I just migrated a ZFS pool to a new machine. I had previously used a filesystem as the Nextcloud data/ directory, containing user directories. I created a new folder, called “old,” and moved everything into it, thinking it would be out of the way. Then I pointed the filesystem mount point /var/subsonic to docker’s /var/www/html . I started the containers and entered my database credentials, and then hit Submit. When I came back, the “old” folder was gone and my zpool was 800 GB lighter.
Mostly a warning, I guess. Is there any chance this data isn’t totally gone?
yes Docker image takes control of the directories you mount into the container /var/www/html:
Custom volumes
If mounting additional volumes under /var/www/html, you should consider:
Confirming that upgrade.exclude contains the files and folders that should persist during installation and upgrades; or
Mounting storage volumes to locations outside of /var/www/html.
Data inside the main folder (/var/www/html) will be overridden/removed during installation and upgrades, unless listed in upgrade.exclude. The additional volumes officially supported are already in that list, but custom volumes will need to be added by you. We suggest mounting custom storage volumes outside of /var/www/html and if possible read-only so that making this adjustment is unnecessary. If you must do so, however, you may build a custom image with a modified /upgrade.exclude file that incorporates your custom volume(s).
not from container`s point of view. but if you have backup or maybe ZFS snapshot it should be fine. As last resourt I think chances exist to "un"delete data from ZFS volume even without this measures - but you should immediately write-protect the volume and only access it in read-only mode using forensic tools.
Thanks for pointing to that documentation, clearly I missed it. Hopefully this will help someone else. I’m looking into data recovery tools and I’ll update this thread with my results.