Several users have reported issues with updating Discourse to versions 2.3.8 and 2.3.9, with the update process claiming failure and requiring a server reboot to complete. peternlewis initially reported this issue, and volas shared a similar experience with error logs showing a Discourse::NotFound error.
After updating to 2.4, peternlewis encountered further issues, including the site going down and requiring a ./launcher rebuild app to recover. This has led to a lack of confidence in the update process, causing stress and uncertainty.
codinghorror suggested that rebuilding twice after each update may not be a significant issue, but peternlewis emphasized the importance of a reliable update process, especially for critical forums. codinghorror shared their own experience of frequent rebuilding without issues, but peternlewis noted that the problems only started with recent versions and that updating should be a bulletproof process.
The 2.3.8 and 2.3.9 updates, both finished with:
Sorry, there was an error upgrading Discourse. Please check the logs below.
Version page at that point shows everything still needs updating.
Logs do not indicate anything wrong.
Reboot the server, and when it comes up, everything is updated.
I don’t know what any of this means.
Is it the maxmind geolocation database issue?
I hope you were asking someone else, since I don’t know what that means. I’ve never heard of “maxmind”.
Same problem (2.3.8 to 2.3.9), /admin :
and /admin/upgrade :
Unfortunately I closed upgrade page with logs, can’t find them anywhere else. Only this in /logs
Discourse::NotFound (Discourse::NotFound)
/var/www/discourse/plugins/docker_manager/app/controllers/docker_manager/admin_controller.rb:89:in `upgrade'
2.4 update failed as well. Basically took the site down.
Rebooted the Digital Ocean instance instance, and it came back up but the forum was entirely blank (as in, the whole contents of the browser window was just black).
Did ./launcher rebuild app and it came back.
But this is making me extremely nervous about updating! I’m going to have to start snapshotting my Digital Ocean instance before each update.
So rebuild twice every time? Doesn’t seem like a huge problem to me.
The problem is the lack of confidence this produces.
My forum is, unsurprisingly, very important to me, and to a lot of other people.
So when I upgrade it, and it goes offline, and I have no idea if a rebuild is going to get it back or not, that is a lot of stress, and a lot of uncertainty.
Odd. I rebuild mine all the time (probably 6-12 times a year, times two instances, on bare metal hosting at a colocation provider) and I’ve never had these experiences 
I’ve never had any problem with rebuilding.
But it is only the last few versions that updating has been failing in various ways.
Anyway, it seems to me that updating should work and be bulletproof, and should not fail and require a rebuild to get the site back up and running.