11-09-2012 03:34 AM
I have a disk pool with 12 disks (from 2 ext. storage sources, limited 2 strems per vol). The disk pool is used in storage unit (24 maximum concurrent jobs).
And the disk pool is used on SLP for capacity managed storage/staging (ret: cap. managed; period: infinity).
The strange thing is that backups tend to prefer 1 disk from the pool (if there is only 1 stream). It would be ok if the disk would change to another after it's full. However it tends to cleanup this disk if it's start getting full, despite all other disks are almost empty (and retention is cap.managed+infinity). Disk pools size is about 10TB and the watermarks are at 99% and 95%.
Can I change this behavior in a way that uses all the disks in the pool equally and/or in a way that does not trigger cleanup on one disk if the pool is not nearly full?
12-13-2012 03:42 PM
Are you saying that backups occur and run concurrently on one drive? That doesn't sound like default NBU behavior. It should use different drives based on space availability and some other factors. How can you tell it is only using one drive, and is this the case with multiple jobs running to the disk pool?