I've just tried deploying a Cassandra cluster and with two nodepools: ringnodes-a and ringnodes-b. It is complaining:
statefulset name "cass-demo-ringnodes-a" did not contain cluster name "demo"
We shouldn't be relying on parsing the name of the statefulset in order to determine anything. We should use labels/annotations instead. This locks us to a particular naming format forever, and if ever we look at 'rolling' statefulsets (like how the Deployment controller does with ReplicaSets), this will become a bit of a nightmare to manage 😄
I've just tried deploying a Cassandra cluster and with two nodepools:
ringnodes-a
andringnodes-b
. It is complaining:We shouldn't be relying on parsing the name of the statefulset in order to determine anything. We should use labels/annotations instead. This locks us to a particular naming format forever, and if ever we look at 'rolling' statefulsets (like how the Deployment controller does with ReplicaSets), this will become a bit of a nightmare to manage 😄
/kind bug /area cassandra