cloudronbackup folder disappears when trying to restore

[3246]

I am trying to pull down a backup for testing but when I do it fails with the following error:

Access denied. Create the directory and run "chown yellowtent:yellowtent /mnt/cloudronbackup" on the server

Great, I'll create it. Again.

Try again.

And... get the same error. The directory I just created disappeared.

This process repeats ad nauseam.

Help!

[3246]

Right, user error. However, I think the error messages don't help.

I use a sub-user and the account had external reachability disabled.. hence why trying to restore off-site didn't work. Yay.

However, the error message and disappearing folder make no sense. I would have expected an error message saying that the connection is refused as my FTP duck did... which led me to check the permissions on the backup server.

@girish @nebulon is there a way we can improve on those messages to help slow sysadmins like myself to find the answers they need sooner?

[nebulon]

I agree this is not ideal. We use systemd-mount here and its not trivial to extract proper error messages from that. Also I believe systemd will manage the mountpoint (folder) then accordingly.

[robi]

@nebulon it would be wise to rely less on systemd and more on Cloudron validations.

[nebulon]

@robi frankly the mount command output, to test it for real, is also terrible for proper error reporting

[robi]

@nebulon agreed, have you considered things like rclone instead?

[girish]

Tested with a variety of cases here in both volumes and backups UI:

Incorrect hostname:

Invalid username/password:

Bad remote path:

Sub account and no samba:

[girish]

This took me a while to figure but when using subaccounts for CIFS, one has to use "/subaccount" as the remote directory. Thanks to https://drechsel.xyz/posts/hetzner-storage-box-how-to-mount-cifs-for-sub-user/

[girish]

I found the issue. Fixed in https://git.cloudron.io/cloudron/box/-/commit/b2d072f2e6d3c8f1b39842c11b70f29ed2cb1231