Once you have multiple clusters running in the same namespace, you can stop and start the databases independently, the
same way you would a single database, by setting the property cluster.clusterSpec.replicas to zero to
stop the database and the correct number of nodes to restart it. For example, the following command stops the
rome cluster without affecting the operator or other clusters in the namespace:
$ helm upgrade rome voltdb/voltdb \ --reuse-values \ --set cluster.clusterSpec.replicas=0
If you want to shutdown and remove the clusters and operator entirely, you must first shutdown and delete the clusters, then delete the operator. The key point is that you cannot delete the Helm release for the operator until all of the releases it manages have been removed. Therefore, the process is:
Shutdown and delete the individual clusters
$ helm upgrade rome voltdb/voltdb --reuse-values \ --set cluster.clusterSpec.replicas=0 $ helm upgrade venice voltdb/voltdb --reuse-values \ --set cluster.clusterSpec.replicas=0 $ helm delete rome $ helm delete venice
Delete the operator
$ # Make sure all pods have been deleted $ kubectl get pods $ # Once all pods are gone, remove the Operator $ helm delete voltoperator
Documentation