Details
Description
It is possible to trigger a full cluster "Rolling Upgrade" to move the Couchbase Server Pods to a new set of Kubernetes Worker Nodes if needed, but this is currently done by making an arbitrary change somewhere in the CouchbaseCluster definition, to make CAO think that all the Pods need to be recreated.
We should support a more explicit method of doing this, i.e. by using per-Pod Annotations (for example cao.couchbase.com/reschedule).