[lug] Daily system crash....

[Zan Lynx]

On Wed, 2008-02-06 at 10:30 -0700, George Sexton wrote:
> 
> Zan Lynx wrote:
> > 
> > You need a VGA or FB console, a serial console or netconsole, all of
> > which have support for polled, no-interrupt, no-DMA output to be used
> > during a crash.
> 
> Good luck. I have a machine that's dying, and it's not sending ANYTHING 
> to the logged in console.

Try booting with the NMI watchdog enabled.  I add "nmi_watchdog=2,panic"
to my kernel boot line.  The value 1 uses an IO-APIC, usually only
available on multi-processor boards, and 2 uses one of the CPU
performance counters but it'll work on single-CPU systems.

If your machine is getting stuck inside a kernel lock of some sort, the
NMI watchdog can catch it.  I've had freezes where it didn't work
though.  Never figured out why either.
-- 
Zan Lynx <zlynx at acm.org>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
URL: <http://lists.lug.boulder.co.us/pipermail/lug/attachments/20080206/e2625bee/attachment.pgp>