From: Jim Ennis (jim@pegasus.cc.ucf.edu)
Date: Thu Feb 27 2003 - 10:41:08 EST
Hello,
We have a Sun 5500 for webct support (online classes). This server has
it's primary storage on a T3BPP and we use an A5200 with Sun's Instant
Image software to update a copy of the primary filesystem so that we can
backup the copy and not impact the primary/production filesystem.
We are seeing recent behaviour where our cron job (perl script) which
umounts the filesystems and performs the Instant Image update has a
problem where the umount can take over an hour on the copy of the
production filesystem (it has to be mounted for the backup software to
run). There should be no I/O or users hanging the volume from being
unmounted, but I am not sure what else to check.
We are pretty recent on all major patches (Solaris 8 02/02 with the -17
kernel) and the storage software is patched up. I used lsof to check for
active I/O but did not see any signs of activity. The filesystem in
question has about 6 million+ files/directories on it, and about 60GB of
disk space in use.
Are there any other things to check when a filesystem umounts so slowly.
This problem is recent, but I can't correlate the problem to any recent
changes except that we had to rebuild the shadow volume due to a disk
failure last week. The umount problems (slowww umounts) seem to have
started after the volume was rebuilt (and mirrored with VM) and a total
copy done on it.
Jim Ennis | jim@pegasus.cc.ucf.edu
Systems Administrator | (407) 823-1701 | Fax: (407) 823-5476
University of Central Florida | Murphy's paradox:
| Doing it the hard way is always easier.
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers
This archive was generated by hypermail 2.1.7 : Wed Apr 09 2008 - 23:25:53 EDT