From: Eric Sandeen <sandeen@redhat.com>
Subject: Re: [BUG] e2fsprogs: mke2fs -S and e2fsck cannot recover on ext4
Date: Thu, 27 Oct 2011 15:02:16 -0500
Message-ID: <4EA9B8C8.7070302@redhat.com>
References: <4EA9048C.8050801@rs.jp.nec.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=ISO-2022-JP
Content-Transfer-Encoding: 7bit
Cc: Theodore Tso <tytso@mit.edu>, adilger.kernel@dilger.ca,
	ext4 development <linux-ext4@vger.kernel.org>
To: Akira Fujita <a-fujita@rs.jp.nec.com>
Return-path: <linux-ext4-owner@vger.kernel.org>
Received: from mx1.redhat.com ([209.132.183.28]:29346 "EHLO mx1.redhat.com"
	rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
	id S1752520Ab1J0UCX (ORCPT <rfc822;linux-ext4@vger.kernel.org>);
	Thu, 27 Oct 2011 16:02:23 -0400
In-Reply-To: <4EA9048C.8050801@rs.jp.nec.com>
Sender: linux-ext4-owner@vger.kernel.org
List-ID: <linux-ext4.vger.kernel.org>

On 10/27/11 2:13 AM, Akira Fujita wrote:
> Hi,
> 
> If all of the superblock and backup superblocks are corrupted,
> mke2fs -S which reinitializes the superblock and
> group descriptors only is a last resort to recover.
> 
> However, on ext4, all of the files are lost with this way.
> Because mke2fs -S for ext4 sets uninit_bg and then later e2fsck regards
> all of block group as uninitialized one so that it can not recover
> with bitmaps.
> 
> Of course ext3 can recover all of the files with mke2fsk -S,
> ext4 should be able to recover as well.
> There is a avoidance (see #3 of the reproduce steps)
> but an appropriate fix would be needed for e2fsck.
> 
> * Reproduce steps and log are as follows:
>   My e2fprogs was e2fsprogs-1.42-WIP
> 
> 1. mkfe2s -t ext4 /dev/sdaX
> 2. create some files on ext4
> 3. mkfe2s -t ext4 -S -b 4096 /dev/sdaX　
>     # This problem does not occur,
>     # when we remove uninit_bg feature flag (-O ^uninit_bg).

It seems to me that the main bug here is that mkfs.ext4 should
never mark blockgroups as uninit, when run with -S.  Marking
them uninit goes against the described "not touching the inode
table  and  the  block  and  inode bitmaps" behavior...

-Eric

> 4. e2fsck /dev/sdaX
> 
> e2fsck 1.42-WIP (16-Oct-2011)
> Backing up journal inode block information.
> 
> /dev/sda8 contains a file system with errors, check forced.
> Pass 1: Checking inodes, blocks, and sizes
> Pass 2: Checking directory structure
> Pass 3: Checking directory connectivity
> Root inode not allocated.  Allocate<y>? yes
> 
> /lost+found not found.  Create<y>? yes
> 
> Pass 4: Checking reference counts
> Pass 5: Checking group summary information
> Block bitmap differences:  -(8483--8486) -(33026--33027)
> Fix<y>? yes
> 
> Free blocks count wrong for group #0 (24286, counted=24285).
> Fix<y>? yes
> 
> Free blocks count wrong (1030071, counted=1030070).
> Fix<y>? yes
> 
> Free inodes count wrong for group #0 (8191, counted=8190).
> Fix<y>? yes
> 
> Directories count wrong for group #0 (1, counted=2).
> Fix<y>? yes
> 
> Free inodes count wrong (262143, counted=262142).
> Fix<y>? yes
> 
> 
> /dev/sda8: ***** FILE SYSTEM WAS MODIFIED *****
> /dev/sda8: 2/262144 files (0.0% non-contiguous), 18506/1048576 blocks
> 
> 
> As I reported before, e2fsck -b also can not recover with
> backup superblock correctly.
> # http://marc.info/?l=linux-ext4&m=131541543931429&w=2
> 
> Is there any ideas to fix these issues?
> 
> Regards,
> Akira Fujita
> --
> To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html