Problem with mounting BTRFS RAID1

    • OMV 4.x
    • Problem with mounting BTRFS RAID1

      Hello!

      I have installed OMV version 4.1.22-1 (Arrakis).

      Faced the following problem.

      After one of my 2000Tb disks included in BTRFS RAID1 began to demonstrate a large number of SMART errors, I removed it from the array and connected a 2000Tb partition on the new 4000Tb disk using standard btrfs tool. After that, the new array is perfectly mounted and works, but it seems that OMV considers it to be unmounted and this leads to the following error when trying to apply any changes to the settings::

      Some commands outputs in attachments.

      Help me please to apply changes. If necessary, I am ready to provide additional information.
      Files
      • fstab.txt

        (1.29 kB, downloaded 7 times, last: )
      • mount.txt

        (5.5 kB, downloaded 8 times, last: )
      • fdisk.txt

        (7.25 kB, downloaded 8 times, last: )
      • btrfs_fi_show.txt

        (3.55 kB, downloaded 9 times, last: )
    • Hi, all!

      For those who may be interested, the problem was solved by physically connecting an already existing disk (used completely, without partitions) AFTER a new disk (divided into two partitions).

      I suspect that realpath ($ outputv) and $ this-> getCanonicalDeviceFile () do not work properly in the file /usr/share/php/openmediavault/system/filesystem/filesystem.inc, line 676.

      Most likely, realpath ($ outputv) returned "/dev/sdb", and $ this-> getCanonicalDeviceFile () - "/dev/sda1", the path to which the link /dev/disk/by-label/2000_raid1 refers.

      After the physical reconnection of the disks, the results returned by the functions probably began to coincide and the problem went away.
    • tkaiser wrote:

      If you think you found a bug please report it including steps to reproduce here github.com/openmediavault/openmediavault/issues

      Hi!

      I'm not quite sure this is a bug. Perhaps this is a feature of my hardware or something else. I think it will be possible to consider that this is a bug, if this problem appears to someone other than me. At the moment, I think that the workaround described by me will suffice.