SUMMARY: Filesystem mount problem on reboot

Thanks to:

Jay Lessert
Kevin Buterbaugh
Mike Kiernan

Mike Kiernan wondered if we had applied the patches (which included
the latest kernel patch) in single-user mode (we hadn't).  He suspects
that applying the patch when not in run-level S could have caused
the problem we saw.  Up to now, we have always applied patches
from a directory that is accessed via NFS so as to minimize the
downtime to only the time needed to reboot the machine (for some
of our older systems, the patching takes several hours).  At the risk
of spawning another discussion thread, how many of you out there
do/don't take the system down to "S" when patching?   We've never noted
any problems up to now.

All three agreed that journaled file systems, via enabling the
"logging" option in /etc/vfstab, is the way to go!  Jay Lessert
wondered why it is not enabled by default.  After seeing our firewall
come up in 90 seconds after a crash, I am a journaled file-system
true-believer.

Kevin Buterbaugh added this caveat:

  "There is one potential issue that I'm aware of.  If you've got database
  servers with the databases in UFS filesystems (as opposed to raw disk),
  then you should mount those filesystems with the forcedirectio option.
  However, forcedirectio and logging are mutually exclusive apparently."

Thanks to all.

--Kevin

______________________________________________________________________
Kevin Tyle, Systems Administrator               **********************
Dept. of Earth & Atmospheric Sciences           ktyle@atmos.albany.edu
University at Albany, ES-235                    518-442-4571 (voice)
1400 Washington Avenue                          518-442-5825 (fax)
Albany, NY 12222                                **********************
______________________________________________________________________