I have a Linux gateway with 2 500gb drives in software RAID 1 for system.

/dev/sda = RAID 1
/dev/sdb = RAID 1
/dev/sdc = RAID 5 (NAS 12 drives iSCSI, /home directories)

While replacing sdb, weird thing happened, NAS mounted as /dev/sdb at reboot, so RAID recovery procedure happily recovered RAID 1 system drive on the NAS, overwriting some data and partition table. The process was interrupded by me, when I saw what is happening.

Now I have end up with 1 broken partiton of 500gb on the 20TB NAS.

Of course I have some users that do not have backups, some 4TB data of high importance is lost somewhere on the NAS, other things can be dumped.

I believe that the data is still there, only partitions are lost.

I have tried testdisk, after 4 days of analysis, I only get this "new" broken partitions:

Disk /dev/sdc - 20 TB / 18 TiB - CHS 2431671 255 63
 Partition               Start        End    Size in sectors
 D Linux                    0   1  1 59999 254 63  963899937 [root]
 D Linux RAID               0   1  1 59999 254 63  963899937 [md0]

I can even read some of the “new” data that I don't want. No sign of real data.

Any ideas what can be done?

  • 1
    RAID is not a backup – Michael May 20 '15 at 6:32
  • I can invite you to explain that to users that are bombarded every day with notes that the files on this array should be backed up manually :( – b4d May 20 '15 at 9:11

I've seen similar cases where the start of a filesystem was mistakenly overwritten. Recovery is sometimes possible by mounting the filesystem with a backup superblock.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.