Nova, Neutron, Cinder, and Ironic are pursuing rolling, minimal-downtime upgrades using oslo.versionedobjects. Keystone and Glance are pursuing rolling, zero-downtime upgrades using database triggers. Swift has had rolling, zero-downtime upgrades forever. Horizon needs to be able to talk to multiple releases of every service it interacts with. Regardless of the implementations, testing real upgrades while measuring and/or asserting against downtime in the gate might be the hardest part. Let's review the state of the union, where we're headed, and what's it going to take from your project for us to deliver the latest version of OpenStack to deployers and end users without all the risk, frustration, and downtime of the past.
References:
* User story
* Upgrades
* Rolling upgrades
* Zero-downtime upgrades
Session Leader(s): Dolph Mathews (dolphm)
Registration Opening Soon
Registration will open for you to purchase passes for the OpenStack Summit very soon!
We're excited to see you there!
