Local server.pid file prevents docker container from starting back up

[angelocordon]

In some instances, when starting the container back up after stopping, a log error pops up:

web_1  | A server is already running. Check /dungeonbuddies/tmp/pids/server.pid.

Running docker-compose ps to list container status returns:

     Name                      Command               State     Ports  
-------------------------------------------------------------------------
dungeonbuddies_db_1    docker-entrypoint.sh postgres    Up       5432/tcp
dungeonbuddies_web_1   bundle exec rails s -p 300 ...   Exit 1            

A quick and dirty solution is to remove the server.pid file in the above directory:

rm ./tmp/pids/server.pid

Also see: https://stackoverflow.com/questions/35022428/rails-server-is-still-running-in-a-new-opened-docker-container

However, logging this as an issue to keep track of how frequent this is happening and that we should look for ways around it so that restarting a container back up is more predictable.

[stain88]

docker/compose#1393

Some more discussion and workarounds.

[angelocordon]

Yeah, I like this comment here: docker/compose#1393 (comment)

[angelocordon]

#24 fixes this issue with

dungeonbuddies/docker-compose.yml

Lines 19 to 20 in 0de8add

	- type: tmpfs
	target: /dungeonbuddies/tmp/pids/

From https://auth0.com/blog/ruby-on-rails-killer-workflow-with-docker-part-1/

The second volume masks the tmp folder in the host OS so that all other project folders are available for the container except the tmp. When you run docker-compose down, this tmp folder within the container is wiped out.

[angelocordon]

Marking as closed in favor of #24