Book Image

IBM DB2 9.7 Advanced Administration Cookbook

Book Image

IBM DB2 9.7 Advanced Administration Cookbook

Overview of this book

IBM DB2 LUW is a leading relational database system developed by IBM. DB2 LUW database software offers industry leading performance, scale, and reliability on your choice of platform on various Linux distributions, leading Unix Systems like AIX, HP-UX and Solaris and MS Windows platforms. With lots of new features, DB2 9.7 delivers one the best relational database systems in the market. IBM DB2 9.7 Advanced Administration Cookbook covers all the latest features with instance creation, setup, and administration of multi-partitioned database. This practical cookbook provides step-by-step instructions to build and configure powerful databases, with scalability, safety and reliability features, using industry standard best practices. This book will walk you through all the important aspects of administration. You will learn to set up production capable environments with multi-partitioned databases and make the best use of hardware resources for maximum performance. With this guide you can master the different ways to implement strong databases with a High Availability architecture.
Table of Contents (21 chapters)
IBM DB2 9.7 Advanced Administration Cookbook
Credits
About the Authors
About the Reviewers
www.PacktPub.com
Preface
Index

Tuning memory utilization


Self-tuning memory is activated on database creation. Unless you are using a partitioned database, we recommend you leave the settings to automatic. You can also run the configuration advisor and do additional tweaking on parameters.

Capacity planning is crucial for performance. If you need to deactivate self-tuning, you will need some planning and preparation in order to split your server's memory into chunks for each instance. 4-8 GB per processor should be enough for most applications.

Each instance will have to go through a design process, so for each instance, we suggest you proceed from the top, at the partition level, the instance (database manager) level, and at the database level. We can consider, next, the agent's private memory:

Getting ready

Monitor your system closely, and gather as much information as possible on current use. Examine db2diag.log for memory-related messages. Once you have your numbers ready, allow for some down time to recycle the database...