We would like to try the Kafka Replicator component on a Kubernetes-based Confluent Platform.
Our Kubernetes cluster’s domain name is not default one, i.e. cluster.local, but let’s say <something_else>.local.
As a result, when deploying the Confluent Platform, the Zookeeper servers fail to run, each one trying to connect to the service of each other Zookeeper server in the <namespace>.svc.cluster.local domain instead of the <namespace>.svc.<something_else>.local one.
I unsuccessfully tried to override the default domain name when deploying the Confluent Operator using helm with the --set k8sClusterDomain=<something_else>.local option.
It looks like the user “anastazoch” is encountering an issue with deploying the Confluent Platform on a Kubernetes cluster with a non-default domain name. The problem specifically involves Zookeeper servers failing to run because they are trying to connect to the service of each other in the default domain instead of the specified one.
To address this issue, you can try the following steps:
Verify Helm Override: Double-check the Helm override command used to deploy the Confluent Operator. Ensure that the --set k8sClusterDomain=.local option is correctly applied. Check for any typos or syntax errors in the command.
Operator Configuration: Check if there are additional configuration options within the Confluent Operator Helm chart that allow you to specify the Kubernetes cluster domain. Refer to the Confluent documentation or Helm chart values for any relevant configuration parameters.
Zookeeper Configuration: If the above steps don’t resolve the issue, investigate if there are specific Zookeeper configuration options that allow you to set the cluster domain. It’s possible that there may be Zookeeper-specific configurations in the Confluent Platform deployment.