Docker integration

Most of the time your application requires external services. Compas supports automatically spinning up those external services via Docker.

Getting started

The docker integration is not enabled by default. You can run the following command or apply the config changes below manually.

compas init docker

This creates or updates the Compas config file in config/compas.json with the following contents:

{
	"dockerContainers": {
		"compas-postgres-15": {
			"image": "postgres:15",
			"createArguments": "-e POSTGRES_USER=postgres -e POSTGRES_PASSWORD=postgres -e PGDATA=/var/lib/postgresql/data/pgdata -v compas-postgres-15:/var/lib/postgresql/data/pgdata -p 5432:5432",
			"runArguments": ""
		}
	}
}

Config

Each container definition has the following properties:

• The image name, in the above example 'compas-postgres-15'. This name should either be unique over all your projects, or the container is reused across projects that specify the same name. This may be the desired behavior if you work on multiple projects that share the same set of services.
• createArguments: Arguments to pass to the docker create command. See the Docker documentation for more information. The --name argument and image are provided by Compas.
• image: The image to create the container from. Anything that Docker supports works. It is advised to develop against the same versions as your production environment will have.

Limitations

• Compas only evaluates the state of the Docker containers on startup and on configuration changes.
• Compas automatically stops all containers that are not required for the current project. However, once the development environment exits, the containers will be kept running.
• Compas assumes that containers with the same name are compatible across projects. Make sure to use unique names when that is not the case.
• Compas does not detect changes to the image, createArguments or runArguments. Manually remove the container with docker container rm $name to let Compas recreate the container.