Users are encountering a Zlib::BufError: buffer error during Discourse rebuilds, particularly after updating Theme Components, with the error often occurring on the first attempt but resolving on subsequent rebuilds. Some users, including Isambard and nathank, report success after retrying the rebuild without changes, while others, like pfaffman, suspect Maxmind integration as the culprit and temporarily disabled it to proceed. itsbhanusharma notes the issue is inconsistent and not easily reproducible. philh confirms the problem occurs with Maxmind enabled but disappears when it’s commented out — though updates via /admin work fine without touching app.yml. Many users are now adopting a 2-container setup to keep sites live during rebuilds. The consensus leans toward this being a transient bug, possibly unrelated to Maxmind, with retrying the rebuild as the most common workaround.
I just tried to update and there is a failure near the end:
I, [2024-07-04T07:17:18.714988 #807] INFO -- : Writing /var/www/discourse/public/assets/scripts/discourse-test-listen-boot-9b14a0fc65c689577e6a428dcfd680205516fe211700a71c7adb5cbcf4df2cc5.js
rake aborted!
Zlib::BufError: buffer error (Zlib::BufError)
...
< more stuff >
...
FAILED
--------------------
Pups::ExecError: cd /var/www/discourse && su discourse -c 'SKIP_EMBER_CLI_COMPILE=1 bundle exec rake themes:update assets:precompile' failed with return #<Process::Status: pid 805 exit 1>
Location of failure: /usr/local/lib/ruby/gems/3.3.0/gems/pups-1.2.1/lib/pups/exec_command.rb:132:in `spawn'
exec failed with the params {"cd"=>"$home", "tag"=>"precompile", "hook"=>"assets_precompile", "cmd"=>["su discourse -c 'SKIP_EMBER_CLI_COMPILE=1 bundle exec rake themes:update assets:precompile'"]}
bootstrap failed with exit code 1
** FAILED TO BOOTSTRAP ** please scroll up and look for earlier error messages, there may be more than one.
But I already have the maxmind account ID in the yaml file.
I guess I could enter the container and re-run this command again, but I wonder in general is there a way to recover from failed rebuilds or do people just try to run the whole thing again?
Yesterday i gave up and commented out maxmind. I then went into the container and added the values to discourse.conf and it pulled the database successfully (a rather complicated work around).
I don’t understand how this could be happening, but it does look like a bug.
I think the only solution now is to do without maxmind.
So I had this issue today however, It was only once, I just issued rebuild again and it went well. so weird there is no consistent way to reproduce this. other containers with similar configuration built just fine.
I’ve rebuilt 3 sites today, and each one tripped up as above at the same time, which seemed to be just after updating the Theme Components.
Each time a 2nd rebuild ran just fine, with me changing nothing at all. I do wonder if Maxmind is a bit of a red herring (or simply a different issue).
I use a 2 container setup so a failed rebuild is no big deal.
I had the same experience. It seems to fail on the first run. Luckily I switched to a 2 container set-up so the site remains live while I run the rebuild twice.
