Find the answer to your Linux question:
Results 1 to 3 of 3
Enjoy an ad free experience by logging in. Not a member yet? Register.
  1. #1

    Software RAID0 mess up, please help!!!

    I need help big time.

    I have a file server running Red Hat Linux 7.3. It was built around 6 months ago and has 3 HDD's, one is for the primary linux installation, the other 2 (IBM Deskstar 120GB) are running Software RAID 0 as a public file store Samba'd to the rest of the home network.

    I just got another 2 drives (the exact make) and I plugged one of the drives into where the CD-ROM was... except my biggest mistake was forgetting to change the new hard drive from Master to Slave And so on booting up (no slip ups from the BIOS) linux prepped the md0 (raid array) and noticed the second drive wasn't quite right, and it decided to modify the Superblock (assumedly on the first drive in the array)... I can't find it in the messages but I was sure it changed the fs from ext3 journaled to ext2, and then a prompt came up asking if I wanted to continue to boot... which came with a "no" from me and shutting the system down.

    I rebooted (without the new drive) and it promptly spent 20 minutes fsck'ing the arrayed drives (where the message log said: "journal inode not in use, but contains data. CLEARED". So now when it tries to mount the array, first it reports "wrong fs type, bad option, bad superblock on device /dev/md0. or too many mounted files systems". Also when the md program (driver version 0.90.0) loads up, it reports that the raid personality 2 cannot be found (I'm guessing when the 2nd drive "disappeared") and that "do_md_run returns -22".

    I cannot remember whether it is a linear raid 0, but is there any tool or command I can run where the raid driver is forced to remap all the data... or anyway to get the backup inode info returned? I'm sitting here with the thought that 210GB of data has disappeared all because of a Master/Master conflict.... arghhhhh! Either way I'm seriously considering lvm, even before this occurred. Any help would be appreciated.

  2. #2
    Linux Guru
    Join Date
    Oct 2001
    Täby, Sweden
    This seems fairly strange. It would seem as if it had destroyed the array, but I can't imagine that it would just do something like that without asking questions. I can explain why no messages are logged though: syslog isn't started at the point where the raid arrays are built.
    What happens if you run fsck.ext2 on /dev/md0 now?

  3. #3


    Thanks Dolda2000, once I started playing around... I eventually got there (with your help ).

    I ended up doing a "fsck.ext3 -f /dev/md0" and it came up with details like, there are plenty of files there and there is data there. Then I did a manual "mount /dev/md0 /public" and it worked, I think I put too much trust in the bootup messages (i.e. it freaking on having anything to do with mounting the raid).

    I've now been able to back up the very important files, and am rebuilding the server... the data is now on a nice simple ext2 fs and is soon going to be transferred back to the lvm'd ext3 drives. :P :P :P

  4. $spacer_open

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts