Lemmy 0.18.3 has been officially released. You can read the official release notes here. The eagle-eyed among you may notice that we are upgrading from 0.18.1 to 0.18.3 and wonder what happened to 0.18.2, and I share you confusion. It never got an official GitHub release so I never upgraded to it.
I plan to update tonight at 2023-07-29 2023-07-30 2023-07-31 04:00 UTC. Expected downtime is slightly longer than usual due to database updates, but should still be less than 10 minutes. As always, downtime may extend to up to an hour if any thing goes wrong.
Edit: There was a glitch with the update and I had to restore from a backup, which took most of the planned hour. I did successfully test the update on a clone of the production VM beforehand so I’m hopeful that simply retrying the upgrade will succeed, but to keep things within my one hour maintenance window I’ll hold off on trying again until the same time tomorrow.
Edit 2: Sorry folks, same result attempting the update tonight. I did at least get a much better capture of the logs as I was prepared this time. I’m indefinitely postponing the update while I investigate and come up with a solution. If I can repeat the update successfully again on a test server I have some ideas on how to use that as a migration path. Will also be reaching out to other instance admins on Matrix to see if this is a know issue. Will keep you all updated with a new post when I have something new to share.
Edit 3: I had an epiphany last night as I was going to bed. I tested it this morning and my guess was correct. Basically I’m just an impatient idiot. This update requires a significant database migration, during which time the lemmy backend doesn’t respond to requests. The frontend is up and running though and is able to display en error message. Being cautious about downtime, I saw the error message and defaulted to getting things to a known-good state withing the one hour maintenance window I had published. I suspect I was multi-tasking a bit while testing the upgrade so I just never got around to loading the test instance until after the database migration had completed. Will proceed with the update tonight with an improved understanding of what to expect.
Edit 4: Update completed. Everything looks good.
Thank you for being a great admin!
Eh, they can’t all go smoothly. You’re doing good work here, and I’ll keep my fingers crossed for the next attempt, for your sake!
I will cross my fingers double harder for the next attempt. I am sorry I failed you last night; I will not make that mistake again.
Thank you, fingers crossed it will succeed next time! 🤞
I posted about updating Villa to 0.18.2 some time ago, I didn’t realize it didn’t get a release on GH, appreciate the clarification.
FYI, I don’t always see local posts (although I’m trying to explicitly take a look at the local tab once per day). If you post stuff that you hope I’m going to see, please @mention me so that I don’t accidentally ignore you!
I did, but apparently only comments with mentions send notifications, not posts. Anyway, thank you again for your hard work!
Ah, that’s interesting. Thanks for letting me know.
Thank you for treating this instance like a production server with a high-availability SLA and not just a hobby project we’re all mooching off of. It’s nice being able to rely on this little server while I read about larger instances having stability issues.
I heard this 0.18.3 release can reduce the server’s disk usage by like 80%! I know you recently had to expand the disk size, hopefully you won’t have to do that again after this upgrade.
Very awesome, thank you!
