Docker, How to replace capistrano tasks in docker

Now is a good time to step back and consider if the benefits of Docker outweigh the added complexity, for your situation. Assuming it is, here are a few suggestions on how to make these components work together better.

While Ansible is a configuration management system, it's also designed for orchestration (that is, running commands across a series of remote machines). This has some cross-over with Capistrano, and as such, you may find it useful to port your Capistrano tasks to Ansible and eliminate a tool (and thus complexity) from your stack. This would likely come about from creating a deploy.yaml playbook that you run to deploy your application.

Docker also overlaps responsibilities with Ansible, but in a different area, configuration. Any part of your system configuration that's necessary for the app can be configured inside the container using the Dockerfile, rather than on a system-wide level using Ansible.

If you have rake tasks that set up the application environment, you can put them in a RUN command in the Dockerfile. Keep in mind, however, that these will only be executed when you build the image, not when you run it.

Generally speaking, I view it this way: Docker sets up a container that has everything required to run one piece of your app (including a specific checkout of your code). Ansible configures the environment in which you run the containers and manages all the work to update them and put them in the right places.