Hi guys
We are using SPS/FCM 8.1.9 (Spectrum Protect Snapshot/Flashcopy Manager) to snapshot a couple of quite large filesystems on a SVC. The snapshot is done on a set of 6 LUNs each 32TiB in size. This creates a CG (consistency group) containing 6 FC maps (using NOCOPY parameter). Then the FC volumes start to fill up and sometimes reach a change percentage of up to 10% before being deleted/cleaned up.
When it's time to delete such a big snaphot (either being deleted by reaching MAX_VERSIONS oder by hand with 'fcmcli -f delete -B <BackupID>') the first thing that happens is the CG needs to be stopped. Now the "delete job" itself returns successfully after a couple of seconds but if the stopping of the CG takes too long (sometimes up to 2 hours depending on the load of the SVC) the CG just gets left behind in a stopped state including the FC volumes going offline. This stays for days without being cleaned up by SPS/FCM - we usually have to do this manually on the SVC.
My questions now are:
- how can we convince SPS/FCM to clean up "it's mess"?
- is there a way to have SPS/FCM look a the storage system and delete any old volumes/CGs/FC maps that it cannot find in the repository anymore
Thanks and Cheers
Peter
------------------------------
Peter Wyder
------------------------------