Continuing the discussion from Site offline after rebuild (4th Feb 2025):
How do I implement “the fix”?
You rebuild.
1 Like
Each of these indicates a failed or missing substitution - I haven’t read over all of the previous posts but I see a LOT of people on Docker version 20.10.5 are having trouble:
This might be a cause of the problem.
Did you take a backup of the server first? Can you revert back to before you first tried the upgrade?
If not, I’m sure all of the data is still there,
Regardless, take a backup of it now.
What do you see if you:
cat /var/discourse/shared/standalone/postgres_data/PG_VERSION
2 Likes
At some point they decided that Support was for help with using a working Discourse site and Installation was for help for self-hosters who don’t have a working discourse site. It’s not help with installation, but help with your installation, whether it exists or not. So it’s installation as a noun rather than a verb. I was confused the first dozen times they did it to me too. 
You can’t. There’s no going back. (Unless, perhaps, you have a backup image of the whole server.)
I thought I’d seen that you’d reverted to some old server image and your site was running again? Or maybe that was someone else.
You can answer the above questions and maybe Michael can help you. I think that it will be much faster, easier and safer to move to a new droplet. If you have a recent discourse backup ,you can just do a fresh install and restore that backup there. If you don’t, you can follow the Move a Discourse site to another VPS with rsync instructions and try the upgrade there. You have nothing to lose since your existing site will keep functioning. If you find that it works, then you can stop your old site, run the rsync command, do it again, and then switch the DNS.
1 Like