I have a Raid10 array using software raid on 4*2-tb drives. My OS on this Raid array froze and when restarting my system via Ubuntu-CD i noticed the Raid array was gone.

I assembled the raid array again using "sudo mdadm -A /dev/md2 /dev/sda3 /dev/sdb3 /dev/sdc3 /dev/sdd3"

But now i noticed files somehow got corrupted; either some files in folders are dissapeared or when i copy some files i get: "You do not have the permissions necessary to view the contents of ..." I would like to backup what is left and am copying the left-overs that are not corrupted.

The drives seem to be working; not broken (1 of 4 disk is OK according to Diskutils but it has 6 bad sectors, the others are OK)

Can i save the seemingly corrupted files? how? should i resync? something else?


| improve this question | | | | |
  • What does /proc/mdstat say? Is one drive faulty or all up and running? You should check the file system with fsck. And after that you should scrub your array, if you don't do this on a regular base. – Marc Stürmer Sep 8 '14 at 20:22
  • ubuntu@ubuntu:~$ cat /proc/mdstat Personalities : [raid10] md2 : active raid10 sda3[0] sdd3[3] sdc3[2] sdb3[1] 3372643968 blocks super 1.2 64K chunks 2 near-copies [4/4] [UUUU] unused devices: <none> – tafelplankje Sep 10 '14 at 14:36

You may have assembled the RAID members in the wrong order. You should have used mdadm --examine /dev/sda3; mdadm --examine /dev/sdb3 etc to check the metadata stored on the members. That will give a clue as to what the order should be.

Unfortunately the assembly you did will have overwritten the old metadata. Perhaps an old copy of /var/log/dmesg or /var/log/kern.log may give some clues.

| improve this answer | | | | |
  • re --examine; what clues would it give me? see shrib.com/XbiANqAP for the output – tafelplankje Sep 10 '14 at 14:34
  • The Device-Role number shows in what order the devices should have been assembled. As you've already assembled them now, they're sequentially numbered starting with 0. Check out link for a lot more useful info. – wurtel Sep 10 '14 at 14:43
  • I now notice that sda3 is broken, in the beginning it said it was ok, but now it is marked faulty, also the bad clusters have increased from 6 to 103. i disconnected the drive. What is the best way to get my raid10 to work again with only 3 drives, so i can backup? thanks, – tafelplankje Sep 10 '14 at 14:57
  • With a RAID 10 over 4 drives you only need 2 drives; the biggest question is which 2, and in what order. After your assembly (and without --assume-clean :-( ) all bets are off. Read the link I gave in my earlier comment, I can't help more than what's there. – wurtel Sep 10 '14 at 15:03

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.