Bug#534545: linux-image-2.6.26-2-openvz-686: server is hangs on ?heavy? disk io on ocfs2

[Jonathan Nieder <jrnieder@gmail.com>]

notfound 534545 2.6.26-24
quit

Hi,

Szabolcs Gyuris wrote:

> I use openvz on ocfs2 fs which is on drbd on sw mirror raid.
>
> Once or twice a day the server froze.
>
> I got a lot of:
> [116049.348676] (10868,0):dlmunlock_common: 128 ERROR: lockres F000000000000000028c8f0c0517267: Someone is calling dlmunlock while waiting for an ast!<3>(10868,0):dlmunlock:685 ERROR: dlm status = DLM_BAD PARAM
> [116049.348676] (10868,0):ocfs2_cancel_convert:3001 ERROR: DLM error -22 while calling ocfs2_dlm_unlock on resource F000000000000000028c8f0c0517267
> [116049.348676] (10868,0):ocfs2_flock_handle_signal:1505 ERROR: status = -22
> [116049.348676] (10868,0):ocfs2_do_flock:79 ERROR: status = -22
>
> but it seem to be non-fatal.
>
> before the server hangs:
> [222288.239258] (4197,3):dlmunlock_common:128 ERROR: lockres F00000000000000002e601561c073e7: Someone is calling dlmunlock while waiting for an ast!<3>(4197,3):dlmunlock:685 ERROR: dlm status = DLM_BADPARAM
[.
> [222310.043131] kernel BUG at fs/ocfs2/dlmglue.c:678!

Sorry for the long silence.  Basic questions:

 - can you still reproduce this?  Are you still using lenny?
 - is the "kernel BUG" line the same every time it happens?
 - any other weird symptoms?  Workarounds?  How have you been
   coping since then?

Sorry for the trouble,
Jonathan