dockerdev

This repository contains a collection of bash functions that make it extremely easy to use Docker containers as local development environments. Using Docker containers this way keeps the environment in which you develop software both deterministic and isolated from the rest of your system. It is a much more powerful form of determinism and isolation than package managers like npm and pipenv, which you can still use inside of the container.

You are meant to copy the bash scripts in this repository into your own software project, write your own scripts that call the helper functions, and use them as part of your workflow.

The functions defined here automate a lot of the tedium of managing Docker containers when using them as local development environments. For example, with a single command you can open a shell inside the container, while ensuring that that the container has been created, that it is running the latest image, and that the internals of the container have been configured correctly.

These scripts are the product of lots of trial and error trying to get a nice workflow with Docker set up. They have no major dependencies other than bash and Docker.

Why is this useful?

I see Docker, and containers in general, as solving two very important but distinct problems:

1. Dev and production environments should match
2. Installing the environment necessary to run the app should be automated, deterministic, and isolated from the host system
   1. in development
   2. more importantly, in production

This repository is meant to facilitate point 2.i.

Features

• Automatically creates a container based on a Docker image you specify, or re-uses a previously-created container with that image to avoid constant container refreshing
• If a container contains an older version of the same image, makes sure to start a new container with the updated image
• Creates a non-root user inside the container matching your user on the host, so that things like node_modules and lock files can be written to the host with the correct permissions using bind mounts
  • Supported for images based on Ubuntu or Alpine
• Keeping node_modules on the host avoids re-installing them every time the container image changes
• Fixes the terminal width and height of the shell inside the container
• Supports callbacks that only run when a new container is started, e.g. for attaching networks to the container
• Also works for containers that live inside of stacks: given a docker-compose.yml file, automatically creates a stack, service, and container if they do not exist
• If the image of the container in the stack is not up to date, it is automatically updated
• If the container in the stack needs to be created or updated, automatically waits until the container is ready to run commands before running a command
• Supports the use of Docker clients within the container by connecting them to the Docker daemon on the host
• Every function is well-documented, and every quirk and workaround is commented

Example Usage

$ cd /path/to/your/projects/git/repo
$ mkdir scripts
$ cd scripts
$ wget https://raw.githubusercontent.com/bdusell/dockerdev/master/dockerdev.bash
$ git add .
$ git commit -m 'Add dockerdev scripts.'
$ vim shell.bash # Write a script to open a shell in the dev environment
$ cd ..
$ bash scripts/shell.bash
> # Now you're in the container

Complete examples are included under examples/.