Docker Introduction by Ryan Houston

Why use Docker?

There has been plenty written elsewhere about Docker in general and the choice of Docker against other tools like Vagrant, Ansible, etc. In short, Docker allows running of containers that provide specific environments and dependencies required to run an application, either in production or in a simple to manage development environment. These containers are generally lightweight, composable, and disposable. This keeps the host system (perhaps your laptop) isolated and clean by not locally installing all kinds of dependencies needed for specific projects or just when trying something new. The dependencies are installed and ran inside the disposable container instead.

Here I’ve made some notes on a quick introduction to Docker by using a container to run Jekyll, which is the tool that generates this site. See more at the Docker docs.

Installing Docker CE

Follow the Ubuntu installation instructions.

Be sure to add your user to docker group to avoid requiring sudo. It is probably safer not to run containers as root unless necessary.

sudo adduser `whoami` docker

Run Jekyll in a Docker container

Simply run:

docker run --rm --label=jekyll --volume=$(pwd):/srv/jekyll \
  -it -p 127.0.0.1:4000:4000 jekyll/jekyll \
  jekyll server --drafts

This will download and run the official Jekyll docker image with all dependencies running at or very close to the same versions as used on Github pages. This means it is only necessary to have Docker installed locally but not Ruby or any other dependencies.

To break this command down a little:

--rm removes the container once the command exits
--label=jekyll gives the container a nice label. I’m not quite sure where this label is used. It seems the --name=jekyll option would be more useful as the name is usable in place of <container-id> in commands like start and stop.
--volume=$(pwd):/srv/jekyll shares the current directory with the Docker container located at /srv/jekyll within the container
-it is shorthand for --interactive and --tty
-p 127.0.0.1:4000:4000 defines a port mapping to expose port 4000 inside the container to port 4000 on the host. This ensures we can access localhost:4000 in a browser and access the jekyll server running inside the container
jekyll/jekyll is the name of the docker image on which the container is based
jekyll server --drafts is the command that is run inside the container to run the jekyll server with drafts published

The --rm option could be left of to preserve the container after jekyll is stopped. You could then restart jekyll without having to wait for the container to rebuild by running

docker start <container-id>

with the container ID that is shown for the jekyll container in docker ps -a. This would run the container but you would not see any output. To see output from a running container, you can run docker attach <container-id>. You can start the container and attach automatically by using the -a option. You could stop the container by running docker stop <container-id>. All pretty simple stuff.

If you did not provide the --rm option, the container can be removed when you are done with it via docker rm <container-id>. Once the image has been downloaded, the container comes up pretty quickly even if the --rm option is used. The main difference in this particular case is that the ruby gems will have to be reinstalled when fresh containers are used.

Docker Compose

The docker run command is kind of long and inconvenient to type every time you want to work on a jekyll site. Docker Compose can be used to simplify this so the single command docker-compose up will create or start the container as defined by a docker-compose.yml file, and run the jekyll server so everything is ready to work on the site.

The docker-compose.yml file for the jekyll container looks something like this:

jekyll:
  image: jekyll/jekyll:pages
  command: jekyll server --drafts
  ports:
    - 4000:4000
  volumes:
    - .:/srv/jekyll