OVH becomes an OpenStack "Infrastructure Donor"

OpenStack has recently and formally announced that OVH is now contributing to the open source project as a cloud resource provider for its continuous integration system. This support strengthens the ties between the OpenStack community and OVH, which operates large-scale solutions and has been one of the members of the project’s Foundation since 2014.

The continuous integration system: a very important tool for OpenStack

OpenStack is currently one of the most important open source projects, both in terms of the level of their contribution and what they intend to achieve: to provide a set of tools for the deployment and operation of cloud infrastructures. Each day, hundreds of patches are returned to the community by developers from all around the world. Patches must undergo numerous tests to verify syntax, integration with the existing code and performance.

OpenStack has put in place a continuous integration system (CI) which automates compilation as well as unit and functional testing to identify bugs and risks of regression early on. Their aim is to move the project forward as quickly as possible. This tool, a “project gating” system based on Zuul, is essential for the community. It is indispensable for good collaboration between the thousands of contributors who receive patches for review and validation after they have successfully passed the tests. The community are so convinced by the efficiency of this CI system that it has extended its use to all project components: documentation, technical governance and more broadly all the important decisions that need to be discussed.

Zuul Status ( let's you view jobs launched by the continuous integration system in real time, in order to automatically test patches that have been returned to the community.

30 000 job launched each day and increasingly growing needs

Within the OpenStack Foundation, the “Infrastructure” team is in charge of the tools available to the community. A large part of this team’s work is, as you probably understand, related to the CI system. As explained by James Blair (Principal Infrastructure Software Developer) in a post on the OpenStack blog, the needs are exponential. 30,000 jobs being launched each day and three cloud providers are contributing to these mass calculations – among them is the OVH Group.

OpenStack over OpenStack

The OVH Group’s teams have been working for several weeks in collaboration with the “Infrastructure” team and the Foundation’s technical board to expand computing capacity within OVH datacenters. The jobs sent to OVH are executed on Public Cloud instances, which are based on an OpenStack platform. Therefore, future OpenStack code is tested…on OpenStack instances!

For each job, a new Public Cloud instance is created. It gets the job, then executes it. Once the job has terminated, the results are returned to the patch manager and the cloud instance is destroyed. This use is a typical case for the Public Cloud, specifically designed to meet the many requests in "on demand" mode.