After you start a data group, you need to monitor the status of the replication processes and replication activity so you can respond to any error conditions. Regular monitoring and timely responses to error conditions significantly reduce the amount of time and effort required in the event that you need to switch a data group. You can access status of replication processes and details about replication errors from the Work with Data Groups display.
When processing the transactions represented by object activity entries, system journal replication processes check for common replication problems and attempt to automatically correct them. The type of detected problem determines how recovery occurs.
-
If a replication process cannot perform its operation because the object is in use or is locked, several attempts to delay and retry the operation are performed.
-
For other types of replication problems, a separate job is started to perform the recovery operation. This occurs when policies permit automatic object recovery.
The replication status of the activity entry is set to failed only when all configured delay retry cycles are exhausted without success or the recovery operation fails. The objects with problems that failed to be recovered are logged in an internal database and have their auditing priority updated so they will be checked by the next prioritized audit. Objects that have one or more failed activity entries need to be manually addressed.
You should periodically review the objects in error, and the associated failed entries, and determine the appropriate action. You may manually retry or delete one or all of the failed entries for an object. You can check the progress of activity entries and take corrective action through the Work with Data Group Activity display and the Work with DG Activity Entries display. You can also subset directly to the activity entries in error from the Work with Data Groups display.
If you have new objects to replicate that are not within the MIMIX name space, you need to add data group entries for them. Before any new data group entries can be replicated, you must end and restart the system journal replication processes in order for the changes to take effect.
The system manager removes old activity entries from the work log on each system after the time specified in the system definition passes. The Keep data group history (days) parameter (KEEPDGHST) indicates how long the activity entries remain on the system. You can also manually delete activity entries. Containers in the data libraries are deleted after the time specified in the Keep MIMIX data (days) parameter (KEEPMMXDTA).