Mounting issue on an LVM - please help!

[kal9001]

Sorry if this has been asked before but I'm stressing quite badly at the moment.

I'm running Openmediavault inside of proxmox. It has a 750GB LVM-thin partition passed to OMV.

OMV appears to have had a bad shutdown and corrupt the partition/file system.

Proxmox says the data is still there with the used percentage being what it was before. I'm now trying to back up the whole image to attempt recovery. and I have some questions.


1, how do I actually get the LVM off the Proxmox machine? and into some more readily readable state like an .iso or a vhd.
2, How would I mount the LVM-thin inside proxmox its self to see if I can access files then I can cp them out to a second disk?
3, The LVM is entirely on a single 1TB SSD, is it possible to just yank the drive and mount it on a separate system.

(Yes i know backups... but of course this would happen the one day I amalgamated all my stuff, finished late, so figured I'd back it all up in the morning...)
(I'm even willing to look at pay-for professional data recovery options if it seems they have half a chance)

Thanks anyone who can help. I've ben scouring the forums and found not very much, directly relating to fixing an issue like this :S
If there is a super obvious answer, please be gentile, I'm having a bit of a stress at the moment!

 

[wolfgang]

Hi,

1, how do I actually get the LVM off the Proxmox machine? and into some more readily readable state like an .iso or a vhd.

For this, you can use "qemu-img" it is part of Proxmox VE

2, How would I mount the LVM-thin inside proxmox its self to see if I can access files then I can cp them out to a second disk?

This is dependent on how you have used the LVM inside the VM.

The LVM is entirely on a single 1TB SSD, is it possible to just yank the drive and mount it on a separate system.

Yes you can. But do not forget to disable the storage before you take the SSD out.

 

[kal9001]

This is dependent on how you have used the LVM inside the VM.

The LVM-thin was requested by me to be 750 GB, but it's actually 840 something GB, I presume due to overheads. but it's only 350 GB full.
Inside the guest it's seen as /dev/sdb which is 750 GB. and I put a single partition with ext4 file system.

I believe it's the file system, or partition which has then become corrupt by the guest itself due to a bad unmount during shutdown of the guest.
Hence the whole problem. Given this condition I'd imagine the LVM-thin layer 'above' that is fine. Indeed proxmox doesn't suggest any issues with the LVM, though I'm not sure exactly how to extensively test that.

Ideally I'm wanting to leave the existing drive as a 'master'. And here is my dilemma. How exactly do I unwrap this onion?

#1 The physical disk (me@proxmox:/dev/sdb) ~1 TB
#2 The LVM-thin pool (Not even sure where this is) ~840 GB
#3 The disk image inside the LVM (I think I know where this is) ~750 GB
#4 the 'physical' disk inside the guest (me@guest:/dev/sdb) ~750 GB

I know for sure tools I have can read #4 as this will just be a standard disk image with no complexity. but it's perhaps more reliable to read and copy one of the higher levels, but then I need more tools to extract and convert the resulting file.

I've set up a second proxmox server to try and dry run the transfer and recovery, the exact same set up though I don't have another SSD, so I used a 500GB HDD perhaps this makes more difference than I'd think because I can't corrupt the thing again... even yanking the power from the whole server with a live file transfer going. So I'm really stumped as to how this could have happened to begin with!
It's possible the SSD only had the file system in cache the power off lost it, or I'm unlucky and some random cosmic ray or perfectly timed peak of RFI or something...

Ether way I'm playing with DMDE and know what I'm doing with it now, so once I get the image copied out I can attempt recovery and hopefully it's all there.