[Update: Failed again] Update to 0.18.1-rc.1 tried and rolled back
We've upgraded lemmy.world to 0.18.1-rc.1 and rolled back that upgrade because of issues.
(If you had posted anything in those 10 minutes between upgrade and rollback, that post is gone. Sorry!)
The main issue we saw is that users can't login anymore. Existing sessions still worked, but new logins failed (from macos, ios and android. From linux and windows it worked)
Also new account creation didn't work.
I'll create an issue for the devs and retry once it's fixed.
Edit Contacted the devs, they tell me to try again with lemmy-ui at version 0.18.0. Will try again, brace for some downtime!
Edit 2 So we upgraded again, and it seemed to work nicely! But then it slowed down so much it was unuseable. There were many locks in the database. People reported many JSON errors. Sorry, we won't be on 0.18.1 any time soon I'm afraid..
@ruud@lemmy.world just wondering if you have considered setting up a second, beta, instance of lemmy.world open to the public?
With all the performance issues with 0.18.1, it's highlighted that there needs to be a way to stress test these updates before applying them to the main instance.
Yes, considering that. But we'll need people to use that when we will do testing...
Well, you have at least one person who would be willing to use it ;D
One more. You find a stable way to notify anout upgrades and get a test sheet to run through and we can generate posts and activity to help test with.
Light the beacons! lemmy.world calls for aid!
I'm not familiar with what the server architecture looks like, but is there a possibility of using a load balancer in front of the instance's server and swapping a "beta" server into the load balancer when you need to do testing? You could basically migrate your traffic with zero downtime, assuming Lemmy's architecture allows for it.
Well that doesn't differ much from what I do. I just copy the files to a second directory and test with that. Easy rollback. Downside is, that all data is lost between upgrade and rollback, which will be the same in the scenario you suggest.
Ah, that's fair. Best of luck either way. This is the tough part of admin work. FWIW, the stability we've had thus far has been pretty impressive!