We are using TD 2580 with 72 AMP's and 10TB space, so each AMP have approx 74000 Cilinders. 74000*1.9M*72 ->10TB
Some days ago we noticed a lagre number of Mini-Cyl-packs, and explored than some AMP's have much less Cilinder than should.
For example vproc 44 have 55300 Avaible Cil's one day and next day there was only 52000 Avaible. We were using ferret SHOWSPACE command to identify this.
Because there about 51000 Cils on 44 vproc is used for perm we have lack of spool space on this AMP
that causes total system perfomance degradation. Support personel advised to use ferret PACKDISK command, but it did't help because it only changed FSP value on existing tables and freed some cil's from perm space, but total number of avaible cil's didn't change.
We've supposed that this is a hardware problems with disks, but DBA's said that everything is fine.
If someone have any ideas or have ever dealed with such a problem, please, help!
Thank's in advanceP
Problem is gone after simple restart, number of avaible cilinders on all AMP-s become to normal in 20 minutes