Allright,
I traced the problem down - rather a lucky coincidence - to the fact that
listing a certain directory's content causes the crash!
Subdirectories of this certain directory are accessible. So the problem
lies in the Reiserfs filesystem modules.
I consulted the users and it seems they were trying to put very large files
on the server. They are not sure of this is the case for the problem dir.
So, for now, we know what causes the problem, and the next thing I will try
is manually upgrade the kernel and reiserfs modules / utils to see if I can
recover from the problem/avoid similar problems in the future.
Thanks for the advice!
Lieven
On 2001.03.12 11:53 A V Flinsch wrote:
> On Monday 12 March 2001 04:21, you wrote:
>
> >
> > The station was set-up about 3 weeks ago and proved until last week to
> > be stable. Since last week the station has begun to ran into several
> > kernel panics a day, first with a lower frequency (say twice a day),
> > now after a few minutes of uptime.
> >
> > OK, so I expected some kind of hardware failure, and replaced the
> > station completely with another, keeping the harddisks in the original
> > configuration.
> >
> > This didn't solve the problem, so now I am thinking of a software
> > failure.
>
> The only parts of your original server happen to be the disks, and it
> still happens. Any possibility that the hardware failure is on the disks
> themselves, or something to do with parameters that impact the disks?
>
> Do you run hdparm during the boot process? If you do, try turning it off,
>
> you will get lower performance, but perhaps a more stable machine.
>
> Do you have any autotune parms in the boot? if so try turning them off
> and see if that fixes the situation.
>
> --
> Alex
> Kernel Panic is General Failure's second in command
>
>
>
