Purging

This topic discusses the need for scheduled purging of old data, and the tools that Clarify provides to do so.

Periodic purging of data is key to good system performance and maintenance practices. Production environments accumulate logs and other data as they operate. Over time, accumulating data can and will degrade performance. It may not be noticeable now, but over time, it will become a problem.

You can delete all obsolete, non-vital log information by using a Purge process; this helps reduce performance problems, and should be part of your scheduled maintenance process.

How purging works

The ExecuteSimpleDatePurge Business Process task purges log records and associated transaction data. The task can be configured to remove logs based on specific dates or number of days. All log entries older than the indicated times frames are purged. Different time frame parameters within the task provide purging details per these classifications.
  • Delete All Data: This deletes all log records older than this parameter.
  • Delete Transaction Data: This deletes all log records older than this parameter, with the exception of connection and interchange logs.
  • Delete Virtual Data: This deletes virtual disk data older than this parameter.

When using these parameters, Clarify calculates the data to be deleted based on the time at which the purge is executed. For example, if the parameter is set to 5 days and the purge is executed on the 20th of the month at 10:05:30 PM, then the process purges the logs prior to the 15th of the month at 10:05:30 PM.

The purge Business Process can be scheduled based on your data retention policies, and you can determine the log type, age, and frequency criteria for purging. In addition, the Admin Console perspective's Auditor view can be used to view and manage logs marked for purge.

Depending on your specific Clarify implementation, certain log data types may hold more importance than others; for example, you might consider the Application log (logs of Ruleset activity) more important than the Process log (logs of launched Business Processes). In this situation, you may wish to apply different purge criteria, perhaps indicating a shorter life-span for the Process logs. Additionally, you may wish to schedule your Process log purge more often than the Application log purge.

Note: Logs that have been purged are removed from the system and are not recoverable. Any transaction data (EDI, XML, Database, File) referenced by a purged log is also permanently removed.

When to purge

While there are many factors involved in deciding how ofter to purge - such as volume, scheduling, timing - it is generally recommended to purge periodically.
Note: Any necessary backups or warehousing of transactional data or logs should be completed prior to purging.

How to purge

The general purge process follows these steps.
  • Create a Business Process and add the ExecuteSimpleDatePurge task.
  • Define the task by setting the purging times and criteria.
  • Schedule the Business Process to execute as required.

Purging Harmony data, if applicable

If including Harmony tables as part of log warehousing, the SimpleDatePurgeproperties file must be edited to add the property purgeHarmonyTables = true. This can be found in the configuration directory of the production server install, and in the CLW server as well if also purging records from that server.
Note: Check the System Options/Messages pane in Harmony to ensure that the Automatically purge logged events settings do not conflict with purge settings from Clarify.