mdadm RAID1 Recovery

KIA · 26 Mar 2011 at 15:33

Hi,

I had to break up a RAID1 array.

When it came to recreating it using mdadm, I accidentally used --create instead of --assemble.

The file system is ext3. When I try to mount the array:

root@debian:~# mount -t ext3 /dev/md2 /root/temp/
mount: wrong fs type, bad option, bad superblock on /dev/md2,
missing codepage or helper program, or other error
In some cases useful info is found in syslog - try
dmesg | tail or so

Am I screwed? Is my data gone?

tntcoder · 26 Mar 2011 at 16:13

Did you get it to assemble correctly?

It has most likely fubared the filesystem headers though im afraid. As it's just RAID1 the actual data should in theory be fine and a recovery program should work quite well.

In any case, I would look at running testdisk on it, it supports partition and data recovery for both ext3 and linux software raid1.

KIA · 26 Mar 2011 at 16:15

tntcoder said:
Did you get it to assemble correctly?

Yes, it will still assemble although the array information refers to the current distro rather than the previous.

I will check out testdisk.

Thanks!

tntcoder · 26 Mar 2011 at 16:26

from the wiki:

testdisk /dev/md0 to repair a filesystem on top of a Linux RAID device.

Might be as simple as that, might well not be though

KIA · 26 Mar 2011 at 18:05

I'm not sure which is the best testdisk option to use. I have gone for the Analyse option. It's going to take a while as the drive is 1.5TB in size.

KIA · 26 Mar 2011 at 21:42

This app saved my life: https://www.runtime.org/nas-recovery.htm

I'll read the man entry before I use mdadm again.

blueacid · 26 Mar 2011 at 22:44

Another +1 for testdisk - that app is awesome beyond all belief.

You could try pointing it directly at the drive (ie /dev/sdb), see if it manages any better / worse.

Equally if you've sufficient space, you could use 'dd' to make an image of an entire drive. Means that if you break anything, you can rest easy.