Db2 Big SQL event monitoring

For a detailed discussion of using event monitors to analyze Db2 Big SQL performance, see Db2 Big SQL Job History Analysis using Event Monitors.

Starting with the Db2 Big SQL 5.0.3, the behavior and scope of Write-to-Table (WTT) event monitors in Db2 Big SQL has changed. They will be allowed to operate only on the Db2 Big SQL head node. The monitoring information collected will be restricted as a result to the head node only. The default Db2 Big SQL table space location for Write-to-Table Event Monitors will be the same as that of LOCAL tables, meaning BIGSQLCATSPACE which spans the single Db2 Big SQL head node. If the user specifies a different table space on the create event monitor statement, this will be ignored. Note that if the user has the Db2 Big SQL High Availability feature enabled, both the event monitor definitions and control (and user) tables will be automatically replicated to the standby node. Explicit activation of the event monitor may be required post failover.

If you have migrated up from an earlier release of Db2 Big SQL to version 5.0.3 or above, and you already have WTT event monitors defined, these will continue to operate unchanged. It is highly recommended, however, that you drop these at the earliest possible time and re-define them so that they will be subject to the new head node restriction. Upgrade will not automatically migrate these.

In order to deactivate and drop a WTT event monitor, all Db2 Big SQL nodes must be up and running. If there are nodes that are down and cannot be started up, then you must first remove the Db2 Big SQL service from them (the Big SQL worker component). If a worker node fails while deactivating and/or dropping a WTT event monitor, the operation will fail. The procedure can be retried once all workers are operational.