MDAMD reshape broken

  • hi


    I had an issue during growing the RAID5 from 4x4TB drives to 5 in openmediavault (debian).


    Now I can't assemble the raid again.


    Since the grow process broke right at the beginning (0%), the backup file is also missing.



    the log tells me the following:


    Code
    md: pers->run() failed ...
    md/raid:md127: reshape_position too early for auto-recovery - aborting.



    I need the data or my wife will kill me since all the familiy data is stored on it.


    mdadm.conf:
    ARRAY /dev/md/nas:data4 metadata=1.2 spares=1 name=nas:data4 UUID=49b...


    HEEELP? ;(


    regards

  • OK I think I've found the solution allthough I'm not sure if my data is still valid.


    FSCK shows the following status:
    data4: sauber, 1485874/732569600 Dateien, 2810248528/2930262016 Blöcke


    The solution was re-Creating the Raid by executing the following command:
    mdadm --create /dev/md127 --assume-clean --level=5 --raid-devices=4 --readonly --metadata=1.2 --data-offset=2048 --layout=ls --chunk=512 /dev/sdc /dev/sdb /dev/sda /dev/sdd


    The parameters I've derived from the old superblocks. Please be sure thata also the data-offset is right since if you don't use this option it could be wrong.


    Please give me some hints checking the file system other then using fsck.ext4 /dev/mapper/data4-data4 -n

Jetzt mitmachen!

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