[Techtalk] mysterious system halts - how to prevent/fix/detect?

[Julie Meloni]

CS> On Wednesday 05 June 2002 08:10 am, Malcolm-Rannirl wrote:
>> On Wednesday 05 June 2002 10:29 am, Mandi wrote:
>> > You may also want to poke around looking for a cracker, nmap scan the
>> > box, see if anything weird is open and listening.
>>
>> I was about to suggest that. We had a cracker get one of the servers I
>> worked on (learned a lot about security from that incident), and installed
>> a kernel module as part of the rootkit. Having someone deliberately crash
>> the machine to force a reboot (and thus a kernel reload) seems plausible.
>>
>> It does sound more like the power supply problem that someone else
>> suggested. (I've had a few PSs die on me with similar results).
 
CS> I third the power supply notion, got a spare to swap in?

Yep,  did that.  Got real excited when we crossed the 12 hour
threshold.  Then it hung at 15 1/2 hours.

Swapped out memory, still  hung.  Set the screen to not blank out,
didn't get any kernel panic or crazy hex stuff, so ... now we're
thinkning it's the MB or the processor.  It's an intel board of some
sort (don't know off the top of my head) and dual P-III 550
processors.

CS> TuffTest Pro is a very nice hardware tester, and cheap, get the 29.95 
CS> version. http://www.tufftest.com/

Thanks for the tip!  I was going to put a fresh new drive in the
machine today, just to eliminate the possibility of it being the
drive.

- Julie