Raid 5 lost after failure of power supply

  • Hello together,


    I'm using Openmediavault for nearly 2 years now with a Raid 5 array (3x WD Red 4 TB).


    Recently, my NAS stopped working and I figured out it was the power supply (external power supply with PicoPSU). During troubleshooting, I also dismounted the mainboard and the SATA Cables. I think I placed the SATA cables in the same order as before, but I'm not 100 % sure.


    After replacing the power supply, the System works again and OMV shows all three drives as physical drives, but no Raid.


    I read many of the threads concerning such problems, but am not 100 % sure what I can do without destroying something (the last Backup doesn't contain all data).


    I retrieved the following information:


    1. cat /proc/mdstat



    Bash
    Personalities : [raid6] [raid5] [raid4]
    unused devices: <none>


    2. blkid


    Bash
    /dev/sda: UUID="7a3a0eca-7615-3cc2-f086-e85a5cba2017" UUID_SUB="f082f604-46c3-9868-87d2-3a812ba76a6d" LABEL="openmediavault:Meins" TYPE="linux_raid_member"
    /dev/sdc: UUID="7a3a0eca-7615-3cc2-f086-e85a5cba2017" UUID_SUB="21a38f7b-9b5b-d2e1-758a-68d1d8d823c1" LABEL="openmediavault:Meins" TYPE="linux_raid_member"
    /dev/sdb1: UUID="8fcb8f05-0087-4943-849b-29809705ae97" TYPE="ext4" PARTUUID="db38b868-01"
    /dev/sdb3: UUID="06b87404-2eba-4acf-8079-24812f838995" TYPE="ext4" PARTUUID="db38b868-03"
    /dev/sdb5: UUID="2fa71961-def1-415d-83ab-c13c2def7e37" TYPE="swap" PARTUUID="db38b868-05"
    /dev/sdd: UUID="7a3a0eca-7615-3cc2-f086-e85a5cba2017" UUID_SUB="81d2fe2f-8f12-8980-6e83-c79a8d0e57ee" LABEL="openmediavault:Meins" TYPE="linux_raid_member"


    3. fdisk -l | grep "Disk "



    Bash
    Disk /dev/sda: 3,7 TiB, 4000787030016 bytes, 7814037168 sectors
    Disk /dev/sdc: 3,7 TiB, 4000787030016 bytes, 7814037168 sectors
    Disk /dev/sdb: 28 GiB, 30016659456 bytes, 58626288 sectors
    Disk identifier: 0xdb38b868
    Disk /dev/sdd: 3,7 TiB, 4000787030016 bytes, 7814037168 sectors


    4. cat /etc/mdadm/mdadm.conf




    5. mdadm --detail --scan --verbose


    This command does absolutely nothing. It results just in the next empty line in the console.


    As I don't want to destroy anything, it would be great if someone can tell me if there's a chance to restore the raid array and what I can try to do so. Thank you! :)



    - Edit: additionally, OMV sent me the following mails just before it stopped working (one for each drive):


  • Hi there,


    no answer since 10th Oct.? Uhm ...


    Did you tried an "assemble" on your drives?


    mdadm --assemble /dev/md0 /dev/sda /dev/sdc /dev/sdd (tries to re-assemble the drives to a raid-array)
    - may be, that will fail - because of invalid eventcount of each drive -> please copy the error message in here
    - you can expand the command for your own after understanding the possible error message(s):
    mdadm --assemble --run /dev/md0 /dev/sda /dev/sdc /dev/sdd
    mdadm --assemble --run --force /dev/md0 /dev/sda /dev/sdc /dev/sdd


    -use with absolute caution, and only if you backupped your data:
    mdadm --assemble --run --force --update=resync /dev/md0 /dev/sda /dev/sdc /dev/sdd


    Sc0rp

  • Re,

    1// you answered in a few minutes several RAID questions

    i'm RAID-addicted ... worked with md-RAID since ages :P



    2// seems like members tend to avoid RAID questions

    Since they are not responsible for the linux-software-RAID-stuff ... so i'll jump in (if i got time).
    Btw. RAID-questions (incl. SnapRAID) are common in the german technikaffe.de-Forum too ;)


    3// you helped me on my own RAID question

    I'm always glad i could help, since i don't earn money with that (it's pure hobby) :D


    Btw.: no answer from thread-creator so far ... i think this is stub.


    Sc0rp

  • Hey,


    thanks for your response :) No, no stub. Sorry for not answering earlier, I just didn't find the time to try your suggestion as I'm currently mostly not at home and the data on the NAS are not that time-critical (and I had almost given up the thread ;) ). So I'm just happy that someone answered my question at all :)
    I didn't try assemble so far, as I wasn't sure which commands I can use without risking losing the data (as you said about the update=resync command).


    I just tried
    mdadm --assemble /dev/md0 /dev/sda /dev/sdc /dev/sdd


    this responds in
    mdadm: /dev/md0 assembled from 1 drive - not enough to start the array.


    Can I just try the other commands like --run or --run --force without destroying something?

  • Re,

    Can I just try the other commands like --run or --run --force without destroying something?

    You can try, but it will fail too - i wouldn't recommend it, because your array is not in a "sync, but not startet" state ...


    What gaves:
    cat /proc/mdstat
    mdadm -D /dev/mdX (take the number (X) from above output)?


    I figured out it was the power supply (external power supply with PicoPSU).

    Uhm, didn't read that part ... thanks to @tkaiser pointing me at that.


    REALLY? PicoPSU? Which one in detail? What was the power source (primary side)?


    Sc0rp

Jetzt mitmachen!

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