Docker compose with shared mounts

Because it makes life happy.

Prerequisites

You'll need the following flux capacitors installed:

• docker-toolbox
• Vagrant
• Virtualbox (or if you have it, vmware is cool too and way faster)
• On Windows, use the git bash shell.

Starting it up

1. Validate that vagrant, docker-machine, and docker-compose run the right commands from your command line.
2. Fork and clone this repo; navigate to it in your shell.
3. Customize docker-compose.yml as desired. The default setup is a super-simple Wordpress theme editing setup.
4. Run vagrant plugin install vagrant vagrant-host-shell
5. Run vagrant up
6. Load up the Docker environment and start your services: eval $(docker-machine env vagrant); docker-compose up
7. Once the machines are up, just go to 172.20.254.2 in your web browser.

You can use the same Vagrant environment for multiple docker compositions; just set up your environmental variables with eval $(docker-machine env vagrant). The only limitation here is that you'll want to make sure your shared ports don't conflict.

Tearing it down

1. Remove the docker-machine configuration: docker-machine rm vagrant
2. Destroy the vagrant box: vagrant destroy -f

Under the hood

Terminology

The outmost environment (Windows, OS X, etc) is what I call the VM host. On it, we run a vagrant guest. The vagrant guest runs Docker so it's also the docker host. The docker host runs a bunch of containers with are... containers.

Notify events in the guest/docker

Things like grunt watch/etc will still need to be run on the VM host. This is because filesystem event notifications aren't passed through the Virtualbox or VMware sharing mechanisms.

Changing shared paths

If you want to change the paths shared with the guest, it's a two-step process:

1. Update your Vagrantfile to share the desired VM host paths with the docker host.
2. Update docker-compose.yml to mount the desired paths on the docker host to the docker container.

Changing IPs

Make sure you change IPs in both places in the Vagrantfile if you need to do so.

Multiple providers caveat

There's a cheap hack to find the SSH key during provisioning. If you have used multiple vagrant providers, completely remove the .vagrant directory from the cloned repository before bringing a new machine up.