CentOS - Jan 2007 - Machine all of a suddens freezes. Any suggestions?

This is a A8N32-SLI Deluxe motherboard with a AMD Athlon(tm) 64 X2 Dual 
Core 4400+ Processor. It freezes and cannot be accessed from the 
network. Keyboard/mouse/display, everything just stuck. It had an uptime 
of 17 days after I re-connected all cables and boards after a similar crash.

My feeling is, this must be a hardware related problem. I inspected the 
logs under /var/log, but nothing.

What would be the best approach to debug this problem? What can be the 
most likely cause? Or should I just accept it (it's a single user 
desktop machine). The machine has FC4 installed. Is there a chance that 
it becomes more stable with Centos (as said, I actually  don't suspect 
the OS). For new machines I started installing Centos (and I'm happy 
with it, especially the fact that my EDA software vendors support it well).
I monitor the machine with Zabbix, but no weird things there. Just 
before the freeze the machine had a processor load of 1 running a user 
simulation.
Any suggestions would be welcome.

Thanks,
Theo

Temperature could indeed be the case. The last time it happened was 
during heavy load. The fans are spinning, that I checked already. I will 
add temperature monitoring as well to the logs. Dirt could be, but this 
machine is not even a year old. I'll check the CPU cooler assembly.

Thanks,
Theo
Dave K wrote:
> This bounced when sent to the list, so I'm copying you directly.
>
> ---------- Forwarded message ----------
> From: Dave K <davek08054 at gmail.com>
> Date: Jan 18, 2007 1:37 PM
> Subject: Re: [CentOS] Machine all of a suddens freezes. Any suggestions?
> To: CentOS mailing list <centos at centos.org>
>
>
> On 1/18/07, Theo Band <theo.band at xanadu-wireless.com> wrote:
>> ... It freezes and cannot be accessed from the network....
>> ...
>> My feeling is, this must be a hardware related problem. I inspected the
>> logs under /var/log, but nothing.
>> ...
>> What would be the best approach to debug this problem? What can be the
>> most likely cause?
>
> The first thing I would suspect is some sort of cooling problem.
> Verify that all of your fans are actually spinning.  If the BIOS has
> the option to always run the fans at full speed, try it for a while
> (if you can stand the noise).  Use a flashlight and check that the
> airspace between the fins of any heatsinks aren't clogged with dust.
> If there is room (and power) for additional fans, try adding one.
>
> The last system I had with "random" freezes runs fine now after a
> replacement CPU fan, an additional case fan, and suitable application
> of half a can of air-duster.
>
> -- 
> Dave K
> Unix Systems & Network Administrator
> Mount Laurel NJ
>
>

Theo Band wrote:
> What would be the best approach to debug this problem? What can be the 
> most likely cause? Or should I just accept it (it's a single user 
> desktop machine). The machine has FC4 installed. Is there a chance that 
> it becomes more stable with Centos (as said, I actually  don't suspect 
> the OS). For new machines I started installing Centos (and I'm happy 
> with it, especially the fact that my EDA software vendors support it well).
> I monitor the machine with Zabbix, but no weird things there. Just 
> before the freeze the machine had a processor load of 1 running a user 
> simulation.
Turn on nmi_watchdog:
http://www.mjmwired.net/kernel/Documentation/nmi_watchdog.txt#34

run memtest86+ for 24 to 48 hours.

Stress your disks:
for i in $(seq -w 20); do cp -ax / /tmp/$i & done

(Be sure to stop the 20 cp processes before your disk(s) fill up)

Mike