Mismatched mirror raids

1. offizieller Beitrag

    I have OMV 6.0.632.

    I added two new disks as a new mirrored raid yesterday.
    When I restarted the server I noticed that an older raid, as well as this new one had started a scan.
    The old raid stopped it's sync and the new one was running.
    But as I had issues accessing the existing raid, I shut down the server again, removed the two new disks and restarted.
    Reason to let the old raid resync get ready with the resynch first.
    When done, I attached the two new disks again and this one got done this afternoon.

    But now I can't add a new filesystem and noticed mismatched names.
    The new one is called "/dev/md124" in the RAID tab. But that name is what the old one has if I open the dropdown in Shared folders.
    And the name the old one has on the RAID tab, "/dev/md126" is missing on "Shared folders" tab.
    And there are nothing to select if I try to add a new filesystem.

    How do I fix this issue? The new raid is completely empty, so I do not mind to reformat that if that helps.
    But I think that there are references in Debian that need to be fixed to sort this out, and I am still too much of a linux rookie     to fix it.
    The risk is that I make it worse.

    • Offizieller Beitrag

    All of the above is very confusing, so ssh into omv as root and run the following;


    cat /proc/mdstat

    blkid

    cat /etc/mdadm/mdadm.conf

    fdisk -l | grep "Disk "


    post each output into a separate code box this symbol </> on the forum thread bar, makes it easier to read

    how do you use the </> symbol?

    Chenbro NR12000 32GB Ram - Xeon E2-1230 V2 @ 3.3GHz - OMV6

    Supermicro X9DRI-LN4F+ 128GB Ram - 2x Xeon E5-2670 V2 @ 2.5GHz - OMV6

    2x Wiwynn Lyra SV315 10SFF 32GB Ram - 2x Xeon E5-2630 v2 @ 2.60GHz - OMV6

    2x Lenovo EMC PX12-400R 32GB Ram - Core i3-3220 @ 3.3GHz - OMV6

    Thanks.
    Here is the results of those commands:




    Removed the instructions and default settings of the file.

    Code
    e:~# cat /etc/mdadm/mdadm.conf
...

# definitions of existing MD arrays
ARRAY /dev/md/ior metadata=1.2 name=nasse:ior UUID=edb0fb6c:b865e482:91005459:1f536e48
ARRAY /dev/md/puh metadata=1.2 name=nasse:puh UUID=c38f46f3:382c20e5:02a93b53:837294a3
ARRAY /dev/md/kangu metadata=1.2 spares=1 name=nasse:kangu UUID=0c98e989:12de39be:be119640:e4f548e0



    I can see that md124 and md126 have the same guid in blkid and that the md124 array is missing in mdadm.conf.
    How do I fix that - without making more mess of things...

    MultiUser "</>" is a button in the editor creating the above boxes.

    • Offizieller Beitrag
    Zitat von 7ore

    How do I fix that - without making more mess of things

    How many actual arrays do you have, what arrays have data on them, this is a first, duplicate uuid's for two different arrays


    This is a mess, mdadm references, that is md127 go in reverse order, md127, md126 etc during creation, there are reasons why mdadm uses md127 rather than md0, then md1 but that is not going to resolve this.


    Therefore md127 would be the first reference

    • Offizieller Beitrag
    Zitat von 7ore

    No, it doesn't

    OK well that's something.......therefore I would suggest you delete it from raid management and subsequently apply the changes


    /dev/md126 is in an auto-read-only state to correct this run mdadm --readwrite /dev/md126  from the cli


    If both of the above work, then the two drives you want to use (I assume /dev/sdj and /dev/sdh) will need to be securely wipe both drives first before a new array is created. Whilst secure wipe does take a long time it has been noted that this can be stopped after approx. 25% completion.


    Caveat: DO NOT REBOOT or SHUTDOWN this can cause drive and mdadm references to change!!

    OK, So I tried to remove the raid from GUI, but could only remove one of the disks in it.
    The other one is still there and the option to remove the complete disk can't be selected.



    But I have formatted both disks now, that worked.
    I have not restarted the machine yet.


    I got a bigger issue now, though.

    One of the shared folders in md126 appears empty. Other folders are intact and the filesystem seems untouched. This seems like another issue outside of this thread, but as it is related I continue here.
    What steps can I do to recover or rescan the filesystem to restore the content of that folder

    Solved the empty folder issue.
    Moved one of the drives in that RAID to another session of OMV (started up another computer with an old version of OMV that I had).
    And the drive had all files on it, so I recover the raid there.

    Then back to this instance and continue to do a cleanup there to remove the extra disk and the other issues.

    Thanks for the help so far. I might have followup questions, but I think that I know the path forward from here.

  • 7ore

    Hat das Label gelöst hinzugefügt.

Jetzt mitmachen!

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

Benutzerkonto erstellen Anmelden