Raid 1 verschwindet beim Plattentausch

    Code
    blkid
/dev/sdb: UUID="3ca82093-31bd-aba6-e3fc-2da78924c663" UUID_SUB="931edf09-fb4e-0c                                                                                                             6c-18d0-3503a69a0410" LABEL="Zangs-NAS:Zangs" TYPE="linux_raid_member"
/dev/sdc: UUID="3ca82093-31bd-aba6-e3fc-2da78924c663" UUID_SUB="6b249649-ad3c-9a                                                                                                             85-6dda-5b49a53af0ae" LABEL="Zangs-NAS:Zangs" TYPE="linux_raid_member"
/dev/sda1: UUID="E620-721A" TYPE="vfat" PARTUUID="9d5b7707-65db-433f-9b6a-c2cbb7                                                                                                             9b5dd8"
/dev/sda2: UUID="0ce16ee2-727d-4ac2-8e28-13fbf48155ab" TYPE="ext4" PARTUUID="5cf                                                                                                             cfa16-fe17-4afb-9424-954b3708bc37"
/dev/sda3: UUID="ec420058-796e-406f-962b-66e7cae4fd39" TYPE="swap" PARTUUID="bab                                                                                                             0249f-4d80-404f-85a8-6a10f9139157"
/dev/md127: LABEL="Raid" UUID="b090d450-8a3c-4b13-845c-b1876a8d7174" TYPE="ext4"


    Code
    mdadm --detail --scan --verbose /dev/md127
ARRAY /dev/md127 level=raid1 num-devices=2 metadata=1.2 name=Zangs-NAS:Zangs UUID=3ca82093:31bdaba6:e3fc2da7:8924c663
   devices=/dev/sdb,/dev/sdc
    Code
    mdadm --stop /dev/md127
mdadm: Cannot get exclusive access to /dev/md127:Perhaps a running process, mounted filesystem or active volume group?

    ?(:S

    • Offizieller Beitrag

    With the raid now up and everything off have you rebooted? to try again.


    If the array cannot be stopped there are two options to sort this;


    1. Use the systemrescuecd, but that would mean using for doing each drive change as you have to shutdown to replace a drive then restart, then go back to systemrescuecd.


    2. Not sure about this, create a clean install of OMV on a new USB flash drive, boot with that check that it's working, shut down plug the raid back in and see if the raid is initialised -> will this work I don't know + you then have to put your previous boot drive back in and hopefully it will all function with the new drives. The problem with this is if doesn't work, you have a new raid set up but your current omv setup/config doesn't work, this is unknown territory for me.


    I have help a few with raid problems and usually don't have a problem, but this is like a tunnel with no light at the end.

    • Offizieller Beitrag
    Zitat von Jahresprogramm

    ASRock J4105M with a Celeron J4105.

    That should not be the problem then, I still believe this related to something running not allowing the raid to be stopped.


    Option 1 above;


    Install systemrescuecd in OMV-Extras, this takes time as it has to be downloaded, read the information under SystemRescueCD (another option is to download System Rescue CD create a CD from the ISO) OMV-Extras option allows you to boot once only, you would have to use this option twice. I have not used this!! This allows you to access OMV via the cli without anything running, I believe that's the idea.


    SystemRescueCD


    Stop the raid mdadm --stop /dev/md127 if this works it will return a confirmation
    Remove a drive from the array /dev/sd[bc] -> mdadm --fail /dev/md127 /dev/sdc cat /proc/mdstat should show (F) next to sdc, mdadm --remove /dev/md127 /dev/sdc cat /proc/mdstat should confirm the removal of the drive.


    This is the point you would shut down remove sdc and install a new drive, then reboot, if using the OMV-Extras route you can check the new drives reference under Storage Disks -> sd? (? being a new letter) make a note, boot into systemrescuecd from OMV-Extras.
    Would you need to shutdown? according to your m'board the ports are Hot Swap, you could remove sdc from the machine and run fdisk -l to see if /dev/sdc is listed, if not hot swap works! Plug in the new drive and run fdisk -l again to find the new drives reference, the drive size here will be the giveaway.


    Whatever the outcome of the above a new drive is installed;


    wipefs /dev/sd? replace ? with the drive reference letter, this will wipe the drive.
    mdadm --add /dev/md127 /dev/sd? replace ? with the drive reference letter
    cat /proc/mdstat this will display the raid status it will either be syncing or it has failed, if it fails it means creating the exact partition on the new drive, today this should not be necessary.
    If it syncs you will need to wait for it to finish, then run cat /proc/mdstat to confirm then mdadm --detail /dev/md127 to confirm the riads' configuration.


    IF by some miracle the above works it will have to be done again to replace the second hard drive, you will then need to grow the array and resize the filesystem. Will this work? I have no idea, I am at a loss as to why the current raid will not stop, something is accessing it. TBH the time spent on this, a new raid setup could have been completed and your data restored, whilst that might be a PIA it's straightforward.


    For 2 drives a better option is one for data and one running rsync, and have a spare should a drive fail.

    Zitat von geaves

    SystemRescueCD


    Stop the raid mdadm --stop /dev/md127 if this works it will return a confirmation....


    Yeah, you're the best! :thumbup:

    Code
    root@sysresccd /root % mdadm --stop /dev/md127
mdadm: stopped /dev/md127


    I now connect mouse, keyboard and monitor so I can get the bootloader.
    That should start from the SystemRescueCD not a problem.

    OK, part one is done!I have of the SystemRescueCD booted.


    RAID is alive

    Code
    cat /proc/mdstat
Personalities : [linear] [multipath] [raid0] [raid1] [raid6] [raid5] [raid4] [raid10]
md127 : active raid1 sdb[2] sda[1]
      1953383512 blocks super 1.2 [2/2] [UU]
      bitmap: 0/15 pages [0KB], 65536KB chunk


unused devices: <none>


    find serial number first



    Code
    /sbin/udevadm info --query=property --name=sda | grep ID_SERIAL
ID_SERIAL=ST2000VN001-1T4164_W520BPCG
ID_SERIAL_SHORT=W520BPCG

    remove sda from RAID


    Code
    mdadm --fail /dev/md127 /dev/sda
mdadm: set /dev/sda faulty in /dev/md127



    Code
    cat /proc/mdstat
Personalities : [linear] [multipath] [raid0] [raid1] [raid6] [raid5] [raid4] [raid10]
md127 : active raid1 sdb[2] sda[1](F)
      1953383512 blocks super 1.2 [2/1] [U_]
      bitmap: 0/15 pages [0KB], 65536KB chunk


unused devices: <none>
    Code
    mdadm --remove /dev/md127 /dev/sda
mdadm: hot removed /dev/sda from /dev/md127
    Code
    cat /proc/mdstat
Personalities : [linear] [multipath] [raid0] [raid1] [raid6] [raid5] [raid4] [raid10]
md127 : active raid1 sdb[2]
      1953383512 blocks super 1.2 [2/1] [U_]
      bitmap: 0/15 pages [0KB], 65536KB chunk


unused devices: <none>


    now the disks are exchanged

    • Offizieller Beitrag


    Hm, bit of smart work there, I would have had a sticker on each drive edge, so I knew which was which.

    ok, hotswap works. I do not have to reboot. the new plate is recognized immediately and runs as sda


    Code
    mdadm --add /dev/md127 /dev/sda
mdadm: added /dev/sda
    Code
    cat /proc/mdstat
Personalities : [linear] [multipath] [raid0] [raid1] [raid6] [raid5] [raid4] [raid10]
md127 : active raid1 sda[3] sdb[2]
      1953383512 blocks super 1.2 [2/1] [U_]
      [>....................]  recovery =  0.3% (6768064/1953383512) finish=204.4min speed=158703K/sec
      bitmap: 0/15 pages [0KB], 65536KB chunk


unused devices: <none>


    Now the sync must be completed. After that, I had recently rebooted and the raid was gone. Now I will change the second disk without reboot, right?

    sorry, had forgotten something. Current status:

    • Offizieller Beitrag
    Zitat von Jahresprogramm

    After that, I had recently rebooted and the raid was gone.

    Are you referring to previous attempts?


    Zitat von Jahresprogramm

    ok, hotswap works. I do not have to reboot. the new plate is recognized immediately and runs as sda

    Ok, that's good, did you wipe it before adding it?


    Zitat von Jahresprogramm

    Now I will change the second disk without reboot, right?

    If the first drive change has worked, yes, but ensure you keep the two old drives just in case, because they are current raid drives.


    The thing to remember is double check everything you on the command line before executing.

    Zitat von geaves

    Are you referring to previous attempts?

    yes


    Zitat von geaves


    Ok, that's good, did you wipe it before adding it?

    yes



    Zitat von geaves

    The thing to remember is double check everything you on the command line before executing.


    Of course. Sorry, if I express myself misleading. Google translate helps me :)



    It will be tomorrow until the sync is done. I will record the status and continue with the exchange of the second disk.

    • Offizieller Beitrag
    Zitat von Jahresprogramm

    Of course. Sorry, if I express myself misleading. Google translate helps me

    :D:D:D:D how do think I'm managing


    Well if this works, then the growing the raid and resizing the file system I hope can be done in the GUI.

    I'm afraid I have a problem again. The pretitioning of the disks is faulty.


    Current status is: I have the second disk after instructions in RAID installed. sda was the first swapped disk, sdb was the second one. However fdisk -l reports error.

    Code
    cat /proc/mdstat
Personalities : [linear] [multipath] [raid0] [raid1] [raid6] [raid5] [raid4] [raid10]
md127 : active raid1 sdb[2] sda[3]
      1953383512 blocks super 1.2 [2/2] [UU]
      bitmap: 0/15 pages [0KB], 65536KB chunk


unused devices: <none>


    I suspect we first need to partition the disks right away with a clean MBR?
    I'm worried that the RAID will be inactive again when I reboot ...

Jetzt mitmachen!

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

Benutzerkonto erstellen Anmelden