Bulk cleanup of NC tables

Use the bulk cleanup process to remove finished tasks from NC tables in the content store.

Batch processing is supported for various tasks and objects. A common example can be a scheduled job that contains report executions. The monitor service stores the job and report execution commands in a task queue in NC tables. Under ideal circumstances, the finished tasks are removed one-by-one by a background thread. However, on busy servers or when the scheduled execution rate exceeds the system capacity there might not be enough time to remove the finished tasks this way. The tasks queue keeps growing and might cause server performance issues. The problem manifests itself as a long list of pending report executions when you view current activities in the administration component.

The following NC tables are affected:

• NC_TASK_ANCESTOR_STOREIDS
• NC_TASK_HISTORY_DETAIL
• NC_TASK_QUEUE
• NC_TSE_STATE_MAP
• NC_TASK_PROPERTY

The bulk cleanup process uses a database-specific BulkFinishedTaskCleanerThread script. This script is initiated by the monitor service during IBM® Cognos® service startup. When enabled, the script first checks if there is a certain (configurable) number of completed tasks in the NC tables. If such tasks are found, they are moved to a temporary table, and the NC tables are cleared all in one transaction. No table needs to be dropped or re-created from NC schema. You do not need to stop the IBM Cognos services either.

The cleanup process is enabled and configured by using the following monitor service advanced settings:

• event.finished.check.active
• event.finished.check.interval
• event.finished.check.threshold

For more information about these settings, see the topic Monitor service advanced settings.

For information about configuring advanced settings, see Advanced settings configuration.