Database tuning is a complex task, and for important databases, it requires
the skills of a database administrator. For the Tivoli® Data Warehouse, a database located
on a single disk using “out of the box" database parameters is suitable
only for small test environments. For all other environments, careful planning,
monitoring and tuning are required to achieve satisfactory performance.
There are a number of sources of database configuration and tuning information
that should be helpful in the planning, monitoring, and tuning of the Tivoli Data Warehouse:
The “Understanding the disk requirements for your database" section
of the IBM Tivoli Monitoring Installation and Setup Guide describes
factors to consider in planning the disk subsystem to support the Tivoli Data Warehouse.
The paper “Relational Database Design and Performance Tuning for DB2® Database
Servers" is available from the Tivoli Open Process Automation Library (OPAL)
by searching for "database tuning" or navigation code "1TW10EC02"
at OPAL. This paper is a concise reference describing
some of the major factors that affect DB2 performance. This document is a good starting
point to use before referencing more detailed information in manuals and Redbooks® devoted to DB2 performance. While DB2 specific, many of the concepts are applicable
to Relational Databases in general, such as Oracle and Microsoft® SQL
Server.
The “Tivoli Data Warehouse tuning" chapter of the Redbook Tivoli Management Services Warehouse and Reporting
(SG24-7443) builds upon the previously referenced OPAL paper, supplementing
it with information about Oracle and Microsoft SQL Server. This chapter is
almost 100 pages in length.
The “Optimizing the performance" chapter of the Redbook IBM® Tivoli Monitoring Implementation and Performance
Optimization for Large Scale Environments (SG24-7443) contains a
section on database tuning considerations for the Tivoli Data Warehouse. This section makes
suggestions about specific tuning parameters for DB2, Oracle and MS SQL. At approximately 12
pages, this section is much shorter than the tuning chapter in the previously
referenced Redbook (item number 3).
The IBM Tivoli Monitoring Problem Determination Guide contains
an item “Summarization and Pruning Agent in large environment" which
lists specific tuning changes that were made to a Tivoli Monitoring V6.2 Tivoli Data
Warehouse supporting a large number of agents.
The remainder of this section is a summarized version of material from
the OPAL paper in number 2 above, which has been supplemented by identifying
specific parameters relevant to Tivoli Data Warehouse and providing some suggested
ranges of values.