[lug] (no subject)

[D. Stimits]

keith.herold at cox.net wrote:
> 
> Howdy!  I have just started using a dual processor machine (2x 1.7G, w/3G RAM) and RH 7.1 .  Install works fine and the smp kernel (2.4.7-10smp ) appears to work, but occasionally (randomly, and not triggered by a particular application) the system goes dead (the monitor displays the 'no signal' warning) and will not wake up (the HD light just blinks, and the power light on the box is off).  I need to be able to run 8 hour+ processes (using both
> CPU's), but it doesn't seem to be stable.
> 
> I worked at another place that had a similar problem (not the same manufacturer though) and they were never able to resolve it.  Can anyone here help?


What chipset? The i840 has a defective IO-APIC that does this by sending
an interrupt to a non-existent interrupt handler at times, I think irq
217. There is often no log message, since the failure locks up IO, but
on dual cpu, sometimes one of the cpu's manages to log the problem prior
to lockup. Search /var/log/messages for "interrupt" (and search older
messages files, since chron will rotate them). Unless it is a known good
chipset like BX, try to pass at the boot prompt the kernel option
"noapic" and see if that helps. In any case, there have been a lot of
bug fixes since that kernel, you may want to just attempt the RH 2.4.18
kernel.

D. Stimits, stimits at idcomm.com

> 
> --Keith
> 
> _______________________________________________
> Web Page:  http://lug.boulder.co.us
> Mailing List: http://lists.lug.boulder.co.us/mailman/listinfo/lug
> Join us on IRC: lug.boulder.co.us port=6667 channel=#colug