Hi managers,
we run a Tru64 5.1A PK1 on a cluster formed by one ES45, two DS20E's and one AS4100, with two pairs of HSG80's in a SAN.
About 6 hours ago, I issued a command:
rmvol /dev/disk/dsk21c PP0S1
The domain PP0S1 has six filesets, one newly added disk (a mirror of two 18 GB disks in HSG pair B) and one old disk dsk21c (a 15 GB HSG partition in HSG pair A). So the data are beeing moved from one HSG pair to another. The data are beeing moved at a pace of about 1 kb (one kilobyte!!!) a minute and in about one hour, our users will start to work hard and I'm afraid that the system will have a very poor response time.
We have a very bad experience with AdvFS and HSG on 5.1A cluster, our AdvFS domains or even whole cluster go into panic almost everytime any change is made in HSG disks (for example failing a disk in HSG mirror).
Does anybody know the reason of this strange behavoir? Is there anything I should do to fasten the rmvol process without corrupting the data in the domain?
With best regards
Martin Dusek
PREGIS
martin.dusek_at_pregis.cz
The actual output of showfdmn:
alpha7# showfdmn -k PP0S1
Id Date Created LogPgs Version Domain Name
3caafcbf.010a792b Wed Apr 3 14:59:43 2002 512 4 PP0S1
Vol 1K-Blks Free % Used Cmode Rblks Wblks Vol Name
-data unavailable- on 256 256 /dev/disk/dsk21c
2 17778192 13266272 25% on 256 256 /dev/disk/dsk60c
Received on Fri Jun 28 2002 - 03:35:56 NZST