Table partitioning in a Db2 pureScale environment

You can use table partitioning in a Db2 pureScale environment to divide large table objects between multiple partitions for better performance.

You can use table partitioning in Db2 pureScale tables; this includes tables that use the PARTITION BY RANGE clause. In addition, the commands associated with table partitioning can be used in a Db2 pureScale environment.

This means, for example, that all of the following operations are supported:

The roll-in and roll-out partition operations available through the ALTER TABLE statement
The PARTITIONED and NOT PARTITIONED clauses for the CREATE INDEX statement
For partitioned indexes, the ON DATA PARTITION clause of the REORG TABLE and REORG INDEXES ALL statements

In addition, the MON_GET_PAGE_ACCESS_INFO table function has been updated to work with partitioned tables. All existing monitoring functions that operate on data partitions will work with Db2 pureScale tables.

If you are already using the Db2 pureScale Feature, you can use table partitioning to help resolve page contention issues. By spreading contention out over a larger range, you can reduce data page contention; similarly, you can reduce contention with index pages by using partitioned indexes.

Note: From a Db2 pureScale performance perspective, the amount of memory used depends on the number of table partitions and the number of indexes. The memory resource used for the partitioning on the member comes from the dbheap configuration parameter. On the CF, the memory resource is defined by the cf_sca_sz configuration parameter.