On Tue, 2003-06-10 at 16:06, Jesse Kline wrote:
> >Hope you have lots of room for email if you join that mailing list :) 
> > It's pretty busy.  You might want to try testing your RAM
> > (http://www.memtest86.com/).  Segmentation faults in processes and kernel
> > panics are signs of possible buggy software, however in my experience it
> > is more often then not a sign of dying or dead hardware.... bad RAM,
> > faulty CPU, motherboard, etc.  Sometimes it is hard to pin down which one
> > but I would definitely try testing the RAM.  You could also try a compile
> > of a kernel (With a .config you know works) and see if it seg faults, if
> > it does try it again and see if it seg faults again in the same place or
> > different place.  If you are able to test the box for a while (Ie, you can
> > have it offline for the day), then you can try the kernel compile thing
> > using a loop and let it compile the kernel over and over again for a few
> > hours.  Sometimes faulty CPUs only show signs of faultiness if they are
> > really pushed for extended periods of time (Ie, they get too hot).
> 
> lmsensors is also very good for seeing if your processor is overheating.
> You can either record it in a log file, or keep an app like gkrellm
> running in X, so that when your box crashes you can see what the
> processor temp. is.
> 
> Jesse

Looks like it was the memory. It seems to be cruising along just fine
now with some new kingston ram in it. At least I've compiled a kernel
with no errors, and transfered some large files with no problems.

Thanks for all the help folks!

-- 
Neil Jolly

(with Yoda-like voice)
"Confrontation leads to anger...  Anger leads to fear...  Fear leads
to using Windows NT in mission-critical combat systems...  And this is
how the ancients fell...

Reply via email to