RAID5 new disk failed while rebuild

  • Hello,


    I have a RAID5 and one disk failed. I changed this disk and the rebuilding started. Now the new disk failed with a lot of failed sectors.


    Now i cannot enable the RIAD5.


    The new disk which failed is sdc (SMART shows a lot of errors)
    The other disks are sda, sdb, sde, sdf, sdg (SMART shows no error)




    Code
    mdadm --run /dev/md127
    mdadm: failed to run array /dev/md127: Input/output error


    Has anyone a good idea to get my files back?


    Greetings,


    Alex

  • Thanks for your answer.


    It was while the rebuilding. Perhaps I used the wrong word. In mdadm --detail were the following infos after the new disk were inserted.


    active sync /dev/sde
    active sync /dev/sdg
    active sync /dev/sdf
    active sync /dev/sdb
    spare rebuilding /dev/sdb
    active sync /dev/sda


    I'm not an expert but i thought only the new disk get rebuilding and my data on the old disk would be save...

  • The disk was new and I checked it bevor


    Could it help if I run


    mdadm --assemble --run --force /dev/md127 /dev/sd[egfba]


    or does I lost everything after this?

  • This was the first issue. After I bought a new disk, I started the recovering again and now the system is in the described state from above.
    Perhaps it is important to know, this morning the recovery was about 50% I think.

  • I thought about changing the mainboard. But befor I do something I would get a tip from someone. If I do something where I think it could help, i would damage the raid. So it is better someone who knows linux raid would help me.


    Perhaps this information are useful too

    • Offizieller Beitrag

    Do you have a sixth disk to replace the failed one? Because you have tried a few things, post the output of:


    cat /proc/mdstat
    blkid
    fdisk -l

    omv 7.0-32 sandworm | 64 bit | 6.5 proxmox kernel

    plugins :: omvextrasorg 7.0 | kvm 7.0.9 | compose 7.0.9 | cputemp 7.0 | mergerfs 7.0.3


    omv-extras.org plugins source code and issue tracker - github


    Please try ctrl-shift-R and read this before posting a question.

    Please put your OMV system details in your signature.
    Please don't PM for support... Too many PMs!

  • Thank you very much for replaing!


    Because the failed disk was a new one, i have no other disk. I've bought a new one which will arrived Saturday or after easter.


    I didn't touch anything since I opened this thread.



    Code
    blkid
    /dev/sde: UUID="8de9f199-c6e4-b147-9b8e-008ddde34fc3" LABEL="openmediavault:SG01" TYPE="linux_raid_member" UUID_SUB="3fe1257c-9c15-ed11-6dee-8036aab43efc"
    /dev/sdb: UUID="8de9f199-c6e4-b147-9b8e-008ddde34fc3" LABEL="openmediavault:SG01" TYPE="linux_raid_member" UUID_SUB="316d0336-36f4-767a-de91-4cf4aac70b5d"
    /dev/sdf: UUID="8de9f199-c6e4-b147-9b8e-008ddde34fc3" LABEL="openmediavault:SG01" TYPE="linux_raid_member" UUID_SUB="e6cd17a4-ce49-590a-a96e-323012e0c847"
    /dev/sdc: UUID="8de9f199-c6e4-b147-9b8e-008ddde34fc3" LABEL="openmediavault:SG01" TYPE="linux_raid_member" UUID_SUB="12e602c8-359e-bd80-cba6-c8c898d4fad3"
    /dev/sdg: UUID="8de9f199-c6e4-b147-9b8e-008ddde34fc3" UUID_SUB="67b89680-a0da-5769-6791-fef5f481858b" LABEL="openmediavault:SG01" TYPE="linux_raid_member"
    /dev/sda: UUID="8de9f199-c6e4-b147-9b8e-008ddde34fc3" UUID_SUB="414f8a14-b470-54b1-8776-682f7d495188" LABEL="openmediavault:SG01" TYPE="linux_raid_member"
    /dev/sdd1: UUID="0eeb730d-69d8-461f-939b-c56e5f0c1081" TYPE="ext4"
    /dev/sdd5: UUID="64622b2c-3b1f-4702-b77e-61c296c057a7" TYPE="swap"


    • Offizieller Beitrag

    If you can I would wait until you get the new drive to rebuild the array. If another drive fails while rebuilding with one missing, you will lose everything.


    If you can't wait, do the following:


    umount /dev/md0
    umount /dev/md127
    mdadm --stop /dev/md0
    mdadm --stop /dev/md127
    mdadm --assemble /dev/md127 /dev/sd[abcefg] --verbose --force


    If the last line doesn't work because sdc if failing, we may have to try something different.

    omv 7.0-32 sandworm | 64 bit | 6.5 proxmox kernel

    plugins :: omvextrasorg 7.0 | kvm 7.0.9 | compose 7.0.9 | cputemp 7.0 | mergerfs 7.0.3


    omv-extras.org plugins source code and issue tracker - github


    Please try ctrl-shift-R and read this before posting a question.

    Please put your OMV system details in your signature.
    Please don't PM for support... Too many PMs!

  • Ok thank you. I could wait but i have a question: sdc have a lot of smart errors (a few thousand uncorredctable sectors, etc.). Is it possible to assemble the raid without sdc?


    mdadm --assemble /dev/md127 /dev/sd[abefg] --verbose --force


    If the raid started, it is degraded but i could do a backup of the important files isn't it?


    After the new disk arrives I would add this. If something get lost after backup this wouldn't be a problem. But I want to save my holiday and family photos as soon as possible. I'm scary about loosing this files.


    Yes crisman, but i thought only the new disk was rebuilded. When one disk failed only the data from this disk should be generated and not the whole data from all disks shouldn't it?

    • Offizieller Beitrag

    Aaron If I understand He already have a failed drive while rebuilding, so probably all data has gone!
    But not sure!


    That is the whole point of raid 5 - to allow one disk to fail without losing data. So, his data should be fine at this point.



    Yes, you should be able to start the array without the failing disk. You can mount and read data from a degraded array. I just recommend against stressing a degraded array (especially that large) because of the risk of an additional failure.

    omv 7.0-32 sandworm | 64 bit | 6.5 proxmox kernel

    plugins :: omvextrasorg 7.0 | kvm 7.0.9 | compose 7.0.9 | cputemp 7.0 | mergerfs 7.0.3


    omv-extras.org plugins source code and issue tracker - github


    Please try ctrl-shift-R and read this before posting a question.

    Please put your OMV system details in your signature.
    Please don't PM for support... Too many PMs!

  • I know it is better to wait until i have a new drive but i couldn't wait :)


    So I used:
    mdadm --assemble /dev/md127 /dev/sd[abefg] --verbose --force
    and than mounted the drive.


    It works. Thank you very very very much!!!!!


    Now I take a Backup of all important Files and after this i shutdown the pc until the new disk arrives.

Jetzt mitmachen!

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