Most of the commonly-deployed OpenStack projects have been using a six-month release cycle for the majority of their existence. In the early days of OpenStack’s lifetime in particular, a relatively short release cycle made a lot of sense due to the rapid pace of feature addition and new projects joining the fold. Six years in, OpenStack is now a much more robust, mature platform that has attracted a much more diverse user base. As the user base has expanded, it’s become apparent that not all operators wish to upgrade their clouds every six months.
Is it really possible to skip upstream releases? Under what circumstances might you consider doing so, and what issues can you expect to encounter? How can you orchestrate a skip-release upgrade?
In this talk, we’ll dive into our experiences with skipping upstream releases on production clouds.
* Our experiences with demand for skip-release upgrades: who’s doing it and why they want it.
* Tradeoffs and gotchas: mechanically, skip-release upgrades introduce some special scenarios operators will need to handle.
* How we do it: an upgrade pattern for doing skip-release deployments and what components we use. We’ll also have a demo of a skip-release upgrade to illustrate what’s happening at various points in the process.
* Other patterns: we’ve often found that the same organizations that want skip-release upgrades also have some adjacent concerns that can be addressed at the same time.
* Tales from the trenches: we’ll discuss what’s happened when we performed skip-release upgrades on production clouds, and lessons we’ve learned in the process.
Registration Opening Soon
Registration will open for you to purchase passes for the OpenStack Summit very soon!
We're excited to see you there!
