e2fsck: Device or resource busy

e2fsck: Device or resource busy

[Christian Kujau]

Hi there,

I'm trying to repair a severely damaged ext4 partition. Actually, I don't 
care any more for the data, I just wanted to see how far I could get, 
because the "damage" here is: the partition was 1/3 of a RAID-0 setup :)

So, if you're interested, read on. If not, please ignore this one...

So, after trying to (forcefully) repair an image of the parition[0], I'm 
getting this now:

--------------------------------------------------------------
# losetup /dev/loop1 /mnt/cifs/backup/sdb2.img 
# losetup -a
/dev/loop1: [0013]:1715743 (/mnt/cifs/backup/sdb2.img)

# file -s /dev/loop1
/dev/loop1: , 44.1 kHz, Stereo

# /opt/e2fsprogs/sbin/e2fsck /dev/loop1
e2fsck 1.41-WIP (27-Apr-2008)
/opt/e2fsprogs/sbin/e2fsck: Superblock invalid, trying backup blocks...
/opt/e2fsprogs/sbin/e2fsck: Bad magic number in super-block while trying to open /dev/loop1

The superblock could not be read or does not describe a correct ext2
filesystem.  If the device is valid and it really contains an ext2
filesystem (and not swap or ufs or something else), then the superblock
is corrupt, and you might try running e2fsck with an alternate superblock:
     e2fsck -b 8193 <device>

# /opt/e2fsprogs/sbin/e2fsck -b 8193 /dev/loop1
e2fsck 1.41-WIP (27-Apr-2008)
/opt/e2fsprogs/sbin/e2fsck: Device or resource busy while trying to open /dev/loop1
Filesystem mounted or opened exclusively by another program?

# losetup -a
/dev/loop1: [0013]:1715743 (/mnt/cifs/backup/sdb2.img)

# lsof -ln | grep loop
loop1      5065        0  cwd       DIR        8,5       4096          2 /
loop1      5065        0  rtd       DIR        8,5       4096          2 /
loop1      5065        0  txt   unknown                                  /proc/5065/exe
--------------------------------------------------------------

Does anyone have an idea why it says "Device or resource busy" as soon as 
I specify an alternate superblock? Details and strace:
http://nerdbynature.de/bits/e2fsprogs.pu/sdb2/2008-06-02/


Thanks,
Christian.

[0] http://nerdbynature.de/bits/e2fsprogs.pu/sdb2/2008-05-25/
-- 
BOFH excuse #47:

Complete Transient Lockout

Re: e2fsck: Device or resource busy

[Theodore Tso]

On Mon, Jun 02, 2008 at 08:29:24PM +0200, Christian Kujau wrote:
> Hi there,
>
> I'm trying to repair a severely damaged ext4 partition. Actually, I don't 
> care any more for the data, I just wanted to see how far I could get, 
> because the "damage" here is: the partition was 1/3 of a RAID-0 setup :)

The key thing that you need to do when trying to recover from RAID
setups where you are missing one or more disks is to align the pieces
so they are in their original places, with the missing pieces replaced
with a /dev/zero device.

There is a starting point here which you may find helpful:

      http://debian.co.nz/Recovery+of+RAID+and+LVM2.html

      http://tldp.org/HOWTO/LVM-HOWTO/index.html

      http://www.tldp.org/HOWTO/Software-RAID-HOWTO.html 

(I'm assuming here you were using software RAID; if this was a
hardware RAID setup, then you need to know what sort of low-level
format was used by your hardware RAID device.)

						- Ted

Re: e2fsck: Device or resource busy

[Christian Kujau]

On Mon, 2 Jun 2008, Theodore Tso wrote:
> The key thing that you need to do when trying to recover from RAID
> setups where you are missing one or more disks is to align the pieces
> so they are in their original places, with the missing pieces replaced
> with a /dev/zero device.

Hm, interesting, I did not consider that yet. Thanks for the links, I'll 
have a look. But again: I don't really care for the data, I was just 
playing around and wondering why it reported "Device or resource busy"
when I expected something like "unable to find superblock".

> (I'm assuming here you were using software RAID; if this was a

Yes, it was software RAID.

Thanks again,
Christian.
-- 
BOFH excuse #112:

The monitor is plugged into the serial port

Re: e2fsck: Device or resource busy

[Theodore Tso]

On Tue, Jun 03, 2008 at 12:23:48AM +0200, Christian Kujau wrote:
> On Mon, 2 Jun 2008, Theodore Tso wrote:
>> The key thing that you need to do when trying to recover from RAID
>> setups where you are missing one or more disks is to align the pieces
>> so they are in their original places, with the missing pieces replaced
>> with a /dev/zero device.
>
> Hm, interesting, I did not consider that yet. Thanks for the links, I'll 
> have a look. But again: I don't really care for the data, I was just 
> playing around and wondering why it reported "Device or resource busy"
> when I expected something like "unable to find superblock".

That will only happen if someone else has opened the device with
O_EXCL.  Are you sure you didn't ^Z an e2fsck or some such and then
started a second one, or some such?  

Is this repeatable?

							- Ted