proper way to restart Zookeeper

[rifler]

Hello!

We are using Istio and from time to time we need to restart deployments/sts/daemonsets under Istio contol, for instance while upgrading to new major Istio version. We need it to apply new version of Istio sidecar, which is added via webhook.

I noticed that I can't restart zookeeper just via kubectl rollout statefulset/zookeeper -n my-namespace, because Operator denied it.

I used to add new env variable DO_RESTART: "1" to ZookeeperCluster, but it is a bit awkward, cause in next helm upgrade it will also restart for no reason (because there is no such env variable).

Now I come to conclusion to scale down operator, then kubectl rollout restart zookeeper, then scale up operator back.

Maybe there is easier way to do it?

[rifler]

Now I come to conclusion to scale down operator, then kubectl rollout restart zookeeper, then scale up operator back.

There is also downside with two restarts:

k scale deploy/zookeeper-operator --replicas=0
k rollout restart sts/zookeeper # first restart
k scale deploy/zookeeper-operator --replicas=1
# then second restart