Hello all, Final step of the upgrade procedure is to run command like: "./bin/kafka-features.sh upgrade --metadata 3.6"
In the Kubernetes world, which works in the desired state configuration (yamls and its the upper level abstractions), this is quite complicated. The first thing I tried to find is if I can call it from another kafka pod while specifying a server to connect to, however it is not possible. It needs to be run from withing the running kafka pod. This leads to executing to the pod or running the kubectl exec (e.g. kubectl exec kafka-0 -n cloud -- bin/kafka-configs.sh ... ) However, this command is also imperative instead of declarative. Question: Is there another approach? E.g. driving the metadata version via ENV variable or file configuration? It seems like it was designed without Kubernetes world in mind.
