Upgrading apps fails (fail2ban issue) and makes them broken (nginx landing page)

My YunoHost server

Hardware: Dedicated server
YunoHost version:
yunohost: 4.2.8.3
yunohost-admin: 4.2.5
moulinette: 4.2.4
ssowat: 4.2.4
I have access to my server : Through SSH
Special context : yes

I’ve updated my server after the bullseye release, and it said stable instead of buster in my apt source (yunohost tools update/upgrade system). Normally it would ask if you want to change the entry in your sources, but mine didn’t. I had some dependency issues so “fixed” those (basically kept digging until I found that python3-yaml wasn’t installed, after installing this one it installed like 1.3k packages or at least a lot). My guess is that somehow it upgraded to bullseye? or were there just this many updates? I updated about a month ago. in /etc/os-release it states buster. Since, I’ve changed the sources files.

Description of my issue

I’ve ran yunohost update/upgrade apps and it has broken netdata, matrix-synapse, funkwhale, nextcloud and mailman. Peertube was broken too, but “fixed”^TM by symlinking /usr/bin/python to /usr/bin/python3

I think it might be fail2ban related, since I’ve seen this log coming back in each upgrade (didn’t keep all logs):

Warning: [Error] Upgrade failed.
Warning: Job for fail2ban.service failed.
Warning: See "systemctl status fail2ban.service" and "journalctl -xe" for details.
Warning: [Error] Sep 07 19:18:06 systemd[1]: Reloading Fail2Ban Service.
Warning: Sep 07 19:18:06 fail2ban-client[299830]:  NOK: ("'FilterPyinotify' object has no attribute 'isAlive'",)
Warning: Sep 07 19:18:06 fail2ban-client[299830]: 'FilterPyinotify' object has no attribute 'isAlive'
Warning: Sep 07 19:18:06 systemd[1]: fail2ban.service: Control process exited, code=exited, status=255/EXCEPTION
Warning: Sep 07 19:18:06 systemd[1]: Reload failed for Fail2Ban Service. 

I think it might be something in python 3.9 with some syntax changed, and if I remember correctly I read something about how fail2ban will give issues when upgrading to bullseye.

I can login and open the main landing page, but when going to the admin and loggin in it give a 500 internal server error about yunohost-api btw. Not sure if related.

If I have to wait for yunohost to release under bullseye, I don’t have an issue doing so. But, if my analysis is wrong and I can fix it, I would be very much taker !

Thanks for reading this far already and have a nice day/night!

If you see “stable” instead of “buster” in /etc/apt/sources.list, and considering that it upgraded 1.3k packages, then yes, you probably upgraded to bullseye. I’m surprised that this didn’t trigger a Yunohost uninstall … But anyway at this point your server is pretty much fucked up

If yunohost backup create works, you could try to do a backup and restore it on a debian buster yunohost and next (force) upgrade all apps to repair them.

surprisingly most stuff keeps working, it’s only those 4-5 services.

to install a new machine would take me days to configure, full disk encryption on a dedicated machine with FDE data disks on btrfs, not 1 2 3 done :wink:

I guess I’ll wait for yunohost bullseye stable then, thanks for the responses! Perhaps I can be a tester for bullseye then (yay, testing in prod).

the weird thing is that the apps that failed, are not appearing anymore as installed? I hope I haven’t lost the data? how can I know and how to fix in the future?