Re: Circular lock / deadlock in kernel client

[Amon Ott <a.ott@m-privacy.de>]

On Thursday 08 December 2011 wrote Amon Ott:
> On Wednesday 07 December 2011 wrote Sage Weil:
> > I pushed a patch to wip-d-lock that may fix this one, but unfortunately
> > don't have time to test this very carefully right now.  Let us know if
> > that helps, or you can wait until next week.
>
> Rebuilding kernel with that patch.

With the patch the deadlock disappears and the kernel does not hang at the 
locks afterwards.

> > The call path that was triggering both of these can be exercised by
> > restarting the ceph-mds daemon.  Try running your client for a bit and
> > the doing that and see if you get any more splats.
>
> What triggered the kernel problem was bug 1047. ceph-mds crashed on all
> nodes with that assert. When the kernel detected that the main mds
> connection was missing, it tried to reconnect and hung.

This problem remains, and without a working mds the whole ceph mount hangs. 
Instead of crashing, mds sometimes goes into a dead loop and uses a cpu core 
at 100%. This also makes the mount hang.

Amon Ott
-- 
Dr. Amon Ott
m-privacy GmbH           Tel: +49 30 24342334
Am Köllnischen Park 1    Fax: +49 30 24342336
10179 Berlin             http://www.m-privacy.de

Amtsgericht Charlottenburg, HRB 84946

Geschäftsführer:
 Dipl.-Kfm. Holger Maczkowsky,
 Roman Maczkowsky

GnuPG-Key-ID: 0x2DD3A649
--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html