Strange leftovers after using copy --refresh

LXD Support
1

I’m currently on LXD 6.4 with BTRFS storage backend. I take nightly snapshots and transfer the container to my backup host using lxc copy ct target:ct --refresh
This works, however I noticed some oddly named subvolumes in my target storage pool, like
containers/migration.1158372195/s20250929001/s20250929001

These don’t show up when listing containers or storage volumes, and I didn’t find any occurence when I grepped the SQL dump of LXD global database. What are these subvolumes? Should I keep or delete them?

2

It seems to be leftovers from failed copies/cleanup based on the name of the subvolume and looking at the code here and there.

It is probably safe do delete them but as always when deleting data, tread carefully and make sure to have working and tested backups.