🚨 DEPRECATION WARNING 🚨

As Docker now has compose built in there's no longer need for this Docker image.

You should use the official Docker image instead.

Supported tags and respective Dockerfile links

• latest (Dockerfile)

Note: There are tags for each build date. If you need to "pin" the Docker image version you use, you can select one of those tags. E.g. tiangolo/docker-with-compose:2021-09-17.

Docker with Docker Compose image

Docker image with Docker Compose installed for CI.

Description

The main purpose of this image is to help in Continuous Integration environments that need the docker binary, the docker-compose binary, and possibly require doing other small things, like running shell scripts or notifying some API with curl.

It includes both programs (docker and docker-compose) and allows to run arbitrary shell scripts (contrary to the official Docker Compose image).

By not having to install docker-compose on top of a docker:latest image it can reduce the building time about 10 / 15 seconds in a cloud data center for each build. In environments in where the Internet connection is less good than a cloud provider, the time saved would be more.

GitHub repo: https://github.com/tiangolo/docker-with-compose

Docker Hub image: https://hub.docker.com/r/tiangolo/docker-with-compose/

Usage

Pull the image:

$ docker pull tiangolo/docker-with-compose

Then run a container of this image mounting the Docker sock as a host volume.

By mounting the Docker sock as a volume you allow the docker client inside of the container to communicate with your Docker (the Docker daemon/service) on your machine directly.

This way, you can send Docker commands, like pulling, running, or building a new Docker image, from inside this container.

You might also want to mount a host volume with the files that you need to use.

For example, let's say you have a Dockerfile like:

FROM nginx

RUN echo "Hello World" > /usr/share/nginx/html/index.html

You could:

• Mount the local directory containing that Dockerfile.
• Mount the local Docker sock.
• Build that Nginx image from inside of container running this image.

$ docker run -v $(pwd):/app -v /var/run/docker.sock:/var/run/docker.sock tiangolo/docker-with-compose sh -c "cd /app/ && docker build -t custom-nginx ."

Problem description

There is an official Docker image that contains the docker binary. And there is a Docker Compose image.

But the Docker Compose image has docker-compose as the entrypoint.

So, it's not possible to run other commands on that image, like installing something, e.g. apt-get install -y curl.

And it's also not possible to run docker commands directly, e.g. docker login -u ci-user -p $CI_JOB_TOKEN $CI_REGISTRY.

This image allows running arbitrary commands like shell scripts, docker commands and also Docker Compose commands like docker-compose build and docker-compose push.

As several Continuous Integration systems allow doing previous steps, like installing packages before running the actual main script, those steps could be used to install Docker Compose. But by downloading and installing Docker Compose every time, the builds would be slower.

For example, a very simple GitLab CI file .gitlab-ci.yml could look like:

# Do not use this file example
image: docker:latest

before_script:
  - apk add --no-cache py-pip
  - pip install docker-compose
  - docker login -u gitlab-ci-token -p $CI_JOB_TOKEN $CI_REGISTRY

ci:
  script:
    - docker-compose build
    - docker-compose up -d
    - docker-compose exec -T tests run-tests.sh
    - docker-compose down -v
    - docker stack deploy -c docker-compose.prod.yml --with-registry-auth prod-example-com

But when the base image has to download and install Docker Compose every time, that's time added to the process. Specifically the lines in:

...

  - apk add --no-cache py-pip
  - pip install docker-compose

...

This image's solution

This image includes Docker Compose and allows you to run any other arbitrary command.

So your GitLab CI .gitlab-ci.yml file could then look like:

image: tiangolo/docker-with-compose

before_script:
  - docker login -u gitlab-ci-token -p $CI_JOB_TOKEN $CI_REGISTRY

ci:
  script:
    - docker-compose build
    - docker-compose up -d
    - docker-compose exec -T tests run-tests.sh
    - docker-compose down -v
    - docker stack deploy -c docker-compose.prod.yml --with-registry-auth prod-example-com

And it would run faster as it doesn't have to install Docker Compose every time.

And you could start that initial GitLab Runner following the GitLab CI runner for CI/CD guide on DockerSwarm.rocks.

The same would apply for Travis, Jenkins or whichever CI system you use.

Release Notes

Latest Changes

• 🗑️ Deprecate this project, use official Docker instead 🎉. PR #47 by @tiangolo.
• ⬆️ Bump actions/checkout from 3.1.0 to 3.3.0. PR #44 by @dependabot[bot].
• 👷 Update Latest Changes token. PR #46 by @tiangolo.
• 👷 Add GitHub Action for Docker Hub description. PR #41 by @tiangolo.
• ⬆️ Upgrade CI OS. PR #40 by @tiangolo.
• 🔧 Add funding config. PR #39 by @tiangolo.
• 👷 Add automatic scheduled CI every monday. PR #38 by @tiangolo.
• 👷 Add automatic scheduled CI every Monday. PR #37 by @tiangolo.
• 📝 Update README, replace bash with shell, as Bash itself is not installed. PR #36 by @tiangolo.
• 👷 Add alls-green GitHub Action. PR #35 by @tiangolo.
• 👷 Do not run double CI for PRs, run on push only on master. PR #34 by @tiangolo.
• ⬆️ Bump tiangolo/issue-manager from 0.3.0 to 0.4.0. PR #28 by @dependabot[bot].
• Bump actions/checkout from 2 to 3.1.0. PR #31 by @dependabot[bot].
• 🐛 Fix deployment. PR #26 by @tiangolo.
• 🐛 Fix GitHub Actions and latest requirements. PR #25 by @tiangolo.
• 👷 Move from Travis to GitHub Actions. PR #23 by @tiangolo.
• ✨ Add external dependencies and Dependabot to get automated upgrade PRs. PR #27 by @tiangolo.
• 👷 Add Latest Changes GitHub Action. PR #24 by @tiangolo.
• Upgrade Python to use version 3.x. PR #15.
• Add curl to the installed and available packages. PR #14 by @stratosgear.
• Add Travis CI. PR #4.
• Upgrade Docker Compose installation. PR #3 by @boskiv.

License

This project is licensed under the terms of the MIT license.