@jo23 still we would need to get more information why it fails. You can also download full logs from the logviewer for the app, what should also include more details on what is going wrong.

@girish can you add another check before the rmdir to log the files left before upgrading to an rm -f ?

This behavior may mean that while deleting everything, something else wrote to the dir during the long operation. So it may need to happen twice as the second try will be much faster.

Not sure of the sequencing too, what else is running that could write to the dir, when it shouldn't be?

a restart dit the trick! thank you.

https://forum.cloudron.io/topic/3661/unable-to-uninstall is the same for another app (mastodon).

@nebulon there should be branching logic after Cloudron 'senses' that it doesn't own that DNS anymore.

Prompt for delete of an unmanaged resource in that state?

in fact in the directory only the dir data is left and it is empty but root:root is the owner

fe2b1d58-3d79-4c28-bc6d-17bc1821095b-image.png

after rmdir /home/yellowtent/appsdata/f266ef68-b01b-4a9e-954e-b8bc68010672/data
the uninstallation went well 😉

@girish wrote:

Are you able to open a request [for this]?

Done, thanks: https://forum.cloudron.io/topic/4306/provide-ui-support-for-choosing-app-package-version-during-installation

And you can follow this doc to restore an uninstalled app from the backup - https://cloudron.io/documentation/backups/#import-app-backup . You have to make sure the app backup is still around though since it will get cleaned up by the backup retention policy.

The issue here was that docker was stuck and not responding to API calls anymore. The issue is known and it happens at random times (no idea what the root cause is).

The fix is to simply systemctl restart docker and if that command gets stuck to just reboot the server.