ran out of disk space, metadata of the pool is full

[Abergr]

running Proxmox 4.4, with a few CT containers among others. Today one of them acted up weird, so i tried to reboot, with ext4 errors. Tried to do a restore from last nights backup, but only got a message that out of disk space, "lvcreate 'pve/vm-101-disk-2' error: Free space in pool pve/data is above threshold, new volumes are not allowed."

And now im a bit stuck, I can see the vm-101-disk-1 disk, but I cannot remove it. Even better would be if I could restore it, without having to delete that disk, but after a few hours of searching here I cannot find a good answer what to do.

Attached a screenshot with a list of all the LV's. I believe i cannot extend the LV, since I think it's using the entire raid, minus root.

Im also unable to remove the 101-disk-1, only state the disk does not exist?

Any pointers what to do?

 

[Alwin]

On the screenshot, I can see some snapshots, if they are not needed, you could delete those to get some free space. You should be able to remove the disk, if your VM is in a non-operating state and you need to restore anyway, then it may help to delete the VM first and try to do a restore. Another option might be to move the disk to a different storage, to free up space. In the end, you might need to get bigger disks.

 

[Abergr]

Thank you Alwin for your response.

The issue was that I could not remove either the main disk or the snapshots. The disks were not found.
After a bit of investigation I found out that during the process of restore an old backup, the config file was removed. Why was the config removed? After that I could not access any disks belonging to that container. Luckily i had backup of all config files, and after I put the config back, I could see the container again, remove the snapshots, and restore the backup. I would consider this as a bug in proxmox, but I let you guys judge that. If you need more info, let me know, otherwise this case is closed.

 

[Alwin]

If you have a reproducible case and you don't find already a bug report on https://bugzilla.proxmox.com/, then please open one.

 

[Abergr]

Im not sure if I should take that serious I obviously don't plan to reproduce such a serious error on a server which is quite important to the company, hence why I tried to describe it well for you.

 

[udo]

running Proxmox 4.4, with a few CT containers among others. Today one of them acted up weird, so i tried to reboot, with ext4 errors. Tried to do a restore from last nights backup, but only got a message that out of disk space, "lvcreate 'pve/vm-101-disk-2' error: Free space in pool pve/data is above threshold, new volumes are not allowed."

And now im a bit stuck, I can see the vm-101-disk-1 disk, but I cannot remove it. Even better would be if I could restore it, without having to delete that disk, but after a few hours of searching here I cannot find a good answer what to do.

Attached a screenshot with a list of all the LV's. I believe i cannot extend the LV, since I think it's using the entire raid, minus root.

Im also unable to remove the 101-disk-1, only state the disk does not exist?

Any pointers what to do?

Hi,
overcommiting is not allways an good idea...

are there free space in the VG?
Look with "vgs"

If not I would do try something like this:
install an good disk in the server.
Use this disk for swap and turn swap off on pve/swap
remove pve/swap (adapt /etc/fstab)
extend the LV data.
remove/move data to another place.

After that perhaps shrink data again an recreate swap again?!

Udo