From: Andreas Dilger <adilger@sun.com>
Subject: Re: Scripting e2fsck: no errors,
 but still exit code 1 "FILE	SYSTEM WAS MODIFIED"
Date: Mon, 19 May 2008 21:58:28 -0600
Message-ID: <20080520035828.GQ3491@webber.adilger.int>
References: <483006F1.6080008@tuxes.nl>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7BIT
Cc: linux-ext4@vger.kernel.org
To: Bas van Schaik <bas@tuxes.nl>
In-reply-to: <483006F1.6080008@tuxes.nl>
Content-disposition: inline
Sender: linux-ext4-owner@vger.kernel.org

On May 18, 2008  12:37 +0200, Bas van Schaik wrote:
> ... there is a slight
> problem with scripting e2fsck: it seems that e2fsck /always/ exits with
> exit code 1 just because of the fact that the snapshot journal has been
> replayed.

> > e2fsck 1.40.8 (13-Mar-2008)
> > /dev/loop1: recovering journal
> > Pass 1: Checking inodes, blocks, and sizes
> > Pass 2: Checking directory structure
> > Pass 3: Checking directory connectivity
> > Pass 4: Checking reference counts
> > Pass 5: Checking group summary information
> >
> > /dev/loop1: ***** FILE SYSTEM WAS MODIFIED *****

You shouldn't get journal replay after a proper LVM snapshot.  The
LVM code should freeze the filesystem temporarily, flush out the journal,
mark the filesystem clean, take the snapshot of the clean fileysstem,
and then unfreeze the filesystem again.

If you are using LVM for snapshots, and this is not happening, then
the snapshot code is broken.  If you are NOT using LVM, then you
should look into the "freeze ioctl" patches that were on linux-ext4 and
linux-fsdevel recently "[RFC PATCH 1/3] Implement generic freeze feature"
from Takashi Sato.

Cheers, Andreas
--
Andreas Dilger
Sr. Staff Engineer, Lustre Group
Sun Microsystems of Canada, Inc.