Hello all,

I've been trying to deploy the openproject 12 docker compose version on our docker server, but I cannot get the web_1 and seeder_1 containers to get active.
I have changed the default port from 8080 since we have a previous container on that port, also not using the '.env' file. All the rest are the ones cloned from the git repo.

The containers start and we can get up to the login page, but we keep having HTTP 503 so frequently that we can't even change the password. The web server keeps restarting.

If I deploy using the single line setup all works ok.

So, where to start debugging this? the LOGS don't seem to show much about the issue, only that PUMA keeps shutting down.

Thanks for the help.

JG

=== raw of web_1 ===

2023-04-25T16:59:46.024691420Z I, [2023-04-25T16:59:46.024575 #55]  INFO -- : duration=6.93 db=1.68 view=5.25 status=404 method=GET path=/api/v3/users/1/avatar params={} host=oproject.dev user=4

2023-04-25T16:59:50.310153494Z [1] === puma shutdown: 2023-04-25 16:59:50 +0000 ===

2023-04-25T16:59:50.310173772Z [1] - Goodbye!

2023-04-25T16:59:50.310177346Z [1] - Gracefully shutting down workers...

2023-04-25T16:59:50.511014471Z Exiting

2023-04-25T16:59:51.148276140Z -----> Setting PGVERSION=13 PGBIN=/usr/lib/postgresql/13/bin PGCONF_FILE=/etc/postgresql/13/main/postgresql.conf

2023-04-25T16:59:58.069082535Z => Booting Puma

2023-04-25T16:59:58.069109250Z => Rails 7.0.4.2 application starting in production

2023-04-25T16:59:58.069113190Z => Run `bin/rails server --help` for more startup options

2023-04-25T17:00:09.653815373Z I, [2023-04-25T17:00:08.481634 #1]  INFO -- : Using schema cache file /app/db/schema_cache.yml

2023-04-25T17:00:09.653855773Z [1] Puma starting in cluster mode...

2023-04-25T17:00:09.653859445Z [1] * Puma version: 6.1.1 (ruby 3.2.1-p31) ("The Way Up")

2023-04-25T17:00:09.653862974Z [1] *  Min threads: 4

2023-04-25T17:00:09.653882042Z [1] *  Max threads: 16

2023-04-25T17:00:09.653892619Z [1] *  Environment: production

2023-04-25T17:00:09.653895444Z [1] *   Master PID: 1

2023-04-25T17:00:09.653907104Z [1] *      Workers: 2

2023-04-25T17:00:09.653910182Z [1] *     Restarts: (✔) hot (✖) phased

2023-04-25T17:00:09.653913634Z [1] * Preloading application

2023-04-25T17:00:09.653958049Z [1] * Listening on http://0.0.0.0:8080

2023-04-25T17:00:09.654092550Z [1] Use Ctrl-C to stop

2023-04-25T17:00:09.669086134Z [1] - Worker 0 (PID: 50) booted in 0.01s, phase: 0

2023-04-25T17:00:09.669830084Z [1] - Worker 1 (PID: 56) booted in 0.0s, phase: 0

The issue was related to the fact that I was modifying the value for:

      test: ["CMD", "curl", "-f", "http://localhost:8080${OPENPROJECT_RAILS__RELATIVE__URL__ROOT:-}/health_checks/default"]

With the UR for the OPJ URL. That seems to cause de web_1 container to shutdown.

Hope this can help others in future.