Maximum amount of physical memory
that a user’s process can use
64 MB minimum
(131,072 512-byte
blocks) or more
Attention: For readability, the table values in this section include commas.
Omit them when setting parameters.
Kernel
parameter
Description Engine tier DB2
minimum
Services
tier
Notes
MSGMAX Maximum
message size in
bytes
8,192 65,536 No
requirement
MSGMNB Maximum bytes
per message
queue
16,384 65,536 No
requirement
Chapter 4. Installation and configuration 51
MSGMNI Maximum queue
system wide
No
requirement
1024 No
requirement
SEMMNI Number of
semaphore
identifiers (system
wide)
1024 1024 No
requirement
SEMMNS Total number of
semaphores
128,000 256,000 No
requirement
SEMMSL Maximum number
of semaphores
per id list
250 250 No
requirement
SEMOPM Number of
operations per
semop call
32 32 No
requirement
SHMALL Maximum total
shared Memory
838608 KB 2511724800
SHMMAX Maximum shared
memory segment
size
307,200,000 32,768 KB 2511724800 If disk caching is
turned on and
DISKCACHE is
larger than 512, this
must be set higher.
SHMMNI Shared memory
identifiers
2000 4096 No
requirement
SHMSEG Maximum number
of shared memory
segments per
process
200
Rlim_fd_max No
requirement
>=8193
Rlim_fd_cur No
requirement
.+8193
MAXUPROC 200 1,000 or
unlimited
NOFILES 1,000 10,000
Kernel
parameter
Description Engine tier DB2
minimum
Services
tier
Notes
52 IBM InfoSphere Information Server Installation and Configuration Guide
4.12.6 Solaris 9 system configuration
This section describes the Solaris kernel parameters.
Solaris kernel parameters
Table 4-14 lists the minimum requirements for UNIX kernel parameters for
Solaris 9 installations of IBM InfoSphere Information Server.
After installing DB2 on Solaris 9, run the db2osconf command to verify the
suggested kernel settings.
Table 4-14 Solaris 9 kernel settings
Attention: For readability, the table values in this section include commas.
Omit them when setting parameters.
Kernel
parameters
Description InfoSphere
Information
Server minimum
WebSphere
Application
Server minimum
Notes
MSGMAP Number of entries in
message map
1026
MSGMAX Maximum message size in
bytes
8,192 65,535
MSGMNB Maximum bytes per
message queue
16,384
MSGMNI Maximum message queues
(system wide)
1024
SEMAEM Maximum adjust-on-exit 16,384
SEMMAP Number of entries in
semaphore map
1024
SEMMNI Number of semaphore
identifiers (system wide)
1024
SEMMNS Total number of
semaphores
128,000 16,384
SEMMNU Maximum number of undo
structures
2048
SEMMSL Maximum number of
semaphores per ID list
1024 100
Chapter 4. Installation and configuration 53
4.12.7 Solaris 10 system configuration
InfoSphere Information Server has not been certified to run in Solaris 10 zones
or projects. If you choose to deploy InfoSphere Information Server on a Solaris
zone or project, any issues raised with IBM Support must first be reproduced in a
native operating system environment.
Solaris 10 kernel parameters
Table 4-15 on page 54 lists minimum requirements for UNIX resource
parameters for Solaris 10 installations of IBM InfoSphere Information Server.
Starting with Solaris 10, many older kernel parameters have been superseded by
resource limits.
After installing DB2 on Solaris 10, run the db2osconf command to verify the
suggested kernel settings.
SEMOPM Maximum number of
semaphore operations
32 100
SEMUME Maximum number of undo
structures per process
256
SHMMAX Maximum shared memory
segment size
307,200,000 4,294967,295 If disk caching is
turned on and
DISKCACHE is
larger than 512,
this must be set
higher.
SHMMNI Shared memory identifiers 2000
SHMSEG Maximum number of shared
memory segments per
process
1024
Kernel
parameters
Description InfoSphere
Information
Server minimum
WebSphere
Application
Server minimum
Notes
Attention: For readability, the table values in this section include commas.
Omit them when setting parameters.
54 IBM InfoSphere Information Server Installation and Configuration Guide
Table 4-15 Solaris 10 kernel settings
4.13 Verifying connectivity and network configuration
IBM InfoSphere Information Server relies on TCP/IP network infrastructure for
communication between:
InfoSphere Information Server clients and InfoSphere Information Server
domain
InfoSphere Information Server clients and Information Sever engine
InfoSphere Information Server domain and metadata repository
InfoSphere Information Server domain and Information Analyzer Repository
InfoSphere Information Server engine and source/target data stores:
– Databases
– File transfer through FTP or named pipes
– WebSphere MQ
– Third-party applications such as SAS, Siebel, PeopleSoft, SAP
Kernel
parameters
Description InfoSphere
Information
Server minimum
WebSphere
Application
Server minimum
Notes
MSGMAX Maximummessage size
in bytes
8,192 65,535
MSGMNB Maximum bytes per
message queue
16,384
SEMMNI Number of semaphore
identifiers (system wide)
1024
SEMMSL Maximum number of
semaphores per id list
1024 100
SEMOPM Maximum number of
semaphore operations
1024 100
SHMMAX Maximum shared
memory segment size
6 GB 5 If disk caching is
turned on and
DISKCACHE is
larger than 512,
this value must
be set higher.
SHMMNI Shared memory
identifiers
2000
Chapter 4. Installation and configuration 55
Across InfoSphere Information Server engines in a clustered or Grid
configuration
InfoSphere Information Server engine and domain Server
InfoSphere Information Server Job Monitor and InfoSphere Information
Server engines
InfoSphere Information Server Performance Monitor and InfoSphere
Information Server engines
InfoSphere Information Server requires access to a number of TCP ports, which
must be open, if necessary, through firewall configuration. Related connectivity
(for example, databases and enterprise applications) might require additional port
assignments. Consult vendor-supplied documentation for their requirements.
For a complete list of network ports used by InfoSphere Information Server, see
“Network ports used by InfoSphere Information Server” on page 162.
InfoSphere Information Server network requirements
InfoSphere Information Server has the following network requirements:
TCP/IP network is required.
All servers must be resolvable by name, between tiers and from clients to
servers.
For optimal performance, all layers of InfoSphere Information Server should
be installed on the same high-speed, low-latency local area network.
In a cluster or grid deployment, all engine servers must be located in the
same physical data center, and should be connected by a dedicated, private
high-speed network connection.
The domain and metadata repository database should be installed on the
same server. If they are installed on separate servers, the domain and
metadata repository database should be connected by a dedicated, private,
high-speed network connection.
Due to the data exchanges that occur between the InfoSphere Information
Server layers, it is strongly suggested that all tiers be located in the local area
network (LAN). When deploying in a WAN configuration, use a network
hosting tool (for example, Citrix or Windows Remote Desktop) to host the
InfoSphere Information Server clients.
56 IBM InfoSphere Information Server Installation and Configuration Guide
4.14 Configuring operating system users, groups, and
permissions
This section highlights system configurations.
4.14.1 Privileged installation user
IBM InfoSphere Information Server requires a privileged user account for
installation.
On UNIX platforms, the installation must be performed by root or by a user
account with root privileges.
On Windows environments, the installation must be run from a local administrator
account. This user must have read/write access to the target installation
directories. The installation cannot be run from a domain administrator.
4.14.2 Required operating system users
InfoSphere Information Server requires a set of operating system user accounts
to install the engine and metadata repository database, which are listed in
Table 4-16. These accounts are used by the InfoSphere Information Server
engine, and internal domain services.
Table 4-16 Operating system users
User account Default user
name
Primary
group
Secondary
group
Notes
DataStage administrator dsadm dstage
DB2 administration server dasusr1 dasadm1 Only needed for DB2.
DB2 instance owner db2inst1 db2iadm1 dasadm1 Only needed for DB2.
DB2 fenced user db2fenc1 db2fadm1 Only needed for DB2.
Metadata repository owner xmeta xmeta DB2 uses OS
authentication.
Information Analyzer
analysis database owner
iauser iauser DB2 uses OS
authentication.
Important: These users can be created by the installer, but this is
not suggested.
Chapter 4. Installation and configuration 57
The security configuration of many operating systems (for example, AIX) requires
new users to log in before an account is activated.
For instructions to create users, see the IBM Information Server Planning,
Installation, and Configuration Guide, GC19-1048-07. You can find a simple
UNIX user setup in “Example user setup for UNIX environments” on page 164.
Operating system user requirements
Because of the way that the InfoSphere Information Server installer parses its
parameters, passwords should not include embedded dollar signs ($).
4.14.3 Domain (WebSphere Application Server) user registry
IBM InfoSphere Information Server users log in and authenticate through the
domain WebSphere Application Server. During install, two domain accounts
must be specified, as listed in Table 4-17.
Table 4-17 Domain accounts
WebSphere Application Server user requirements
WebSphere Application Server has the following user requirements:
During the InfoSphere Information Server installation, WebSphere Application
Server can be configured to authenticate using an internal registry or using
operating system users. This option can be changed later through the
WebSphere Application Server Administration Console.
When using OS authentication, user accounts must be created and activated
before running the InfoSphere Information Server installation.
WebSphere Application Server does not support the use of NIS as a
user registry. The supported user registries are LDAP, OS, or an internal
user registry.
When installing in an LDAP environment, choose an internal user registry.
LDAP authentication is configured after the InfoSphere Information Server
installation.
LDAP Servers supported by InfoSphere Information Server are those
supported by WebSphere Application Server 6.0.2. A complete list of
User account Default user name Notes
WebSphere administrator wasadmin
InfoSphere Information
Server administrator
isadmin Should be different from the
WebSphere administrator
58 IBM InfoSphere Information Server Installation and Configuration Guide
supported LDAP Servers can be found in the list of WebSphere Application
Server software requirements at:
http://www.ibm.com/support/docview.wss?rs=180&uid=swg27007256
Because of the way that the InfoSphere Information Server installer parses its
parameters, passwords should not include embedded dollar signs ($).
4.14.4 Engine (DataStage) user setup
The IBM InfoSphere Information Server engine must have at least one operating
system user defined. DataStage and QualityStage jobs and Information Analyzer
jobs run on the engine server using operating system user permissions.
When using the WebSphere Application Server internal user registry,
InfoSphere Information Server users must be mapped to at least one
operating system user.
When using the WebSphere Application Server OS or LDAP configuration,
InfoSphere Information Server can be configured to share the user registry
with the engine (DataStage/QualityStage) registry. This eliminates the need to
individually map each InfoSphere Information Server user to an operating
system or DataStage user.
DataStage supports four basic categories of users:
Managers
Developers
Operators
Super operators
These are implemented as InfoSphere Information Server roles that can be
assigned to each user. The InfoSphere Information Server Console is used to
assign either the DataStage Admin or DataStage User to each user. This
allows the DataStage Administrator Client to assign each user to a particular
role (operator, super operator, developer, and production manager) for a
particular project.
Engine (DataStage) user setup on UNIX
DataStage can be administered on a UNIX platform by a special non-root user.
This is
dsadm by default, but you can specify a different administrative user at
install. Set up this user before installing DataStage.
Each user is then allocated to the product manager, developer, operator, or super
operator role (but not to more than one role per project). You can then use the
DataStage Administrator to assign the appropriate DataStage user role to the
user. For more information, see "User Roles on UNIX Systems" in the IBM
Chapter 4. Installation and configuration 59
Information Server Administrator Guide, SC18-9929-02. Operators cannot use
the DataStage Designer and only see released jobs in the DataStage Director.
Neither operators nor developers can create protected projects or add anything
to them.
4.14.5 Engine (DataStage) user setup on Windows
On the Windows 2003 Server, DataStage must be installed by a user that has
local administrator rights. This user must also have read/write permission to the
target directory used to install the DataStage server.
If you are logged into a domain account, it must be part of the local
administrators group on the server that you are installing, and you must have
network access to the Windows domain controller for authentication.
4.15 Verifying and installing C++ compiler and runtime
libraries
To develop parallel jobs with DataStage, install the C++ compiler that is specific
to your platform on the same server as the InfoSphere Information Server
engine. In development environments, DataStage uses the C++ compiler to
generate parallel transformer and BuildOp components.
Many compilers are licensed on a per-user basis. DataStage only invokes the
C++ compiler when the developer compiles a parallel job with transformer stages
or compiles a BuildOp component. That is, the maximum number of
simultaneous DataStage developers determines the maximum number of
concurrent C++ compiler licenses.
On each platform, only one C++ compiler is supported. Check the InfoSphere
Information Server system requirements for information about supported
compilers and compiler installation requirements.
IBM InfoSphere Information Server system requirements
For InfoSphere Information Server system requirements, go to:
http://www.ibm.com/support/docview.wss?rs=14&uid=swg21315971#dqx1
windowsdiskspace
More information: See "Setting Up Security" in the IBM Information Server
Administration Guide, SC18-9929-02.
60 IBM InfoSphere Information Server Installation and Configuration Guide
For deployment (production) systems, a C++ compiler is not required. However,
certain platforms might require C++ runtime libraries to be installed. This
information is also contained with the InfoSphere Information Server system
requirements.
4.16 Verifying InfoSphere Information Server connector
requirements
InfoSphere Information Server uses connectors to access source and target data
stores. Each connector (for example, ODBC, DB2, and Teradata) has specific
requirements for setup and configuration that might have platform-specific
steps. Consult the connector documentation and release notes for specific
connector requirements.
4.17 Downloading and installing InfoSphere Information
Server
This section explains the process for downloading and installing InfoSphere
Information Server and required fix packs.
The installation packages for InfoSphere Information Server are available
through IBM Passport Advantage. If you do not already have an account (login)
for Passport Advantage, you will need to create one and make sure that it is
associated with your customer ID to access all downloads for which you are
licensed.
Get IBM Passport Advantage Software Downloads from:
http://www-306.ibm.com/software/howtobuy/passportadvantage/index.html
On certain platforms (specifically, 64-bit installations of RedHat Linux, SUSE
Linux, and HP-UX) you might also need to download WebSphere Application
Server and DB2 packages. You can find details about specific packages required
by platform in the IBM InfoSphere Information Server system requirements at:
http://www-01.ibm.com/support/docview.wss?rs=14&uid=swg21315971#dqx1win
dowsdiskspace
IBM InfoSphere Information Server 8.0.1 installer
IBM InfoSphere Information Server 8.0.1 fix pack installer
If applicable: WebSphere Application Server v6.0.2 installer
If applicable: WebSphere Application Server fix pack installer
Chapter 4. Installation and configuration 61
If applicable: DB2 Enterprise Server Edition v9.1 installer
If applicable: DB2 Enterprise Server Edition v9.1 fix pack installer
In addition, product-specific patches might be required for issues discovered
after the latest fix pack release. Contact IBM Support for details on obtaining
product patches.
4.18 Performing complete system backup
Before beginning an InfoSphere Information Server installation, it is important to
have a complete system backup if the installation fails or you need to bring the
system back to the state it was in before the installation began.
This backup should be performed by the system administrator on all target
installation servers and should include both operating system and user
directories.
4.19 Identifying and configuring file systems
DataStage requires file systems and space to be available for the
following elements:
Software Install Directory
DataStage executables, libraries, and pre-built components
DataStage Project (Repository) Directory
Data Storage
– DataStage temporary storage (scratch, temp, buffer)
– DataStage parallel data set segment files
– Staging and archival storage for any source files
By default, each of these directories (except for file staging) is created during
installation as subdirectories under the base DataStage installation directory.
Important: Each storage class should be isolated in a separate file system to
accommodate its different performance and capacity characteristics and
backup requirements.
62 IBM InfoSphere Information Server Installation and Configuration Guide
The default installation is generally acceptable for small prototype environments.
Figure 4-4 illustrates how you might configure the file systems to satisfy the
requirements of each class of DataStage storage.
Figure 4-4 Suggested DataStage installation, projects, data, and scratch file systems
Notes about Figure 4-4:
All file systems used by DataStage for installation must have read and write
permissions for the primary DataStage users and groups.
The DataStage installation directory should be reserved for installation of
executables and libraries only. For cluster/grid implementations, it is generally
best to share the installation mount point across servers.
Notice that there is a projects file system under the DataStage directory. This
allows optimal control of the DataStage project storage while still maintaining
the default path used by the administrator client. Project naming standards
include the deployment phase (dev, it, uat, prod) as a prefix. Ensuring that the
phase is part of the project directory name will isolate projects in the same
environment. For cluster/grid implementations, it is generally best to share the
projects mount point across servers.
/As centia l
File systems are
highlighted in blue with
dashed outli nes
Ins tall FS
...
...
Scratch File Systems
Data File Systems
Install File System
...
..
......
/Scratch_<phase>0 /Scratch_<phase>N
/Data_<phase>0 /Data_<phase>N
...
/Projects
1 Gigabyte
/patches
/D at aStage
/DSEngine
/PXEngine
/Configurations
/Project_A
/Project_Z
/Project_A
/Pr oject _Z
/P rojec t_A
/P rojec t_Z
/Project_A
/Project_Z
.
/<phase>_Project_A
/<phase>_Project_Z
Where “<phase>” is the deployment
phase (dev, it, uat, prod)
Chapter 4. Installation and configuration 63
Consider creating separate file systems for each scratch and data resource
partition to scale DataStage I/O, naming the file systems in accordance with
the partition numbers in your parallel configuration file. This standard practice
advocates creating subdirectories for each project for each scratch and
disk partition.
File systems should be expandable without requiring destruction
and recreation.
In this document, the base install directory is referred to as $DSROOT, and we
refer to the example installation directory as /opt/IBM/InformationServer/.
4.19.1 Software installation directory
The software installation directory is created by the installation process and
contains the DataStage software file tree. The installation directory grows little
over the life of a major software release. Therefore, the default location ($HOME
for dsadm, for example, /home/dsadm) might be adequate.
The installation of InfoSphere Information Server requires the following minimum
disk space for the listed environments:
1.3 GB for WebSphere Application Server
500 MB for DB2
1.4 GB for the InfoSphere Information Server components
2.5 GB for the metadata repository database
1 GB for the InfoSphere Information Analyzer analysis database
2 GB of temporary space during the installation
For cluster or grid implementations, it is generally best to share the installation
file system across servers (at the same mount point).
4.19.2 DataStage Projects (repository) directory
The DataStage Projects subdirectory contains the repository (universe database
files) of job designs, design and runtime metadata, logs, and components.
Project directories can grow to contain thousands of files and subdirectories
depending on the number of projects, the number of jobs, and the volume of
logging information retained about each job.
Note: The DataStage installer attempts to rename the installation directory to
support later upgrades. If you install directly to a mount point, this rename will
fail, and several error messages will be displayed. The installation will succeed
but the messages might be confusing.
64 IBM InfoSphere Information Server Installation and Configuration Guide
During the installation process, the projects subdirectory is created in the
DataStage install directory. By default, the DataStage Administrator client
creates its projects in this projects subdirectory.
For cluster or grid implementations, it is generally best to share the projects file
system across servers (at the same mount point).
Creating the projects file system
On most operating systems, you can create separate file systems at non-root
levels, which is illustrated in Figure 4-5 on page 67, as a separate file system for
the projects subdirectory within the DataStage installation. Use the following
guidelines:
Create a separate file system and mount it over the default location for
projects, the $DSROOT/Projects directory. Mount this directory after installing
DataStage but before projects are created.
The projects directory should be a mirrored file system with sufficient space
(minimum 100 MB per project).
For cluster or grid implementations, share the project file system across
servers (at the same mount point).
Effective management of space is important to the health and performance of a
project, and as jobs are added to a project, new directories are created in this file
tree, and as jobs are run, their log entries multiply. These activities cause file
system stress. For example, they result in more time to insert or delete
DataStage components and longer update times for logs. Failure to perform
routine project maintenance (for example, remove obsolete jobs and manage log
entries) can cause project obesity and performance issues.
The name of a DataStage Project is limited to a maximum of 54 characters. The
project name can contain alphanumeric characters and underscores (_).
Important: It is not a good practice to create DataStage projects in the default
directory within the install file system, as disk space is typically limited.
Projects should be created in their own file system (Figure 4-5).
Important: Monitor the project file system to ensure that adequate free space
remains. If the project file system runs out of free space during DataStage
activity, the repository might become corrupted, requiring a restore from
backup.
Chapter 4. Installation and configuration 65
Maintain project names in unison with source code control. As projects are
promoted through source control, the name of the phase and the project name
should reflect the version, in the form:
<Phase>_<ProjectName>_<version>
Where Phase corresponds to the phase in the application development life cycle
(Table 4-18).
Table 4-18 Development life cycle
Project recovery considerations
Devising a backup scheme for project directories is based on the following three
core issues:
Will there be valuable data stored in Server Edition hash files?
DataStage Server Edition files located in the DataStage file tree might require
archiving from a data perspective.
How often will the UNIX file system containing the entire DataStage file tree
be backed up? When can DataStage be shut down to enable a cold snapshot
of the universe database and the project files? A complete file system backup
while DataStage is shut down accomplishes this backup.
How often will the projects be backed up? Keep in mind that the grain of
project backups will represent the ability to recover lost work should a project
or a job become corrupted.
At a minimum, a UNIX file system backup of the entire DataStage file tree should
be performed at least weekly with the DataStage engine shut down, and each
project should be backed up with the manager at least nightly with all users
Phase name Phase description
Dev Development
IT Integration test
UAT User acceptance test
Prod Production
Note: The use of Server Edition components in an Enterprise Edition
environment is discouraged for performance and maintenance reasons.
However, if Server Edition applications exist, their corresponding objects
might need to be taken into consideration.
66 IBM InfoSphere Information Server Installation and Configuration Guide
logged out of DataStage. This is the equivalent of a cold database backup and
six updates.
If your installation has valuable information in server hash files, increase the
frequency of your UNIX backup
or write jobs to unload the server files to
external media.
4.19.3 Data set and sort directories
The DataStage installer creates the following two subdirectories within the
DataStage installation directory:
The Datasets/ subdirectory stores individual segment files of DataStage
parallel data sets.
The Scratch/ subdirectory is used by the DataStage framework for temporary
files for such things as sort and buffer overflow.
Try not to use these directories, and consider deleting them to ensure that they
are never used. This is best done immediately after installation, but be sure to
coordinate this standard with the rest of the team.
DataStage parallel configuration files are used to assign resources (such as
processing nodes, disk, and scratch file systems) at run time when a job is
executed. Parallel configuration files are discussed in detail in Chapter 5,
“Parallel configuration files” on page 123.
The DataStage installer creates a default parallel configuration file
(configurations/default.apt) that references the datasets and scratch
subdirectories within the install directory. The DataStage administrator should
consider removing the default.apt file altogether, or at a minimum updating this
file to reference the file systems that you define.
Data and scratch file systems
It is not a good practice to share the DataStage install and projects file systems
with volatile files such as scratch files and parallel data set segment files.
Resource, scratch, and sort disks service different kinds of data with opposite
persistence characteristics. Furthermore, they compete directly with each other
for I/O bandwidth and service time if they share the same path.
Optimally, these file systems should not have any physical disks in common and
should not share any physical disks with other applications. While it is often
impossible or impractical to allocate contention-free storage, it must be noted
that at large data volumes and in highly active job environments, disk arm
contention can and usually does significantly constrain performance.
Chapter 4. Installation and configuration 67
For best performance, and to minimize storage impact on development activities,
create separate file systems for each data and scratch resource partition. This
practice advocates creating subdirectories for each project within each scratch
and disk partition (Figure 4-5).
On clustered and grid environments, data and scratch file systems should not be
shared across servers. Each server contains its own subset of the data.
On systems where multiple phases are shared on the same server, consider
separating data and scratch storage to different file systems for each deployment
phase to completely isolate each environment. Figure 4-5 illustrates this (where
<phase> is dev, it, uat, or prod). Use of the phase name is not required when
environments are not shared on the same system.
Figure 4-5 Suggested data and scratch file systems
Data sets
Parallel data sets are used for persistent data storage in parallel, in native
DataStage format. The DataStage developer specifies the location of the data
set header file, which is a small pointer to the actual data segment files that are
created by the DataStage engine, in the directories specified by the disk
resources assigned to each node in the parallel configuration file. Over time, the
data set segment file directories will grow to contain dozens to thousands of files,
depending on the number of DataStage data sets used by DataStage jobs.
The need to archive data set segment files depends on the recovery strategy
chosen by the DataStage developer, the ability to recreate these files if the data
sources remain, and the business requirements. Whatever archive policy is
chosen should be coordinated with the DataStage administrator and developers.
Note: For optimal performance, create file systems in high-performance,
low-contention storage. The file systems should be expandable without
requiring destruction and recreation.
Data File Systems
..
...
/Data_<phase>0 /Data_<phase>N
...
/Project_A
/Project_Z
/Project_A
/Pr oject _Z
.
...
Scratch File Systems
...
.. .
/Scratch_<phase>0 /Scratch_<phase>N
/Project_A
/Project_Z
/Project_A
/Project_Z
68 IBM InfoSphere Information Server Installation and Configuration Guide
If data set segment files are archived, careful attention should be made to also
archive the corresponding data set header files.
Sort space
As a suggested practice, isolate DataStage scratch space from data sets and flat
files, and DataStage sort space, because temporary files exist only while a job is
running, and they are warm files (that is, they are being read and written at above
average rates). Certain files created by database stages persist after job
completion. For example, the Oracle .log, .ctl and .bad files will remain in the
first scratch resource pool after a load completes.
Sizing DataStage scratch space is somewhat difficult. Objects in this space
include lookups and intra-process buffers. Intra-process buffers absorb rows at
run time when stages in a partition (or all partitions) cannot process rows as fast
as they are supplied. In general, there are as many buffers as there are stages
on the canvas for each partition. As a practical matter, assume that scratch
space must accommodate the largest volume of data in one job. There are
advanced ways to isolate buffer storage from sort storage, but this is a
performance-tuning exercise, not a general requirement.
Maintaining parallel configuration files
DataStage parallel configuration files are used to assign resources (such as
processing nodes, disk, and scratch file systems) at run time when a job is
executed. For more information about parallel configuration files, see Chapter 5,
“Parallel configuration files” on page 123.
Parallel configuration files can be located within any directory that has suitable
access permissions, defined at run time through the environment variable
$APT_CONFIG_FILE. However, the graphical configurations tool within the
DataStage clients expects these files to be stored within the configurations
subdirectory of the DataStage install. For this reason, it is suggested that all
parallel configuration files be stored in the configurations subdirectory, with
naming conventions to associate them with a particular project or application.
The default.apt file is created when DataStage is installed, and references the
datasets and scratch subdirectories of the DataStage install directory. To
manage system resources and disk allocation, the DataStage administrator
Note: The sort space must accommodate only the files being sorted
concurrently, and, assuming that jobs are scheduled non-concurrently, only
the maximum of those sorts. There is no persistence to these temporary sort
files, so they do not need to be archived.
Chapter 4. Installation and configuration 69
should consider removing this file, creating separate configuration files that are
referenced by the $APT_CONFIG_FILE setting in each DataStage project.
At a minimum, the DataStage administrator should edit the default.apt
configuration file to reference the newly created data and scratch file systems
and to ensure that these directories are used by any other parallel
configuration files.
4.19.4 Extending the DataStage project for external entities
Create another directory structure, referred to as Project_Plus, to integrate all
aspects of a DataStage application that are managed outside of the DataStage
Projects repository. The Project_Plus hierarchy includes directories for secured
parameter files, data set header files, custom components, Orchestrate schema,
SQL, and shell scripts. It might also be useful to support custom job logs and
reports.
The Project_Plus directories provide a complete and separate structure in the
same spirit as a DataStage project, organizing external entities in a structure
that is associated with one and only one corresponding DataStage project.
This provides a convenient vehicle to group and manage resources used by
aproject.
It is common for a DataStage application to be integrated with external entities,
such as the operating system, enterprise schedulers and monitors, resource
managers, other applications, and middle ware. The Project_Plus directory
provides an extensible model that can support this integration through directories
for storing source files, scripts, and other components.
Project_Plus and change management
Project naming conventions recommend naming a project with a prefix to
indicate the deployment phase (dev, it, uat, prod). Following this naming
convention will also separate the associated files within the corresponding
Project_Plus hierarchy.
However, to completely isolate support files in a manner that is easy to assign to
separate file systems, an additional level of directory structure can be used to
enable multiple phases of application deployment (development, integration
testing, user acceptance testing, and production) as appropriate. If the file
system is not shared across multiple servers, not all of these development
phases might be present on a local file system.
70 IBM InfoSphere Information Server Installation and Configuration Guide
Project_Plus file system
The Project_Plus directory is often stored in the /usr/local home directory (for
example, /usr/local/dstage), but this can be in any location as long as
permissions and file system access are permitted to the DataStage developers
and applications.
Project_Plus directory structure
Figure 4-6 illustrates typical components and structure of the Project_Plus
directory hierarchy.
Figure 4-6 Project_Plus directory structure
Note: The file system where the Project_Plus hierarchy is stored must be
expandable without requiring destruction and recreation.
as needed
Project_Plus Directory Hierarchy
Project_Plus
/dev
/it
/u at / prod
/b in
/src
/d at asets
/bin
/b in
/bin
/src /src /src
/d oc /d oc /d oc / doc
/sql
/d at asets /datase ts
/datase ts
/l ogs /lo gs
/l ogs / log s
/p arams /p arams
/para ms /param s
/schemas /schemas
/schema s /schemas
/scrip ts /scripts /scrip ts /scripts
/sql
/sq l / sql
/d ev_Pro ject_Z /it_Pro ject_Z /u at _Proj ect_Z /prod_ Proje ct_ Z
/prod_Project_A/u at _Proj ect_A/i t_ Proje ct_ A/dev_Project_A
Chapter 4. Installation and configuration 71
Table 4-19 provides a description of the Project_Plus directory items.
Table 4-19 Project_Plus directory descriptions
Project_Plus environment variables
The Project_Plus directory structure is made to be transparent to the DataStage
application, through the use of environment variable parameters used by the
DataStage job developer. Environment variables are a critical portability tool,
which enable DataStage applications to be deployed through the life cycle
without any code changes.
Directory Description
Project_Plus Top-level of directory hierarchy.
/dev Development phase directory tree (if applicable).
/dev_Project_A Subdirectory created for each DataStage project (the actual
directory name dev_Project_A should match the corresponding
DataStage project name).
/bin Location of custom programs, DataStage routines, BuildOps,
utilities, and shells.
/doc Documentation for programs in /bin subdirectory.
/src Source code and Makefiles for items in /bin subdirectory.
Note: Depending on change management policies, this
directory might only be present in the /dev development phase
directory tree.
/datasets Location of parallel data set header files (.ds files).
/logs Location of custom job logs and reports.
/params Location of parameter files for automated program control, a
backup copy of dsenv, and backup copies of
DSParams:$ProjectName project files.
/schemas Location of Orchestrate schema files.
/it Integration test phase directory tree (if applicable).
/uat User acceptance test phase directory tree (if applicable).
/prod Production phase directory tree (if applicable).
72 IBM InfoSphere Information Server Installation and Configuration Guide
In support of a Project_Plus directory structure, configure the user-defined
environment variable parameters (Table 4-20) for each project using the
DataStage Administrator, substituting your Project_Plus file system and project
name in the value column:
Table 4-20 User-defined environment variables
Name Type Prompt Value
PROJECT_PLUS_DATASETS String Project +
dataset
descriptor dir
Project_Plus/devProject_
A/datasets/
PROJECT_PLUS_LOGS String Project + log
dir
/Project_Plus/devProject
_A/logs/
PROJECT_PLUS_PARAMS String Project +
parameter
file dir
Project_Plus/devProject_
A/params/
PROJECT_PLUS_SCHEMAS String Project +
schema dir
Project_Plus/devProject_
A/schemas/
PROJECT_PLUS_SCRIPTS String Project +
scripts dir
/Project_Plus/devProject
_A/scripts/
Note: The Project_Plus default values include a trailing directory separator to
avoid having to specify in the stage properties. This is optional, but whatever
standard the administrator chooses, it should be set and consistently deployed
across projects and job designs.
Chapter 4. Installation and configuration 73
Figure 4-7 depicts the Project_Plus environment variables.
Figure 4-7 Project_Plus environment variables
Using Project_Plus with grid or cluster deployments
When deploying a DataStage application in cluster or grid environments, or
when configuring for high-availability and disaster recovery scenarios,
careful consideration should be made when sharing the Project_Plus file
system configuration.
In general, the custom components, data set header files, and other components
of the Project_Plus directory should be visible to all members of the cluster or
grid, using the same mount point on all servers. Creation of small individual
mounts points is generally not desirable.
Mount this directory on all members of the cluster after installing DataStage, but
before creating any DataSets.
74 IBM InfoSphere Information Server Installation and Configuration Guide
4.19.5 File staging
Use a separate staging file system and directory structure to store, manage, and
archive various source data files (Figure 4-8).
Figure 4-8 DataStage staging directories
Project naming conventions recommend naming a project with a suffix to indicate
the deployment phase (dev, it, uat, prod). Following this naming convention will
also separate the associated files within the corresponding staging hierarchy.
However, to completely isolate support files in a manner that is easy to assign to
separate file systems, an additional level of directory structure can be used to
enable multiple phases of application deployment (development, integration test,
user acceptance test, and production), as appropriate. If the file system is not
shared across multiple servers, not all of these development phases might be
present on a local file system.
In support of the staging directory structure, the user-defined environment
variable parameters (Table 4-21) should be configured for each project using the
DataStage Administrator, substituting your staging file system and project name
in the value column.
Table 4-21 User-defined environment variable parameters
Name Type Prompt Example value
STAGING_DIR String Staging directory /Staging/
PROJECT_NAME String Project name devProject_A
DEPLOY_PHASE String Deployment phase dev
as needed
/dev
/it
/u at / prod
/dev_Project_Z /it_Project_Z /uat_Project_Z /prod_Project_Z
/prod_Project_A/u at_Proj ect_A/i t_ Proje ct_ A/dev_Project_A
/Staging
/archive /archive /archive /archive
/archive /archive
/archive
/a rchiv e
.
.
.
.
.
.
.
.
.
.
.
.
Chapter 4. Installation and configuration 75
The Project_Name and Deploy_Phase variables are used to properly
parameterize the directory location within job designs.
Within each deployment directory, files are separated by project name
(Table 4-22).
Table 4-22 Deployment directory files
4.19.6 File system sizing example
File systems should be expandable to accommodate growth from prototype to
full-scale development to test and deployment (as applicable) without requiring
destruction and recreation.
Accurate capacity planning requires detailed requirements for persistence,
failover and archive, growth volumes, and projections across all applications to
be deployed in the environment.
An effective method of specifying the storage requirements is a table such as
Table 4-23 on page 76. In this example, two logical processing nodes are
Note: The STAGING_DIR default value includes a trailing directory separator
to avoid having to specify in the stage properties. This is optional, but
whatever standard the administrator chooses, it should be set and
consistently deployed across projects and job designs.
Directory Description
Staging Top-level of directory hierarchy
/dev Development phase directory tree (if applicable)
/dev_Project_A Subdirectory created for each DataStage project (The actual
directory name dev_Project_A should match the corresponding
DataStage Project Name.)
Location of source data files, target data files, and error and reject
files
/archive Location of compressed archives created by archive process of
previously processed files
/it Integration test phase directory tree (if applicable)
/uat User acceptance test phase directory tree (if applicable)
/prod Production phase directory tree (if applicable)
76 IBM InfoSphere Information Server Installation and Configuration Guide
configured to process 100 GB of raw data. Our space calculation is per logical
node, so for data, the formula would be 100 x 2.5 / 2 or 125 GB, assuming that
the data is fairly evenly distributed between nodes. We describe our
requirements in terms of contention objectives wherever possible.
Table 4-23 Storage requirements
File system
type
Size
(GB)
Mount point Use and contention requirement
Data set 125 /RaidVol1/Data0 One-half the volume of DataStage DataSets. This file
system should be expandable without removal and
recreation, should not share physical disks with other
DataSets or database files, should be highly
available, and should have as many physical disks as
possible, with a preferred minimum of 4.
Data set 125 /RaidVol1/Data1 Same as above.
Scratch 125 /RaidVol1/Scratch0 One-half the volume of DataStage scratch (working)
space. This file system should be expandable without
removal and recreation, should not share physical
disks with other DataSets or database files, should be
highly available, and should have as many physical
disks as possible, with a preferred minimum of 4.
Scratch 125 /RaidVol1/Scratch1 Same as above.
Install 2 /usr/dsadm DataStage software. This space must be highly
available.
Projects 2 /usr/dsadm/Ascential/
DataStage/Projects
Project_Plus 1 /export/dsadm DataStage utilities and shared files. This space must
be highly available.
Staging 125 /datavol1 Source data. This file system should be expandable
without removal and re-creation, should be highly
available, and should have as many physical disks as
possible, with a preferred minimum of 4.
Chapter 4. Installation and configuration 77
4.20 Connectivity and network configuration
DataStage relies on network infrastructure for communication between:
DataStage clients and DataStage server
DataStage server nodes and source/target data stores (for example,
databases, file through FTP or named pipes, third-party applications such as
SAS, Siebel, PeopleSoft, and SAP)
Across DataStage server nodes in a clustered or grid configuration
For optimal performance, each of these network connections should be on
high-speed connections with low latency, particularly the server-to-server
network connections for clustered/grid configurations and access to remote
sources/targets.
4.20.1 Network port usage
DataStage requires access to a number of TCP ports, which must be opened (if
necessary, through firewall configuration). Related connectivity (for example,
databases and enterprise applications) might require additional port
assignments. Consult vendor-supplied documentation for their requirements.
Further details on firewall configuration can be found in “Windows XP Service
Pack 2 firewall configuration” on page 147 and “Network ports used by
InfoSphere Information Server” on page 162.
Table 4-24 lists ports with descriptions.
Table 4-24 Ports and descriptions
Port Use Description or notes
31538 DSRPC daemon
(client/server) connectivity
Default for v7.0 and later.
>= 10,000 Conductor/section leader
messages
At job startup, tries to allocate port starting from
$APT_PM_STARTUP_PORT (default 10,000) until an
unused port is located.
>= 11,000 Player - player data transfer On cluster/grid configurations, tries to allocate a range of
ports starting at $APT_PLAYER_CONNECTION_PORT
(default 11,000) based on unused ports. Number of ports
required varies based on job complexity (number of
stages) and degree of parallelism.
2000 RTI Agent Static port assignment (default).
78 IBM InfoSphere Information Server Installation and Configuration Guide
Network port usage on UNIX System Services platforms
On UNIX System Services platforms, if you are using an rsh connection (a tightly
coupled environment) and want to collect process metadata via the job monitor,
you will need to assign two TCP ports to DataStage.
The default numbers for these ports are 13400 and 13401. The numbers are
specified at installation time and are easily modified afterwards.
4.20.2 UNIX NIS configuration
If you are using NIS, the master /etc/services file must be updated manually to
add the following entry:
dsrpc 31538/tcp
The local /etc/passwd file must be updated to include a dummy entry for the
DataStage administration user (for example, dsadm) for the installation to work.
>1024
assigned
by OS
RTI Agent Dynamic Port Assignments. For example, RTI Agent
opens a port to allow notification messages to be sent
from DataStage Jobs. When an RTI Server sends data to
the RTI Agent, both have additional ports open to handle
that exchange.
1099 and
8083
RTI - Jboss Application Server EJB Server.
1476 RTI - Jboss Application Server Internal Database Server.
8090 and
8091
RTI - Jboss Application Server JMS Services.
4444 RTI - Jboss Application Server JMX Service.
8080 RTI - Jboss Application Server HTTP.
2379 MetaStage - Listener Default (can be configured).
4379 MetaStage - Process
Metabroker
Default (can be configured).
9090 MetaStage - Browser Default (can be configured).
Port Use Description or notes
Chapter 4. Installation and configuration 79
4.20.3 Windows network configuration
On Windows 2003 Server environments, the default port assignments for the
Microsoft Telnet Server are the same as those used by the DataStage Telnet
server and the MKS Telnet server. This can cause the DataStage install to hang
when starting its services.
On Windows 2003 systems, you should reconfigure these services so
that the ports do not overlap. Or you can selectively disable startup of the
conflicting services:
MKS Telnet Server is not needed or used by DataStage.
DataStage Telnet Server is only used for remote debugging of the internal
DataStage repository engine, and is not necessary for normal operation.
4.21 Configuring OS users, groups, and associated
permissions
Before beginning a DataStage installation, operating system groups and users
must be selected or implemented. DataStage supports three basic categories of
users:
Managers
Developers
Operators
These are implemented as the secondary group for each category of user. That
is, the primary group ID (GID) of every DataStage user must be the same as the
DataStage super-user, and the secondary GID of a user in one of these
categories must be one of these groups. The DataStage Client Administrator
function is used to assign roles to those users who are members of each group.
Security is best supported when there are four operating system groups
associated with DataStage (Table 4-25).
Table 4-25 DataStage groups
Group Function
dstage The primary group of the DataStage super-user. By convention, this user
is dsadm.
dsdev The secondary group of DataStage developers. No users in this group are
members of any other DataStage group.
80 IBM InfoSphere Information Server Installation and Configuration Guide
Minimally, the only group that must be created is the primary group of the
DataStage super-user. This group and the super-user (dsadm) are used to
secure files and perform software maintenance.
An effective method of specifying the groups and users is a table, such as the
one depicted in Table 4-26. On our example system, users and groups in bold
already exist. The remaining groups and users are to be created. We describe
our requirements in terms of business functions wherever possible and identify
employees, contractors, and virtual users, such as dsadm.
Table 4-26 Groups and users
dsmgr The secondary group of DataStage managers. No users in this group are
members of any other DataStage group.
dsopr The secondary group of DataStage operators. Typically, this group is the
primary group for schedule management software user IDs, for example,
asys. No users in this group are members of any other DataStage group.
Note: The DataStage administrator must have unrestricted access to the
super-user (dsadm).
Group Function
Primary group Secondary
group
User Role
dstage dsmgr dsadm DataStage software and project manager. This user ID
manages the software and project population and must
exist on all systems (development, test, and production
systems). The password will be closely controlled.
dsmgr johnl DataStage project manager. This user is an employee
who manages the development and testing project
environments and is responsible for project backup and
recovery, job migration to and from development, source
control, and production, and must exist on both
development and production systems.
dsmgr paulm Paul is an employee who is John's back up. He has the
same roles as John, and must have access to both
production and development systems.
dstage dsopr,
autosys
asys This user ID runs the jobs in the system integration and
production environments. Autosys is a popular job
scheduler for UNIX and is used as an example.
Chapter 4. Installation and configuration 81
4.21.1 UNIX user configuration
DataStage can be administered on a UNIX platform by a special non-root user,
which is dsadm by default. However, you can specify a different administrative
user at installation. You must set up this user before installing DataStage. All
DataStage users should belong to the same UNIX group, and this should be
the administrative user’s primary group. We suggest that you name the
group
dstage.
If you want to set up the system so that it distinguishes between product
managers, developers, and operators, set up secondary groups for each class of
user. Each user is then allocated to the product manager, developer, or operator
secondary group (but not to more than one secondary group). You can then use
the DataStage administrator to assign the appropriate DataStage user role to the
secondary groups. For more information, see “User Roles on UNIX Systems” in
the DataStage Administrator Guide, SC18-9929-02.
Operators cannot use the DataStage Designer and only see released jobs in the
DataStage Director. Neither operators nor developers can create protected
projects or add anything to them. This configuration option requires that
DataStage run in impersonation mode. This is the default for a root install but
needs to be manually configured for a non-root install.
On UNIX installations of DataStage, you can, if required, use an authentication
mechanism other than the standard UNIX one. To this end, DataStage supports
pluggable authentication modules (PAM). This provides a way of keeping
underlying authentication technologies separate from application code, thus
eliminating the need to update the application every time that the authentication
mechanism is changed. To use PAM authentication, configure DataStage after
dstage dsdev,
users
georgeh DataStage developer. This user is an employee who
needs access to the development and test systems.
dsdev,
users
richards DataStage developer. This user is an employee who
needs access to the development and test systems.
dsdev,
users
mannie DataStage developer. This user is a contractor who needs
access to the development and test systems.
dsdev,
users
moe DataStage developer. This user is a contractor who needs
access to the development and test systems.
dsdev,
users
jack DataStage developer. This user is a contractor who needs
access to the development and test systems.
Primary group Secondary
group
User Role
82 IBM InfoSphere Information Server Installation and Configuration Guide
you install. For more information, see “Configuring for Use with PAM” in the
DataStage Install and Upgrade Guide, Part No. 00D-018DS60.
4.21.2 Windows user configuration
On Windows 2003 Server, DataStage must be installed by a user who has local
administrator rights. This user must also have read/write permission to the target
directory used to install the DataStage server.
If you are logged into a domain account, it must be part of the local
administrator’s group on the server that you are installing, and you must have
network access to the Windows domain controller for authentication.
The administrative user that you are logged into during installation of DataStage
will become the owner, or administrator, of the DataStage installation.
After installation of DataStage, you can use the Windows user management
tools to define groups of associated users to restrict access to individual
DataStage projects. By default, every login user (the
everyone group) is given
access to newly created projects. However, you can use the Windows user
management tools to create multiple Windows groups with assigned users, and
then use the DataStage Administrator to restrict access on a per-project basis.
For more information, see “User Roles on UNIX Systems” in the DataStage
Administrator Guide, SC18-9929-02.
4.22 C++ compiler and runtime library requirements
To develop and deploy DataStage parallel jobs, you need to install the C++
compiler that is specific to your platform. Every DataStage development server
must have this specific C++ compiler installed (in order to compile BuildOp
components or parallel jobs with transformer stages).
For deployment (production) systems, a C++ compiler is not required. However,
certain platforms might require that C++ runtime libraries be installed.
Always double-check the specific C++ compiler versions and any associated
patches that are required for your specific operating system platform. This
information is contained in the DataStage release notes.
Many compilers are licensed on a per-user basis. DataStage only invokes the
C++ compiler when the developer compiles a parallel job with transformer stages
or compiles a BuildOp component. That is, the maximum number of
Chapter 4. Installation and configuration 83
simultaneous DataStage developers determines the maximum number of
concurrent C++ compiler licenses.
4.22.1 Development systems
To develop parallel jobs, you need the C++ compiler specific to your platform.
When installing the C++ compiler for your machine, ensure that all packages
are installed.
Table 4-27 lists the supported compilers.
Table 4-27 Supported C++ compilers
AIX compiler configuration
An IBM VisualAge® 6 include file might cause some DataStage Paralle
Transformer jobs to fail to compile with the following message or similar:
"/usr/vacpp/include/stdlib.h", line 123.13: 1540-0040 (S) The text
"undef llabs" is unexpected. "undef" may be undeclared or ambiguous.
Important: Only the following compilers and versions are compatible with
DataStage. IBM certifies DataStage for specific compiler releases for a
given platform.
Platform Required C++ compilers
AIX Visual Age 6, XL C++ Enterprise Edition v7.0
or v8.0
HP-UX (Itanium) HP Itanium ANSI C++ 6.0
HP-UX (PA-RISC) HP ANSI C++ A.03.63
LINUX - Red Hat AS 4 32 bit GCC 3.4
LINUX - Red Hat AS 4 64 bit GCC 3.4.2
LINIX - Red Hat AS 5 32 bit and 64 bit GCC 4.1.2
LINUX - SUSE 9.0 kernel 2.6 GCC C++ 3.3.3
Solaris Sun Studio 10, 11, 12
Windows 2003 Server Microsoft Visual Studio.NET 2003 (The
standalone Microsoft C++ compiler does not
work. You must install the complete Visual
Studio.NET 2003.)
84 IBM InfoSphere Information Server Installation and Configuration Guide
This is as per IBM Case 25168:
http://www.developer.ibm.com/tech/faq/individual?oid=2:25168
The solution is to edit the /usr/vacpp/include/stdlib.h file, by looking for line
123:
undef llabs
Replace line 123 with the following text:
#undef llabs
If you are using the DataStage Parallel pallet and therefore using the IBM
VisualAge C++ compiler, you must make sure that the appropriate runtime
libraries for the specific version of the compiler are installed on all systems to
which parallel jobs will be deployed.
The VisualAge C++ for AIX V6.0 Runtime Libraries can be downloaded from the
IBM website at the following address:
http://www.ibm.com/support/docview.wss?rs=0&uid=swg24001467
Red Hat LINUX compiler configuration
Be aware of the following compiler configurations:
32-bit RedHat Enterprise Linux Advanced Server 4 requires compiler 3.4.
64-bit RedHat Enterprise Linux Advanced Server 4 requires compiler 3.4.2.
32-bit and 64-bit RedHat Enterprise Linux Advanced Server 5 require
compiler 4.1.2.
To determine the version of the compiler that is currently configured as the
default, run the following command:
g++ --version
4.22.2 Deployment systems
For systems where you are only required to run pre-compiled parallel jobs, a full
C++ compiler is not required, but certain runtime libraries are required.
Platforms not mentioned in this section already include the required runtime
libraries as part of the operating system, and no further action is required (as
long as the runtime libraries have not been explicitly removed).
Chapter 4. Installation and configuration 85
AIX 5.1 and 5.2 C++ Runtime Libraries
The VisualAge C++ Version 6 Runtime Libraries for AIX 5.x. The runtime library
filesets for AIX 5.x can be installed on AIX 5.x using the AIX system
administrative tool smit. The libraries are available from the March 2003 PTF at:
http://www.ibm.com/support/docview.wss?rs=0&q=xlC.rte&uid=swg24004427&l
oc=en_US&cs=utf-8&cc=us&lang=en
Red Hat LINUX Runtime Libraries
On Red Hat LINUX, the legacy runtime libraries can be installed separately from
the install CDs. Insert Install CD1 and accept the prompt to autorun. Under the
Development category select the Legacy Software Development sub-option.
Continue with Forward to install the libraries.
4.23 Checking product release notes
With any release of DataStage always check the release notes, which are
included on the server installation CD in the readme directory. Use the
server-installation media, because the client installation might have an older
release note directory.
4.24 Installing DataStage/Parallel Framework
When prerequisites are complete, installation of DataStage is a straightforward
process, as documented in the DataStage Install and Upgrade Guide, Part No.
00D-018DS60. As explained in this section, you must consider several options
before running the server installation.
4.24.1 Installing multiple DataStage Servers on UNIX
Starting with release 7.5.1, DataStage supports installing multiple DataStage
installations on a single UNIX server. Multiple server installations are not
supported on Windows 2003 server environments.
For information about installing, upgrading, and managing multiple server
instances, see “Installing and configuring multiple server instances” on page 130.
86 IBM InfoSphere Information Server Installation and Configuration Guide
4.24.2 Installing plug-ins
DataStage provides a number of plug-ins that are used to provide connectivity to
particular database and data sources, and to perform other special functions.
The main install program offers you a selection of plug-ins. In general, unless you
are completely confident of your future connectivity needs, it is a good idea to
select all plug-ins on the installation. You can determine the available plug-ins by
examining the packages directory on the install media.
If you do not select all the required plug-ins at initial install, you can rerun the
install in maintenance mode to select additional ones (see "Reinstalling and
Maintenance Menu" in the DataStage Install and Upgrade Guide, Part No.
00D-018DS60). You can also download plug-ins from the web and install them
using the package installer. See "Installing DataStage Packages" in the
DataStage Administrator Guide, SC18-9929-02, for details.
4.24.3 UNIX install requirements
DataStage installation requires that the user performing the installation be logged
into the system as
root, unless a non-root install is being performed.
The DataStage installation requires that the administration user (
dsadm by
default) is configured to run with the
sh or ksh shell. Using an alternative shell
such as
csh will cause failures during the installation.
If the installation program detects plug-ins that were installed at a previous
release, it upgrades them automatically. If the installation program cannot detect
plug-ins that were installed (this is typically because they have been installed by
the dspackinst command rather than by the server installation program), these
plug-ins will not be upgraded automatically. These plug-ins can still be upgraded
by the installation program, but the user must select them specifically.
Back up existing DataStage projects before attempting to upgrade DataStage.
We suggest that you use the DataStage Manager client to export all job designs
and definitions from each project. This can be achieved by selecting the Whole
project radio button in the Export dialog when exporting jobs.
Important: Only the plug-ins that are installed by the DataStage Installer
(install or maintenance mode) are automatically upgraded with future
releases. Also certain plug-ins (especially the PACKs) are not automatically
installed into new projects, even if they were installed prior to project creation.
Important: If you are already using GCI, contact IBM Support.
Chapter 4. Installation and configuration 87
4.24.4 Windows installation requirements
DataStage installation on Windows 2003 Server requires that the user
performing the installation be logged into the server as the local administrator, or
a domain user with local administrator rights for the server.
When installing DataStage Server on Windows 2003 Server you must have a
valid network connection when attached to a domain. If your primary domain
controller (PDC) cannot be found across the network, it causes setup to fail when
licensing the DataStage engine. If you experience problems, install the product
by using a local admin on the machine, or be sure that you are connected to the
network. Workgroup users are not affected.
There are known issues when using DataStage on Windows 2003 and also using
InfoSphere Information Server 6.0. Symptoms include clients experiencing
disconnections and connection error 81022, and occasionally these can occur
when InfoSphere Information Server 6.0 is not running. If you are experiencing
these symptoms, using either of the following workarounds prevents them from
occurring:
Set the DSRPC service to Interact with the Desktop, specified in the Log On
properties of the DSRPC Service in the Services administrative control panel.
Set Microsoft InfoSphere Information Server 6.0 to run in InfoSphere
Information Server 5.0 isolation mode.
The InfoSphere Information Server 5.0 isolation mode is the preferred
workaround, but if InfoSphere Information Server 6.0 features are required, then
Interact with Desktop mode must be used.
4.25 Verifying the installation log file
On UNIX Systems, always verify the installation log file after a DataStage
installation. It might containt error messages that are not captured by the
installer. Although it is always a good idea to review the DataStage installation
log file, if the UNIX installer appears to hang (that is, if a long list of periods is
printed on the console but no progress appears to be made), the installation log
file most likely has an error condition.
Text output generated during the installation process is written to a disk-based
log file in the /tmp/dsinstall/logfiles directory. Each log file contains
information specific to a particular installation instance. Log files are generated
for both clean and upgrade installations, as well as maintenance mode functions.
These log files can be viewed during the installation process using standard
UNIX tools, such as tail.
88 IBM InfoSphere Information Server Installation and Configuration Guide
Log file names are of the form dsinstall.log.MMDDYY.HHMMSS, where:
MMDDYY equals the two-digit month, day, and year.
HHMMSS equals the two-digit hour, minute, and seconds in a 24-hour format.
Consider the following example:
dsinstall.log.082201.162431
Using standard UNIX utilities to examine the install log file, perform a
case-insensitive search for keywords such as
error and warning.
4.26 Installing DataStage patches
In general, install the official maintenance release (distributed on CD). If you
encounter a problem, contact IBM InfoSphere Information Server support to see
whether a resolution (for example, a workaround or patch) is available.
In certain instances, when technical issues are discovered after a release of
DataStage, patches might be available to correct the problems before the next
official maintenance release (which includes fixes for issues discovered after the
previous release).
Patches are specific to the version of DataStage for a particular platform
(OS/hardware) and are available through IBM InfoSphere Information Server
support. A patch is identified by an
eCase (engineering case) number.
Because there is no automated delivery or update mechanism for patches to
DataStage, maintain a list of patches provided to you by IBM support for future
installations of a specific release.
Patch list for InfoSphere Information Server 8.1
Fix packs, patches, and refreshed install packages might be available for
InfoSphere Information Server and its components. Periodically check the
following web page for the current listing:
http://www.ibm.com/software/data/infosphere/support/info-server/
download.html
Important: The DataStage installer creates temporary files and stores
settings in the /tmp/dsinstall directory. If you need to restart a failed
installation, it is best to remove this directory before re-running the
DataStage installer.
Chapter 4. Installation and configuration 89
4.27 Installing and configuring optional components
DataStage provides additional connectivity and functionality through new
stages, which can be provided through a number of APIs (such as plug In,
BuildOp, and Operators), and Packaged Application Connectivity Kits
(DataStage
PACKs).
Each component type provides different ways of installation and configuration. If
you have purchased or built any optional components, consult their
documentation for installation and configuration.
4.28 Configuring post-installation operating system
settings
Certain operating systems (notably Windows 2003 Server) require additional
configuration after DataStage has been installed. Where applicable, these
post-install configurations are detailed in this section.
4.28.1 Securing JobMon ports
Entries should be made in the /etc/services file to protect the sockets used by
the job monitor. The default socket numbers are 13400 and 13401, and entries in
this file might look similar to the following:
13400 tcp dsjobmon
13401 tcp dsjobmon
These entries prevent use of these sockets by applications other than the
job monitor.
4.28.2 Post-installation configuration of Windows 2003 Server
Windows 2003 Server uses different security policies from those used by other
Windows platforms. To allow access to a DataStage server on this platform,
additional setup steps are required, as outlined below.
You must be logged on as an administrator to be able to set the following access
and permissions.
Note: Install DataStage first, and then perform the post-installation
configuration as outlined in this section.
90 IBM InfoSphere Information Server Installation and Configuration Guide
Altering the account for DataStage services
By default a Windows server runs the DataStage services using the LocalSystem
account, but you can set up another user to run the services if required. The
DataStage services are:
DSRPC
DataStage engine resource
DataStage Telnet
To alter the user associated with these services:
1. Open the Services dialog box in the control panel.
2. In the Properties dialog box for each of the DataStage services, go to the Log
On tab.
3. Under Log on as, select the This account option to specify the user and
supply the password.
The user that you choose must have the following privileges:
Log on local.
Act as part of the OS.
Replace a process level token.
Create a token object.
Two sets of instructions are given below, depending on whether the Windows
2003 Server is a domain controller.
Configuring the Nondomain Controller Windows 2003 Server
post-installation
To configure the Nondomain Controller Windows 2003 Server after installation:
1. Allow log on locally:
a. In Explorer, select Control Panel Administrative Tools.
b. Start Local Security Policy.
c. Select Local Policies User Rights Assignment.
d. Select Allow Log on Locally.
e. Select Action Properties from the menu.
f. Select Add User or Group.
g. Click Locations, the select your local machine.
h. Click OK.
i. Click Advanced.
j. Click Find Now.
k. Click Authenticated Users, then select OK.
l. Click OK.
m. Click OK.
n. Exit the Local Security Policy.
Chapter 4. Installation and configuration 91
2. Create group:
a. In Explorer, select Control Panel Administrative Tools.
b. Start Computer Management.
c. Select System Tools, Local Users & Groups.
d. Select Groups.
e. Select Action New Group from the menu.
f. Enter a name for the group (for example, DataStage Users).
g. Click Create.
h. Click Close.
3. Add users:
a. In Explorer, select Control Panel then Administrative Tools.
b. Start Computer Management.
c. Select System Tools Local Users & Groups.
d. Select Groups in the tree.
e. Select the required group (for example, DataStage Users).
f. Select Action Add To Group from the menu.
g. Select Add.
h. Select Locations.
i. Select your local machine name, then click OK.
j. Select Advanced.
k. Select Find Now.
l. Select users to be added to the group, including authenticated users, then
click OK.
m. Click OK.
n. Click OK.
o. Close Computer Management.
4. Set permissions on the DataStage folder:
a. In Explorer, locate the DataStage solder (for example,
c:\Ascential\DataStage)
b. Select File Properties from the menu.
c. Select the Security tab, then click Add.
d. Select Locations.
e. Select your local machine name, then click OK.
f. Select Advanced.
g. Select Find Now
.
h. Select your group name (for example, DataStage users).
i. Click OK.
j. Click OK.
k. Select your group.
l. Check the Modify and Write check boxes in the Allow column.
m. Click OK.
92 IBM InfoSphere Information Server Installation and Configuration Guide
Configuring the Domain Controller Windows 2003 Server
post-installation
To configure the Domain Controller Window 2003 Server after installation:
1. Allow log on locally:
a. In Explorer, select Control Panel Administrative Tools.
b. Start the Domain Security Policy.
c. Select Local Policies User Rights Assignment.
d. Select Allow Log on Locally.
e. Select Action Properties from the menu.
f. Select Add User or Group.
g. Select Browse.
h. Select Advanced.
i. Click Find Now.
j. Select Authenticated Users, then click OK.
k. Click OK.
l. Click OK.
m. Click OK.
n. Exit the Domain Security Policy.
Repeat these steps for the Domain Controller Security Policy application.
2. Create a group.
It is not possible to add the
built-in authenticated users group to a group that
we create in steps 2 and 3, so you might prefer to skip to step 4 and use the
authenticated users group directly.
a. In Explorer, select Control Panel Administrative Tools.
b. Start Active Directory and Computers.
c. Select Users in the current domain.
d. Select Action New Group from the menu.
e. Enter a name for the group (for example, DataStage Users).
f. Leave the Group scope as Global and Group type as Security.
g. Click OK.
3. Add users:
a. In Explorer, select Control Panel, Administrative Tools.
b. Start Active Directory and Computers.
c. Select Users in the current domain.
d. Select the required group (for example, DataStage Users).
e. Select Action Properties from the menu.
f. Select the Members tab.
g. Click Add.
h. Click Advanced, then click Find Now.
i. Select users to add to the group (authenticated users not available).
j. Click OK.
Chapter 4. Installation and configuration 93
k. Click OK.
l. Click OK.
m. Close the application.
4. Set permissions on the DataStage folder:
a. In Explorer, locate the DataStage folder (for example,
c:\Ascential\DataStage).
b. Select File Properties from the menu.
c. Select the Security tab, then click Add.
d. Click Advanced.
e. Click Find Now.
f. Select a group name, such as DataStage Users or Authenticated Users.
g. Click OK.
h. Click OK.
i. Select your group.
j. Check the Modify and Write check boxes in the Allow column.
k. Click OK.
Cluster or grid configuration
In addition to single-server (SMP) installations, DataStage also supports
clustered deployment of parallel jobs, allowing a single parallel job to run across
multiple servers. Clustered deployment is supported for UNIX, LINUX, and
Windows releases of DataStage Parallel Engine, but not z/OS Edition.
In a clustered scenario, one server is designated the primary or
conductor node.
This is the node that (typically) DataStage clients connect to, and it is also where
the Server Edition engine, repository engine, and job monitor components run.
Grid deployments take the clustered configuration one step further, offering
totally dynamic configuration and deployment of clustered DataStage parallel
jobs. A grid deployment requires the DataStage Grid Toolkit. Its configuration
and setup are described in the accompanying documentation.
In a clustered configuration, the DataStage Server Edition engine, repository
engine, parallel engine, and job monitor must be installed on at least one server,
which typically serves as the conductor node at run time. The remaining servers
only need the parallel engine components.
Note: In a clustered configuration, only parallel components can run across
machines. Because Server Edition components run only on the single
conductor node, it is better not to include Server Edition components in
scalable parallel job designs.
94 IBM InfoSphere Information Server Installation and Configuration Guide
Clustered configuration of DataStage
The clustered configuration of DataStage requires the following:
All servers in the cluster must be running on the same hardware platform and
operating system configuration. You cannot run a single parallel job across a
combination of UNIX and Windows servers, nor can you run a single parallel
job across different UNIX operating systems, or even different release levels
of the same operating system.
DataStage parallel framework is available on each server in the cluster. It can
be installed on each server in the cluster, or installed on a shared file system
mount point with the same absolute path on all servers.
DataStage parallel job components are available across all servers in the
cluster. Depending on your situation, one of the following will apply:
– The DataStage project directory is located on a shared file system mount
point with the same absolute path on all servers.
– The environment variable $APT_COPY_TRANSFORM_OPERATOR is
set on the first job run for each job, to copy compiled transform objects:
Any compiled BuildOp objects and custom components are copied into
the LIBPATH (LD_LIBRARY_PATH or SHLIB_PATH) of each node on
the cluster.
– If the BuildOp components were defined and compiled within DataStage
client tools, then the project is on a shared mount point across servers.
DataStage is configured to log in to all servers in the cluster (from each
direction) using rsh or ssh without requiring a password.
Users (and their corresponding groups) that will be running jobs must be
created on all servers in the cluster, and must have privileges to run rsh (or
ssh) to each remote server in the cluster.
A parallel configuration file is created that specifies node resources for each
server (
fastname) in the cluster.
4.28.3 UNIX cluster configuration
In this section we discuss and describe the UNIX cluster configuration.
Specifying the remote shell
To find rsh on a processing node, the DataStage Parallel engine searches for the
following executables in the order shown:
1. $APT_ORCHHOME/etc/remsh (if it exists)
2. /user/lpp/ssp/rcmd/bin/rsh (AIX only)
3. /usr/ucb/rsh
Chapter 4. Installation and configuration 95
4. /usr/bin/remsh
5. /bin/remsh
6. /usr/bin/rsh
Where $APT_ORCHHOME is the top-level directory of your parallel engine
installation.
If the parallel engine does not find your rsh command, you must specify its
location. To do so, copy or rename the supplied file:
$APT_ORCHHOME/etc/remsh.example to install_dir/etc/remsh
This file contains the following shell script:
#!/bin/sh
# Example apt/etc/remsh
exec /usr/bin/rsh "$@"
As written, this shell script invokes /usr/bin/rsh. Edit the last line of this script to
invoke your specific remote shell command. The script should be executable by
all users. Use chmod to ensure that it is:
chmod 755 script-filename
Test this by running rsh on each node:
rsh nodename uptime
Allowing user execution of RSH without a password
This process is performed differently according to the type of system that you are
running. For example, you can either edit /etc/hosts.equiv or create a .rhosts
file for each user. In both cases, add the host name of each parallel processing
node to /etc/hosts.equiv or .rhosts, one host name per line. The host name
that is included in this file must correspond to the setting of the node's fastname
parameter in the parallel configuration file. For information about the fastname
configuration option, see the section "Node Names" in the DataStage Parallel
Job Developer's Guide, LC18-9891.
If you choose to edit the /etc/hosts.equiv file, the file must be owned by root
and must grant read/write access to root and no access to any other user (file
mode of 600).
If you choose to create an .rhosts file for each user, it must meet the following
criteria:
Be located in the home directory of each parallel user.
Be owned by the user.
96 IBM InfoSphere Information Server Installation and Configuration Guide
Grant read/write access to the user and no access to any other user (file
mode of 600).
To check that users can use rsh without a password, issue the following
command on each node:
rsh hostname uptime
Here hostname is the name of a processing node that you use with the parallel
engine. If hostname is accessible, this command prints a message displaying the
time that it has been up.
Configuring the Parallel Framework to use ssh instead of rsh
This assumes that ssh servers have been installed on all the machines where
DataStage is to run. SSH needs to be configured such that you can launch a
command from the conductor node to all other nodes without a password, but
instead authenticate via public key encryption. Carry out the following process for
each user that will run DataStage parallel jobs. Assume that the conductor node
is etlnode and the remote node is dbnode.
1. Generate a public/private DSA key pair on the conductor, using:
etlnode% ssh-keygen -b 1024 -t dsa -f ~/.ssh/id_dsa
The identification keys have been saved in ~/.ssh/id_dsa. When you are
asked for a passphrase, leave it empty. Now send the public key to the
remote node. The scp command is the secure version of the rcp command,
as follows:
etlnode% cd .ssh
etlnode% scp id_dsa.pub user@dbnode:~/.ssh
2. Log in to dbnode and add the public key to the list of authorized keys using
the following commands:
dbnode% cd .ssh
dbnode% cat id_dsa.pub >> authorized_keys2
dbnode% chmod 640 authorized_keys2
dbnode% rm -f id_dsa.pub
The filename is
authorized_keys2, not authorized_keys. You should now be
able to ssh from etlnode to dbnode without a password. For example, the
following should work from etlnode:
etlnode% ssh dbnode ls
3. Create a /apps/Ascential/DataStage/PXEngine/etc/remsh file, which
contains:
#!/bin/sh
exec /usr/bin/ssh "$@"
Chapter 4. Installation and configuration 97
You can find a similar example of this file at:
$APT_ORCHHOME/etc/remsh.example
Installing the parallel engine on a remote node on UNIX
If you plan to use DataStage in a cluster across multiple servers, you need to
ensure that the parallel engine components are installed on all the nodes. This
can be done either of the following ways:
Install DataStage to a shared file system that is mounted (with the same
absolute path) across all servers in the cluster.
Use the maintenance menu of the DataStage installer to copy over the
$APT_ORCHHOME directories to new nodes. To use this facility, you must
configure rsh (or ssh) access to remote servers before the installation. This
feature uses the command-line utility orchcopydist, which can also be run
directly from the UNIX command line.
Configuring AIX clusters
If you are installing the parallel engine on an AIX cluster, you must verify the
setting of the network parameter
thewall. The value of this parameter can greatly
affect the performance of the parallel engine.
Set
thewall to at least 25% of each node's physical memory, or the maximum
allowed on your system, if that is less than 25% of memory. The maximum value
of
thewall is AIX-version dependent. The main page for the network option’s (no)
command contains the details and system default values. The value of
thewall is
specified in kilobytes. For example, if each node on your system has 256 MB
(262,144 KB) of physical memory, set
thewall to 65,536.
To set
thewall, use the following steps:
1. Determine the amount of physical memory on a node (the value of
realmem is
the amount of physical memory on the node in KB), by using the following:
lsattr -E -l sys0 | grep realmem
2. Determine the current setting of
thewall for a particular node or workstation:
/usr/sbin/no -a | grep thewall
3. Set
thewall by doing one of the following:
a. On a specific node, execute the following command with root privileges:
/usr/sbin/no -o thewall=65536
b. Set
thewall on all nodes of an AIX system by executing the following dsh
command from the control workstation, with root privileges:
dsh -a no -o thewall=65536
98 IBM InfoSphere Information Server Installation and Configuration Guide
4.28.4 Windows cluster configuration
This section highlights the Windows cluster configuration.
Configuring RSH access on Windows 2003 Server
In a clustered configuration, configure rsh on each of the Windows 2003 servers:
1. On each remote player node, you must create a user with the same login ID
and password as the user running the job on the conductor node.
2. From the control panel, double-click the System icon, select the Advanced
tab, and click Environment Variables.
3. In the User Variables section, change the value of the HOME environment
variable from %HOMEDRIVE%%HOMEPATH% to C:/Documents and
Settings/username. Click OK to accept the settings, click OK to exit from the
Environment Variables dialog, and click OK to exit from system properties.
4. Open the Notepad application and create a text file called .rhosts in the
C:/Documents and settings/username directory. Edit the file and place the
following line of text in the file:
+ + (press enter <cr> after the + +)
There should be a space between the plus signs. Make note of the exact
machine names for each of the player nodes, as these names will be needed
when defining the multi-node DataStage configuration file.
Enabling rsh authentication on Windows 2003
On the master (conductor) Windows 2003 server, you must enable rsh
authentication. The rsh mechanism on the conductor node (usually the same
computer running the DataStage server) must be set up for proper authentication
of rsh commands as follows:
1. Be sure that you are logged in to Windows with the user ID of the user who
will run DataStage jobs.
2. Run the rsetup command via the MKS shell.
3. Enter your password.
Note: Windows domain logins are not supported for clustered deployment
of DataStage jobs.
Local user accounts must be created on each server in the
Windows cluster.
Chapter 4. Installation and configuration 99
4. To test that rsh is working to node PC_n, enter the command:
rsh PC-n ls
This command sends the ls command to use the domain and user name of the
currently logged-on user.
Installing DataStage parallel engine on a remote node
To configure a clustered deployment of DataStage for Windows, you must install
DataStage 8.1 Server for Windows on each of the Windows 2003 servers. Unlike
UNIX, there is no maintenance facility (or corresponding orchcopydist
command) to distribute the installation to remote servers in the cluster.
4.29 Configuring the DataStage environment and
default settings
This section highlights the DataStage environment and default settings.
4.29.1 Setting the DataStage environment
DataStage provides a number of operating system environment variables to
enable and disable product features and to fine-tune job performance.
Although operating system environment variables can be set in multiple places,
there is a defined order of precedence that is evaluated when a job's actual
environment is established at run time:
1. The daemon for managing client connections to the DataStage server engine
is called
dsrpcd. By default (in a root installation), dsrpcd is started when the
server is installed and should start whenever the machine is restarted (unless
this machine is part of a high-availability (HA) configuration). dsrpcd can also
be manually started and stopped using the $DSHOME/uv -admin command.
(For more information, see the DataStage Administrators Guide, LC18-9895.)
By default, non-UNIX System Services DataStage jobs inherit the dsrpcd
environment, which on UNIX platforms is set in the etc/profile,
$DSHOME/dsenv, and ds.rc scripts. On Windows, the default DataStage
environment is defined in the registry. Because client connections are forked
from DSRPC, they do not pick up per-user environment settings from their
$HOME/.profile script.
2. Environment variable settings for particular projects can be set in the
DataStage Administrator client. Any project-level settings for a specific
environment variable will override any settings inherited from dsrpcd.
100 IBM InfoSphere Information Server Installation and Configuration Guide
Within DataStage Designer, environment variables can be defined for a
particular job using the Job Properties dialog box. Any job-level settings for a
specific environment variable will override any settings inherited from dsrpcd or
from project-level defaults.
4.29.2 Altering the DataStage dsenv on UNIX
On UNIX Systems, the DataStage server has a centralized file for storing
environment variables called
dsenv. It resides in $DSHOME, where $DSHOME
identifies the DataStage main directory (for example,
/opt/IBM/InformationServer/Server/DSEngine).
The dsenv file is a series of Bourne shell arguments that are referenced during
DataStage server startup and can be referenced by interactive users or other
programs or scripts. You are likely to want to add new environment variables as
you configure DataStage to connect to different databases using plug-ins or
ODBC drivers (see "Configuring plug-ins" and "Configuring ODBC Access" in the
DataStage Install and Upgrade Guide, Part No. 00D-018DS60).
To emulate the DataStage server environment, in a Bourne shell execute the
following command from the $DSHOME/DSEngine directory:
. ./dsenv
You must include any environment variable required by the DataStage server for
all projects in the dsenv file.
Certain plug-ins require shared libraries to be loaded, and you need to include
the library path in an environment variable. The names of the library path
environment variables are platform dependent (Table 4-28).
Table 4-28 Library path environment variables
Important: When migrating projects between machines or environments,
project-level environment variable settings are not exported when a project
is exported. These settings are stored in the project directory in the
DSPARAMS file. Any project-level environment variables must be set for new
projects.
Platform Library path environment variable
AIX LIB PATH
HP-UX (PA-RISC) SHLIB_PATH
HP-UX (Itanium), LINUX, Solaris
LD_LIBRARY_PATH
Chapter 4. Installation and configuration 101
For any changes to, or addition of, any environment variables in dsenv to
become effective, the DataStage server should be stopped and restarted
as follows:
1. To stop the server, use the following:
$DSHOME/bin/uv -admin -stop
2. To start the server, use the following:
$DSHOME/bin/uv -admin -start
4.29.3 Suggested default settings for all projects
The following project-level default settings are suggested for all DataStage
projects. Default settings can be made through the DataStage Administrator,
using the Environment button on the General tab.
$APT_DUMP_SCORE (found in the Parallel/Reporting category)
By default, the DataStage Job Monitor uses a time-based job monitor that can
introduce intermittent problems on certain platforms. Time-based job monitoring
can be disabled in favor of size-based job monitoring by altering the following
environment variables (in the Parallel category):
$APT_MONITOR_TIME = unset
$APT_MONITOR_SIZE = 100000
Default settings on AIX
On AIX, the environment variable $LDR_CNTRL is sometimes used (in
combination with other variables) to increase available memory to Server Edition
jobs (see “Increasing DataStage Server Edition memory on AIX” on page 143).
However, setting this variable can cause jobs to fail on AIX.
If you are upgrading an install of DataStage Server Edition or are running both
Server and Parallel jobs on AIX platforms, make sure that the environment
variable LDR_CNTRL is not set by default in either the dsenv file or as a default
environment variable in administrator.
Default settings on Solaris
On Solaris platforms, if you need to create and read very large parallel data sets
(where the underlying files are greater than 2 GB), you must define the
environment variable $APT_IO_NOMAP.
Note: HP-UX 11.23 shared libraries are suffixed .so instead of .sl.
102 IBM InfoSphere Information Server Installation and Configuration Guide
4.30 Configuring the DataStage administrator
environment
This section explains how to configure the DataStage administrator environment.
4.30.1 Setting the UNIX and LINUX administrator environments
On UNIX and LINUX systems, a number of command-line utilities are provided to
administer the DataStage engine and to manage objects (such as configuration
files and parallel data sets) used and created by DataStage.
To use these utilities, the default login profile should be altered for the users who
are administering DataStage to include the environment variables listed in
Table 4- 29.
Table 4-29 Default login profile environment variables
4.30.2 Setting the Windows 2003 environment
On Windows 2003, environment variables are set on a per-user or system-wide
basis through the Windows control panel. To alter or create new environment
variables on a Windows 2003 server:
1. From the Windows Start menu, select Settings Control Panel.
2. From the control panel, select System to display the system dialog.
3. From the System dialog, select the Advanced tab.
4. Select Environment Variables to display the Environment Variables dialog.
5. To edit an existing environment variable, select the variable and click Edit.
Environment variable Setting Example
DSHOME Identifies top level of the
Server Edition directory
/opt/IBM/InformationServer/Server/DSEngine/
APT_ORCHHOME Identifies top level of the
parallel engine directory
/opt/IBM/InformationServer/Server/PXEngine/
APT_CONFIG_FILE Location of the specified
parallel configuration file
/opt/IBM/InformationServer/Server/Configura
tions/default.apt
PATH Search path for
executable directories
$PATH:$DSHOME/bin:$APT_ORCHHOME/bin
LIBPATH or
LD_LIBRARY_PATH or
SHLIB_PATH
Search path for object
libraries
$LIBPATH:$DSHOME/lib:$APT_ORCHHOME/lib
Chapter 4. Installation and configuration 103
6. To add a new environment variable, select New.
7. To delete a new environment variable, select Delete.
8. Click OK to accept the changes.
4.31 Configuring and verifying database connectivity
DataStage Parallel palette provides a number of pre-built components (stages)
for accessing data in popular Relational Database Management Systems
(RDBMS).
On most platforms, native parallel connectivity is provided to IBM DB2 (with
DPF), IBM Informix®, ODBC, Oracle, Sybase, and Teradata. Specific
configuration instructions are provided in the DataStage Install and Upgrade
Guide, Part No. 00D-018DS60.
In addition, some database connectivity is provided through DataStage
plug-ins. Documentation for installing and configuring plug-ins is provided in
the DataStage Plug-in Installation and Configuration Guide. See "Installing
DataStage Packages" in the DataStage Administrator Guide, SC18-9929-02,
for details.
In cases where DataStage is deployed as a 32-bit application, 32-bit database
libraries are used. Otherwise, 64-bit libraries are used.
Where necessary, this section provides supplemental database configuration
instructions that are not in the install guides.
4.31.1 DB2 configuration for Enterprise stage
To connect to DB2 (DB2 Enterprise Server Edition with DPF) on UNIX, you must
install or crossmount DataStage Parallel engine on each node of the DB2 cluster.
In essence, you are creating a DataStage cluster, as described in “Cluster or grid
configuration” on page 93.
To connect to DB2 databases using DataStage, run the script
$APT_ORCHHOME/bin/db2setup.sh from the UNIX command line to configure
DataStage to access the DB2 database. The file must be called once for each
DB2 database to be accessed by DataStage users. Pass the database name as
Note: Specific, detailed instructions for configuring DataStage against a DB2
database are included in “Configuring remote DB2” on page 131.
104 IBM InfoSphere Information Server Installation and Configuration Guide
an argument. For example, the following command calls db2setup.sh to configure
DataStage to access the database db2_8:
db2setup.sh db2_8
You must grant privileges to each user who will run jobs containing the DB2
Enterprise Stages by running the following script once for each user:
$APT_ORCHHOME/bin/db2grant.sh
You need DBADM privileges to run this script. The syntax of the command is:
db2grant.sh database_name user_name
DB2 configuration for z/OS
DB2 for z/OS Versions 7.1 and 8.1 are supported by DataStage. The
following are considerations and actions to be taken, depending on your
particular z/OS environment:
The DSNTIJCL job from the SDSNSAMP library must be run. This job binds
the default CLI packages and plan needed by the DB2 stage.
For DB2 for z/OS Version 8, PTF UQ89056 (from APAR PQ88085) must be
applied. This PTF defines a character set conversion between 367 and 1208,
which is needed for Version 8.
Make sure that these exits have been implemented (DSN3@SGN,
DSN3@ATH, and DSNX@XAC) if access to DB2 is controlled via IBM
RACF®. Refer to the DB2 Administration Guide for your version of z/OS.
Authorization to use the DB2 load utility if that is to be used with DataStage.
Generally, if you are running this from the IBM MVS™ environment today, you
should be fine.
The DB2 bind needs to be run to allow DataStage connection. For this, refer
to the DataStage installation instructions.
If DB2 loads are to be done, verify the applicability of DB2 resource definition
changes in the configuration file (to set the high-level qualifier (HLQ) for the
temporary datasets).
4.31.2 Informix configuration
Testing has shown that it is not possible to configure a UNIX environment to
connect to both Sybase ASE and Informix IDS databases at the same time. This
means that you cannot construct a job that contains both the Sybase OC stage
and the Informix CLI stage.
Chapter 4. Installation and configuration 105
In addition, specific configuration of the environment needs to be done for the
Informix CLI stage to connect to an Informix IDS database. You must ensure that
the following files are configured correctly:
etc/hosts
etc/services
$INFORMIXDIR/etc/sqlhosts
$DSHOME/.odbc.ini
$DSHOME/../Projects/<Project Name>/uvodbc.config
Also, in the $DSHOME/dsenv file, verify that the environment variables are set
appropriately:
INFORMIXDIR
INFORMIXSERVER
INFORMIXBIN
INFORMIXC
THREADLIB
The PATH environment variable should have $INFORMIXDIR/bin appended to it.
The environment variable LIBPATH (LD_LIBRARY_PATH or SHLIB_PATH on
some platforms) should have $INFORMIXDIR/lib, $INFORMIXDIR/lib/cli, and
$INFORMIXDIR/lib/esql appended to it. Ensure that these three are after the
DataStage directories in LIBPATH (LD_LIBRARY_PATH or SHLIB_PATH).
Informix configuration on AIX
The following is an example of the settings in the $DSHOME/dsenv file for AIX:
#Informix sdk 2.8 UC2-1
#
INFORMIXSERVER=<server name>; export INFORMIXSERVER
#
INFORMIXDIR=<Informix directory path>; export INFORMIXDIR
LIBPATH=`cat
/.dshome`/lib:$LIBPATH:$INFORMIXDIR/lib:$INFORMIXDIR/lib/cli:
$INFORMIXDIR/lib/esql; export LIBPATH
INFORMIXC=CC; export INFORMIXC
THREADLIB=POSIX;export THREADLIB
PATH=$PATH:$INFORMIXDIR/bin; export PATH
Within your INFORMIXSERVER definition, set the protocol to onsoctcp.
Note: The /.dshome file might not exist in iTag installations (see “Installing and
configuring multiple server instances” on page 130). In these instances, you
might need to explicitly set $DSHOME.
106 IBM InfoSphere Information Server Installation and Configuration Guide
Informix configuration on Solaris
On Solaris, when DataStage is installed, additional configuration needs to be
performed to avoid conflicts. Following is an example $DSHOME/dsenv file. Note
the sequence of entries in the LD_LIBRARY_PATH:
1. #Informix sdk 2.8 UC1
2. #
3. INFORMIXSERVER=<server name>; export INFORMIXSERVER
4. #
5. INFORMIXDIR=<Informix directory path>; export INFORMIXDIR
6. INFORMIXBIN=$INFORMIXDIR/bin; export INFORMIXBIN
7. LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$INFORMIXDIR/lib:$INFORMIXDIR/lib/cl
i:
$INFORMIXDIR/lib/esql:$APT_ORCHHOME/lib:$APT_ORCHHOME/
DSCAPIOp
:
$APT_ORCHHOME/osh_wrappers:$APT_ORCHHOME/usr_osh_wrappers:
$APT_ORCHHOME/etc; export LD_LIBRARY_PATH
8. INFORMIXC=CC; export INFORMIXC
9. THREADLIB=POSIX;export THREADLIB
10.PATH=$PATH:$INFORMIXDIR/bin; export PATH
Within your INFORMIXSERVER definition, the protocol should be set to
ontlitcp.
4.31.3 Oracle configuration for Enterprise stage or connector
To configure the Oracle Enterprise stage with DataStage, install and
configure the Oracle Database Utilities and Oracle Network software on your
DataStage server.
You must also complete the following step (often these are set in the dsenv file):
1. Create the user-defined environment variable ORACLE_HOME and set this
to the $ORACLE_HOME path (for example, /disk3/oracle9i).
2. Create the user-defined environment variable ORACLE_SID and set this to
the correct service name (for example, ODBCSOL).
3. Add ORACLE_HOME/bin to your PATH and ORACLE_HOME/lib to your
LIBPATH, LD_LIBRARY_PATH, or SHLIB_PATH.
Chapter 4. Installation and configuration 107
In addition, for Oracle connectivity you must:
Have login privileges to Oracle using a valid Oracle user name and
corresponding password. These must be recognized by Oracle before you
attempt to access it.
Have SELECT privilege on:
– - DBA_EXTENTS
– - DBA_DATA_FILES
– - DBA_TAB_PARTITONS
– - DBA_TAB_SUBPARTITIONS
– - DBA_OBJECTS
– - ALL_PART_INDEXES
– - ALL_PART_TABLES
– - ALL_INDEXES
– - SYS.GV_$INSTANCE
Only if Oracle Parallel Server is used. If not, you might need to specify the
environment variable APT_ORACLE_NO_OPS to prevent OPS checks
from being performed.
To ease administration, you can create a role that has the appropriate SELECT
privileges, as follows:
CREATE ROLE DSXE;
GRANT SELECT on sys.dba_extents to DSXE;
GRANT SELECT on sys.dba_data_files to DSXE;
GRANT SELECT on sys.dba_tab_partitions to DSXE;
GRANT SELECT on sys.dba_tab_subpartitions to DSXE;
GRANT SELECT on sys.dba_objects to DSXE;
GRANT SELECT on sys.all_part_indexes to DSXE;
GRANT SELECT on sys.all_part_tables to DSXE;
GRANT SELECT on sys.all_indexes to DSXE;
After the role is created, grant it to users who will run parallel DataStage jobs,
as follows:
GRANT DSXE to <oracle userid>;
Note: $APT_ORCHHOME/bin must appear before $ORACLE_HOME/bin in your
$PATH.
108 IBM InfoSphere Information Server Installation and Configuration Guide
Oracle configuration on AIX
On the AIX platform, users must have SELECT access to the
sys.gv_$instance and sys.v_$cache tables. Issue the following SQL statements
to grant this access:
GRANT select ON sys.gv_$instance TO public;
GRANT select ON sys.v_$cache TO public;
If you do not have Oracle OPS on these platforms, set the
$APT_ORACLE_NO_OPS environment variable to disable OPS checking on the
Oracle Enterprise stage.
Oracle configuration on HP-UX
By default, the DataStage dsenv file will be installed with support for Oracle 9i
and 10g connectivity. If support for an Oracle 8i client is required, the
environment must be modified to something similar to the following. PA_RISC2.0
must be replaced by PA_RISC:
SHLIB_PATH=`dirname
$DSHOME`/branded_odbc/lib:$DSHOME/lib:$DSHOME/uvdlls
$DSHOME/java/jre/lib/PA_RISC:$DSHOME/java/jre/lib/PA_RISC/hotspot
$ORACLE_HOME/lib;export SHLIB_PATH
SHLIB_PATH=:$DSHOME/java/jre/lib/PA_RISC:
$DSHOME/java/jre/lib/PA_RISC/hotspot:$SHLIB_PATH; export SHLIB_PATH
To run Oracle9i or 10G, you need:
SHLIB_PATH=`dirname
$DSHOME`/branded_odbc/lib:$DSHOME/lib:$DSHOME/uvdlls
$DSHOME/java/jre/lib/PA_RISC2.0:$DSHOME/java/jre/lib/PA_RISC2.0/hotsp
ot Export SHLIB_PATH
To run Oracle 9i on HP-UX 11i, you must add $ORACLE_HOME/lib32 preceding
$ORACLE_HOME/lib in the SHLIB_PATH environment variable.
4.31.4 Sybase configuration
The Sybase operators are built against Version 12.5 of the Sybase Open Client.
If you are using an earlier version of the Sybase Open Client, you will need to
upgrade. Version 12.5 of Sybase Open Client is compatible with earlier versions
of the Sybase ASE server.
Chapter 4. Installation and configuration 109
Using Version 12.5 of the Sybase Open Client, DataStage 7.5.2 also supports
release 15 of Sybase ASE with ASE 12.5 level of functionality (new ASE 15
features are not supported). If you are using the Sybase ASE 15 client, a
configuration change is required, as documented in this section.
Configuring Sybase Open Client 12.5
Sybase open client software has to be installed on the DataStage server side.
The configuration details are:
1. Create the user-defined environment variable SYBASE and set this to the
$SYBASE path that specifies the Sybase home directory (for example, export
SYBASE=/disk3/Sybase).
2. Create the user-defined environment variable SYBASE_OCS and set this to
the Sybase open client software installation directory (for example, export
SYBASE_OCS=OCS-12_5).
3. Interfaces file: Add the details about the database server (database name,
host machine name or IP address, and port number) to the interfaces file
located in the $SYBASE directory.
4. Add SYBASE/bin to your PATH and SYBASE/lib to your LIBPATH,
LD_LIBRARY_PATH, or SHLIB_PATH.
5. Have login privileges to Sybase using a valid Sybase user name and
corresponding password, server name, and database. These must be
recognized by Sybase before you attempt to access it.
When accessing Sybase Databases with NLS, the following steps are required:
1. Create a database using a collation of the language that you are going to test
(for example, create database <<database path>> COLLATION 932JPN for
Japanese (Shift_JIS) database).
2. Install the DataStage server in that particular language (for example,
Japanese (Shift_JIS)). Upgrading the existing DataStage server will not work,
as you will not have an option to select support of other languages. You need
to uninstall the existing server and install it with the language that you want.
3. The language that you want to test should be the default setting on your OS
(desktop) (that is, the machine on which you are going to test through the
DataStage client). Select the language through Control Panel Regional
settings, and the keyboard input should also be set to that language.
Note: $SYBASE/$SYBASE_OCS/bin must appear first in your PATH. This is
to ensure that the script $SYBASE/$SYBASE_OCS/bin/isql is always
executed when the user runs the isql command.
110 IBM InfoSphere Information Server Installation and Configuration Guide
Configuring the Sybase ASE 15 client on UNIX
Changes in the naming of Sybase ASE 15 client libraries make this release
incompatible with DataStage when used directly. It is suggested that you install
Sybase Open Client 12.5, which is compatible with Sybase ASE 15 and earlier.
With the following workaround on the DataStage server machine, it is possible to
configure Sybase ASE 15 client to work with DataStage 7.5.x.
Create symbolic links for the Sybase 15 library files in the /OCS-15_0/lib/*.so
path (so that they appear to DataStage 7.5.2 as having the previous expected
names):
libct.so libsybct.so
libtcl.so libsybtcl.so
libcs.so libsybcs.so
libcomn.so libsybcomn.so
Configuring the Sybase ASE 15 client on Windows
Changes in the naming of Sybase ASE 15 client libraries make this release
incompatible with DataStage when used directly. It is suggested that you install
Sybase Open Client 12.5, which is compatible with Sybase ASE 15 and earlier.
With the following workaround on the DataStage server machine, it is possible to
configure Sybase ASE 15 client to work with DataStage 7.5.x.
Rename the Sybase 15 ASE Open Client Libraries as follows (to reflect their
previous names at release 12.5):
libct.dll libsybct.dll
libtcl.dll libsybtcl.dll
libcs.dll libsybcs.dll
libcomn.dll libsybcomn.dll
4.31.5 Teradata configuration for Enterprise Stage
You must install the Teradata Utilities Foundation on all nodes that will run
DataStage parallel jobs. Refer to the installation instructions supplied by
Teradata. (You need system administrator status for the install.)
Client setting: Using the NLS tab in the Enterprise stage, you must select the
language that you want to test. For example, if your OS has Japanese as the
default, then in the DataStage client the project default will be Shift_JIS, which
you do not need to select for every job that you run.
Chapter 4. Installation and configuration 111
You must set up a Teradata database user (this is the user who will be referred to
by the DB options property in the Teradata stage). The user must be able to
create tables and insert and delete data. The database for which you create this
account requires at least 100 MB of PERM space and 10 MB of SPOOL. Larger
allocations might be required if you run large and complex jobs. You need
database administrator status to create the user and database.
The example that follows shows you how to create the orchserver account. The
user information is stored in the terasync table. The name of the database in this
example is userspace. The following four commands for BTEQ are used to set up
the account:
CREATE USER orchserver FROM userspace AS
PASSWORD = orchserver
PERM = 100000000
SPOOL = 10000000
Once the account is set up, issue the following command:
GRANT select ON dbc TO orchserver;
Creating a database server
If you want to use a pre-existing Teradata user, you only need install a database
server and configure it to use a new database. Install the new database server
with the same PERM and SPOOL values as shown above.
The following is an example of creating a database server called devserver using
table userspace:
CREATE DATASBASE devserver FROM userspace AS
PERM = 100000000
SPOOL = 10000000
GRANT create table, insert, delete, select ON devserver TO orchclient;
GRANT create table, insert, delete, select ON devserver TO orchserver;
Teradata is optimized for batch operations with large volumes of data. Teradata
offers multiple utilities for extracting and loading data, with differing usage rules.
For this reason, DataStage offers multiple Teradata stages to choose for the
source and target of a given data flow. All data flows must be constructed on the
parallel canvas within the DataStage. It is important to understand that although
you are deploying parallel jobs, all stages from a Teradata perspective will not run
in parallel.
The Teradata Enterprise stage is intended for maximum parallel performance for
sources or targets in parallel. This interface is flexible, and for Teradata instances
with a large number of AMPs (VPROCs), it might be necessary to set the
112 IBM InfoSphere Information Server Installation and Configuration Guide
optional SessionsPerPlayer and RequestedSessions in the DBOptions string in
the Teradata Enterprise stage.
4.31.6 Netezza connectivity
DataStage supports Netezza® Performance Server (NPS®) targets on AIX,
Linux Red Hat, Linux SuSE, and Solaris platforms.
Documentation for the Netezza Enterprise stage is installed with the DataStage
client referenced in the documentation bookshelf.
The Netezza Enterprise stage is a write stage. The stage takes bulk data from a
data source and writes that data to a specified destination table in NPS. You can
write data to NPS using two available load methods (Table 4-30).
Table 4-30 Netezza load methods
To write data to NPS using the Netezza Enterprise stage, you must install the
required components. Then you must configure the stage and your system
appropriately. The following is a list of the installation and configuration
requirements:
Install the Netezza server and client. The Netezza Enterprise stage supports
Netezza Server 2.05 Patch 4 and later. You must install the Netezza client if
you want to use the nzload load method.
Install and configure the nzload utility and ODBC driver provided by NPS. The
Netezza Enterprise stage uses this ODBC driver to retrieve metadata for the
destination table or file. Ensure that you install the 2.05 version of this ODBC
driver. This driver is in conformance with 3.0 ODBC specifications.
Install and configure DataDirect's ODBC driver manager (see 4.32,
“Configuring and verifying ODBC connectivity” on page 114).
Obtain explicit LOAD privileges for the target table in NPS. For more
information or help for installing and configuring the above, see the
documentation that accompanies the above software.
Load method Description Requirements
Netezza load Uses NPS nzload utility to load
directly to target NPS table.
LOAD privileges for the target table. Data in the
source database is consistent, contains no default
values, has single-byte characters only, and uses a
predefined format.
External table Writes to an external table
within NPS. Data is then
streamed into the target table.
If the data source contains default values for table
columns and uses variable format for data encoding
such as UTF-8.
Chapter 4. Installation and configuration 113
Create the odbc.ini file in $DSHOME. To create the odbc.ini file, you must first
configure the .odbc.ini file located in $DSHOME by adding necessary
information, such as database name, host name, user name, and password.
Then copy the contents of the .odbc.ini file to the odbc.ini file in the same
location. Alternatively, you can create a soft link to the .odbc.ini file.
First, add the following entries in the .odbc.ini file:
1. ODBC Data Sources]
2. NZSQL = NetezzaSQL
3. NZSQL]
4. Driver = [Enter the driver path]
5. Description = NetezzaSQL ODBC
6. Servername =
7. Port = 5480
8. Database = [Enter the database name.]
9. Username = [Enter the user name to connect to the database.]
10.Password = [Enter the password to connect to the database.]
11.CancelAsFreeStmt = false
12.CommLog = false
13.Ksqo = false
14.Lie = false
15.Optimizer = false
16.Parse = false
17.PreFetch = 256
18.Protocol = 7.0
19.ReadOnly = false
20.ShowSystemTables = false
21.Socket = 16384
22.DateFormat = 1
23.TranslationDLL = [Enter the appropriate variable value.]
24.TranslationName = [Enter the appropriate variable value].
Now add the environment variable NZ_ODBC_INI_PATH and have it point to the
parent directory of the odbc.ini file. This odbc.ini file can be a copy of, or a soft
link to, the .odbc.ini file.
Set user-defined and general environment variables appropriately. Table 4-31 on
page 114 provides information about how to set user-defined and general
environment variables for Netezza.
114 IBM InfoSphere Information Server Installation and Configuration Guide
Table 4-31 Netezza environment variables
4.32 Configuring and verifying ODBC connectivity
In this section we describe how to configure and verify ODBC connectivity for
Enterprise stage and connector.
DataStage provides access to any database that supports ODBC using
database-specific ODBC drivers, which are included on the installation media.
The ODBC Drivers are an OEM version of the Data Direct ODBC Driver package.
These drivers are licensed solely for use with DataStage and require certain
connection parameters to be set in order to function properly. Do not try to use
these drivers with other applications, as licensing errors will occur.
ODBC driver packs are often updated between major releases of DataStage. You
are strongly advised to check your release notes for more up-to-date information
about ODBC drivers.
The ODBC drivers are one of two types, depending on the database being
connected to and your platform type:
Non-wire protocol drivers. These drivers require you to install the
database-specific client software for the database on the DataStage server.
(The drivers use the API supplied by the database client.)
Wire protocol drivers. These drivers do not require database client software.
(They communicate with the database directly.)
Environment variable Setting Description
$NETEZZA [ path] Specifies the Netezza home
directory.
$NZ_ODBC_INI_PATH [filepath] Points to the location of the
.odbc.ini file.
$PATH Should include
$NETEZZA/bin
Search the path for executable
files.
$LIBPATH,
$SHLIB_PATH, or
$LD_LIBRARY_PATH
Should include
$NETEZZA/lib
The actual environment variable
name depends on the platform to
set the search path for library files.
Chapter 4. Installation and configuration 115
4.32.1 Configuring ODBC access on UNIX
On UNIX systems, the DataDirect ODBC drivers are installed in the
$DSHOME/../branded_odbc directory. These drivers must be installed and
configured before they can be used by DataStage.
You will need to edit three files to set up the required ODBC connections, and
they are as follows:
dsenv
.odbc.ini
uvodbc.config
All three files are in the $DSHOME directory.
In addition, each DataStage project has a uvodbc.config file that can be used to
override or extend the uvodbc.config settings in $DSHOME. This allows access to
be customized by project for security and virtualization requirements.
Non-wire drivers require different setup information to wire drivers. Non-wire
drivers require information about the location of the database client software, and
wire drivers require information about the database itself.
Within your dsenv file:
The shared library path should be modified to include
$DSHOME/../branded_odbc/lib.
The ODBCINI environment variable will be set to $DSHOME/.odbc.ini.
Add $APT_ORCHHOME/branded_odbc to your PATH.
Add $APT_ORCHHOME/branded_odbc /lib to your LIBPATH,
LD_LIBRARY_PATH, or SHLIB_PATH.
The ODBCINI environment variable must be set to the full path of the
odbc.ini file (which by default is the hidden $DSHOME/.odbc.ini file, but can
be any syntactically correct file).
Further details on ODBC configuration can be found in the DataStage Install and
Upgrade Guide, Part No. 00D-018DS60.
Important: A newer version (v5.1/6.0) of the bundled DataDirect ODBC
Drivers might be available for download through the IBM eService support
site (as part of your support contract). Check this site to verify availability for
your platform.
116 IBM InfoSphere Information Server Installation and Configuration Guide
4.32.2 Setting up DSNs on UNIX
The DataDirect Drivers Reference manuals provide specific information for
configuring the ODBC environment for a particular data source. These manuals
are installed on your DataStage server in the $DSHOME/../branded_odbc/books
path.
Copy these files (which are in PDF format) from the server to your client machine
for review.
4.32.3 Configuring ODBC access on Windows 2003 Server
The ODBC driver pack that ships with the Windows version of DataStage is
available as a separately installable module. The drivers are located in the
DataStage ODBCDrivers directory on the DataStage CD.
4.32.4 ODBC readme notes
After installation refer to the readmeODBC.txt file located in the branded_odbc
directory (UNIX) or the drivers directory (Windows) for further information
regarding ODBC driver configuration and use. This file is also linked to the HTML
release notes.
4.33 Creating and verifying project location
By default, DataStage Administrator creates its projects (repositories) within the
projects directory of the DataStage installation directory.
In general, do not create DataStage projects in the default directory for the
following reasons:
Disk space is typically limited in product installation file systems.
Backup and restore policies on product installation file systems are often less
frequent than user or database directories.
Standards for many production environments dictate that product installation
directories should not be used for actual development storage.
With a default project location, careless developers can mistakenly fill up the
file system containing DataStage and corrupt not only their project, but also
other projects and components in the DataStage installation directory.
Chapter 4. Installation and configuration 117
For these reasons, specify a different file system (directory) when creating new
DataStage projects.
DataStage projects should be stored outside of the installation directory on a
redundant file system with sufficient space (minimum 100 MB per project). For
cluster/grid implementations, it is often better to share the project file system
across servers.
Creating a separate project mount on UNIX
It is suggested that a separate file system be created and mounted over the
default location for projects, the $DSHOME/Projects directory. Mount this directory
after installing DataStage but before creating projects.
To implement a separate file system, archive the contents of $DSHOME/Projects
with a utility such as tar, delete the contents of the directory, have the system
administrator create and mount the new file system, and restore the contents.
An example set of UNIX commands to accomplish this task is:
%etlhost:dsadm /usr/dsadm/Ascential/DataStage >
/bin/tar -cvf /dev/rmt0 /usr/dsadm/Ascential/DataStage/Projects
%etlhost:dsadm /usr/dsadm/Ascential/DataStage >
rm /usr/dsadm/Ascential/DataStage/Projects/*
%etlhost:root /usr/dsadm/Ascential/DataStage >
mkfs /usr/dsadm/Ascential/DataStage/Projects 1024
%etlhost::root /usr/dsadm/Ascential/DataStage >
mount /usr/dsadm/Ascential/DataStage/Projects
%etlhost::root /usr/dsadm/Ascential/DataStage >
chgrp dstage /usr/dsadm/Ascential/DataStage/Projects
%etlhost:root /usr/dsadm/Ascential/DataStage >
chown dsadm /usr/dsadm/Ascential/DataStage/Projects
%etlhost::dsadm /usr/dsadm/Ascential/DataStage >
/bin/tar -xvf /dev/rmt0 /usr/dsadm/Ascential/DataStage/Projects
Creating a project creates a new directory that appears under $DSHOME/Projects.
After creating a project, a file system can be created to contain that project and
be mounted over the $DSHOME/Projects/ProjectName directory using the
Important: The project file system should be monitored to ensure that
adequate free space remains. If the project file system runs out of free space,
this might corrupt the entire repository, requiring a restore from backup.
Note: This file system must be expandable without requiring destruction and
re-creation.
118 IBM InfoSphere Information Server Installation and Configuration Guide
technique previously described. Creating such file systems is not suggested if
$DSHOME/Projects is itself a separate file system.
4.34 Verifying project security settings and roles
By default, DataStage project file and directory ownership is set to the user and
group of the os-mapped user IS that creates them from within the administrator
client. Within administrator, project security can be specified by group
membership, InfoSphere Information Server role, or individual.
If you want to set up the system so that it distinguishes between product
managers, developers, and operators, set up groups for each class of user. Each
user is then allocated to the product manager, developer, or operator secondary
group. You can then use the DataStage Administrator client to assign the
appropriate DataStage groups to each project.
4.35 Configuring and verifying client installations
Once the DataStage server has been installed, the DataStage client should be
installed on each client workstation. This section provides specific requirements
and notes regarding the DataStage client.
The version of the DataStage client is tightly tied to the version of the
corresponding DataStage Server. The product release notes will detail which
client versions are compatible with a particular server.
For example, DataStage for Windows 8.1 requires and is only compatible with
Version 8.1 of the DataStage client. Furthermore, DataStage client release 8.1
should only be used against a Version 8.1 DataStage server on Windows. (This
should not be confused with release 8.1 for UNIX and UNIX System Services
platforms.)
For this reason, it is often necessary to install and maintain multiple DataStage
client versions on a single workstation. This is particularly the case when
developing against different server platforms and when performing a DataStage
server upgrade.
Important: You must always match the version numbers of the DataStage
client and server. This eliminates any potential compatibility issues between
the DataStage server and plug-in GUI clients or DataStage clients.
Chapter 4. Installation and configuration 119
4.35.1 DataStage Multi-Client Manager
The DataStage Multi-Client Manager (MCM) allows multiple versions of the
DataStage client to be installed on a single workstation. Only one version can be
active at any time, but the MCM allows switching between installed versions.
The DataStage Multi-Client Manager is now part of the default client install.
If you do not want to installed or configured the MCM on your workstation,
select a custom client install. This provides an option to not install the
Multi-Client Manager.
Otherwise, the DataStage Multi-Client Manager installs a Windows service.
During installation, it prompts you for the user name and password of a Windows
administrator to install and start this service. This service needs to be run using a
user account that is part of the administrator group on the local machine,
because the service must alter settings in the Windows registry. It cannot be the
built-in local administrator. Consider the following:
Even for local administrator accounts, you must supply the fully
qualified Windows login name and password. This is of the form
DOMAIN\USER, where DOMAIN is the name of your Windows server
for local (non-domain) logins.
Before the DataStage Client installer is run, the user must exist and be in that
group. Also, the local security policy must be set up so that the user account
has the
logon as a service user rights assigned to it.
If a user account that is in the administrators group on the local machine is not
specified, then the service will not be placed or installed in the service control
manager, and so the MCM will not work.
If a valid user account is specified but does not have the
logon as a service
user rights, then the service will be installed into the service control manager,
but it will not be started. The user would have to manually go to the service
control manager, re-enter the user account and password for the service, and
start the service. At this point the service control manager would automatically
assign the correct user rights to the account. Rebooting the machine does not
solve the problem.
Note: The Multi-Client Manager installs the DataStage Multi-Client Manager
Windows service using the administrative user and password that you specify.
If you change the password for this account, you must update the service
startup properties to use the new password.
120 IBM InfoSphere Information Server Installation and Configuration Guide
4.35.2 WAN development considerations
The DataStage client is a four-tier client/server application. As such, its
internal communication protocol includes many messages. Therefore, it is not
intended for networks where the latency (roundtrip transmission time for
messages) is large.
When using DataStage for remote or distributed development across a wide area
network (for example, developers across the world communicating with a central
server), it is better to configure a centralized Windows
terminal server using
Microsoft Remote Desktop, CITRIX, VNC, or similar technologies. In these
configurations, the DataStage client would be installed on the Windows machine
that is co-located with the DataStage Servers.
4.35.3 Secure client installation considerations
Implementing secure installation requires that users other than dsadm be
restricted from administrative functions. In addition to InfoSphere Information
Server roles, this is accomplished by performing a custom installation and
de-selecting the DataStage Administrator client on all workstations other than
those authorized to use dsadm. Figure 4-9 depicts a custom client install panel.
Figure 4-9 Custom client installation
Chapter 4. Installation and configuration 121
This step is important in securing your installation. The workstations on which the
administrator function is installed should be password secured and listed in a
client installation inventory.
An effective method of specifying which workstations are to receive which
functions is a table such as Table 4-32. We describe our requirements in terms of
the business function that the user supports wherever possible. The installer
must certify that the administrator function is installed only on the workstations
that are specified.
Table 4-32 Workstation administrator functions
4.35.4 Enterprise Application PACKs
If you want to install and run any of the DataStage Enterprise Application PACKs,
you must first uninstall the previous version of the DataStage client (if installed).
Failure to do so will result in the stage editors for these PACKs failing to load in
the Designer client.
Fix packs, patches, and refreshed install packages might be available for
InfoSphere Information Server and its components. Periodically check the
following web page for the current listings:
http://www.ibm.com/software/data/infosphere/support/info-server/
download.html
Workstation User Administrator Server
Version
Comments
ws127 johnl Yes No Yes John is the primary project
administrator responsible for all
areas of the project.
ws324 paulm Yes No Yes Paul is the secondary project
administrator responsible for all
areas of the project.
ws718 georgh No No No
ws817 richards No No No
ws887 mannie No Yes No Mannie is a contractor who will
develop server routines.
ws888 moe No Yes No As above.
ws889 jack No Yes No As above.
122 IBM InfoSphere Information Server Installation and Configuration Guide
© Copyright IBM Corp. 2011. All rights reserved. 123
Chapter 5. Parallel configuration files
This chapter highlights the IBM InfoSphere DataStage parallel configuration files.
A configuration file tells the DataStage Parallel Engine how to use underlying
system resources, such as processing, temporary storage, and data set storage.
In more advanced environments, it can also define other resources such as
databases and buffer storage. At run time, DataStage first reads the
configuration file to determine what system resources are allocated to it, and
then distributes the job flow across those resources.
When you modify a system by adding or removing nodes or disks, you must
modify the configuration files accordingly. Because the Parallel Engine reads a
configuration file every time that it runs a job, it automatically scales the
application to fit the system without having to alter the job design.
There is not necessarily one ideal configuration file for a system because of the
high variability between the way different jobs work and the varying performance
for different jobs and applications. For this reason, use multiple configuration files
to optimize overall throughput and to match job characteristics to available
hardware resources and business needs. At run time, the configuration file is
specified through the environment variable $APT_CONFIG_FILE.
The information in this section supplements the product documentation. For
more information about setting up and maintaining a configuration file, see the
IBM InfoSphere DataStage and QualityStage Designer Client Guide,
SC18-9893-02.
5
124 IBM InfoSphere Information Server Installation and Configuration Guide
DataStage provides a configuration file editor in the Designer client (Figure 5-1).
This graphical tool displays all configuration files located in the default directory
(
Configurations under the DataStage install directory). You can view, edit, and
verify created configuration files.
The CHECK option in this dialog box verifies connectivity to the servers and
resources listed in the file. It also warns if any operating system limits are too low
to handle basic resource allocations for a given configuration file. Because the
configuration file editor is not associated with a particular job design, it cannot
determine overall resource needs that depend on the size and complexity of a
given job.
Figure 5-1 Configuration file editor
Chapter 5. Parallel configuration files 125
An alternate way of checking a parallel configuration file is using the orchadmin
command, which verifies the configuration file pointed to by the environment
variable $APT_CONFIG_FILE:
orchadmin check
The dsadm administrator should use an alias to allow easy access to the
configurations directory. Typically, this alias (and others) is added to the login
profile of dsadm, for example:
A.1.54 alias cdcf="cd
/usr/dsadm/Ascential/DataStage/Configurations"
126 IBM InfoSphere Information Server Installation and Configuration Guide
© Copyright IBM Corp. 2011. All rights reserved. 127
Appendix A. Configurations and
checklists
This appendix provides sesveral example configurations and checklists to aid in
the installation of IBM InfoSphere DataStage in various environments.
A
128 IBM InfoSphere Information Server Installation and Configuration Guide
Installation and configuration checklist
The checklist in Table A-1 outlines the steps necessary to install and configure
DataStage. For information about each step, see the sections later in this
appendix. See also the DataStage Install and Upgrade Guide, Part No.
00D-018DS60.
Table A-1 Installation and configuration checklist
Complete Task
Verify operating system configuration and resource limits.
Verify RAID/SAN configuration.
Verify/configure file systems, available space.
Verify connectivity and network configuration.
Configure operating system users, groups, and associated permissions.
Verify C++ compiler and runtime requirements.
Check product release notes.
Install DataStage/Parallel Framework.
Verify install log.
Install DataStage patches (if applicable).
Install and configure optional DataStage components.
Configure post-install operating system settings (if applicable).
Verify cluster/grid configuration (if applicable).
Configure and verify DataStage environment and default settings.
Configure DataStage administrator environment (command line).
Configure and verify database connectivity.
Configure and verify ODBC connectivity.
Create/verify DataStage projects.
Verify project permissions and security settings/roles.
Configure and verify client installs.
Create and verify configuration files.
Appendix A. Configurations and checklists 129
DataStage administrator UNIX environment
The information summarized in this section includes suggested shortcuts and
environment settings for the DataStage super-user account (default dsadm).
They are listed in Table A-2, Table A-3, and Table A-4 on page 129.
Table A-2 Environment variables
The actual name of $LIBPATH is platform-dependent, as shown in Table A-3.
Table A-3 Platforms
The shell shortcuts (aliases) in Table A-4 can be used to move between major
DataStage directories.
Table A-4 Aliases
Environment variable Setting
$APT_ORCHHOME Location of PXEngine subdirectory in DataStage install (for
example, /usr/dsadm/Ascential/DataStage/PXEngine)
$DSHOME Location of DSEngine subdirectory (for example,
/usr/dsadm/Ascential/DataStage/DSEngine)
$PATH $APT_ORCHHOME/bin:$DSHOME/bin:$PATH
$LIBPATH * $APT_ORCHHOME/lib:$DSHOME/lib:$LIBPATH
$DSROOT Root directory of DataStage installation (for example,
/usr/Ascential/DataStage)
$APT_CONFIG_FILE $DSROOT:/Configurations/default.apt
Platform Library path environment variable
AIX LIB PATH
HP-UX (PA-RISC) SHLIB_PATH
HP-UX (Itanium), LINUX, Solaris LD_LIBRARY_PATH
Alias (shell
shortcut)
Description Setting
cdpj cd to Project dir cd /usr/dsadm/Ascential/DataStage/Projects
cdcf cd to Config dir cd
/usr/dsadm/Ascential/DataStage/Configurations
130 IBM InfoSphere Information Server Installation and Configuration Guide
Installing and configuring multiple server instances
Starting with release 8.1, it is now possible to install several instances of IBM
InfoSphere Information Server on a single UNIX or LINUX platform. This means
that you can install a new release while maintaining your current release, and
install multiple instances of the new release. This allows you, for example, to set
up different permissions for different users and effectively limit them to a
particular DataStage server instance and associated projects.
Installing or configuring additional InfoSphere Information Server component
instances on a single machine works much like a regular installation, but you
must take special care to not accept any default values for paths or ports. Every
port used by InfoSphere Information Server will need to be manually configured
for each additional instance.
When performing additional installs, be sure to choose the New Install and
Custom Install options during the installer interrogation process to ensure that
no ports or paths are re-used. Failure to do so can result in severe damage to the
existing installations, possibly rendering them inoperable or unrecoverable.
Prerequisites
Before installing:
Shut down all InfoSphere Information Server processes on the machine that
you will be working on before attempting the install.
Remove the /.dshome file.
When the install is completed, bring only the new instance online. If
everything is working fine, bring up other instances. If anything stops working,
most likely a path or port is doing double duty.
For instances that support the same RDBMS for XMETA, you can use the
same RDBMS instance for those instances, if desired.
Remove all of the auto-start logic from the RC.D directories. The
directories/logic used for auto-start varies by platform, so consult the system
administrator for your platform.
cdds cd to
DataStage root
cd /usr/dadm/Ascential/DataStage
Alias (shell
shortcut)
Description Setting
Appendix A. Configurations and checklists 131
If you want to run multiple DSRPC daemons on the same server, you will
have to choose a new iTag for each additional instance. You should be
prompted for this value if you choose the custom install option. An iTag is a
3-digit hexadecimal value used to identify the DSRPC instance. The default
value is ADE. We suggest using a version number such as 810, 801, 811,
and so forth.
Finally, as always, take a
full system backup before attempting this.
Configuring remote DB2
This section provides background information and the steps required to configure
connectivity for the WebSphere DataStage DB2 Enterprise Stage.
As a native parallel component, the DB2 Enterprise Stage is designed for
maximum performance and scalability. These goals are achieved through tight
integration with DB2 Enterprise Server Edition on UNIX, including direct
communication with each DB2 database node, and reading from and writing to
DB2 in parallel (where appropriate) using the same data partitioning as the
referenced tables.
The DB2 Enterprise Stage requires, and provides tight integration with, DB2
Enterprise Server Edition with Data Partitioning Facility (DPF) on UNIX. Both
WebSphere DataStage and DB2 Enterprise Server Edition must be running on
the same operating system and version.
This section does not highlight configuration requirements for DataStage Stage
types that provide connectivity to other DB2 platforms, including:
DB2 API
DB2 Load
Dynamic RDBMS
ODBC/Enterprise Stages
DB2 Enterprise Stage architecture
This section outlines the high-level architecture of the native parallel DB2
Enterprise stage, providing relevant background to understanding its
configuration, as detailed in the remaining sections of this document.
The DataStage Parallel Engine provides a remote DB2 configuration, separating
the primary ETL server (conductor node) from the primary DB2 server
(coordinator or node zero) using the native parallel DB2 Enterprise Stage.
Because DataStage is tightly integrated with the DB2 servers, routing of data to
132 IBM InfoSphere Information Server Installation and Configuration Guide
individual nodes based on DB2 table partitioning, configuration is provided by a
combination of DB2 client and DataStage clustered processing.
The primary ETL server (conductor node) must have the 32-bit DB2 client
installed and configured to connect to the remote DB2 server instance. This is
the same DB2 client that DataStage uses to connect to DB2 databases through
the DB2 plug-in stages (DB2 API, DB2 Load, Dynamic RDBMS) for reading,
writing, and import of metadata.
The native parallel DB2 Enterprise stage of DataStage uses the DB2 client
connection to pre-query the DB2 instance and determine partitioning of the
source or target table. This partitioning information is then used to
read/write/load data directly from and to the remote DB2 nodes based on the
actual table configuration.
This tight integration is provided by routing data within the DataStage engine to
DS nodes configured on the DB2 instance servers. This requires a clustered
configuration of DataStage, as detailed in the DataStage Install and Upgrade
Guide, Part No. 00D-018DS60, and in “Cluster or grid configuration” on page 93.
The DB2 client does not have to be installed in the same location on all servers,
as long as all locations are included in the $PATH and $LIBPATH environment
variable settings.
The actual connectivity scenario for the DB2 Enterprise stage is:
1. The DataStage primary (conductor) node uses DB2 environment variables to
determine DB2 instance. If defined, the environment variable
$APT_DB2INSTANCE_HOME can be used to specify the location on the DataStage
conductor server where the remote DB2 server's db2nodes.cfg has been
copied. The db2nodes.cfg file must reside in a subdirectory named sqllib
within $APT_DB2INSTANCE_HOME.
2. DataStage reads the db2nodes.cfg file from the sqllib subdirectory of the
specified DB2 instance. This file allows DataStage to determine the individual
node names of the DB2 instance.
3. DataStage scans the current parallel configuration file (specified by the
environment variable $APT_CONFIG_FILE) for node names whose fastname
Note: As with any clustered DataStage Enterprise Edition configuration, the
engine and libraries must be installed in the same location on all ETL and DB2
servers in the cluster. This is most easily achieved by creating a shared mount
point on the remote DataStage and DB2 nodes through NFS or similar
directory-sharing methods.
Appendix A. Configurations and checklists 133
properties match the node names provided in db2nodes.cfg. DataStage must
find each unique DB2 node name in the configuration file or the job will fail.
4. The DataStage conductor node queries the local DB2 instance via the DB2
client to determine table partitioning information. The results of this query are
then used to route data directly from and to the appropriate DB2 nodes.
5. DataStage starts up jobs across all ETL and DB2 nodes in the cluster. This
can be easily verified by setting the environment variable $APT_DUMP_SCORE to
true and examining the corresponding job score entry placed in the job log
within DataStage Director.
Prerequisites
Note the following prerequisites:
DataStage must be installed on all ETL servers and on each DB2 node in the
DB2 cluster.
The hardware and operating system of the ETL server and DB2 nodes must
be the same.
A DB2 32-bit client must be installed on the primary (conductor)
DataStage server.
The database must be DB2 Enterprise Server Edition with the Data
Partitioning Facility (DPF) option installed.
The DB2 database schema to be accessed by DataStage must not have any
columns with user defined types (UDTs).
TIP: Use the db2level command on the ETL server to identify the version
of the database.
TIP: Use the db2level command on the DB2 server to identify the version
of the database.
TIP: Use the db2 describe table [table-name] command on the DB2
client for each table to be accessed to determine if UDTs are in use.
Alternatively, examine the DDL for each schema.
134 IBM InfoSphere Information Server Installation and Configuration Guide
Setting up DB2 connectivity for remote servers
In the following simplified example configuration, two separate AIX servers are
configured:
db2_server as the DB2 database server
etl_server as the primary DataStage server
Figure A-1 illustrates this.
Figure A-1 DataStage DB2 example systems
The steps in this section are demonstrated using the DataStage administrator,
which, by default, is dsadm. The administration dsadm does
not have to be the
local database instance owner.
1. Perform the following actions on
all members of the cluster before installing
DataStage on the ETL server:
a. Create the primary group to which the DataStage users will belong (in this
document, this group is the suggested default dstage) and ensure that this
group has the same UNIX group ID (for example, 127) on all the systems.
b. Create DataStage users on all members of the cluster. Make sure that
each user has the same user ID (for example, 204) on all the systems, and
that every user has the correct group memberships, minimally with dstage
as the primary group, and the DB2 group in the list of secondary groups.
Important: Do not attempt to connect the DB2 Enterprise Stage to a remote
database by cataloging the remote database in the local instance. If you
attempt to use the stage in this way, you might experience data duplication
when working with partitioned instances because the node configuration of the
local instance might not be the same as the remote instance.
Host: db1_server
DB2 Database: db2_dpf1_db
Host: etl_server
Appendix A. Configurations and checklists 135
c. Add the following users to the DB2 database and ensure that they can log
in to DB2 on db2_server. At this step, we are on the DB2 server, and
not
the ETL server. If you fail here, contact your DB2 DBA for support. This is
not a DataStage issue.
/db2home/db2inst1@db2_server>
. /db2home/db2inst1/sqllib/db2profile
/db2home/db2inst1@db2_server>
db2 connect to db2_dpf1_db user dsadm using db2_psword
A.1.55
A.1.56 Database Connection Information
A.1.57
A.1.58 Database server = DB2/6000 8.2.2
A.1.59 SQL authorization ID = DSADM
A.1.60 Local database alias = db2dev1
2. Enable the rsh command on all servers in the cluster. The simplest way to do
this is to create a .rhosts file in the home directory of each DataStage user
who has the host name or IP address of all members of the cluster, and then
setting the permissions on this file to 600. This must be done for each user on
all members of the cluster. Modern security systems might prohibit this
method, but it will serve as an adequate example of the requirement. Contact
the system administrators for the cluster for assistance.
The commands to be performed on each node of the example system to
implement the rhosts method are:
echo "etl_server dsadm" > ~/.rhosts
echo "db2_server dsadm" >> ~/.rhosts
chmod 600 ~/.rhosts
An example of the validation of the etl_server is:
/home/dsadm@etl_server>
rsh db2_server date
A.1.61 Wed Jan 18 15:40:51 CST 2006
3. Install a 32-bit DB2 client if one is not installed on the primary ETL server (the
server on which DataStage is installed and on which the DS repository
resides, also known as the
conductor node).
a. Make dsadm the owner of the client. While the software will be installed in
/usr, management directories and components appear under the home
directory of this owner, the top of which is ~/sqllib. For dsadm on our
sample AIX system, this is /home/dsadm/sqllib.
b. Comment out the call to ~/sqllib/db2profile that the client install puts
into the .profile of dsadm. If you do not, DataStage will not operate. It will
find DB2 libraries before it finds DataStage libraries.
136 IBM InfoSphere Information Server Installation and Configuration Guide
c. Edit ~/sqllib/db2profile to export INSTHOME, DB2DIR, and
DB2INSTANCE.
4. The DB2 DBA must now catalog all the databases that you want to access on
the DB2 server into this instance of the DB2 client.
Ensure that dsadm can log in to DB2 on the db2_server. At this step, we are
on the ETL server, and
not the DB2 server. If you fail here, contact your DB2
DBA for support. This is
not a DataStage issue.
/home/dsadm@etl_server>
. /home/dsadm/sqllib/db2profile
/home/dsadm@etl_server>
db2 connect to db2dev1 user dsadm using db2_psword
A.1.62
A.1.63 Database Connection Information
A.1.64
A.1.65 Database server = DB2/6000 8.2.2
A.1.66 SQL authorization ID = DSADM
A.1.67 Local database alias = db2dev1
5. Ensure that the remote database is cataloged.
A.1.68 home/dsadm@etl_server> db2 "LIST DATABASE DIRECTORY"
A.1.69
A.1.70 Database alias = db2dev1
A.1.71 Database name = db2_dpf1_db
A.1.72 Node name = db2_server
A.1.73 Database release level = a.00
A.1.74 Comment =
A.1.75 Directory entry type = Remote
A.1.76 Authentication = SERVER
A.1.77 Catalog database partition number = -1
6. Log out of the ETL server and log back in to reset all the environment
variables to their original state. Edit $DSHOME/dsenv to include the following
information. (The bolded underlined items should be substituted with
appropriate values for your configuration.) We assume that the $DB2DIR
directory is the same on all nodes in our cluster. This ensures that $PATH
and $LIBPATH are correctly set for the remote sessions, as well as the local
session, without resorting to individual files on each member of the cluster.
On operating systems other than AIX (our example system), $LIBPATH might
be $SHLIB_PATH or $LD_LIBRARY_PATH.
A.1.78 ################################################
A.1.79 # DB2 Setup section of dsenv
A.1.80 ################################################
A.1.81 #DB2DIR is where the DB2 home is located
Appendix A. Configurations and checklists 137
A.1.82 DB2DIR=/usr/opt/db2_08_01; export DB2DIR
A.1.83
A.1.84 #DB2INSTANCE is the name of the DB2 client where
A.1.85 #the databases are cataloged
A.1.86 DB2INSTANCE=db2inst1; export DB2INSTANCE
A.1.87
A.1.88 #INSTHOME is the PATH of where the client instance is located
A.1.89 #(usually the home directory of the instance owner.
A.1.90 I NSTHOME=/home/db2inst1; export INSTHOME
A.1.91
A.1.92 #Include the sqllib, bin, adm and misc to the PATH
A.1.93 PATH=$PATH:$DB2DIR/bin; export PATH
A.1.94 THREADS_FLAG=native; export THREADS_FLAG
A.1.9
A.1.96 #Include the DB2 libraries into the LIBPATH AIX
A.1.97 #or LD_LIBRARY_PATH for SUN and Linux
A.1.98 LIBPATH=$LIBPATH:$DB2DIR/lib; export LIBPATH
7. Copy the db2nodes.cfg file from the remote instance to the primary
DataStage server. If you create a user on the DataStage server with the same
name as the DB2 remote instance owner (for example, db2inst1), then the
db2nodes.cfg can be placed in that user's "home directory/sqllib" on the
DataStage server (unless that user is already the owner of a db2 instance on
the ETL server). Otherwise, create a user-defined environment variable
APT_DB2INSTANCE_HOME in the DS administrator, add it to a test job, and
have it point to the location of the sqllib subdirectory where the
db2nodes.cfg has been placed. Avoid setting this at the project level so that
other DB2 jobs that are connecting locally do not pick up this value.
In our example, the DB2 server has four processing nodes (logical nodes),
the instance owner is db2inst1, the db2nodes.cfg file on the DB2 server is
/home/db2inst1/sqllib/db2nodes.cfg, and this file has the following
contents:
A.1.99 0 db2_server 0
A.1.100 1 db2_server 1
A.1.101 2 db2_server 2
A.1.102 3 db2_server 3
In our example, the ETL server client is owned by dsadm, the
APT_DB2INSTANCE_HOME environment variable has been set to
Important: The DataStage libraries must be placed before the DB2 entries
in $LIBPATH ($SHLIB_PATH or $LD_LIBRARY_PATH). DataStage and
DB2 use the same library name "librwtool".
138 IBM InfoSphere Information Server Installation and Configuration Guide
"/home/dsadm/remote_db2config", and this file was copied to
/home/dsadm/remote_db2config/sqllib/db2nodes.cfg on the ETL server.
8. Ensure that dsadm can connect to the instance using the values in
$DSHOME/dsenv instead of ~/sqllib/db2profile. Log out of the ETL server and
log back in to reset all the environment variables to their original state:
/home/dsadm@etl_server> cd `cat /.dshome`/dsenv
/home/dsadm@etl_server> . ./dsenv
/home/dsadm@etl_server> db2 connect to db2dev1 user dsadm using
db2_psword
The following is the database connection information:
Database server = DB2/6000 8.2.2
SQL authorization ID = DSADM
Local database alias = db2dev1
9. Implement a DataStage cluster (refer to the Install and Upgrade Guide, Part
No. 00D-018DS60, for more details). In this example, /etl/Ascential is the
file system that contains the DataStage software system, and it is
NFS-exported from the ETL server to the DB2 server, and NFS-mounted
exactly on /etl/Ascential, a file system owned by dsadm on the DB2 server.
10.Verify that the DB2 operator library has been properly configured by making
sure that the link
orchdb2op exists in the $APT_ORCHHOME/lib directory.
Normally, this link is configured on install, but if it does not exist, you must run
the script $APT_ORCHHOME/install/install.liborchdb2op. You will be
prompted to specify DB2 Version 7 or 8 (in our case, Version 8).
11.The db2setup.sh script located in $PXHOME/bin/ can run without reporting
errors even if they occur, and if there are errors, DataStage will not be able to
connect to the databases. Run the following commands and ensure that no
errors occur:
/home/dsadm@etl_server> db2 connect reset
/home/dsadm@etl_server> db2 connect terminate
/home/dsadm@etl_server> db2 connect to db2dev1 user dsadm using
db2_psword
/home/dsadm@etl_server>
db2 bind ${APT_ORCHHOME}/bin/db2esql.bnd blocking all grant public
/home/dsadm@etl_server> cd ${INSTHOME}/sqllib/bnd
/home/dsadm@etl_server>
Note: On an iTag install (see “Installing and configuring multiple server
instances” on page 130), the /.dshome file might not exist, or it might point
to the wrong DataStage instance. In this case, set $DSHOME manually,
then use the cd (change directory) command to change to that directory.
Appendix A. Configurations and checklists 139
db2 bind @db2bind.lst datetime ISO blocking all grant public
/home/dsadm@etl_server> db2 bind @db2cli.lst datetime ISO blocking
all grant public
/home/dsadm@etl_server> db2 connect reset
/home/dsadm@etl_server> db2 connect terminate
/home/dsadm@etl_server> db2 connect to db2dev1 user dsadm using
db2_psword
/home/dsadm@etl_server>
db2 grant bind, execute on package dsadm.db2.esql to group dstage
/home/dsadm@etl_server> db2 connect reset
/home/dsadm@etl_server> db2 connect terminate
12.The db2grant.sh script located in $PXHOME/bin/ can run without reporting
errors even if they occurred. If there are errors, DataStage will not operate
correctly. Run the following commands and ensure that no errors occur.
Grant bind and execute privileges to every member of the primary DataStage
group (in our case dstage).
/home/dsadm@etl_server> db2 connect to db2dev1 user dsadm using
dsadm_db2_psword
/home/dsadm@etl_server>
db2 grant bind, execute on package dsadm.db2.esql to group dstage
/home/dsadm@etl_server> db2 connect reset
/home/dsadm@etl_server> db2 connect terminate
13.Create a clustered DataStage configuration file that includes nodes to be
used for ETL processing and also includes one node entry for each server in
the remote DB2 instance.
Unless ETL processing is to be performed on the remote DB2 server nodes,
the entries pools "" should be removed from the default node pool. Each
node in the DB2 instance should be part of the same node pool (for example,
pools "db2"). Figure A-1 on page 134 shows an example configuration file.
Example A-1 Configuration file
{
node "node1"
{
fastname "etl_server "
pools ""
resource disk "/worknode1/datasets" {pools ""}
Note: Datetime ISO currently prevents this bind from succeeding.
Omit this option when issuing the bind until this issue has been resolved
by development.
140 IBM InfoSphere Information Server Installation and Configuration Guide
resource scratchdisk "/worknode1/scratch" {pools ""}
}
node "db2node1"
{
fastname "db2_server"
pools "db2"
resource disk "/worknode/datasets" {pools ""}
resource scratchdisk "/worknode/scratch" {pools ""}
}
}
14.Restart the DataStage server.
15.Test server connectivity by trying to import a table definition within DataStage
Designer (or DataStage Manager) using the DB2 API plug-in (Server plug-in).
If this fails, you do not have connectivity to the DB2 server and need to revisit
all the previous steps until this succeeds.
If this succeeds, check the imported table definitions to be sure that the data
types are legitimate.
16.Create a user-defined variable $APT_DB2INSTANCE_HOME in the DataStage
project using the administrator client for use in jobs that access DB2. Avoid
setting this at the project level so that other DB2 jobs that are connecting locally
do not inherit this value. Set this variable in each job to the location of the
sqllib/db2nodes.cfg
file (in our case,
/home/dsadm/remote_db2config
).
Configuring multiple DB2 instances in one job
Although it is not officially supported, it is possible to connect to more than one
DB2 instance within a single job. Your job must meet one of the following
configurations. (The use of the word
stream refers to a contiguous flow of data
from one stage to another within a single job.)
Single stream: Two instances only
Reading from one instance and writing to another instance with no other DB2
instances. (It has not been determined how many stages for these two
instances can be added to the canvas for this configuration for lookups.)
Two streams: One instance per steam
Reading from instance A and writing to instance A and reading from instance
B and writing to instance B.
Multiple streams with N DB2 sources with no DB2 targets
Reading from one to n DB2 instances in separate source stages with no other
downstream DB2 stages.
Appendix A. Configurations and checklists 141
To get this configuration to work correctly, you must adhere to all of the directions
specified for connecting to a remote instance and the following:
You must not set the APT_DB2INSTANCE_HOME environment variable.
Once this variable is set, DataStage will try to use it for each of the
connections in the job. Because a db2nodes.cfg file can only contain
information for one instance, this creates problems.
For DataStage to locate the db2nodes.cfg file for each of the separate DB2
stages, you must build a user on the DataStage server with the same name
as the instances to which you are trying to connect. DataStage's default logic
assumes that the instance corresponds to a UNIX user and that the
db2nodes.cfg file will exist in a sqllib subdirectory in that ID’s home
directory. Therefore, create a sqllib subdirectory for each remote instance
and place the remote instance’s db2nodes.cfg there. Because the
APT_DB2INSTANCE_HOME is not set, DataStage defaults to these
directories to find the config file for the remote instance.
Troubleshooting
If you experience problems, consider the following troubleshooting approaches:
If you get an error while performing the binds and grants, make sure that user
dsadm has privileges to create schema, select on the sysibm.dummy1 table,
and bind packages (see installation documentation for the DB2 grants
necessary to run the scripts).
There are several errors while trying to view data from the DB2 Enterprise
plug-in that do not represent the actual issue:
– If you log into DataStage with a user name (for example, dsadm) and try to
view data with a different user in the plug-in (user name and password
inside of the plug-in), you might get a failed connection. This is because
the user name and password inside of the stage are only used to create a
connection to DB2 via the client, and then the job runs using the
DataStage user (the user name is used to log into DataStage either from
the Designer or the Director).
– The user does not have permission to read the catalog tables
The user ID used to access the DB2 remote servers has to be set in each of
the servers. For example, the dsadm user has to be set up as a UNIX user in
the ETL server and in all of the DB2 nodes. Also, make sure that the groups
are set correctly because the db2grant.sh script only grants permission to the
group (in our example, dstage or as an example bd2group).
The DB2 instance is a service that needs to be running before you can
connect to any of the cataloged databases.
142 IBM InfoSphere Information Server Installation and Configuration Guide
The permission on the resource disk or scratch must be set correctly (mainly
for performing a load). When using the load, make sure that the resource disk
and scratch are open to dstage, as well as the DB2 instance owner where the
data is going to be loaded. Usually, the groups are different, so the permission
needs to be set to 777.
Performance notes
In some cases, when using user-defined SQL without partitioning against large
volumes of DB2 data, the overhead of routing information through a remote DB2
coordinator might be significant. In these instances, it might be beneficial to have
the DB2 DBA configure separate DB2 coordinator nodes (no local data) on each
ETL server (in clustered ETL configurations). In this configuration, DB2
Enterprise stage should not include the client instance name property, forcing the
DB2 Enterprise operators on each ETL server to communicate directly with their
local DB2 coordinator.
Summary of settings
The DB2 libraries must come after the DataStage libraries because both
products have libraries with identical names. The DB2 client alters the .profile of
the DB2 owner, and this must be removed or DataStage will not function.
The last four lines of the .profile for user dsadm on the ETL server are:
A.1.103 home/dsadm @ etl_server >> tail -4 .profile
A.1.104
A.1.105 # The following three lines were added by UDB and removed by
IBM IIS.
A.1.106 # if [ -f /home/dsadm/sqllib/db2profile ]; then
A.1.107 # . /home/dsadm/sqllib/db2profile
A.1.108 # fi
Environment variables set by /home/dsadm/sqllib/db2profile must be supplied
after the native DataStage environment variables. This is done with the dsenv file
for the DataStage server.
The last lines of the dsenv file with DB2 setup information added are:
/etl/Ascential/DataStage/DSEngine @ etl_server >> tail -8 dsenv
A.1.109
A.1.110 # DB2 setup section
A.1.111 DB2DIR=/usr/opt/db2_08_01; export DB2DIR
A.1.112 DB2INSTANCE=dsadm; export DB2INSTANCE
A.1.113 INSTHOME=/home/dsadm; export INSTHOME
Appendix A. Configurations and checklists 143
A.1.114 PATH=$PATH:$DB2DIR/bin; export PATH
A.1.115 THREADS_FLAG=native; export THREADS_FLAG
A.1.116 LIBPATH=$LIBPATH:$DB2DIR/lib; export LIBPATH
The contents of the db2nodes.cfg file located in /home/dsadm/remote_db2config
/sqllib are:
A.1.117 /home/dsadm/remote_db2config/sqllib @ etl_server >> cat
db2nodes.cfg
A.1.118
A.1.119 0 db2_server 0
A.1.120 1 db2_server 1
Increasing DataStage Server Edition memory on AIX
When running DataStage Server jobs on AIX platforms, it might be necessary to
increase the available memory usage to process very large numbers of records
in memory. This appendix details the process of increasing this memory
allocation and outlines a process for ensuring that these changes will not
adversely affect DataStage parallel jobs running on the same environment.
AIX implements a segmented shared memory model. The environment variable
settings and changes to the DataStage Server engine configuration increase the
amount of shared memory available to DataStage Server jobs.
To increase available DataStage Server memory usage on AIX:
1. Change the following settings in the uvconfig file located in the home
directory (DSEngine) of DataStage. These setting change the shared memory
address points. (Edit the file as user dsadm.)
DMEMOFF 0x90000000
PMEMOFF 0xa0000000
2. While logged in as user dsadm, apply these changes to the DataStage Server
engine by running the following UNIX commands:
cd DSEngine
. ./dsenv
bin/uv -admin -stop
bin/uv - admin -regen
bin/uv -admin - start
144 IBM InfoSphere Information Server Installation and Configuration Guide
For each DataStage Server job, add the following job-level environment variable
and setting. (The equal sign is part of the environment variable value and is
required.)
LDR_CNTRL = MAXDATA=0x30000000
Using HP-UX 11 memory on Windows
Memory windows allow applications to use more than the 1.75 GB limit imposed
on 32-bit processes by HP-UX 11 and 11i. To use this feature for DataStage
parallel jobs, one first needs to understand exactly how much memory is being
used by the job. For parallel jobs, this can be estimated by calculating the total
data size of all lookup tables used by a particular job.
Let us take the example of a 32-processor machine with 32 GB of memory,
called hptest. The default configuration file for this machine would contain 16
nodes, each with the fast name entry
hptest. If the jobs that are going to be run
on this will use a maximum of 8 GB, we would want to use at least eight memory
windows, as each window can hold up to 1 GB of memory. Below are the
instructions for how to set up the system and the parallel engine to use eight
memory windows.
1. Change the kernel tunable max_mem_window from 0 to 40 and reboot
the machine.
2. Create seven new entries in the /etc/hosts file, all pointing to the current
machine. For example, if the machine is called hptest, the original /etc/hosts
file probably looks similar to the following:
A.1.121 # @(#)hosts $Revision: 1.9.214.1 $ $Date: 96/10/08 13:20:01 $
A.1.122 #
A.1.123 # The form for each entry is:
A.1.124 # <internet address> <official hostname> <aliases>
A.1.125 #
A.1.126 # For example:
A.1.127 # 192.1.2.34 hpfcrm loghost
A.1.128 #
A.1.129 # See the hosts(4) manual page for more information.
A.1.130 # Note: The entries cannot be preceded by a space.
A.1.131 # The format described in this file is the correct
format.
A.1.132 # The original Berkeley manual page contains an error in
Important: The LDR_CNTRL setting must only be applied to Server Edition
jobs. Setting this value for parallel jobs causes these jobs to fail.
Appendix A. Configurations and checklists 145
A.1.133 # the format description.
A.1.134 #
A.1.135
A.1.136 134.168.56.29 hptest
Add the following line:
A.1.137 134.168.56.29 hptest1 hptest2 hptest3 hptest4 hptest5
hptest6 hptest7
A.1.138
3. If your machine is set up to trust a limited number of machines, you will need
to add each of the new host names (hptest1, … , hptest7) to your ~/.rhosts
or /ext/hosts.equiv file.
4. Create an entry in the /etc/services.window for each host name alias being
used. For example:
A.1.139 #
A.1.140 # /etc/services.window
A.1.141 #
A.1.142 # The format of this file is name followed by a space/tab
followed
A.1.143 # by a unique user key. A memory window application uses the
A.1.144 # getmemwindow(1M) command to extract the user key and then
passes
A.1.145 # that key to the setmemwindow(1M) command. Using the same
key
A.1.146 # places those applications in the same memory window.
A.1.147 #
A.1.148 # For example:
A.1.149 #
A.1.150 # winid=getmemwindow database1
A.1.151 # setmemwindow -i $winid startup_script arg1 arg2
arg3.
A.1.152 #
A.1.153 hptest 10
A.1.154 hptest1 11
A.1.155 hptest2 12
A.1.156 hptest3 13
A.1.157 hptest4 14
A.1.158 hptest5 15
A.1.159 hptest6 16
A.1.160 hptest7 17
5. Once you have done this, modify your 16-node $APT_CONFIG_FILE, so the
fastname entries are divided evenly across hptest through hptest7.
146 IBM InfoSphere Information Server Installation and Configuration Guide
6. Ensure that the user can run remsh to the current node. For example, try the
following command:
remsh hptest ls
7. Add the following file (named startup.apt) in the $APT_ORCHHOME/etc
directory and make sure that it is executable. If at any time you want to
disable the use of memory windows in the future, you can do so by setting the
$APT_NO_STARTUP_SCRIPT environment variable.
A.1.161 #!/bin/sh
A.1.162 shift 2
A.1.163 winid=$(getmemwindow $8)
A.1.164 echo "hostname="$8 "windowid="$winid
A.1.165 setmemwindow -i $winid -p $$
A.1.166 exec $*
8. Test this new configuration with a sample parallel job and the new
$APT_CONFIG_FILE. (A simple column generator to peek will do).
Estimating the size of a parallel data set
For the advanced user, this section provides a more accurate and detailed way
to estimate the size of a parallel data set based on the internal storage
requirements for each data type, as listed in Table A-5.
Table A-5 Data type sizes
Note: When using the memory windowing technique for large lookup tables,
you must hash partition the incoming data and all reference tables using the
same key columns. The default “Entire” partitioning will not use memory
windowing.
Data type Size
Integers 4 bytes
Small integer 2 bytes
Tiny integer 1 byte
Big integer 8 bytes
Decimal (precision+1)/2, rounded up
Float 8 bytes
Appendix A. Configurations and checklists 147
For the overall record width:
add (# nullable fields)/8 for null indicators
one byte per column for field alignment (worst case is 3.5 bytes per
field)
Using the internal DataStage C++ libraries, the method
APT_Record::estimateFinalOutputSize() can give you an estimate for a given
record schema, as can APT_Transfer::getTransferBufferSize(), if you have a
transfer that transfers all fields from input to output.
Windows XP Service Pack 2 firewall configuration
Starting with Service Pack 2, Microsoft has introduced the Windows Firewall—a
replacement for the internet connection firewall (ICF) in the previous version of
Windows XP. Windows Firewall is designed to discard unsolicited incoming
traffic, providing a level of protection for computers against malicious users or
programs. Starting with SP2, the Firewall is enabled on all network connections
by default. This new behavior impairs server communications when the server is
hosted on XP computers. In this section we describe how to configure the
Windows Firewall to work with server components of the IBM Information
Integration Solutions suite.
VarChar(n) n + 4 bytes for non-NLS data
2n + 4 bytes for NLS data (internally stored as UTF-16)
Char(n) n bytes for non-NLS data
2n bytes for NLS data
Time 4 bytes
8 bytes with microsecond resolution
Date 4 bytes
Timestamp 8 bytes
12 bytes with microsecond resolution
Data type Size
Disclaimer: Windows XP is not intended to support server applications, but it
can be configured to do so. IBM does not officially support running many of
the DataStage Server components on the Windows XP platform. XP is only a
supported platform for the DataStage client. If you are running an IBM
Information Integration server component on Windows XP, plan your move to
a supported platform.
148 IBM InfoSphere Information Server Installation and Configuration Guide
To begin, execute the following steps:
1. Determine whether you have Windows XP Service Pack 2 installed. From the
Windows desktop, right-click My Computer, and then select Properties.
2. Select the General tab and look under the System heading.
There are three possible scenarios, depending on or your service pack level,
as listed in Table A-6.
Table A-6 Service pack and action
3. Choose a method for configuring the firewall (Table A-7).
Table A-7 Configuration methods
Appears on Properties
bar
Action
Windows XP and Service
Pack 2 and later
Move to step 3.
Windows XP Only Forward this document to your IS group so that they are aware of the
implications when they roll out SP2.
Windows 2000, Server
2003
Do not worry if you are not running Windows XP and thus do not have the
new Windows Firewall installed.
Method Notes
Disable the firewall entirely. Easiest but least secure in an open environment. Your corporate
firewall might already protect your computer.
Open the firewall on a per-application
basis.
Easy, and more secure than total disablement. Possibly
redundant if you sit behind a corporate firewall.
Open the firewall on a per-port basis. The most traditional method of working with a firewall. Possibly
redundant if you sit behind a corporate firewall.
Appendix A. Configurations and checklists 149
4. Open the Windows Firewall Configuration page (Figure A-2).
Consult the table in step 3 and then choose either step 5, 6, or 7. You only
need to perform one of the steps.
Figure A-2 Firewall configuration
5. Disable all firewall functionality. Once you have the firewall configuration
window open select Off and then choose OK.
Note: No matter which method you choose you will have to open the
firewall configurations screen. You can find it by clicking Start
Settings Control Panel Windows Firewall.
150 IBM InfoSphere Information Server Installation and Configuration Guide
6. Disable the firewall per IBM application. Once you have the firewall
configuration window open, make sure that the Do not allow exceptions box is
not selected, then click the Exceptions tab (Figure A-3).
Figure A-3 Firewall exceptions
Appendix A. Configurations and checklists 151
7. Use the Add Programs button to add the appropriate IBM InfoSphere
Information Server applications (Figure A-4).
Figure A-4 Add applications
You might need to browse to the location of the IBM application. A table of all
IBM application names can be found in the N-Network Ports used section.
The exact location depends on the directory location that you selected when
installing the products. You can change the scope of the network use of the
opened application. Access can be limited to certain networks or IP
addresses. The scope options are the same for application or port method.
8. Enable only certain ports on the firewall.
Once you have the Firewall Configuration panel open (Figure A-2 on
page 149), click the Allow exceptions box, then click the Exceptions tab.
Use the Add Ports button to add the appropriate IBM port applications.
152 IBM InfoSphere Information Server Installation and Configuration Guide
More information
For more information about Windows XP SP2, consult the following resources:
New Networking Features in Windows XP Service Pack 2 (the January 2004
Cable Guy article)
http://technet.microsoft.com/en-us/library/bb877964.aspx
Changes to Functionality in Microsoft Windows XP Service Pack 2
http://www.microsoft.com/downloads/en/details.aspx?FamilyID=7bd948d7
-b791-40b6-8364-685b84158c78&DisplayLang=en
Deploying Windows Firewall Settings for Microsoft Windows XP with Service
Pack 2
http://www.microsoft.com/downloads/en/details.aspx?FamilyID=4454e0e1
-61fa-447a-bdcd-499f73a637d1&displaylang=en
Troubleshooting Windows Firewall in Microsoft Windows XP Service Pack 2
http://www.microsoft.com/downloads/en/details.aspx?familyid=a7628646
-131d-4617-bf68-f0532d8db131&displaylang=en
DataStage ports used in Windows platforms
This section lists all DataStage ports. You can change the scope of the network
use of the opened ports. Access can be limited to certain networks or IP
addresses. The scope options are the same for application and port method and
are covered in “Defining the scope for a program or port” on page 154. Table A-8
summarizes the DataStage ports.
Table A-8 DataStage ports
Component/protocol Default port
number
Configurable? Information
server tier
IBM DB2 database for the metadata repository
(default)
50000 Yes Services
IBM DB2 database for the analysis database
(default)
50000 Yes Services, engine,
client
DHTML reports server 16581 Yes Client
IBM Information Server web-based clients 9080 Yes Services, engine,
client
IBM Information Server web-based clients -
HTTPS
9443
a
Yes Client
Appendix A. Configurations and checklists 153
Table A-9 lists the engine tier ports.
Table A-9 Engine tier ports
WebSphere Application Server administrative
console (redirects to HTTPS)
9060
b
Yes Client
WebSphere Application Server administrative
console (HTTPS)
9043 Yes Client
IBM Information Server services (RMI/IIOP) 2809, 9100,
9401-9403
Yes Services, engine,
client
IBM WebSphere Information Services Director
services with JMS bindings
c
7276, 7286,
5558, 5578
Ye s
a. Used only when using HTTPS to access web clients.
b. Used only if you need access to the WebSphere Application Server administrative console.
c. Used only when publishing services using a JMS binding. The port needs to be accessible to
services consumers.
Component/protocol Default port
number
Configurable? Information
server tier
Component/protocol Default port number Configurable? Information
server tier
IBM InfoSphere Information Server
ASB agent
31531, and a random port
number greater than 1024
a
Yes Services
IBM InfoSphere Information Server
logging agent
31533 Yes Engine
IBM WebSphere DataStage and
QualityStage services
31538 Yes Engine, client
Parallel job monitors 13400 (port 1) and 13401
(port 2)
Ye s E n g i n e
b
Parallel engine
(APT_PM_STARTUP_PORT)
Multiple ports, uses a port
number of 10000 or
greater
Ye s E n g i n e
Parallel engine remote process
startup (rsh/ssh, multiple nodes only)
22514 Engine
Parallel engine
(APT_PLAYER_CONNECTION_PO
RT, multiple nodes only)
Multiple ports, uses a port
number of 11000 or
greater
Ye s E n g i n e
154 IBM InfoSphere Information Server Installation and Configuration Guide
Table A-10 lists the miscellaneous ports.
Table A-10 Miscellaneous ports
Defining the scope for a program or port
You have three options when defining the scope for a program or a port:
Any computer (including those on the internet)
Excepted traffic is allowed from any IPv4 address. This setting might make
your computer vulnerable to attacks from malicious users or programs on
the internet.
My network (subnet) only
Excepted traffic is allowed only from an IPv4 address that matches the local
network segment (subnet) to which the network connection that received the
traffic is attached. For example, if the network connection is configured with
an IPv4 address of 192.168.0.99 with a subnet mask of 255.255.0.0,
excepted traffic is only allowed from IPv4 addresses in the range 192.168.0.1
to 192.168.255.254.
a. Can be fixed to a specific port by specifying agent.objectport=# in the
C:\IBM\InformationServer\ASBNode\conf\agent.properties file after you complete the
installation. After designating a specific port, restart the logging agent and the ASB agent so that
the change takes effect.
b. Access to port 1 is required only from the conductor node. Access to port 2 is required from the
conductor node and the node where the IBM WebSphere DataStage and QualityStage engine is
installed, if that node is not the same as the conductor node.
Component/protocol Default port
number
Configurable? Information server tier
FTP Any port number Yes Engine
SFTP Any port number Yes Engine
SSH Any port number Yes Engine
Telnet Any port number Yes Engine
SMTP Any port number Yes Engine
ODBC Any port number Yes Engine
JDBC Any port number Yes Engine
CLI (database platform specific) Any port number Yes Engine
BW Pack
3200-3615 Ye s E n g i n e
Appendix A. Configurations and checklists 155
The My network (subnet) only scope is useful when you want to allow access
to a program or service for the computers on a local home network that are all
connected to the same subnet, but not to potentially malicious internet users.
Custom list
You can specify one or more IPv4 addresses or IPv4 address ranges
separated by commas. IPv4 address ranges typically correspond to subnets.
For IPv4 addresses, type the IPv4 address in dotted decimal notation. For
IPv4 address ranges, you can specify the range using a dotted decimal
subnet mask or a prefix length.
When you use a dotted decimal subnet mask, you can specify the range as
an IPv4 network ID (such as 10.47.81.0/255.255.255.0) or by using an IPv4
address within the range (such as 10.47.81.231/255.255.255.0).
When you use a network prefix length, you can specify the range as an IPv4
network ID (such as 10.47.81.0/24) or by using an IPv4 address within the
range (such as 10.47.81.231/24).
The following is an example custom list:
10.91.12.56,10.7.14.9/255.255.255.0,10.116.45.0/255.255.255.0,172.16
.31.11/24, 172.16.111.0/24
You cannot specify a custom list for IPv6 traffic. Once the program or port is
added, it is disabled by default in the Programs and Services list.
All of the programs or services enabled from the Exceptions tab are enabled for
all of the connections that are selected on the Advanced tab.
Pre-installation checklist
Table A-11 outlines the areas that you must review and the steps that you must
complete before you install IBM InfoSphere Information Server. For details about
each step, see this paper, the release notes, and the IBM Information Server
Planning, Installation, and Configuration Guide, GC19-1048-07.
Table A-11 Pre-installation checklist
Complete Task
1) Review release notes (InfoSphere Information Server, WebSphere
Application Server, DB2).
2) Review
IBM Information Server Planning, Installation and Configuration
Guide, GC19-1048-07.
156 IBM InfoSphere Information Server Installation and Configuration Guide
Installation and configuration checklist
The checklist in Table A-12 outlines the steps to install and configure IBM
InfoSphere Information Server. For details about each step, see the release
notes and the IBM Information Server Planning, Installation, and Configuration
Guide, GC19-1048-07.
Table A-12 Installation and configuration checklist
3) If migrating from previous versions of DataStage or QualityStage,
review IBM publication Migrating to IBM Information Server Version 8,
SC18-9924.
4) Choose and validate installation architecture/topology.
5) Validate system requirements for all tiers (engine, domain, repository,
client, documentation).
6) Verify domain (WebSphere Application Server) requirements.
7) Verify database requirements for metadata repository.
8) If applicable: Verify database requirements for Information Analyzer
analysis database.
9) Verify and configure disks, volume groups, and file systems.
10) Verify and configure operating system and resource limits.
11) Verify connectivity and network configuration.
12) Configure operating system users, groups, and associated
permissions.
13) Verify and install C++ compiler or runtime libraries, or both.
14) Verify InfoSphere Information Server Connector requirements.
15) Download and install fix pack packages (InfoSvr, WebSphere, DB2).
16) Perform a complete system backup.
Complete Task
Complete Task
1) Complete all items on the Pre-Install Checklist.
2) (If migrating from earlier DataStage or QualityStage) Complete the
pre-installation migration tasks.
Appendix A. Configurations and checklists 157
3) Create and configure the metadata repository ("xmeta") (if not using
supplied DB2).
4) Create and configure the analysis ("iadb") database (required for
Information Analyzer).
5) Install and configure the domain (WebSphere AS) server (if not using
the supplied WebSphere Application Server).
6) Install IBM InfoSphere Information Server.
7) Review all install logs.
8) (If applicable) Install IBM InfoSphere Information Server fix packs and
patches.
9) Review fix pack and patch installer log files.
10) Install and configure optional IBM InfoSphere Information Server
components.
11) (If applicable) Configure post-install operating system settings.
12) (If applicable) Verify the cluster/grid configuration.
13) Configure and verify the InfoSphere Information Server environment
and defaults.
14) Verify the InfoSphere Information Server administration (shell)
environment.
15) Configure information server users, roles, and permissions.
16) Configure and verify InfoSphere Information Server Connectors.
17) Configure and verify client installs.
18) (Where applicable) install client fix packs, patches, and optional
components.
19) Review client install, client fix pack, and client patch install log files.
20) Create and verify parallel configuration files.
21) Create and verify InfoSphere Information Server projects.
22) Verify project permissions and security settings and roles.
Complete Task
158 IBM InfoSphere Information Server Installation and Configuration Guide
InfoSphere Information Server installation settings
The information in Table A-13 summarizes the various settings and values
supplied to the IBM Information Server installer, in the order that they are
supplied to the installer. Not all options in Table A-13 are displayed. Whether an
option is displayed depends on the options that are chosen in the installer.
Table A-13 InfoSphere Information Server installation settings
Configuration topic Installation option Default value Install value
Installation response file /root/is_install.rsp
Installation directory /opt/IBM/Information
Server/
Components Engine True
Domain True
Metadata repository True
Documentation True
License file Name and location root/license.xml
Product module and
component selection
(Depends on license)
Metadata server True
Business glossary True
Information analyzer True
Federation server True False
DataStage and QualityStage True
WebSphere Information
Services director
Tru e
Install type Typical or custom Typical
DataStage server Install new or upgrade Install new
Appendix A. Configurations and checklists 159
Metadata server database
connection
Type (DB2 v8, DB2 v9,
Oracle 10, SQLServer)
DB2 v9
Host Localhost
TCP-IP port 50000
Database name xmeta
Database owner xmeta
Password xmeta1
WebSphere Application
Server
Install type: New or configure
existing
New
WebSphere Application
Server profile (for existing
install)
default
Destination directory /opt/IBM/WebSphere/
AppServer
Metadata server user registry Internal versus local OS
registry
Internal registry
AppSvr administrator user wasadmin
AppSvr administrator pass
Information Server Suite
administrator user
isadmin
Information Server Suite
administrator password
DataStage projects Project name and path
Information Analyzer
database connection
Type (DB2 v8, DB2 v9,
Oracle 10, SQLServer)
DB2 v9
Host Localhost
TCP-IP Port 50000
Database name IADB
Database owner iauser
Password
Configuration topic Installation option Default value Install value
160 IBM InfoSphere Information Server Installation and Configuration Guide
DB2 configuration Language selection (English always
installed)
Install location /opt/IBM/db2/V9
Admin server user dasusr1
Admin server password
Admin server group dasadm1
Admin server home /home/dasusr1
Instance owner user db2inst1
Instance owner password
Instance owner group db2iadm1
Instance owner home /home/db2inst1
Instance port 50000
Fenced user db2fenc1
Fenced password
Fenced group db2fadm1
Fenced home /home/db2fenc1
DataStage configuration DataStage administrator dsadm
DataStage instance tag ade
DataStage RPC port 31538
Install NLS for DataStage FALSE TRUE
MQ plug-in enabled FALSE
MQ plug-in type Server
Oracle operator enabled FALSE
Oracle operator version 10g
Legacy SAS operator
enabled
FALSE
Legacy SAS operator version 8.0
Configuration topic Installation option Default value Install value
Appendix A. Configurations and checklists 161
Online documentation and link summary
Complete documentation for IBM InfoSphere Information Server is available only
through the product installation. Additional and more current documentation is
available online. Table A-14 summarizes the online references that are required
for the pre-installation tasks. These references are listed throughout this paper.
Table A-14 Online references for pre-installation tasks
For IBM employees, the latest InfoSphere Information Server platform support
and InfoSphere Information Server connectivity matrix are on the Xtreme
Leverage Sales Portal at:
http://w3-103.ibm.com/software/xl/portal/viewcontent?type=doc&srcID=DM&
docID=U585697X60278U90
Description Web page
IBM Passport Advantage
software downloads
http://www.ibm.com/software/howtobuy/passportadvantage/index.
html
InfoSphere Information Server
Information Center
http://publib.boulder.ibm.com/infocenter/iisinfsv/v8r0/index.
jsp
InfoSphere Information Server
release notes
http://publib.boulder.ibm.com/infocenter/iisinfsv/v8r0/topic/
com.ibm.swg.im.iis.productization.iisinfsv.nav.doc/containers
/cont_iisinfsrv_rnote.html
InfoSphere Information Server
system requirements
http://www.ibm.com/support/docview.wss?uid=swg27008923
WebSphere Application Server
6.0 Info Center
http://publib.boulder.ibm.com/infocenter/wasinfo/v6r0/topic/c
om.ibm.websphere.base.doc/info/welcome_base.htm
WebSphere Application Server
6.0.2 release notes
http://publib.boulder.ibm.com/infocenter/wasinfo/v6r0/topic/c
om.ibm.websphere.base.doc/info/aes/ae/v6rn.html
WebSphere Application Server
6.0.2 hardware requirements
http://www.ibm.com/support/docview.wss?rs=180&uid=swg27007250
WebSphere Application Server
6.0.2 software requirements
http://www.ibm.com/support/docview.wss?rs=180&uid=swg27007256
DB2 v9.1 Information Center http://publib.boulder.ibm.com/infocenter/db2luw/v9/index.jsp
DB2 v9.1 release notes http://publib.boulder.ibm.com/infocenter/db2luw/v9/topic/com.
ibm.db2.udb.doc/doc/c0023859.htm
DB2 v9 system requirements http://publib.boulder.ibm.com/infocenter/db2luw/v9/topic/com.
ibm.db2.udb.uprun.doc/doc/r0025127.htm
162 IBM InfoSphere Information Server Installation and Configuration Guide
Network ports used by InfoSphere Information Server
Table A-15 summarizes the list of network ports used by InfoSphere Information
Server.
Table A-15 Network ports used by InfoSphere Information Server
Components Description Default port
numbers
Metadata database - DB2 listener port When using DB2 for xmeta repository 50000
Metadata Database - Oracle listener
port
When using Oracle for xmeta repository 1521
InfoSphere Information Server agents InfoSphere Information Server administration 31531,
31533
InfoSphere Information Server admin
console
InfoSphere Information Server administration 9080 *
DataStage RPC daemon DataStage client listener 31538 *
DataStage job monitor Job Monitor listener port 1 3500
Job Monitor listener port 2 13501
InfoSphere Information Server parallel
engine
Engine conductor/section leader
($APT_PM_START_PORT)
>= 10000
Engine player-to-player data transfer for
cluster/grid
($APT_PLAYER_CONNECTION_PORT)
>= 11000
Appendix A. Configurations and checklists 163
Domain server ports (WebSphere
Application Server)
Administrative console port (WC_adminhost) 9060
HTTPS transport port
(WC_defaulthost_secure)
9443
Administrative Console secure port
(WC_adminhost_secure)
9043
Bootstrap port (BOOTSTRAP_ADDRESS) 2809 *
Bootstrap port for deployment manager
(BOOTSTRAP_ADDRESS)
9809
SOAP connector port
(SOAP_CONNECTOR_ADDRESS)
8880
ORB listener port
(ORB_LISTENER_ADDRESS)
9100 *
SSL listener ports 9401-9403 *
High availability manager communication port
(DCS_UNICAST_ADDRESS)
9353
Service integration port
(SIB_ENDPOINT_ADDRESS)
7276
Service integration secure port
(SIB_ENDPOINT_SECURE_ADDRESS)
7286
MQ transport port
(SIB_MQ_ENDPOINT_ADDRESS)
5558
MQ transport secure port
(SIB_MQ_ENDPOINT_SECURE_ADDRESS)
5578
Components Description Default port
numbers
164 IBM InfoSphere Information Server Installation and Configuration Guide
Glossary of terminology and abbreviations
Table A-16 provides terminology and abbreviations.
Table A-16 Terminology and abbreviations
Example user setup for UNIX environments
Table A-17 provides the UNIX commands that you can use to create the groups
and users that are required to install and configure IBM InfoSphere Information
Server.
Table A-17 User names and groups for specific user accounts
Term Definition
Domain Application server (WebSphere Application Server) and
deployed InfoSphere Information Server services
Engine DataStage runtime engine (server and parallel)
Layer A self-contained component of InfoSphere Information Server,
for example, engine, domain, metadata repository, client
Metadata repository Database used to store InfoSphere Information Server
("xmeta") design, configuration, and runtime metadata
Metadata server Metadata repository and domain
Tier A physical hardware server that might have one or more
InfoSphere Information Server layers installed
User account Default
user name
Primary
group
Secondary
group
Notes
DataStage administrator dsadm dstage
DB2 administration server dasusr1 dasadm1 Only needed for DB2.
DB2 instance owner db2inst1 db2iadm1 dasadm1 Only needed for DB2.
DB2 fenced user db2fenc1 db2fadm1 Only needed for DB2.
Metadata repository owner xmeta xmeta DB2 uses OS authentication.
Information Analyzer
analysis database owner
iauser iauser DB2 uses OS authentication.
Appendix A. Configurations and checklists 165
To install and configure the listed groups and users, use the following commands:
Groups
groupadd db2iadm1
groupadd db2fadm1
groupadd dasadm1
groupadd dstage
Users
useradd -gdb2fadm1 -pdb2fenc1 -m -d /home/db2fenc1 db2fenc1
useradd -gdb2iadm1 -Gdasadm1 -pdb2inst1 -m -d /home/db2inst1
db2inst1
useradd -gdasadm1 -pdasusr1 -m -d /home/dasusr1 dasusr1
useradd -gdstage -pdsadm -m -d /home/dsadm dsadm
useradd -piasuer -m -d /home/iauser iauser
useradd -pxmeta100 -m -d /home/xmeta xmeta
166 IBM InfoSphere Information Server Installation and Configuration Guide
®
REDP-4596-00
INTERNATIONAL
TECHNICAL
SUPPORT
ORGANIZATION
BUILDING TECHNICAL
INFORMATION BASED ON
PRACTICAL EXPERIENCE
IBM Redbooks are developed by
the IBM International Technical
Support Organization. Experts
from IBM, Customers and
Partners from around the world
create timely technical
information based on realistic
scenarios. Specific
recommendations are provided
to help you implement IT
solutions more effectively in
your environment.
For more information:
ibm.com/redbooks
Redpaper
™
IBM InfoSphere Information
Server Installation and
Configuration Guide
Pre-installation
checklists for a fast
start implementation
Guidelines for
planning and
configuring your
installation
Detailed product and
platform information
This IBM Redpaper publication provides suggestions, hints and
tips, directions, installation steps, checklists of prerequisites, and
configuration information collected from several IBM InfoSphere
Information Server experts. It is intended to minimize the time
required to successfully install and configure the InfoSphere
Information Server.
The information in this document is based on field experiences of
experts who have implemented InfoSphere Information Server. As
such, it is intended to supplement, and not replace, the product
documentation.
Discover the proven choices and combinations for installing
InfoSphere Information Server that have been the most successful
for the IBM InfoSphere Center Of Excellence. This paper includes a
broad range of customer needs and experiences, with a focus on
the following areas:
Information Server architecture
Checklists
Prerequisites
Configuration choices that work well together
This paper is based on thousands of hours of production systems
experience from which you can now reap significant benefits.
Back cover
