Deploying an OpenStack cloud in a production environment is a great achievement and sometimes overwhelming task in itself. Keeping that cloud up to date though, even within a release cycle or two behind the latest, can sometimes feel impossible.
This presentation will cover the steps that were taken to plan, test, and perform OpenStack upgrades spanning four releases (Mitaka to Queens) during single maintenance windows.
Covering the mistakes made, lessons learned, and tips for operators dreading their next, or even first, major upgrade.
Production OpenStack clouds seem to quickly become out of date and unable to keep pace with the release cycle. Upgrading is often seen as an intimidating, and sometimes impossible, task.
As the deployed release falls further behind, it only seems more daunting to attempt to catch up. Especially with the added pressure of when a release stops receiving security updates.
Deploying a new greenfield cloud and coordinating workloads to move is often seen as a solution, but can be a huge hassle for both the operations team and the application owners, still leave the "old cloud" running, and lead to same situation as the next release, or three, come.
This session will help operators by showing that upgrades are possible, are greatly improved in recent releases, and do not have to be limited to a single release cycle.