Loss of the main disk (proxmox), how can I recover my VMs/CTs without corrupting them?

[HuhuG]

Hello,

I've come to you because I've had a power cut on my DL360p Gen8 machine, and when I try to start it up I get the error “Non-System disk or disk error” (see photo). The previous version installed of Proxmox was the 8.2.

I've tried everything, already checked elsewhere, and unfortunately I can't see any other solution than to redo a fresh install on this disk. It's a RAID1, so under 2 300GB disks. There's another 8TB RAID50 for VM/CT data.

I'd like to know what I can do to boot despite this problem, or do a fresh install of the OS on this Raid1 but find my VMs/CTs on the other RAID without any loss?

Thank you very much for your future help.

 

[waltar]

Looks like a defect raid controller. Power off and on again and try to get into raid menue before "booting OS" (which isn't possible yet).
Could you see there your 10 disks ?

 

[simonhoffmann]

Since your title says "how to recover the VMs" I just want to throw this in here to maybe calm your nerves a bit: if you know how many cores, memory etc every VM had you can just create the VMs with the same ID again (after the reinstall), but without a hard disk, mount the existing raid5 and add it as a datastore, then issue "qm rescan" on the command line and it will add all the VM volumes back to the VMs as unused disk. You then just need to attach them and edit the boot config of each VM.

Had this problem a few months ago because both boot drives broke within a few days before I could replace the first. After the reinstall I had them up and running in about 30 minutes.

 

[HuhuG]

Hello,

Thank you for your answers.
No problems with the RAID card.

I've tried to check with several people but nothing works, the disk won't initialize...

 

[waltar]

As you see in picture 1 it's NOT a raid1 of 300GB disk as there individual one. So your single boot disk is high probably broken and
your 4TB disk even looks not good anymore as made table isn't readable also.
Edit: My fault as I just look at size 29.... as sda and 29... as sda3, sorry, it's looks it's too late ...

 

[simonhoffmann]

As you see in picture 1 it's NOT a raid1 of 300GB disk as there individual one. So your single boot disk is high probably broken
and your 4TB disk even looks not good anymore as made table isn't readable also.

Where do you see that the boot disks are not a raid? To me in the picture it clearly looks like a raid.

 

[simonhoffmann]

Hello,

Thank you for your answers.
No problems with the RAID card.

I've tried to check with several people but nothing works, the disk won't initialize...

How did you get those Screenshots if Proxmox does not boot?

 

[waltar]

You are right as it's a raid1, I looked wrong as it's too warm, sorry.

 

[Falk R.]

It looks like you have the disk cache enabled on your raid controller. The disk cache is the RAM on the HDDs and must never be used with Raid5/6/10/50/60 and should not be used with Raid1.
Your raid is apparently OK because the raid controller does not know that the user data is corrupt. The error messages are quite clear.
I hope you have a backup that you can restore.

 

[waltar]

Still assuming you don't have 2 broken disks at once in your raid1 and your OS is still iO, we changed a couple of defect HP p"xyz" controllers in the past. I would first remove all the actual disks inside the dl360p g8 and label in which slot they were !! Then take any 2 other (old) disks and try to build a new raid1 of these, install any os and reboot. Does that still work ?
If the ctrl is ok that will work (which is bad to your actual situation)
otherwise you need another (new/used) p"xyz" ctrl. which just must be exchanged without any further reconfig as all this is onto the disks.

 

[Falk R.]

I don't think the disks are defective either. But the data.
The disk position does not matter, the maximum message is: Drive position changed.
You can of course try to save data with a fresh OS, but it is very likely that the last written data is defective.