[SOLVED] Problem with disk/raid

elrodri · 04-16-2012, 10:04 AM

I had a raid 1 working with two devices sdc1 and sdd1, 2TB each. I needed some space, so I decided to undo the raid. I must say I'm too new with raids, and I successfully removed sdc1 with "fail" and "remove" mdadm commands, but not with sdd1, so I left the raid working and mounted the filesystem as follows:

/dev/sdc1 -> /data
/dev/md0 -> /data2

But a few days ago this "workaround" end up with a very bad and serious message forcing the server a reboot and rescue mode working.

I successfully removed the raid and mounted /dev/sdc1 into /data, but when I try to mount sdd1 into /data2 I get:

Code:

mount: you must specify the filesystem type

So, I exec:

Code:

mount -t ext4 /dev/sdd1 /data2

And I get the same error message. I tried with ext2 and ext3, but had the same result.

I also get this on dmesg:

VFS: Can't find ext3 filesystem on dev sdd1.

When I execute "fsck /dev/sdd1" I get the following:

Code:

fsck 1.39 (29-May-2006)
e2fsck 1.39 (29-May-2006)
Couldn't find ext2 superblock, trying backup blocks...
fsck.ext3: Bad magic number in super-block while trying to open /dev/sdd1




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>

"fdisk -l /dev/sdd" reveals the following:

Code:

WARNING: GPT (GUID Partition Table) detected on '/dev/sdd'! The util fdisk doesn't support GPT. Use GNU Parted.


Disco /dev/sdd: 2000.3 GB, 2000398934016 bytes
255 heads, 63 sectors/track, 243201 cylinders
Unidades = cilindros de 16065 * 512 = 8225280 bytes

Device                 Begin      End      Blocks  Id  System
/dev/sdd1               1      243201  1953512032   83  Linux

I successfully executed "mdadm --zero-superblock /dev/sdd1", but I got "mdadm: Unrecognised md component device - /dev/sdd1"

I tried everything my expertise could, but with no positive results.

Please, note that sdc1 and sdd1 have now different data, so I can't create a new raid 1 and put these disks into this raid.

Kustom42 · 04-16-2012, 10:56 AM

The only thing I could suggest would be a file system check specifying a new superblock as indicated in the error message. Try running dumpe2fs /dev/sdd | grep -i superblock and see what your results are.

elrodri · 04-16-2012, 11:04 AM

Quote:

Originally Posted by Kustom42

The only thing I could suggest would be a file system check specifying a new superblock as indicated in the error message. Try running dumpe2fs /dev/sdd | grep -i superblock and see what your results are.

Thank you.

This is what i get:

Code:

dumpe2fs bad magic number in super-block while trying to open /dev/sdd
Couldn’t find valid filesystem superblock.

I'm afraid the only solution is dumping /dev/sdd1 into /dev/sdc1 with 'dd', but I'll need to migrate the whole system into a new one if I don't want to loose all sdd1's data

Kustom42 · 04-16-2012, 12:49 PM

Do a dd to a flat file then try running a file system check, with a bad super-block a fsck should correct it but you never know with fsck.. in the Linux world fsck is closely associated with another similar four letter word that is usually spoke when having to run an fsck.

kitgerrits · 04-17-2012, 02:58 PM

Are you sure that re-mounting sdc1 after removing it from the RAID array went right the first time around (no mkfs in between?)
I would guess that sdd1 does not start with a superblock, but with some devicemapper metadata.

Have you tried re-assembling the array with a single disk and mounting that?

Maybe something along the lines of:
mdadm --assemble --scan
or:
mdadm --assemble /dev/md4 /dev/sdd1

rrdonovan · 04-17-2012, 06:19 PM

dumpe2fs bad magic number in super-bl..............

Whoa! You are out of magic dude.......! That's it man, it has hit the fan........ About the best thing to do is start over as the data is most probably corrupt. Raid under linux can be a horror story. Once up and running, don't mess with it. Do backups to an external HD. Even a shadow copy would be good. I use Raid 0. Good speed, but a heartbreak when it breaks. If you use Raid 1, then if one drive breaks, you should be able to replace it. Join it to the array, and it will rebuild. So what happened to sda and sdb?

Master Rod

elrodri · 04-17-2012, 06:21 PM

Hi, guys. Thank you for your anwers. Finally I decided to run mkfs.ext3 and loose all disk data. I already tried all mdadm commands and testdisk to recover this disk, but no luck. So I formatted it and started to use it.