Raid-5 Array deleted and formatted using OMV

shatayen · 26. Dezember 2017

So I tried to use OMV to manage my 3 HDDs (each 4 TB) in a raid-5 Array. Everything worked fine until the system was shutdown because of powerloss. When I was booting up again my raid was gone. But it was not only gone, as far as I understand the logs it was deleted and all disks were formatted... How could this happen?

Is there a way to recover the raid/data somehow? I tried mdadm --assemble and so on of course but it won't work because the superblocks are gone and the devices can't be assigned to any array. My idea would be to just recreate the raid again from scratch and run some data recovery software on it but I'm not sure if this works or if there is another solution.

Here some background information:

OMV guest is running on VMWare Workstation on a Windows Host. All 3 disks were passed through physically.

To copy the data first, I created a degraded raid-5 array first:

Code

sudo mdadm --create --verbose /dev/md0 --level=5 --raid-devices=3 /dev/sda /dev/sdb missing

Then I mounted the raid using OMV (in Web-GUI) and copied the data. Afterwards I recovered the array with the 3rd disk and it was stated clean.

After shutdown everything was gone and formatted:

cat /proc/mdstat

Code

Personalities : [linear] [multipath] [raid0] [raid1] [raid6] [raid5] [raid4] [raid10]
unused devices: <none>

blkid:

Code

/dev/sda1: UUID="afa8270c-6e80-432e-900e-ed5d7b395d6c" TYPE="ext4"
/dev/sda5: UUID="c49b3b4f-c9d2-4cd0-9cf7-a9abcfbcf6c6" TYPE="swap"
/dev/sdb: PTUUID="277dc189-b7ca-42fe-8397-2fccfd1ffe58" PTTYPE="gpt"
/dev/sdc: PTUUID="53072c8a-b8d7-43f4-8de2-eb7b7550efee" PTTYPE="gpt"
/dev/sdd: PTUUID="283a0b71-1a5f-449a-ab90-f75dac4cd04a" PTTYPE="gpt"

In the mdadm.conf the raid is listed:

Code

# definitions of existing MD arrays
ARRAY /dev/md0 metadata=1.2 spares=1 name=OMV:0 UUID=6a6b37ae:ef0a4ba9:3c8a3dc4:44868fa0

pierewoehl · 26. Dezember 2017

Tips for the next Time:

Backup on another Place
Disks trough Windows to VMware WS and then to OMV ? serious man...
UPS for Data Critical Systems is critical!

Can you post the Logs ?

roswitina · 26. Dezember 2017

I don't really understand much, but when I look at the hard drives, I think that sda is the system disk and
sdb, sdc and sdd should be the raid.

Do you know these links?
https://superuser.com/question…vate-my-mdadm-raid5-array

https://www.thomas-krenn.com/d…ync#cite_note-raid-wiki-1

https://raid.wiki.kernel.org/i…ction#recovery_and_resync

Rosi

Translated with http://www.DeepL.com/Translator

henfri · 30. Dezember 2017

Well, when using OMV in a virtual machine: is there an alternative to the scenario outlined above?

shatayen · 30. Dezember 2017

So i got everything back I just had to use the --create array again.

Code

sudo mdadm --create --verbose /dev/md0 --level=5 --raid-devices=3 /dev/sdb /dev/sdc /dev/sdd

Thanks for the tips but the problem is I have to use Windows as Host because I have a new j4205-ITX board which is not supported by ESXI or other OSs. I want to use my NAS as media server and use the physical audio/video output which only seems possible with Windows. So I though to set up OMV in a VM to manage my RAID but this seems problematic.

However I could reproduce the error. When shutting down OMV in the Web GUI everything is fine when restarting.

As soon as I shutdown OMV in the console (shutdown now) or shutdown the VM Workstation, the RAID gets totally deleted somehow and can't be reassembled with any mdadm command except for recreating the RAID (of course with the excact same settings).

Really disappoiting, guess I have to use Xpenology or something else. Any other tips for workarounds?

Jetzt mitmachen!