I recently upgraded SMSQL on a SQL instance from 5.0 to 5.1. I have a job scheduled to run every hour to backup transaction logs. The log backups are spread across 5 mount points. It was working fine in 5.0. Now, in 5.1, the job is no longer deleting old log snapinfo snapshots. Before the upgrade, the log files showed the following which was working:
[09:21:40.775] DELETE SAME TIME LOG/SNAPINFO SNAPSHOT
[09:21:40.776] Deleting snapshot copy sqlinfo__sql_08-23-2011_09.19.52 of LUN J:\sninfo1\ [09:21:41.342] Deletion of snapshot copy succeeded. [09:21:41.343] SNAPSHOT SNAPINFO DRIVE
[09:21:41.344] Creating Snapshot of LUN(s): [09:21:41.345] LUN [J:\sninfo1\], [J:\sninfo2\], [J:\sninfo3\], [J:\sninfo4\], [J:\sninfo5\] [09:21:41.345] Snapshot name: sqlinfo__sql_08-23-2011_09.19.52 [09:22:03.866] [SnapDrive Error]: Failed to create a consistent Snapshot copy of one or more LUN(s). Please check the application event log for more information.
(SnapDrive Error Code: 0xc00402c1)
Why is SMSQL creating the sninfo snapshots ok then turning right around and deleting just the first one (the DELETE SAME TIME LOG/SNAPINFO SNAPSHOT)?!! And then tries to create the 5 snapshots again? Crazy. I think because it's erroring out, it never gets to the process of deleting log snapinfo snapshots.
I installed SMSQL 5.1P2 this morning. However, the problem still exists. It seems the job never gets the chance to delete snapinfo snapshots because of the line that says "DELETE SAME TIME LOG/SNAPINFO SNAPSHOT". Why is SMSQL doing this? It makes no sense to create a snapshot successfully and then turn around to delete what it just created.
After some testing, it seems SMSQL 5.1 doesn't like performing transaction log backups at the same time for multiple volumes/LUNs. This is why I'm seeing the "DELETE SAME TIME LOG/SNAPINFO SNAPSHOT" behavior. This was not a problem with SMSQL 5.0. So now I have to create a separate transaction log job for each volume. Any chance this new feature can be fixed?