I mistakenly jumped from NC22 to NC24 with my docker image (separate pgsql). It of course ‘refused’ to upgrade and logged that multi-rev upgrades are not supported. Problem is whatever it did stamped NC24 on something and spinning up an NC23 instance against the db also complains since downgrades are not supported and it thinks NC24 is the last release I ran. Grrrrr…
What can I do to fix this? (yes I can recover another way, I want to judge which one is less of a time waster)
best would be restore from backup and follow the right procedure.
for other cases search this forum there are lot discussions for this topic.
The short answer for me was erase the nextcloud_html persistent volume, rebuild the container w. NC23 and everything after that was part of the upgrade process as opposed to the recovery.
Upgrading was definitely wacked: upgrade failed w/ an occ error: ‘redisexception noauth required’, so disable the redis password auth that’s been working just fine up til now. Then php occ upgrade within the container. And then I find out that user backend OC_User_SSH disappeared so had to re-install user_external. Then logins didn’t work and I decided to jump to NC24 because why not. It behaved the same exact way, but I managed to figure out it was php session saving issues thanks to this issue: Cannot login after fresh install · Issue #13090 · nextcloud/server · GitHub and
php -i | grep session.save_path spit out a redis path w/ authentication. I re-enabled redis password auth and it all works again.
Bit of a crazy rollercoaster