What type of hardware are you using: VPS bought online What YunoHost version are you running: 11.3.0.2 How are you able to access your server: The webadmin
SSH
Describe your issue
Hi all,
I wanted to migrate my Yunohost 11 / debian 11 to Yunohost 12 / debian 12. I have updated all packages and APPs (as far as possible). Under the CLI, I run the command yunohost tools migrations run but there was nothing to do: Info: Keine Migrationen durchzuführen.
But I m on debian 11 and yunohost 11. I checked this. So I run yunohost tools migrations run 0027_migrate_to_bookworm --force-rerun and than yunohost tools migrations run 0027_migrate_to_bookworm --force-rerun --accept-disclaimer Than the process starts.
Fehler: Migration 0027_migrate_to_bookworm gescheitert mit der Ausnahme Still on YunoHost 11.x at the end of the migration, eh? Sounds like the migration didn't really complete!?: Abbruch
Info: Der Vorgang'Migrationen durchführen' konnte nicht abgeschlossen werden.
Than I checked the OS:
lsb_release -a
No LSB modules are available.
Distributor ID: Debian
Description: Debian GNU/Linux 12 (bookworm)
Release: 12
Codename: bookworm
I’ve seen similar situation which was something along the lines of libluajit being stuck in an old version because of minetest(?) … but forcing the upgrade and then force-upgrading minetest to have everything back on track worked well
Let’s dig in with : sudo apt install yunohost yunohost-admin moulinette ssowat (note that yes, it’s install here because that’s the way to tell apt to upgrade those specific packages)
It may complain that it will remove libluajit which is fine … and maybe some <app>-ynh-deps which should be okay-ish too, just keep in mind that you should force-upgrade the app once yunohost is upgraded, using : yunohost app upgrade <app> --force
@Aleks Thank you very much for your help and suggestions. I have executed the commands so far and it seems to have worked somehow. Somehow there were still two or three errors with the updates, but I solved them myself.
Afterwards there was a 500 error from nginx, but after a reboot it was also gone.
I then started one or two updates and updated the APPs with the --fore option. Then there were two new migrations that failed due to funkwhale. As I had only installed this APP for testing purposes, I deleted Funkwhale without further ado and then it worked again.
Somehow the whole thing was now quite a back and forth of commands. Could the system now be in an unstable state? A state that looks quite good at first, but then leads to problems with the next or the next but one update / action? After all these difficulties, everything in me is screaming to completely rebuild the entire system in order to get a guaranteed ‘clean’ state.
Nah that sounds all good to me, your server is not “dirty”, please dont fall in the reinstallation spiral, forget the dream of a “clean” server whatever this may mean
@Aleks Thank you for your encouraging words and the link to the article on ‘the reinstallation spiral’. Yes, there is some truth in that! I’ve experienced time and again that you can’t always reinstall the server as soon as something goes wrong. Sometimes important things are running there and the effort to completely reinstall everything is sometimes immense.
I’ve been working with servers for almost 30 years now and, although not 24/7, I’m always out in the digital swamp with both hands up to my shoulders solving technical conflicts. Nevertheless, I still have the feeling of being a beginner and only understanding half of it.
A ‘clean’ server seems like something I can understand and repair. If the server gets too messy, then it sometimes requires tricks that reach into areas that I can no longer comprehend. At least that’s my concern. Nevertheless, I have managed to solve everything so far. However, the effort involved was immense. And the desk has many a bite mark
