So I screwed things up trying to run addon tests...

[malvim]

Hey, guys!

So I talked on the other thread about running tests for addons on the raspberry pi, and got some support from @girish when trying to run the postgresql addon.

It turns out I ran it on my production cloudron to see if it would run correctly (if there was something about the test that needed some cloudron "infrastructure" that was not properly set up on my laptop, maybe) and didn't notice it would kill my PRODUCTION postgresql container. Hah! It crashed my nextcloud and wallabag instances. No worries, I'm the only user of these services, so okay.

I restored both from the latest backups and rebooted the machine so box would set everything up again. I THOUGHT everything seemed okay afterwards, nextcloud was responding, wallabag as well.

Except now they're down again, and docker ps on the server shows no postgresql container.

Any idea what I should do? Thanks!

[girish]

@malvim have you tried services -> restart postgres? It will rebuild the container if it is missing

[malvim]

@girish I had, and it kept red for a while, but just did again and it is okay now.

So one thing I noticed was that it was trying to run the test image, I think because it overwrote the "postgresql" name on docker, when trying to run the tests. So it was using cloudron/postgresqltestadon image instead of the correct one. I removed all the test containers and images, restarted the server, postgres container was still not there, I restarted via cloudron interface, it didn't work, but now it did.

postgresql was showing logs from the last time it was up, a few days ago, and never showed logs from the times I tried restarting. Maybe cloudron was still bringing services up, or something like that, the first time.

It's all good now. Not 100% sure what happened, but it seems to be okay now.

[malvim]

So, it should be obvious for someone who has worked in IT for a while like me, but... Kids, DON'T EVER run tests on your production servers. Ever!