Custom ZFS dataset names have automatically been prepended with projectname_

Currently on lxd 3.23.
Dataset name was: zfspool1/custom/drive1
and without my intervention it is now: zfspool1/custom/default_drive1. Existing snapshosts have also taken on the name change.

What is best practice for naming of custom data? Do I create zfspool1/custom/default_drive2 or will that ultimately end up as zfspool1/custom/default_default_drive2? What actually triggers an automatic rename to included the project name? If the custom data is not part of a profile does that change the behavior?

You should never create any of those by hand, only though lxc storage volume create, then LXD will figure out the right name to use in your zpool.

The dataset name now includes the project name as storage volumes can now be isolated in projects and we needed to make sure there could be no name conflicts across projects.