Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 7 Mar 2007 05:02:18 +0200
From:      Kostik Belousov <kostikbel@gmail.com>
To:        freebsd-stable@freebsd.org
Subject:   Re: LOR #193
Message-ID:  <20070307030218.GP10453@deviant.kiev.zoral.com.ua>
In-Reply-To: <45EDE0C5.1010305@netfence.it>
References:  <45EDE0C5.1010305@netfence.it>

next in thread | previous in thread | raw e-mail | index | archive | help

--LwbuP8dfxhLLLUfV
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Tue, Mar 06, 2007 at 10:44:37PM +0100, Andrea Venturoli wrote:
> Hello.
>=20
> I'm experiencing the above mentioned LOR on a 6.2p1/amd64 box (running=20
> gmirror and SMP if that matters).
>=20
> With reference to your question on=20
> http://lists.freebsd.org/pipermail/freebsd-stable/2006-November/031048.ht=
ml:
>=20
> >What application you run that triggers the LOR ?=20
>=20
> Bacula, I guess.
> I'm taking filesystem snapshots, running the backup job and deleting the=
=20
> snapshots.
> In fact I've always seen some problems with some files in the snapshots=
=20
> not being accessible to bacula-fd.
>=20
> In a previous (quite old) thread it was in fact suggested I might be=20
> seeing some LOR, but only recently I activated all the debugging stuff.
The (usual) consequence of the LOR is lock up.
>=20
> What's the risk of running the suggested patch on a (quite critical)=20
> production server?
It shall be safe unless you run filesystems compiled as modules, that where
not built against patched kernel (patch changes the kernel binary
interface).
>=20
> BTW: Sometimes, upon reboot, delayed fscks start and say that the=20
> filesystem cannot be fixed with -p and I should run a full fsck. If I=20
> reboot in single user mode and run a full fsck, it will find no errors.
>=20
> Also, I have a couple of other boxes on which I run bacula this way and=
=20
> I never experienced this problem: they are respectively i386/gmirror/UP=
=20
> and amd64/hardware RAID/SMP; so, might the combination of amd64/gmirror=
=20
> or gmirror/SMP be in the way?
I doubt that gmirror could affect this.

--LwbuP8dfxhLLLUfV
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (FreeBSD)

iD8DBQFF7is5C3+MBN1Mb4gRAg/VAJ4qVA+YIxdjjtQMkQb0vYY0JC4bygCfSJA+
52CqP1chn1sFmd6pNoEMfrU=
=P9Y4
-----END PGP SIGNATURE-----

--LwbuP8dfxhLLLUfV--



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20070307030218.GP10453>