Help!
Upgraded server to Ubuntu 25.10 due to impending EOL for 25.04. Now when I attempt update Discourse it fails with this error consistently:
What is causing this?
AI-generated summary
Andro encountered an issue updating Discourse after upgrading their server to Ubuntu 25.10, encountering a consistent error (shown in an attached image). They later resolved it by performing a git pull and rebuilding the app via launcher rebuild app Andro. pfaffman advised sticking to LTS versions for server stability, noting upgrades every 5 years are sufficient — though itsbhanusharma clarified that new LTS releases arrive every 2 years, and skipping one can complicate upgrades. pfaffman recommended migrating to a new server for OS upgrades to avoid downtime, especially using a static IP switch. itsbhanusharma countered that in-place Ubuntu upgrades can be quick (2–3 minutes) with proper staging tests, though this approach is harder with colocated hardware lacking flexible provisioning. pfaffman expressed surprise at the short downtime and considered trying in-place upgrades again for their own colocated systems.
Answering my own question. I proceeded to do a git pull and a launcher rebuild app and all is back to normal. I’ll leave the query here in case anybody else hits this issue.
It’s generally recommended to go with the LTS versions so you have to upgrade only once every 5 years. Though for many years I ran the latest available Ubuntu on my desktop/laptop, there’s not much advantage of living on the bleeding edge for a server.
isn’t that a stretch? new lts is available every 2 years, but jumping 2 releases at once will be a lot to chew at once no?
I recommend moving to a new server when it’s time to upgrade as an in-place OS upgrade requires lots of downtime and might have issues. So I mostly don’t move folks to a new server until the 2nd LTS release (i.e., I’ve sites on 22.04 that I’ll upgrade this year, likely waiting until 26.04.1 gets released in the summer.
Moving to a clean install requires zero downtime (especially if you can do it with a static ip that gets moved to the new server once it’s writing) working).
I’ve done quite a few of them on ubuntu, and barely ever had a site out of comission for more that 2-3 minutes. Granted I’ve tested the process on staging to ensure there are no blockers but managing a fleet of two dozen or so servers justifies the extra steps.
That’s a good strategy but only works on systems that have that option, most of what I have is colocated hardware with no luxury of spinning up new machines at will.
Wow. I guess it’s been a decade or more since I’ve done one of those. I had no idea it was 2-3 minutes of downtime.
I’ve got a few of those too. Maybe I’ll try an OS upgrade next time.
Thanks!