I setting up 2 schema registry on different k8s cluster and using nodeport service for communication between 2 clusters. I followed steps here for multi-datacenter schema registry: https://docs.confluent.io/platform/6.0.0/schema-registry/multidc.html.

When the secondary schema registry (with leader.eligibility: false) trying to forward new schema registration to primary schema registry, it use internal IP of primary schema registry which end up failing. How do I able to tell my secondary schema registry to use the accessible hostname and port number to connect primary schema registry?

I using image confluentinc/cp-schema-registry:6.0.0.

Primary schema registry configuration:

host.name=10.X.X.X (internal IP of schema registry pod)
kafkastore.group.id=my-schema-registry
kafkastore.bootstrap.servers=PLAINTEXT://my-cluster-kafka-bootstrap:9092 (Primary Kafka Broker Internal Hostname)
avro.compatibility.level=NONE
schema.compatibility.level=NONE

Secondary schema registry configuration:

host.name=10.X.X.X (internal IP of schema registry pod)
kafkastore.group.id=my-schema-registry
kafkastore.bootstrap.servers=PLAINTEXT://10.X.X.X:30292 (Primary Kafka Broker IP)
avro.compatibility.level=NONE
schema.compatibility.level=NONE
leader.eligibility=false

Error throw in schema registry logs when forwarding the registration request.

ERROR Failed to send HTTP request to endpoint: http://10.X.X.X:8081/subjects/XXX/versions (io.confluent.kafka.schemaregistry.client.rest.RestService)
java.net.SocketTimeoutException: connect timed out
   at ... 

Note: internal port 8081 is not accessible from another cluster, so I had expose different port number using NodePort Service for both schema registry on different cluster. This is the reason I want override the port number accessing to master from secondary schema registry.