[lug] Daily system crash....

[Nate Duehr]

Stephen Queen wrote:
> On Jan 25, 2008 12:12 PM, David Morris <lists at morris-clan.net> wrote:
>> I have a strange problem with my system I'm hoping for input
>> on.
>>
>> I run Debian Testing (lenny) 64-bit (AMD64).  After a recent
>> update to the latest version of available packages my system
>> started crashing once a day.
>>
>> The crash is a complete system lockup.  No response locally
>> or when I attempt to ssh in from a remote computer.  System
>> logs contain no indication of the cause of the problem.  At
>> the time of the crash I have always been actively using the
>> system (not surprising given I work from home), however I
>> was only typing at the time.  The only other processes
>> running are XMMS, Pidgin, VIM editors, ssh sessions, and
>> sometimes a few with 'tail -f' running on them.  At every
>> crash I have been typing in a VIM editor in a terminal
>> window (mrxvt) either on a local or remote system, but then
>> I do that constantly throughout the day.
>>
>> Ordinarily I would suspect a hardware problem (Motherboard,
>> CPU, or RAM), however two things lead me to believe this
>> might not be the case:
>>
>> - The first instance was a few hours after a big update of
>>   Debian packages.
>> - Not reproducible on demand, but randomly once every 12-36
>>   hours which usually indicates a software problem.
>>
>> I plan to run memtest86, however this will need to wait
>> until tonight.
>>
>> Any suggestions on how I might glean more insight into the
>> cause of the problem given nothing shows up in the syslog?
> 
> I had a similar issue with an AMD 64 machine at work. I suspected it
> had something to do with the network cards. I grab the latest kernel
> from kernel.org, compiled and installed it, and have not had a crash
> since.

I also had similar problems with Debian's kernels on a machine with a 
VIA-chipset on the motherboard (they're kinda notorious for having 
problems anyway).

I put a hardware reboot watchdog device on it and ignore it.  It craps 
out about once a month now for unknown reasons, but isn't worth 
replacing the hardware, yet.

I have a Dell server to replace it with, but haven't been able to find 
any of my "round 'tuits" out in my garage to bring in to get the project 
done, yet.  :-)

Nate