`WORKER TIMEOUT` on Docker 20.x

1. Starting with the sample docker-compose files (either sqlite or postgres), set any SECRET_KEY and set ALLOWED_HOSTS=* (the latter is only required for gitpod, see below, but left here for consistency).
2. docker-compose up
3. Point the browser at the babybuddy instance, at the login page (or any other page) hit refresh 3-4 times. Most of the time it’ll be sub-second latency, but occasionally it’ll take about 30 seconds.
4. When this occurs, the logs will (after 30 seconds) say the following (with the numbers 13,14 incrementing each time)

   [1] [CRITICAL] WORKER TIMEOUT (pid:13)
   [13] [INFO] Worker exiting (pid: 13)
   [14] [INFO] Booting worker with pid: 14
   

Setting DEBUG=True and setting log-level to debug adds no information. This also occurs with https://github.com/linuxserver/docker-babybuddy.

Tested on (❌ means reproduces, ✔️ means doesn’t-reproduce and works fine)

• ❌ Ubuntu 20.04.2 on WSL, Docker 20.10.6
• ❌ Ubuntu 20.04.2 on bare-metal, Docker 20.10.6
• ✔️ Ubuntu 20.04.2 on https://gitpod.io, Docker 19.03.15

This has been discussed in https://gitter.im/babybuddy/Lobby?at=60d4adb16c992105fd994cee, and I’m formatting it as an issue at the request of @cdubz.

I’ve performed some more investigation, I’ll add in a comment below.

[cdubz]

Yay! @lutzky would you be willing to open a PR to add the changes to the example compose files? I think that’s all that is needed here…

[cdubz]

@lutzky interesting… that leads to some other potentially relevant information:

• Configuring Gunicorn for Docker
• GIScience/openpoiservice/issues/30