Auto-update to 8.3 - various apps down - database issue

[timconsidine]

Not very happy that 8.3 rolled itself out when it doesn’t seem stable
And it had to happen when I am in the middle of projects, which means I don’t have time to research fixes.
So, seeking to leverage the ppower of the community )not being lazy of course) is there a standard fix path for issues like Error : - Error setting up postgresql. Status code: 500 message: the database system is in recovery mode ?
I have 10 apps down, some of which are core productivity apps like paperless and gitlab.

[timconsidine]

Seems that :

• some can fixed with a simple Retry Task
• some with Retry Task then Restart app/container
• some stay as not responding despite those steps

Will have to investigate tomorrow

[timconsidine]

Update : after manually doing retry task and restart on all apps with error (sometimes multiple times), and restarting postgres and the relevant redis, I could resolve all except :

• keycloak : a new installation so uninstalled and will re-install later
• onlyoffice : don’t use much so uninstalled and trying new installation

So going to close this
But worrying it happened.

UPDATE : don’t see how to close - can @staff please do so

[fbartels]

https://forum.cloudron.io/topic/13412/8.3.0-postgres-upgrade-failure

I ran into the same issue. For me all postgres databases were empty and the fix was to restore these specific apps from a backup from before the upgrade.

[robi]

One of mine upgraded with no issues.

[imc67]

Last night my big production Cloudron (out of 3) was updated but I'd noticed that on forehand so after reading issues I increased postgres memory a lot and made sure all apps were updated before. Luckily it went all good but it would've been a disaster if this one had longer downtime.

[robi]

@imc67 @staff perhaps that's a new feature Cloudron upgrades should have, especially when upgrading major database versions and schema updates which may need a bump in the memory requirements temporarily.