kernel panic after upgrading to 2.6

kernel panic after upgrading to 2.6

[David Dougall]

I am running kernel 2.6.18 after trying 2.6.17.6 and 2.6.17.13 with
similar results.  I continue to have nfs servers kernel panic on a very
regular basis, every few days.  Sometimes I need to do an xfs_repair to
get the filesystem to remount.  I got the following kernel panic message
from the syslog this time:
Message from syslogd@nfs09 at Tue Oct 24 09:35:14 2006 ...
nfs09 kernel: Oops: 0000 [#1]

Message from syslogd@nfs09 at Tue Oct 24 09:35:14 2006 ...
nfs09 kernel: CPU:    0

Message from syslogd@nfs09 at Tue Oct 24 09:35:14 2006 ...
nfs09 kernel: EIP is at xfs_alloc_fix_freelist+0x20/0x486

Message from syslogd@nfs09 at Tue Oct 24 09:35:14 2006 ...
nfs09 kernel: eax: f7dbb000   ebx: 0009a000   ecx: 00000016   edx:
f8cc8c20

Message from syslogd@nfs09 at Tue Oct 24 09:35:14 2006 ...
nfs09 kernel: esi: 00000016   edi: edab7c90   ebp: d8fd8b64   esp:
edab7bd0

Message from syslogd@nfs09 at Tue Oct 24 09:35:15 2006 ...
nfs09 kernel: ds: 007b   es: 007b   ss: 0068

Message from syslogd@nfs09 at Tue Oct 24 09:35:15 2006 ...
nfs09 kernel: Process nfsd (pid: 4837, ti=edab6000 task=eda8f030
task.ti=edab6000)

Message from syslogd@nfs09 at Tue Oct 24 09:35:15 2006 ...
nfs09 kernel: Stack: 00000000 00001a9a 00000009 00000200 00000002 c026fffe
f0995b80 f7dbb000

Message from syslogd@nfs09 at Tue Oct 24 09:35:15 2006 ...
nfs09 kernel:        d8fd8b64 f8cc8c20 c0225507 fffffffe 00000000 0000000b
00000000 00000000

Message from syslogd@nfs09 at Tue Oct 24 09:35:15 2006 ...
nfs09 kernel:        d8fd8b64 f7dbb000 c0226fef 00000001 00000001 00000000
00000000 0000000b

Message from syslogd@nfs09 at Tue Oct 24 09:35:15 2006 ...
nfs09 kernel: Call Trace:

Message from syslogd@nfs09 at Tue Oct 24 09:35:16 2006 ...
nfs09 kernel: Code: 24 04 46 06 00 00 e9 9f fe ff ff 55 57 56 53 81 ec 8c
00 00 00 89 c7 89 54 24 10 8b 40 04 89 44 24 1c 8b 57 0c 89 54 24 24 8b 2f
<80> 3a 00 0f 84 e7 02 00 00 c7 84 24 88 00 00 00 00 00 00 00 8b

Message from syslogd@nfs09 at Tue Oct 24 09:35:16 2006 ...
nfs09 kernel: EIP: [<c022742e>] xfs_alloc_fix_freelist+0x20/0x486 SS:ESP
0068:edab7bd0



I am guessing this is an xfs bug, but I am not sure.  Would anybody please
confirm this form me.
I am running nfs-utils-1.0.9 and xfsprogs 2.8.10.
If there is any more information you need, please let me know.
Please advise.
--David Dougall

Re: kernel panic after upgrading to 2.6

[Eric Sandeen]

David Dougall wrote:
> I am running kernel 2.6.18 after trying 2.6.17.6 and 2.6.17.13 with
> similar results.  I continue to have nfs servers kernel panic on a very
> regular basis, every few days.  Sometimes I need to do an xfs_repair to
> get the filesystem to remount.  I got the following kernel panic message
> from the syslog this time:
> Message from syslogd@nfs09 at Tue Oct 24 09:35:14 2006 ...
> nfs09 kernel: Oops: 0000 [#1]

If you ran 2.6.17 < 2.6.17.7, there was a bug:

http://oss.sgi.com/projects/xfs/faq.html#dir2

I would suggest getting xfsprogs v. 2.8.10 or newer, as referenced in
the faq, repair your filesystem, and see if the problem is resolved.

-Eric

Re: kernel panic after upgrading to 2.6

[David Dougall]

I actually did see that bug report.  That is why I have tried upgrading
now a few times.  I have run xfs_repair on all of my file systems since
upgrading this past time and I still get kernel panics.  I already have
2.8.10 as far as I can tell.  2.8.10-1 on debian.
--David Dougall


On Tue, 24 Oct 2006, Eric Sandeen wrote:

> David Dougall wrote:
> > I am running kernel 2.6.18 after trying 2.6.17.6 and 2.6.17.13 with
> > similar results.  I continue to have nfs servers kernel panic on a very
> > regular basis, every few days.  Sometimes I need to do an xfs_repair to
> > get the filesystem to remount.  I got the following kernel panic message
> > from the syslog this time:
> > Message from syslogd@nfs09 at Tue Oct 24 09:35:14 2006 ...
> > nfs09 kernel: Oops: 0000 [#1]
>
> If you ran 2.6.17 < 2.6.17.7, there was a bug:
>
> http://oss.sgi.com/projects/xfs/faq.html#dir2
>
> I would suggest getting xfsprogs v. 2.8.10 or newer, as referenced in
> the faq, repair your filesystem, and see if the problem is resolved.
>
> -Eric
>
>
>

Re: kernel panic after upgrading to 2.6

[David Dougall]

Here is another kernel panic this morning on another server.  I can't even
tell if they are related at all.
--David Dougall



Oct 24 12:00:51 nfs08 kernel: Filesystem "dm-0": XFS internal error
xfs_trans_ca
ncel at line 1138 of file fs/xfs/xfs_trans.c.  Caller 0xc026030c
Oct 24 12:00:51 nfs08 kernel:  [<c026edb4>] xfs_trans_cancel+0xf1/0x112
Oct 24 12:00:51 nfs08 kernel:  [<c026030c>]
xfs_iomap_write_allocate+0x417/0x56b
Oct 24 12:00:51 nfs08 kernel:  [<c026030c>]
xfs_iomap_write_allocate+0x417/0x56b
Oct 24 12:00:51 nfs08 kernel:  [<c025eac6>] xfs_iomap+0x616/0x835
Oct 24 12:00:51 nfs08 kernel:  [<c027c028>] xfs_map_blocks+0x39/0x50
Oct 24 12:00:51 nfs08 kernel:  [<c027ce92>]
xfs_page_state_convert+0x5ed/0xee1
Oct 24 12:00:51 nfs08 kernel:  [<c0118463>] __do_softirq+0x3e/0x86
Oct 24 12:00:51 nfs08 kernel:  [<c027d88a>] xfs_vm_writepage+0x57/0xdd
Oct 24 12:00:51 nfs08 kernel:  [<c027d833>] xfs_vm_writepage+0x0/0xdd
Oct 24 12:00:51 nfs08 kernel:  [<c0167e92>] mpage_writepages+0x212/0x3c1
Oct 24 12:00:51 nfs08 kernel:  [<c034cf60>] dst_output+0x0/0x7
Oct 24 12:00:51 nfs08 kernel:  [<c027d833>] xfs_vm_writepage+0x0/0xdd
Oct 24 12:00:51 nfs08 kernel:  [<c013469f>] do_writepages+0x23/0x3b
Oct 24 12:00:51 nfs08 kernel:  [<c012f2b6>]
__filemap_fdatawrite_range+0x66/0x72
Oct 24 12:00:51 nfs08 kernel:  [<c01300a6>] filemap_fdatawrite+0x23/0x27
Oct 24 12:00:51 nfs08 kernel:  [<c01f9143>] nfsd_sync+0x3e/0x96
Oct 24 12:00:51 nfs08 kernel:  [<c01f9d66>] nfsd_open+0x117/0x15d
Oct 24 12:00:51 nfs08 kernel:  [<c01f9efb>] nfsd_commit+0x80/0xb7
Oct 24 12:00:51 nfs08 kernel:  [<c0201abd>]
nfs3svc_decode_commitargs+0x0/0xc1
Oct 24 12:00:51 nfs08 kernel:  [<c02003f4>] nfsd3_proc_commit+0xde/0xf7
Oct 24 12:00:51 nfs08 kernel:  [<c0201abd>]
nfs3svc_decode_commitargs+0x0/0xc1
Oct 24 12:00:51 nfs08 kernel:  [<c01f66db>] nfsd_dispatch+0x76/0x1b7
Oct 24 12:00:51 nfs08 kernel:  [<c0384447>] svc_process+0x3d5/0x68e
Oct 24 12:00:51 nfs08 kernel:  [<c0385864>] svc_recv+0x22b/0x47f
Oct 24 12:00:51 nfs08 kernel:  [<c01f6996>] nfsd+0x17a/0x304
Oct 24 12:00:51 nfs08 kernel:  [<c0102c02>] work_resched+0x5/0x16
Oct 24 12:00:51 nfs08 kernel:  [<c01f681c>] nfsd+0x0/0x304
Oct 24 12:00:51 nfs08 kernel:  [<c0101005>] kernel_thread_helper+0x5/0xb
Oct 24 12:00:51 nfs08 kernel: xfs_force_shutdown(dm-0,0x8) called from
line 1139
 of file fs/xfs/xfs_trans.c.  Return address = 0xc02854d0
Oct 24 12:00:51 nfs08 kernel: Filesystem "dm-0": Corruption of in-memory
data de
tected.  Shutting down filesystem: dm-0
Oct 24 12:00:51 nfs08 kernel: Please umount the filesystem, and rectify
the prob
lem(s)
Oct 24 12:00:52 nfs08 kernel: BUG: unable to handle kernel NULL pointer
derefere
nce at virtual address 0000030c
Oct 24 12:00:52 nfs08 kernel:  printing eip:
Oct 24 12:01:55 nfs08 kernel: c026e723
Oct 24 12:03:49 nfs08 syslogd 1.4.1#18: restart.


On Tue, 24 Oct 2006, David Dougall wrote:

> I actually did see that bug report.  That is why I have tried upgrading
> now a few times.  I have run xfs_repair on all of my file systems since
> upgrading this past time and I still get kernel panics.  I already have
> 2.8.10 as far as I can tell.  2.8.10-1 on debian.
> --David Dougall
>
>
> On Tue, 24 Oct 2006, Eric Sandeen wrote:
>
> > David Dougall wrote:
> > > I am running kernel 2.6.18 after trying 2.6.17.6 and 2.6.17.13 with
> > > similar results.  I continue to have nfs servers kernel panic on a very
> > > regular basis, every few days.  Sometimes I need to do an xfs_repair to
> > > get the filesystem to remount.  I got the following kernel panic message
> > > from the syslog this time:
> > > Message from syslogd@nfs09 at Tue Oct 24 09:35:14 2006 ...
> > > nfs09 kernel: Oops: 0000 [#1]
> >
> > If you ran 2.6.17 < 2.6.17.7, there was a bug:
> >
> > http://oss.sgi.com/projects/xfs/faq.html#dir2
> >
> > I would suggest getting xfsprogs v. 2.8.10 or newer, as referenced in
> > the faq, repair your filesystem, and see if the problem is resolved.
> >
> > -Eric
> >
> >
> >
>