[OpenAFS] Linux + 32bit files + butc writing to file = obnoxiousness

[Robin Yamaguchi]

Thank you Derrick for your help thus far.

A multitheaded core dump patch doesn't seem to be available for 2.4.20,
which is what Redhat 9 ships with.  I installed 2.4.18 from source, using
my current config, and installed the patch.  After booting 2.4.18, I
recompiled the afs module from the openafs-kernel-src.rpm.

Of course, AFS hasn't crashed yet.  butc has been eating 99% cpu for the
past 48 hours, and AFS hasn't flinched.  I've also tried tarring the entire
filesystem (these two things brought down afs numerous times before on our
system), and no amount of cpu/disk usage is bringing afs down.

Strangely, over the weekend, when we were still running 2.4.20, butc ran at
99% cpu w/ out afs crashing.  I'm banging my head against the wall...

Coinciding with this issue, we've been using an Adaptec IDE raid card that
has been giving us grief.  Using raid-5, we had two drives fail last week,
which later tested ok and working.  The IRQ's on the raid card were the
same as the NIC and video card.  Adaptec suggested configuring the raid car
to use it's own IRQ.  We did that, yet a drive failed again after
rebuilding the array, and I got AFS to crash.  I don't think the problems
with the raid card are related to AFS, but I thought I'd mention it...

Despite all of the above, we're still pushing a deadline to get this server
into production.  Along with getting a new raid card, we're thinking about
running redhat 7.3, and openafs 1.2.8 - we used this os and afs version on
our test server for a few months without any of the issues i've been
outlining.  I'd love to figure out what the hell is going on w/ out
current situation, but time is running out for us.

Before we proceed, do you guys have any more ideas, suggestions, warnings
involving using 1.2.8?

thanks much,
Robin

--------------------------
Robin H. Yamaguchi
Physics Computing Services
UC Santa Barbara
805-893-8366
--------------------------

On Fri, 9 May 2003, Derrick J Brashear wrote:

> On Fri, 9 May 2003, Robin Yamaguchi wrote:
>
> > > b) you patch your kernel to give useful cores
> >
> > I searched around and couldn't find the appropriate patch.  Can you point
> > me in the right direction?
>
> Here's the more current version:
> http://hypermail.idiosynkrasia.net/linux-kernel/archived/2002/week50/att-0264/01-tcore-2418-061802.pat
>
> _______________________________________________
> OpenAFS-info mailing list
> OpenAFS-info@openafs.org
> https://lists.openafs.org/mailman/listinfo/openafs-info
>