Hardware: RPI 4 at home YunoHost version: 11.1 I have access to my server : SSH/Webadmin Are you in a special context or did you perform some particular tweaking on your YunoHost instance ? : no If yes, please explain:
Description of my issue
After upgrading to the latest Yunohost version and upgrading all packages trough the web-ui I saw that Mastodon required an update, i clicked it , it took about 2 hours during wich both Mastodon and the web-ui were unresponsive
After that, it failed in the background and mastodon + web-ui came back up , I tried to run the update trough the CLI to understand a bit more about the issue and here’s what it says :
Could you go in webadmin > tools > logs and find the log of this upgrade operation ?
Click on the green button “Share with yunopaste” and copy the link here.
LoadError: Could not open library '/var/www/mastodon/live/vendor/bundle/ruby/3.0.0/gems/blurhash-0.1.6/lib/../ext/blurhash/encode.so': /var/www/mastodon/live/vendor/bundle/ruby/3.0.0/gems/blurhash-0.1.6/lib/../ext/blurhash/encode.so: cannot open shared object file: No such file or directory
Puis la restauration a échouée parce que
/var/cache/yunohost/app_tmp_work_dirs/app_fn_ktt9s/restore: line 123: ld_preload: unbound variable
Pour en savoir plus sur la raison de l’échec de restauration, il serait utile de voir le contenu du fichier _common.sh qui définit cette variable ld_preload
tar -axf /home/yunohost.backup/archives/mastodon-pre-upgrade1.tar apps/mastodon/settings/scripts/_common.sh -O
On pourra alors peut être réparer l’archive pour restaurer mastodon.
Update : After trying different things, reinstalling mastodon, still no data anywhere and the original backup has disappeared.
I’m gonna try and pull from my remote Borg backup to see if I can salvage anything.
But I’m getting a bit worried about this, it’s fine if I never recover Mastodon and my data but it shouldn’t be normal for Yunohost to “blow up” on it’s own. I’ll let you know how I recover if I manage to
→ I tried restoring from a “mastodon-pre-upgrade2.tar” that was created in it’s place. I have ruby errors on Mastodon-Web when running it (unavailable right now to copy/paste since I’m doing step 3) and a 502 error when loading the web interface
→ After searching trough the entire forum’s search result for Mastodon I’m now trying to upgrade this “failed” restore to the latest Mastodon version to see if it restores access
→ If that doesn’t work can I pull an old backup from my borg repo and restore everything ?
→ Is there a way to surgically pull my followers/media/toots from an old borg repo backup and nuke my server to start all over again clean ? I know this isn’t a popular thing to do but the forums have no answers , I’m apparently the only one to have had such a catastrophic failure with Mastodon and I’m at the limits of my competence on the issue.
→ Does Borg even keep old version or did I loose everything ?
Bottom line for everyone combing trough the forum : Please backup your Mastodon profile/followers before doing a routine update, my server blew up on it’s own
So, I finally learned about how Borg Backup work, logged in trough my MacOS desktop and pulled an archive of Mastodon from a day before the explosion.
I downloaded it straight from the borg repo and created a .tar archive of it
However when I use the Yunohost CLI to import this backup I get the following error : Error: Could not load info for archive ‘/home/yunohost.backup/archives/mastodo040223.tar’… The info.json file cannot be retrieved (or is not a valid json).
Absolutely no idea what format Json this thing expect, there’s close to 0 information available both in docs and in the forums
Update : Managed to pull a BorgBackup mastodon backup from before the accident but it still fails with some Ruby error when restoring mastodon exactly like before. Probably because my Yunohost version is now higher than it was before the accident