Re: random errors with bzip2

[Ville Herva <vherva@mail.niksula.cs.hut.fi>]

On Tue, Jun 19, 2001 at 06:11:48PM +0200, you [André Dahlqvist] claimed:
> Rodrigo Ventura <yoda@isr.ist.utl.pt> wrote:
> 
> > - it could be a memory problem, but if it were, lots of kernel
> > oops were expected, right?
> 
> This certainly sounds like a memory problem. I experienced almost the same
> behaviour with a box some years ago, and it turned out to be memory. The
> kernel didn't oops, and I actually had to run several kernel compiles at
> the same time to have gcc die.
> 
> Try memtest86 on the suspect box.

Seconded.

Exactly the same symptoms (bzip2); the culprit turned out to be memory.

That's when I wrote memburn (http://v.iki.fi/~vherva/memburn.c) for quick
testing without a boot (it did find the problem) and I then verified the
problem with memtest86 (http://reality.sgi.com/cbrady_denver/memtest86/).
You do have to run either for hours, propably for days to be sure. 

The box has now ran perfectly for a year or so with the BadRam patch from
Rick van Rein (http://rick.vanrein.org/linux/badram/).


-- v --

v@iki.fi