sous

Sous is a tool for building, testing, and deploying applications, using Docker, Mesos, and Singularity.

View documentation in the doc/ directory. Specifically, see the getting started file.

Planned features

Multi-datacentre deployment orchestration (coming very soon)
Declarative YAML-based DSL to define deployments (coming very soon)
Safely deploy source code to production Global event log HTTP API to interrogate, and instigate changes to global state Run projects locally in a simulated production environment
Runs on Mac and Linux (Windows support not currently planned) - Use the same tool for local development and in your CI pipeline
Easily distribute shared configuration using the built-in sous server
Automatically adds rich metadata to your Docker images - Run executable contracts against any Docker image, to ensure it behaves appropriately for your platform.
Define platform contracts in terms of application interactions
Automatically build NodeJS and Go code using a multi-stage build process that eliminates build-time dependencies from your production containers. (Java, C#, Ruby, and other languages coming soon.)

Ethos

Sous is designed to work with existing projects, using data they already contain to determine how to properly build Docker images. It is designed to ease migrating existing projects onto The Mesos Platform, using sensible defaults and stack-centric conventions. It is also designed with operations in mind, tagging Docker images with lots of metadata to ease discovery and clean-up of images.

Sous works on your local dev machine, and on CI servers like TeamCity, Jenkins, etc., in the same way, so you can be sure whatever works locally will also work in CI.

Installation

Sous is written in Go. If you already have Go 1.6 set up on your machine, and have your GOPATH set up correctly, you can install it by typing

$ go get -u -v github.com/opentable/sous

(for development guides, see here

Requirements

Sous shells out to your system to interact with Git and Docker. This is a design decision, as it enables you to easily repeat the commands Sous issues. That means that when they fail, as they sometimes do, you have the power to re-play what happened, and figure out the issue.

You will need:

Git >=2.2
Docker >=1.10

On Mac, we recommend installing Docker by installing docker-machine via the Docker Toolbox available at https://www.docker.com/toolbox

LICENSE