Hi Steve,
On Thu, Jul 08, 2004 at 11:54:23AM +1200, Steve Wray wrote:
> ...
> This is a little way off yet so I have an opportunity to prepare some
> testing; we will be doing a full backup and rebuild so I can arrange
> for some destructive testing.
>
> If theres anyone on list who wants to help out with advice, debugging
> etc please let me know.
Could you post a summary of your problems please? Also adding
information about the things you being done when problems arose,
points where you ran repair and what it found, the points where
you were reading from the block device concurrently, with which
kernel versions, what problems still remained after stopping
that (whether you repaired inbetween), etc, etc. The xfs_info
output on the affected filesystem(s) is also very useful.
It became quite difficult to follow what state your filesystem
might have been in during your earlier posts, so there was not
much anyone could do to help out there.
If you can find test cases from the problems you experienced,
that show going from a filesystem in known state A, applied a
sequence of operations X, Y and Z, and ended up with filesystem
or file data in bad state B - that is the sort of thing I can
best help out with, because I can then reproduce and analyse it
locally and figure out if XFS is at fault, etc.
thanks.
--
Nathan
rom owner-linux-xfs@xxxxxxxxxxx Wed Jul 7 17:47:27 2004
Received: with ECARTIS (v1.0.0; list linux-xfs); Wed, 07 Jul 2004 17:47:36
-0700 (PDT)
Received: from mail2.catalyst.net.nz (godel.catalyst.net.nz [202.49.159.12])
by oss.sgi.com (8.12.10/8.12.9) with SMTP id i680lPgi022096
for <linux-xfs@xxxxxxxxxxx>; Wed, 7 Jul 2004 17:47:26 -0700
Received: from leibniz.catalyst.net.nz ([202.49.159.7]
helo=shankara.wgtn.cat-it.co.nz)
by mail2.catalyst.net.nz with asmtp
(Cipher TLSv1:RC4-MD5:128) (Exim 3.35 #1 (Debian))
id 1BiN4C-00062U-02; Thu, 08 Jul 2004 12:47:12 +1200
From: Steve Wray <stevew@xxxxxxxxxxxxxxx>
Reply-To: stevew@xxxxxxxxxxxxxxx
To: Nathan Scott <nathans@xxxxxxx>
Subject: Re: debugging xfs issue (maybe)
Date: Thu, 8 Jul 2004 12:47:01 +1200
User-Agent: KMail/1.6.1
Cc: linux-xfs@xxxxxxxxxxx
References: <200407081154.25556.stevew@xxxxxxxxxxxxxxx>
<20040708103041.D1946083@xxxxxxxxxxxxxxxxxxxxxxxx>
In-Reply-To: <20040708103041.D1946083@xxxxxxxxxxxxxxxxxxxxxxxx>
MIME-Version: 1.0
Content-Disposition: inline
Content-Type: Multipart/Mixed;
boundary="Boundary-00=_FmJ7AdN8Hy8jlcr"
Message-Id: <200407081247.11376.stevew@xxxxxxxxxxxxxxx>
X-System-Filter-Id: mail2.catalyst.net.nz 1BiN4C-00062U-02
X-Virus-Scanned-By: Amavis with CLAM Anti Virus on mail2.catalyst.net.nz
X-archive-position: 3611
X-ecartis-version: Ecartis v1.0.0
Sender: linux-xfs-bounce@xxxxxxxxxxx
Errors-to: linux-xfs-bounce@xxxxxxxxxxx
X-original-sender: stevew@xxxxxxxxxxxxxxx
Precedence: bulk
X-list: linux-xfs
--Boundary-00=_FmJ7AdN8Hy8jlcr
Content-Type: Text/Plain;
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
On Thursday 08 July 2004 12:30, Nathan Scott wrote:
> Hi Steve,
>
> On Thu, Jul 08, 2004 at 11:54:23AM +1200, Steve Wray wrote:
> > ...
> > This is a little way off yet so I have an opportunity to prepare
> > some testing; we will be doing a full backup and rebuild so I can
> > arrange for some destructive testing.
> >
> > If theres anyone on list who wants to help out with advice,
> > debugging etc please let me know.
>
> Could you post a summary of your problems please?=20=20
Well... the first thing we notice is that a nightly cron job returns things=
like;
/etc/cron.daily/standard:
find: /var/lib/postgres/data/base/141159285/141159709: Unknown error 990
find: /var/lib/postgres/data/base/141159285/141159709: Unknown error 990
find: /var/lib/postgres/data/base/141159285/141159709: Unknown error 990
and we look in dmesg and see things like;
Filesystem "dm-0": corrupt dinode 67279401, extent total =3D 1025, nblocks =
=3D 1. Unmount and run xfs_repair.
0x0: 49 4e 81 b4 01 02 00 01 00 00 04 10 00 00 00 64
Filesystem "dm-0": XFS internal error xfs_iformat(1) at line 475 of file fs=
/xfs/xfs_inode.c. Caller 0xc024630a
[<c023d26a>] xfs_error_report+0x3a/0x3c
[<c023d363>] xfs_corruption_error+0x3b/0x48
[<c024630a>] xfs_iread+0x132/0x23c
[<c0245172>] xfs_iformat+0x1e6/0x558
[<c024630a>] xfs_iread+0x132/0x23c
[<c024630a>] xfs_iread+0x132/0x23c
[<c0243c08>] xfs_iget_core+0x26c/0x5c0
[<c0243fe8>] xfs_iget+0x8c/0x164
[<c025e677>] xfs_dir_lookup_int+0x63/0xc8
[<c0263052>] xfs_lookup+0x3e/0x68
[<c026da8f>] linvfs_lookup+0x3f/0x80
[<c015c819>] real_lookup+0x59/0xcc
[<c015ca8d>] do_lookup+0x45/0x84
[<c015d137>] link_path_walk+0x66b/0x944
[<c015d7b5>] path_lookup+0x185/0x18c
[<c015d904>] __user_walk+0x28/0x40
[<c0158af2>] vfs_lstat+0x16/0x44
[<c015910b>] sys_lstat64+0x13/0x30
[<c01061d9>] error_code+0x2d/0x38
[<c010574f>] syscall_call+0x7/0xb
> Also adding=20
> information about the things you being done when problems arose,
So far we've just sucked in our guts and kept at it; the machine in question
is fairly key here and the corruption hasn't so far impacted anything we=20
couldn't sort out. The main thing we've been doing is reboot into single us=
er
and run xfs_repair over all the xfs filesystems.
I have scripted this and generated logs of the output of xfs_repair.=20
I'll tar them up and attach them.
> points where you ran repair and what it found, the points where
> you were reading from the block device concurrently, with which
> kernel versions,=20
various 2.6 kernels, currently its on 2.6.7
As for the other details we havn't had a chance to do any *real* work
getting this sort of info or running any real tests to see exactly what kind
of behavior produces the problems. Its busy here, very busy.
> what problems still remained after stopping=20
> that (whether you repaired inbetween), etc, etc. The xfs_info
> output on the affected filesystem(s) is also very useful.
I'll include this with the logs from the last xfs_repair
> It became quite difficult to follow what state your filesystem
> might have been in during your earlier posts, so there was not
> much anyone could do to help out there.
>
> If you can find test cases from the problems you experienced,
> that show going from a filesystem in known state A, applied a
> sequence of operations X, Y and Z, and ended up with filesystem
> or file data in bad state B - that is the sort of thing I can
> best help out with, because I can then reproduce and analyse it
> locally and figure out if XFS is at fault, etc.
So far we havn't been able to reliably reproduce the problem on demand.
Its a very active box. We reboot it, xfs_repair everything,
bring it up multiuser and sometimes the same day, sometimes weeks=20
later we find corruption. One time we went over a month without any
problems and thought the latest kernel upgrade had fixed it, then
it came back.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)
iD8DBQFA7JmOmVx2hyhuTucRAl18AKDOyY1o+QIIdWj7JQ7mB8N9RVKX/gCfeQT4
OdtTslYqXmJyUb/Vp0tAijM=3D
=3DpKUh
-----END PGP SIGNATURE-----
--Boundary-00=_FmJ7AdN8Hy8jlcr
Content-Type: application/x-tbz;
name="xfs.logs.tar.bz2"
Content-Transfer-Encoding: base64
Content-Disposition: attachment;
filename="xfs.logs.tar.bz2"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--Boundary-00=_FmJ7AdN8Hy8jlcr--
|