From: Jim Ennis (jim@pegasus.cc.ucf.edu)
Date: Tue Apr 22 2003 - 16:07:46 EDT
Hello,
I am using ufs (with logging) on an E5500 with T3B partner pair for
storage. Recently we have been working on a problem where the umounts of
our large filesystem for central webct site is taking a long time (30-40
minutes) even though fuser and lsof show no I/O. We are leaning towards
the theory that the ufs logging is hanging up the umount process. The OS
is Solaris 8 02/02 at the -17 kernel patch (we will be patching the system
up in a few weeks in our semester break).
We use Sun's Instant Image 3.0 software to snapshot the production
filesystem to a shadow volume and then run the backup (Netbackup 3.4) on the
shadow volume to minimize performance impacts to the users. This worked pretty
well until a couple of months ago when something changed (no OS changes,
but we did have a broken shadow volume to fix).
I am looking at moving to vxfs and drop ufs for this filesystem (100 GB,
about 7.5 million files and directories). I did some tests last year and
saw a slight improvement with vxfs in terms of file I/O, but I was
wondering if anyone had experience with using vxfs rather than 'ufs with
logging' for a 'large number of files' filesystem.
Does anyone have any experience to share with me about this?
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 - 23:26:14 EDT