>On Wed, Oct 18, 2006 at 10:41:58PM +0200, Lasse Hillerĝe Petersen wrote:
>> * Am I correct in assuming that the RAM is not necessarily to blame here, IOW, 
>> can memory faults occur due to other reasons than bad RAM?
>
>ksh prints "Memory fault" where sh and bash print "Segmentation fault". It
>is the same thing.

When I had the original 512MB RAM block, I consistently got "Memory fault"
most of the time. I was under the impression that this error was related to
physical memory, whereas Segmentation fault was "less serious" "pointer
errors", related to virtual memory, so to speak.

Since replacing the memory, I have had at most one "Memory fault".

>Such random segmentation faults are most often caused by some hardware
>problem (bad memory, overheat ...)

Of course I can't rule that out completely...

I am more inclined to suspect the OS, as this error - despite some randomness
- only occurs under high load. After all, I had bash complete the operation
successfully. Using top I saw 1800+ MB shown under "File" - I suppose that
means memory mapped files?

I also used the machine to compile hundreds of packages from pkgsrc, so it
can handle a big load fairly well. I had problems with /bin/sh still giving
Segmentation Faults now and then, but replacing it with /bin/ksh seemed to
solve these problems.

Someone else has suggested that there might be a problem with having that
large amount of RAM? Is this an issue in 4.0BETA?

What I find most confusing is the seemingly different behaviour observed with
the PATA disk versus the SATA disk. This indicates - to me - some driver
problem?

At the moment I am torn between removing a RAM block and run NetBSD 3,
install amd64, or install current - as was suggested by another person.

-Lasse