On 14 Jan, Harry Putnam wrote:
> I'm experiencing a problem where during an X session the machine
> inexplicably freezes up... networking along with mouse keyboard etc.
> No access is then possible via ssh or any other way other than a hard
> reboot.
>
> I've attempted to debug the problem by first searching the logs. But
> not finding anything I recognize as a clue.
>
> I've inserted a line like this in /etc/syslog.conf
> cat /etc/syslog.conf
> [...]
> *.* -/var/log/debug.log
> [...]
>
> and then created a tiny script called by cron every 5 minutes that
> simply pinged an external host and reported the results by way of
> piping to `logger', guaranteeing it would appear in the system logs .
>
> Then after a freeze up and reboot, search the logs for the most recent
> successful ping and see what happened after that. (The freeze up freezes
> networking too so the ping would then fail). And Hopefully I'd see
> something of note between the last successful ping and the reboot.
>
> But I see absolutely nothing of note.
>
> The freeze does not appear to obey any particular time frame after a
> boot, only that sooner or later a freeze up will occur.
>
> Apparently there is nothing happening that merits a log entry.
>
> Its been going on for a goodly while and I've just been rebooting as
> needed ... I haven't been really active on the machine for a while so
> it was easy to get along with, but now I need to do more work and so
> the problem is a major stumbling block.
>
A few months, ago, I had similar symptoms.
It turned out to be a defective AMD Phenom CPU
(a very tricky error since all hardware tests passed)
I guess it's a cache coherence problem (the Phenom has 4 cores).
Luckily it was quite a new machine so my dealer
replaced the CPU. The problem didn't occur ever since.
While searching the net for a reason of failure I came accross
messages about failing graphics chips / or graphics drivers.
Perhaps you try an open source driver for your graphics card
and see if the problem still occurs.
(Before I suspected the CPU, I have tortured memory and I have
even replaced the motherboard)
To "exclude" a software problem, try booting from a
SystemRescue-CD
http://www.sysresccd.org/Main_Page
It's Gentoo based and sports an X-Server.
That are the problems that makes one crazy!
Helmut.
--
Helmut Jarausch
Lehrstuhl fuer Numerische Mathematik
RWTH - Aachen University
D 52056 Aachen, Germany