I have a transactional replication with 2 subscriptions:
1 is located at the same instance as the publisher, the other is the remote server. The distributor is at another instance. Both publisher and distributor are SQL Server 2005, the local subscriber is SQL Server 2005 too, the remote subscriber is SQL Server 2008.
we have a purge job running every day to purge the data older than 8 days in publisher.
The replication is working fine till the publisher and distributor migrated to another data center. After migration, it is detected that the data were purged periodically because of the 'del_cmd' were configured as default value to allow delivery delete statement.
This incorrect configuration was corrected later, and after monitoring for some time, we dont find the delete is still replicated. while it seems that the data is decreasing on the local subscriber since after the distributor agent restarted 3 days ago when there was a big queue of undistributed commands to remote subscriber.
Does anyone have any idea on this strange situation?
Thanks a lot
Best Regards, Lisa J