hub has spec.strategy.type: Recreate
This causes the hub pod to be deleted then rebuilt, causing a, brief, downtime whenever PAWS is updated.
This would be better as a rolling update, potentially some other strategy.
Description
Description
Event Timeline
Comment Actions
the hub pod has some in-memory state, and (worse) some shared state with the proxy pod. https://github.com/jupyterhub/zero-to-jupyterhub-k8s/issues/1364 is main upstream issue on this, I think.
Comment Actions
Good to know. Sounds like hub goes and finds a (and only one) proxy pod, then keeps its IP as state. As such it never uses more than one proxy even if there is more than one to be had. Wouldn't this cause more problems on the proxy being refreshed, and hub remained the same (now attempting to talk to a deleted proxy)?