[K12OSN] Strange Lockups

[Les Mikesell]

On Thu, 2005-02-17 at 20:23, Daniel Loomis wrote:

> I suspect it is a hardware problem, since the logs give no indication of
> the problem. It just locks up without warning. 

Based on nothing in particular, I always suspect power supplies
first when I have mysterious crashes - and they are usually
easy to swap.

> Are there any diagnostics in Knoppix or one of the bootable CDROMS that
> might work with a dual cpu Tyan mb?

There is a memory test on fedora/k12ltsp install CD's but it
takes a long time to run.

> My suspicions at this point is that the IBM SCSI harddrive is getting
> flaky.  It works fine and hasn't started showing up with any errors. It
> is a 10,000 rpm 36gb model.

You'll generally see scsi errors in the log (dmesg or
/var/log/messages) with mildly flakey drives.  I'd expect that
model to have SMART capability.  Are you running smartd?  As a
brute-force check you can 'cat /dev/sda >/dev/null' and
watch for error messages on the console or logs.

> But, if it shares any of the electronics
> with their famous "Death Star" ide drives, that would explain a lot.

They made a bad batch of 72gig scsi's a while back but I don't
think there were ever generic problems with 36gig models.

> will probably order a new scsi drive just to be on the safe side, but if
> it is the MB going South, I would like to know that up front.

Running RAID1 is the safest option, but smartd should warn you
if a drive is failing slowly.

-- 
  Les Mikesell
   les at futuresource.com