SCSI timeouts on internal disks?...

[Alderman, Sean]

I've got a Dell 2650 on it's last legs of hardware support.  Lately
we've been having scsi timeouts on the internal PERC controller that
impact only the root filesystem by making it read only.  We run RHEL 3
on this particular 2650, and we have the two internal disks configured
in a mirror at the PERC controller level, so the OS only sees one disk.
What's strange is that we do not see issues with other filesystems.  A
reboot and fsck seems to resolve the issue but only temporarily, it does
not appear that the time outs happen in conjunction with any cron
scheduled activities, and typically these timeouts occur when the system
is operating at 99% idle.

Does this sound like an issue I should pursue with my hardware
maintenance?  I find it very strange that only the root filesystem is
affected, all local filesystem share the same mirrored disk set, the
applications are functional 24/7 but only used during business hours,
this particular system does not store data so we do not back it up
during off hours.  With out hardware specific diagnostics, I do not see
any significant messages at the OS level which indicate a real hardware
failure, but I'm at a loss for a direction to turn otherwise.

TIA for your thoughts and opinions.

--
Sean M. Alderman
Senior Unix Systems Administrator
MedPlus Inc. <http://www.medplus.com>
A Quest Diagnostics Company
513.204.2704

Fast, Reliable, Cheap... Pick any two!

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.us.dell.com/pipermail/linux-poweredge/attachments/20071128/6c725e7d/attachment.htm