[SOLVED] Failing to restore LXC's on ZFS

[mokaz]

Hi all,

Long story short, I migrated/re-setup my PVE with ZFS, everything looks fine. Restoring previously LVM backed up VM's from either direct storage backups or PBS works fine. Although, restoring LXC's always fails with the following error (edited for privacy):

recovering backed-up configuration from 'pbs.domain.local_cryptds:backup/ct/100/2025-10-27T15:41:30Z'
TASK ERROR: unable to restore CT 100 - zfs error: cannot mount 'rpool/data/subvol-100-disk-0': no mountpoint set

Any help would be much appreciated.

Thanks,
Regards,
m.

 

[Kingneutron]

Unless you have a specific reason (like running a business and need to maintain uptime) to install proxmox to zfs boot/root, which should then be a mirror rpool with 2 different make/model disks (otherwise what's the point) - I would recommend possibly adding another storage disk as lvm-thin or a separate zpool.

Something got messed up with your pve install, or you're trying to restore to a storage that no longer exists is my guess. Check the Storage dropdown when you go to restore the LXC from gui

 

[mokaz]

Hi there,

In fact I "wanted" to have full disk encryption over two nVME drives set in ZFS RAID1 at setup time, as said this works fine, the only current issue is the one above.

Restoring LXC's to "local" (/var/lib/vz) works although when attempting to restore to local-zfs (rpool/data) it fails -- are there any known limitation on "thin provisioned" zfs pool in regards to LXC perhaps? Asking this because as said, VM restore went through with no discussions...

The storage dropdown on the restore GUI window propose's : local | local-zfs | TCP/IP configured storage

My guess is that perhaps I'm having a zfs mountpoint mapping issue, I'll try to figure this out.

Thanks,
m.

 

[mokaz]

We can close this thread -- I most probably broke something at setup time while troubleshooting some boot issues.
I did just re-setup my node the way I wanted it and now I can restore container on local-zfs.

Thanks