If a concurrency bottleneck had occurred on the item tables, increasing
the number of them will in turn increase the throughput of the messaging engine.
Procedure
- Relevant performance monitoring tools show that the throughput
of a messaging engine is inefficient.
- Use your database performance monitoring tools to examine lock
statistics for the item tables for evidence of a bottleneck. Consult database
documentation on how to interpret the locking statistics.
- Create tables and increase data store attributes.
You can only increase the number of permanent tables or temporary tables,
not decrease them.
- Stop and restart the WebSphere Application Server so that configuration
changes take effect. The extra tables are used when the messaging engine restarts.
- Observe the effect on throughput and lock statistics by checking
performance monitoring tools. Consider whether any improvement is sufficient
and modifying the data store attributes
further would be beneficial