Update to 8.0.3 left some apps in a non-responding state

[msbt]

Hey there, just a quick fyi: Tonight one of my bigger Cloudron instances got updated to the latest version automatically and it didn't manage to automatically remount two external volumes. Attached to those are two Nextcloud instances, so naturally they were down and in a restart loop. Weirdly enough, at least two LAMP instances also didn't come back up, not until the Nextclouds were running again. It looks like some cert action was going on and got stuck because the others didn't finish.

The solution was easy enough, manual remount for the volumes, then everything came back up again. So this post is just an fyi, but if you want to investigate further, let me know.

[girish]

We started rolling things out slowly. Yours is probably one of the early instances. I wonder what caused the volumes to get unmounted though. Can you tell us how they are mounted (as in, mount provider)?

[msbt]

@girish ah ok, not unexpected then It's 3 different mountpoints (2 volumes, 1 backup, all were unmounted), Hetzner storageboxes via SSHFS.

[imc67]

Same here with an auto-update of Cloudron the backup volume was disconnected (Hetzner)

[girish]

Not clear why this happens but with Cloudron 9 we will just add logic to automount before a backup. Ideally, the kernel/systemd should have flags for this, but it doesn't.

[msbt]

Sounds good! It was also odd that some other apps were affected by the restart-looping Nextcloud instances.