Missing RAID array - please help

  • I've got an HP Microserver running an admittedly older version of OMV. There was a power cut in our area yesterday, and when the mains returned and the server fired up it was missing it's main storage array. The missing array is held on 2 x WD Red 3TB disks, and it's a RAID 1 mirror.


    Here is the diagnostic info that was listed in the pinned message about raid issues:



    I also ran a quick badblock check on sda


    I guess this means that the sda drive is toast, but I'm hoping that as the volume was a mirror I might be able to get the array back up with sdb only so I can recover the data. Please can anyone tell me what commands I should use to go about that?

    • Offizieller Beitrag

    Check the state of /dev/sda in smart, to get the raid back, from the command line mdadm --assemble --verbose --force /dev/md0/ /dev/sd[ab] it will probably come back as clean/degraded due to the missing drive but worth a shot using both drive references.

  • I've checked both drives with smartctl, and as far as I can see neither is showing any errors there:



    I've tried the commands to assemble the raid, and got the following:



    Not sure what the output for assembling with sda means... but I'm guessing that sdb might need the raid superblock to be recovered from a backup? Any ideas?

    • Offizieller Beitrag

    Any ideas?

    Whilst I forgot the obvious, my bad, I clearly stated mdadm --assemble --verbose --force /dev/md0/ /dev/sd[ab] I did not say to assemble with just one drive, although it would work I needed to see the error!!
    So;
    mdadm --stop /dev/md0 followed by the above.


    Then post the output of the first five from here

  • Thanks for sticking with me!


    The assemble still didn't work:


    openmediavault:~$ sudo mdadm --stop /dev/md0
    mdadm: stopped /dev/md0
    openmediavault:~$ sudo mdadm --assemble --verbose --force /dev/md0/ /dev/sda /dev/sdb
    mdadm: looking for devices for /dev/md0/
    mdadm: /dev/md0/ is an invalid name for an md device. Try /dev/md/md0/


    Here's the output of the first items from the other thread:


    • Offizieller Beitrag

    Here's the output of the first items from the other thread:

    I've looked back at your first post and compared it to your latest which is what I needed, put simply the Raids toast, the fdisk output in both is missing information about all the drives.


    Your post 3 attempts to assemble but can't locate /dev/sda and /dev/sdb throws this mdadm: /dev/sdb has no superblock - assembly aborted the superblock is fixable if there is a working drive in an array. Post 5 mdadm --detail --scan --verbose and cat /proc/mdstat return nothing once the raid is stopped.


    You could try booting with a SystemRescueCD and try the commands again as this ignores OMV, and try fsck on each drive may help may not, but ultimately it doesn't look good.


    Can the data be recovered? one way to find out is to mount one drive on another system and see.

Jetzt mitmachen!

Sie haben noch kein Benutzerkonto auf unserer Seite? Registrieren Sie sich kostenlos und nehmen Sie an unserer Community teil!