Siebel Installation Guide for Microsoft Windows > Configuring the RDBMS > Configuring an IBM DB2 UDB Database for Siebel Applications >
Guidelines for Managing DB2 UDB Fragmentation
This topic is part of Configuring an IBM DB2 UDB Database for Siebel Applications.
No strict guidelines can be offered as to which tables and indexes may be fragmented due to the variety in application and customer operation variables at any given customer site. However, DBAs must pay attention to the status of large or heavily used tables, because fragmentation of these tables can affect performance significantly. For a list of these Siebel tables, see Siebel Deployment Planning Guide.
Do not reorganize
S_APSRVR_REQ, and all
%% tables (where
% is a digit), because these tables are defined to be in append mode.
Use the following strategy to manage table fragmentation:
REORGCHK on heavily used tables, and then review the resulting reports and extract list of any fragmented objects.
- Based on the results of
REORGCHK, reorganize any tables, as needed, by running
REORG TABLE. For details on how to reorganize tables or indexes, see:
- 477378.1 (Article ID) on My Oracle Support. This document was previously published as Siebel FAQ 2072.
- 477402.1 (Article ID) on My Oracle Support. This document was previously published as Siebel FAQ 2073.
- After table reorganization, update statistics by using the
runstats utility on any reorganized tables with the following minimum parameters:
runstats on table
tablename with distribution and detailed indexes all shrlevel change
You may add other parameters as required, but use the
shrlevel change parameter to allow concurrent access to your tables while
CAUTION: Because the
runstats utility overwrites statistics loaded by Siebel applications, if you use
runstats, always execute loadstats.sql afterwards, using either DB2 CLP or
odbcsql. Otherwise, valuable statistics will be lost.
To run loadstats.sql using
odbcsql, use the following command: