Version 6.0: elasticsearch not configured

[ggruening]

Infos

• Podman version: 3.4.4 (the problem is not related to podman, zammad stack v9.0.1 (zammad 5.4.1-29) is working perfectly
• Podman-compose version: 1.0.3 ((the problem is not related to podman-compose, zammad stack v9.0.1 (zammad 5.4.1-29) is working perfectly
• Operating system (Docker host): Ubuntu 22.04 LTS

Expected behavior

• elasticsearch is configured and connected to zammad

Actual behavior

• elasticsearch seems to be gone after zammad-init is finished. zammad-init does sucessfully rebuilding searchindexes,...

rebuilding es searchindex...
I, [2023-06-27T20:39:57.025814 #23]  INFO -- : ActionCable is using the redis instance at redis://zammad-redis:6379.
I, [2023-06-27T20:39:57.048064#23-6180]  INFO -- : Using memcached as Rails cache store.
I, [2023-06-27T20:39:57.048155#23-6180]  INFO -- : Using the Redis back end for Zammad's web socket session store.
W, [2023-06-27T20:39:57.150043#23-6180]  WARN -- : Setting "timezone_default" is empty. Using UTC instead. Please set system timezone.
I, [2023-06-27T20:39:58.551720#23-6180]  INFO -- : Setting.set('models_searchable', ["Chat::Session", "KnowledgeBase::Answer::Translation", "Organization", "Ticket", "User"])
Dropping indexes... done.
Deleting pipeline... done.
Creating indexes... done.
Creating pipeline... I, [2023-06-27T20:40:00.822773#23-6180]  INFO -- : Setting.set('es_pipeline', "zammad0f43dd30-76f2-4e5c-9e9b-f34d2d6c7cbb")
done.
Reloading data... 
  - Chat::Session... 
    done in 0 seconds.                    
  - Cti::Log... 
    done in 0 seconds.                    
  - Group... 
    done in 0 seconds.                    
  - KnowledgeBase::Answer::Translation... 
    done in 0 seconds.                    
  - KnowledgeBase::Category::Translation... 
    done in 0 seconds.                    
  - KnowledgeBase::Translation... 
    done in 0 seconds.                    
  - Organization... 
    done in 0 seconds.                    
  - StatsStore... 
    done in 0 seconds.                    
  - Ticket::Priority... 
    done in 0 seconds.                    
  - Ticket::State... 
    done in 0 seconds.                    
  - Ticket... 
    done in 0 seconds.                    
  - User... 
    done in 0 seconds.                    

• but when the system is up and running, you can't trigger a rebuild anymore. zammad-elasticsearch is still up and running without errors, but it seems, that elasticsearch is gone somehow:

$ podman-compose exec zammad-scheduler bundle exec rake zammad:searchindex:rebuild
I, [2023-06-27T20:56:54.746174 #30]  INFO -- : ActionCable is using the redis instance at redis://zammad-redis:6379.
I, [2023-06-27T20:56:54.751568#30-6180]  INFO -- : Using memcached as Rails cache store.
I, [2023-06-27T20:56:54.751618#30-6180]  INFO -- : Using the Redis back end for Zammad's web socket session store.
Elasticsearch is not configured.

• I tried to connect elasticsearch to zammad, it gives me this:

  $ podman-compose exec zammad-scheduler rails r "p Setting.set('es_url', 'http://zammad-elasticsearch:9200')"
  I, [2023-06-27T20:59:14.397657 #33]  INFO -- : ActionCable is using the redis instance at redis://zammad-redis:6379.
  I, [2023-06-27T20:59:14.402765#33-5380]  INFO -- : Using memcached as Rails cache store.
  I, [2023-06-27T20:59:14.402811#33-5380]  INFO -- : Using the Redis back end for Zammad's web socket session store.
  /opt/zammad/app/models/setting.rb:38:in `set': Can't find config setting 'es_url' (RuntimeError)
      from /usr/local/bundle/gems/railties-6.1.7.3/lib/rails/commands/runner/runner_command.rb:45:in `<main>'
      from /usr/local/bundle/gems/railties-6.1.7.3/lib/rails/commands/runner/runner_command.rb:45:in `eval'
      from /usr/local/bundle/gems/railties-6.1.7.3/lib/rails/commands/runner/runner_command.rb:45:in `perform'
      from /usr/local/bundle/gems/thor-1.2.1/lib/thor/command.rb:27:in `run'
      from /usr/local/bundle/gems/thor-1.2.1/lib/thor/invocation.rb:127:in `invoke_command'
      from /usr/local/bundle/gems/thor-1.2.1/lib/thor.rb:392:in `dispatch'
      from /usr/local/bundle/gems/railties-6.1.7.3/lib/rails/command/base.rb:69:in `perform'
      from /usr/local/bundle/gems/railties-6.1.7.3/lib/rails/command.rb:48:in `invoke'
      from /usr/local/bundle/gems/railties-6.1.7.3/lib/rails/commands.rb:18:in `<main>'
      from /usr/local/bundle/gems/bootsnap-1.16.0/lib/bootsnap/load_path_cache/core_ext/kernel_require.rb:32:in `require'
      from /usr/local/bundle/gems/bootsnap-1.16.0/lib/bootsnap/load_path_cache/core_ext/kernel_require.rb:32:in `require'
      from bin/rails:5:in `<main>'

Steps to reproduce the behavior

• Do a clean install according to Install with docker and try the commands mentioned above.

Thanks in advance: for any hint and your fantastic work!

Gregor

[fleverest]

I can confirm this is the case on Ubuntu 22.04/Docker version 24.0.2, build cb74dfc/Docker Compose version v2.18.

However, I notice that accessing the rails console doesn't seem to access the production database:

$ docker compose exec zammad-railsserver rails r "p Ticket.count"
I, [2023-06-28T23:39:27.950173 #2514]  INFO -- : ActionCable is using the redis instance at redis://zammad-redis:6379.
I, [2023-06-28T23:39:27.971176#2514-5380]  INFO -- : Using memcached as Rails cache store.
I, [2023-06-28T23:39:27.971698#2514-5380]  INFO -- : Using the Redis back end for Zammad's web socket session store.
0

I believe this is because the entrypoint doesn't initialize the appropriate environment variables:

$ docker compose exec zammad-railsserver bash -c "echo \$ELASTICSEARCH_HOST"

$

It could be a case of adding an exec "$@" line to the end of docker-entrypoint.sh?

I could be wrong, of course :smile:

[fleverest]

The following seems to work:

docker compose exec zammad-scheduler bash -c "source /docker-entrypoint.sh && rake zammad:searchindex:rebuild
"

So this confirms the console environment doesn't have the required environment variables runs when you add the entrypoint variables manually . I'm not sure if this is the case for previous versions, but I'm certain the following command used to work but now doesn't:

docker compose exec zammad-scheduler rake zammad:searchindex:rebuild

[fliebe92]

Hi there! Recently, there was a small change that could solve the reported behavior here: https://github.com/zammad/zammad/commit/6dc6b0fae5089a7a81b1a5f107f972b9ca279247

[mgruner]

Should be fixed by #366