[wplug] Machine is locking up without a trace

[Vanco, Don]

>-----Original Message-----
>From: wplug-bounces+don.vanco=agilysys.com at wplug.org 
>[mailto:wplug-bounces+don.vanco=agilysys.com at wplug.org] On 
>Behalf Of wmoran at potentialtech.com
>Sent: Tuesday, November 23, 2004 1:21 PM
>To: Scott F. Kiesling
>Cc: wplug at wplug.org
>Subject: Re: [wplug] Machine is locking up without a trace
>
>
>Quoting "Scott F. Kiesling" <kiesling at pitt.edu>:
>
>> Folks I have a mysterious problem which I think is
>> hardware-based but I can't find the problem. I've looked
>> around but haven't a lead.
>>
>> Running Gentoo, Kernel 2.6.7 on a Dell Inspiron 5000 (~4.5
>> years old).
>>
>> In short, the machine just stops. This happened first while
>> running X, but it has also happened when an X session has
>> not been started.
>>
>> I've looked in /var/log/messages and the X logs but haven't
>> found anything at all.
>>
>> e2fsck and reiserfsck show that my fs are OK.
>>
>> I'd appreciate any suggestions on where to look next.
>>
>> And yes, I have backed things up.
>
>These kind of issues are always a pain!
>
>Have you tried memtest86 to ensure the RAM is good?  I 
>recommend running
>it all night, as I've seen dodgy RAM that will pass sometimes, and fail
>others.
>
>There's also a program called cpuburn that will stress your CPU to see
>if it can make things fail.
>
>Aside from that, you may have to go on an Easter Egg Hunt (i.e. remove
>parts, test to see if it's still freezing up, rinse and repeat)
	See also cerberus (http://sourceforge.net/projects/va-ctcs/)  I
also have a bootable tool called Check-It Pro that has proven on at
least one occasion to find bad memory where other "native OS" based
tests could not.

I would:
Look to see if chipsets are really hot
Check all cards / chips for proper seating; falling that
remove cards one at a time (including DIMMS)

...given the age of the box it may be that it's got a part going bad.
Have you parsed the entire log for issues?  Any chance it's related to
Power Management?  That age is a prime candidate for that period of time
when things weren't really stable from a HW / PM standpoint (like the
Intel GX chipset perhaps?)

Don