# Podman Compose ## [![Tests](https://github.com/containers/podman-compose/actions/workflows/test.yml/badge.svg)](https://github.com/containers/podman-compose/actions/workflows/test.yml) An implementation of [Compose Spec](https://compose-spec.io/) with [Podman](https://podman.io/) backend. This project focuses on: * rootless * daemon-less process model, we directly execute podman, no running daemon. This project only depends on: * `podman` * [podman dnsname plugin](https://github.com/containers/dnsname): It is usually found in the `podman-plugins` or `podman-dnsname` distro packages, those packages are not pulled by default and you need to install them. This allows containers to be able to resolve each other if they are on the same CNI network. This is not necessary when podman is using netavark as a network backend. * Python3 * [PyYAML](https://pyyaml.org/) * [python-dotenv](https://pypi.org/project/python-dotenv/) And it's formed as a single Python file script that you can drop into your PATH and run. ## References: * [spec.md](https://github.com/compose-spec/compose-spec/blob/master/spec.md) * [docker-compose compose-file-v3](https://docs.docker.com/compose/compose-file/compose-file-v3/) * [docker-compose compose-file-v2](https://docs.docker.com/compose/compose-file/compose-file-v2/) ## Alternatives As in [this article](https://fedoramagazine.org/use-docker-compose-with-podman-to-orchestrate-containers-on-fedora/) you can setup a `podman.socket` and use unmodified `docker-compose` that talks to that socket but in this case you lose the process-model (ex. `docker-compose build` will send a possibly large context tarball to the daemon) For production-like single-machine containerized environment consider - [k3s](https://k3s.io) | [k3s github](https://github.com/rancher/k3s) - [MiniKube](https://minikube.sigs.k8s.io/) For the real thing (multi-node clusters) check any production OpenShift/Kubernetes distribution like [OKD](https://www.okd.io/). ## Versions If you have legacy version of `podman` (before 3.1.0) you might need to stick with legacy `podman-compose` `0.1.x` branch. The legacy branch 0.1.x uses mappings and workarounds to compensate for rootless limitations. Modern podman versions (>=3.4) do not have those limitations, and thus you can use latest and stable 1.x branch. If you are upgrading from `podman-compose` version `0.1.x` then we no longer have global option `-t` to set mapping type like `hostnet`. If you desire that behavior, pass it the standard way like `network_mode: host` in the YAML. ## Installation ### Pip Install the latest stable version from PyPI: ```bash pip3 install podman-compose ``` pass `--user` to install inside regular user home without being root. Or latest development version from GitHub: ```bash pip3 install https://github.com/containers/podman-compose/archive/main.tar.gz ``` ### Homebrew ```bash brew install podman-compose ``` ### Manual ```bash curl -o /usr/local/bin/podman-compose https://raw.githubusercontent.com/containers/podman-compose/main/podman_compose.py chmod +x /usr/local/bin/podman-compose ``` or inside your home ```bash curl -o ~/.local/bin/podman-compose https://raw.githubusercontent.com/containers/podman-compose/main/podman_compose.py chmod +x ~/.local/bin/podman-compose ``` or install from Fedora (starting from f31) repositories: ```bash sudo dnf install podman-compose ``` ## Basic Usage We have included fully functional sample stacks inside `examples/` directory. You can get more examples from [awesome-compose](https://github.com/docker/awesome-compose). A quick example would be ```bash cd examples/busybox podman-compose --help podman-compose up --help podman-compose up ``` A more rich example can be found in [examples/awx3](examples/awx3) which have - A Postgres Database - RabbitMQ server - MemCached server - a django web server - a django tasks When testing the `AWX3` example, if you got errors, just wait for db migrations to end. There is also AWX 17.1.0 ## Tests Inside `tests/` directory we have many useless docker-compose stacks that are meant to test as many cases as we can to make sure we are compatible ### Unit tests with unittest run a unittest with following command ```shell python -m unittest pytests/*.py ``` # Contributing guide If you are a user or a developer and want to contribute please check the [CONTRIBUTING](CONTRIBUTING.md) section