We have implemented directory container pools in TSM version 7.1.4 and are happy with it. All new backups are written to the DEDUP pool, and old backups are being migrated from the VTL to the DEDUP pool through a process of replicating existing nodes' files to a DEDUP pool on a target server, then replicating them back to the DEDUP pool on the source server. That process works well.
There is an urgency to empty the storage previously used on the VTL so it can be re-purposed. To hurry the migration to the DEDUP pool along we devised a strategy of running FULL backups (COPY MODE: ABSOLUTE) on certain small servers in policy domains whose destination was the DEDUP pool. That would promote all of a node’s previous backups on the VTL to Inactive status. And, since the pertinent RETAIN EXTRA VERSIONS setting was ’30’, we expected within 30 days all inactive versions would be expired and removed from the VTL. It’s 30 days later, and the strategy did not work perfectly. There are thousands of files remaining on the VTL for nodes which had a COPY MODE: ABSOLUTE backup. What am I misunderstanding about COPY MODE: ABSOLUTE? I had understood it would force a 100% full backup, with no mitigation by include and exclude statements. Apparently that’s not the case. Could someone clarify how it works? With many thanks, Keith Arbogast Indiana University
