Announcing a new cloud provider for OpenStack’s CI system: OVH

The content below is taken from the original (Announcing a new cloud provider for OpenStack’s CI system: OVH), to continue reading please visit the site. Remember to respect the Author & Copyright.

The OpenStack Project Infrastructure Team has added OpenStack Compute instances from OVH to the project’s continuous integration system.

There are a lot of ways to participate in an Open Source community like the OpenStack project and OVH is doing so in a way that best matches their strengths: contributing cloud resources to the project to drive our test and automation systems — resources which are critical to ongoing development.

OpenStack runs up to 30,000 automated jobs each day to test proposed changes and build documentation and software artifacts. The system responsible for this is itself a large multi-cloud application. It utilizes OpenStack Compute instances from Rackspace, HP Cloud, and now OVH to provide this service to developers. Each of those instances runs a single job and is then deleted in order to ensure the next job starts with a clean slate. This puts an extraordinarily high demand on our providers.

We are very excited by the addition of OVH, not only because it will help us scale to meet the needs of a growing project, but also because it demonstrates one of OpenStack’s key features: cross-cloud compatibility. When a developer uploads a proposed change to an OpenStack project, available instances from any of our contributing cloud providers will be used interchangeably to test it.

We are able to do this with the help of a number of OpenStack projects. First, we create a base image to use on all of our providers using Diskimage-builder. This ensures that all of our tests operate in an identical environment. Our test resource manager, Nodepool, uses the shade library to upload each of these images to our providers using the OpenStack Image service. Each of our providers is configured slightly differently, but shade handles the details so that Nodepool can focus on the business logic. Nodepool then identifies demand from our automation system, Zuul, to create OpenStack Compute instances as needed.

The Infrastructure Team thanks OVH for their contribution as well as our existing cloud providers, and we are excited about sharing our experience using this very demanding OpenStack application with all of them.