Cloudron 9.0.8 Backup to site sshfs rsync hetzner storagebox

[luckow]

I regocnized on two instances a `BoxError: SSH connection error: ssh exited with code 255 signal null". Manual backup trigger result: same error.

[luckow]

re-saving the backup config solved the problem.

[girish]

There was a regression around sshfs in 9.0.8 . 9.0.9 has the fix .

[SansGuidon]

I came here after facing this issue for the first time
I first rebooted and retried.
Then I found this thread, I saved the config, remount the storage, re-entered the private ssh key each time on each screen just in case.
It seems to have solved the issue

fortunately there is the forum, otherwise it's easy to panic

I got no email for backup failed error, is that expected? @girish I have backup failed "enabled" in the notification settings

[luckow]

@girish

Mmh? Cloudron version 9.0.10

Nov 18 03:02:22 Exiting with code 0
Nov 18 03:02:22 BoxError: SSH connection error: ssh exited with code 255 signal null
Nov 18 03:02:22 at runBackupUpload (/home/yellowtent/box/src/backuptask.js:204:15)
Nov 18 03:02:22 at process.processTicksAndRejections (node:internal/process/task_queues:105:5)
Nov 18 03:02:22 at async uploadAppSnapshot (/home/yellowtent/box/src/backuptask.js:370:34)
Nov 18 03:02:22 at async backupAppWithTag (/home/yellowtent/box/src/backuptask.js:393:26)

[imc67]

@girish same issue here, totally unexpected and unannounced 1of 3 Cloudrons updated itself to 9.0.11 last night and the first backup at 8:00h failed with same errors.

Opening the config and safe it (and remount) solved it.

Compared to previous backups it's really slow, before 2 minutes for complete backup, now at 30% after 30 minutes. Hope this gets better 🤨

[avatar1024]

Same here when upgrading from 8.3.2 to Cloudron 9.0.12. Remounting storage solves it, but still not optimal as a default behaviour.

[avatar1024]

@joseph how come as this been marked as solved? There have been no response and it clearly wasn't as of 9.0.12.

[joseph]

I thought this was fixed, but I see there is a new fix by @nebulon in https://git.cloudron.io/platform/box/-/commit/c2378d33b453a441709c585ab235102d61bcc1d4 . @nebulon should this issue be considered fixed?

[nebulon]

Yes the first was the fix to the backup and second then the fix to the cleanup, which I forgot initially

[avatar1024]

Just for info, another instance auto updated last night from 8.3.2 to 9.0.13 and something weird happened: I got the same error on the backup site (with two failed backup notifications) AND that backup site got duplicated into a new one but for some reasons as a tar archive instead of rsync like the original site. No idea what happened there.

[girish]

@avatar1024 in the previous versions, one could change the backup format between tgz and rsync . We used to track the format of each backup (for restore from old backups to work). in Cloudron 9, you cannot change the backup format of a site once set. instead, you can just create a new backup site if you want another format.

The migration code probably detected that you have backups in both formats. So, it would create two sites - one tgz and one rsync. That's what you are seeing .