[OpenAFS-devel] du/find hang running in AFS space

[Garance A Drosihn]

At 3:06 PM +0200 10/17/01, Touretsky, Gregory wrote:
>Hi,
>
>    we get a problem with Open AFS 1.2.2 running on RedHat 6.2 machine
>(Linux iwsl008 2.2.17-14smp #1 SMP Thu May 17 13:26:42 PDT 2001 i686 unknown).
>
>Both find and du stuck when run in AFS space. Any ideas why this happens?
>There is no such problem with IBM AFS. I don't know if this happens in older
>versions of OpenAFS.
>
>I can see that the process doesn't return from system calls 106/107:

Recommendation #1:
     startup linux without afs, unmount the partition which is your
     AFS cache, remake the filesystem on it (blowing away the current
     cache), reboot with afs.

If your AFS cache gets corrupt, it can cause hanging problems like you are
seeing.  I'm wondering if we (here at RPI) should take the performance hit
and turn 'sync' on the filesystem holding the AFS cache.

Recommendation #2:
     (particularly since you have an SMP kernel there) Rebuild the
     system with Redhat 7.1 and use the version of openAFS for that.
     I realize this is a bit more painful, but here at RPI we were
     having a lot of problems with openAFS hanging up on a SMP system,
     and rebuilding the system has made it a lot better.

-- 
Garance Alistair Drosehn            =   gad@eclipse.acs.rpi.edu
Senior Systems Programmer           or  gad@freebsd.org
Rensselaer Polytechnic Institute    or  drosih@rpi.edu