Unexplained Segmentation Faults

[Brian Kelsay]

Like Chris said, check to see that the /var is not full.  If you have a
hardware failure it is more likely to be in the chipset on the motherboard.
I think it is the North Bridge that contains the hard drive controller.
Put another drive in the system with a fresh OS load.  Any OS.   And see if
you get write errors still.
I had this problem with a users' Pentium 4 PC (6 months old) running Windows
XP.  I replaced the drive and ghosted a fresh image to the PC.  On first
reboot after ghost I got a blue screen of Death with kernel errors.   I
tried to load from CD to eleminate the networking element and got the same
result.   I replaced the motherboard and it was doing the same thing.  Safe
mode didn't help, but did point out that the last driver loaded was a module
for the AGP set.   So next I replaced the motherboard again and also the
video card.   I should have replaced one thing at a time to prove which it
was, but I was in a hurry and had the parts from the OEM so one of the two
things fixed it.
You can also remove all cards, but the video and unplug CD and floppy and
any non-essential devices, add one thing at a time until the problem
reappears.   This method can help with any hardware troubleshooting.
Hope it helps.
Brian

----- Original Message -----
From: "James Colannino" <email2jamez at covad.net>

> Hey everyone.  I've had some problems with an old machine of mine for a
> while.  I have a P200 MMX running Slackware 8.1 that's worked very well
> so far.  The BIOS is able to recognize all 13GB of my HD which I've
> found to be quite rare with such an old machine, and it seemed for a
> while to be pretty stable.
>
> However, everytime I try to run "updatedb" so that I can locate files,
> it works for a few minutes and then always without fail will give me a
> segmentation fault.  Then, if I've attempted that command, everytime I
> halt or restart the system, it'll give me a segmentation fault as it's
> unmounting all remote filesystems (I don't actually have any -- I'm
> planning on removing all references to nfs from the system.)  If I
> haven't tried to updatedb, it halts and restarts without any problems.
>
> After I've gotten one segmentation fault from running the command
> "updatedb," from now on, until I reboot the system, it doesn't even work
> the Hard Drive at all before I get the same error, this time saying
> something about a NULL Pointer.
>
> This is really frustrating me.  I'm pretty sure from other observations
> I've made that excessive writing (and/or reading -- not sure) to the
> disk is what's producing these errors.
>
> Now for my question.  Do you think maybe the processor is bad, or do you
> think maybe there's something wrong with the IDE controllers?  This is
> kind of an odd problem.  I'm certain that the disk is good.  I've had
> very good luck with it thus far (of course I'll probably plunk it in
> another system and give it a good diagnostical checkup to make sure.)  I
> tried running the processor at 166 Mhz as opposed to 200 Mhz in a vain
> attempt to see if maybe for some reason the CPU was overheating, but
> that I see is most likely not the problem.  I did over clock it once to
> 233 Mhz, and had the segmentation faults start occuring (this was before
> I'd used the system pretty much at all though,) so I originally thought
> that maybe I over heated the CPU and thus set it back to 200.  It seemed
> to work alright, but it's still producing these errors (only when I use
> the disk a great deal though.)  Very strange.  Any ideas?
>
> James