[SOLVED] Delete stuck snapshot

[fiona]

Is there a service like updatedb running within the container or anything else accessing the .zfs directory within the container? If yes, this might be what keeps the snapshot busy. What is the snapdir ZFS property on your datasets, e.g. zfs get snapdir ZFS/PATH/TO/subvol-XYZ-disk-N?

 

[Unspec]

Is there a service like updatedb running within the container or anything else accessing the .zfs directory within the container? If yes, this might be what keeps the snapshot busy. What is the snapdir ZFS property on your datasets, e.g. zfs get snapdir ZFS/PATH/TO/subvol-XYZ-disk-N?

No updatedb service.

rpool/data/subvol-110-disk-0 snapdir hidden default

 

[copec]

I was struggling with this too, but it only seems to be happening on server hardware that proved to be problematic. It would report an unusually high load after migrating the same containers on identical hardware to it, and then it would always fail to clean up the snapshot at backup. I never got to the bottom of the issue with this hardware.

If you can, migrate the containers to a different host and see if the problem still occurs.

 

[TobiasV]

I have the same issue from time to time on different machines. It seems to happen more often on nodes with lower performance. I also didn't find any hidden mounts or processes which using the snapshot. I'm already using the kernel 6.14.

It occurs only during backups when the vzdump snapshot can not be destroyed. I wonder why it doesn't happen on zfs snapshots for replication. The are created much more often, i.e. 100 times per day but the vzdump only one or two times.

To make the backup working again I just rename the vzdump snapshot (zfs rename ...), otherwise a reboot is needed.