This is a docker container that I use for developing fig. It's probably the most complicated docker container that I use, since it actually requires running docker itself.

The container provides the things needed to develop fig. In essence, that means Python 2.x (2.7) and Docker. Also included are setuptools and pip.

Run bash and share fig using a volume.

docker run -ti -v /path/to/fig:/fig d11wtq/fig-dev bash

When the container starts, docker should be running (try docker info) and you should be in bash. I don't include fig in the image itself, so that needs mounting as a volume. This is simply because I want to edit the source and manage git on the host itself.

You should be able to run the tests:

cd /fig
python setup.py test

The work is based on dind by @jpetazzo, though it has been modified to use a system init script to manage the docker daemon, which is triggered on login through the ~/.bashrc. It is very important that docker is stopped before the container exits too, otherwise eventually the host will run out of loopback devices that cannot be easily reclaimed. This is also managed through an exit trap in ~/.bashrc.

Rather than use aufs (which would be faster), this container uses devicemapper so that there is no need to mount a volume to /var/lib/docker, which uses disk space on the host machine even after the container has exited. Devicemapper depends on the ability to create loopback devices mounted to /var/lib/docker in the container, however, so we have to do some additional work to ensure these devices are unmounted when the container exits.

The init script also prepares the necessary cgroup mounts if they do not yet exist.

See etc/init.d/docker and .bashrc if you're interested in the code.