Mapping MapR Data Science Refinery Ports for External Access

To access MapR Data Science Refinery through the internet, when it is running as a Kubernetes service, you must map the ports using nodePort. This topic provides a sample that shows you how to do this.

Assume you have the following configMap settings for ports:

apiVersion: v1
  kind: ConfigMap
  metadata:
    name: dsr-confs
    namespace: dsr-ns
  data:
    ZEPPELIN_SSL_PORT: "9995"
    LIVY_RSC_PORT_RANGE: "30000~30009"
    SPARK_PORT_RANGE: "30010~30019"

The following defines mappings for these ports:

apiVersion: v1
kind: Service
metadata:
  name: dsr-svc
  namespace: dsr-ns
spec:
  type: LoadBalancer
  ports:
    - { name: dsr-ui, port: 9995, nodePort: 32000 }

  selector:
    app: dsr-app

---
apiVersion: v1
kind: Service
metadata:
  name: dsr-svc-ports
  namespace: dsr-ns
spec:
  type: NodePort
  selector:
    app: dsr-app
  ports:
   
    - { name: dsr-livy-0, port: 30000, nodePort: 30000 }
    - { name: dsr-livy-1, port: 30001, nodePort: 30001 }
    - { name: dsr-livy-2, port: 30002, nodePort: 30002 }
    - { name: dsr-livy-3, port: 30003, nodePort: 30003 }
    - { name: dsr-livy-4, port: 30004, nodePort: 30004 }
    - { name: dsr-livy-5, port: 30005, nodePort: 30005 }
    - { name: dsr-livy-6, port: 30006, nodePort: 30006 }
    - { name: dsr-livy-7, port: 30007, nodePort: 30007 }
    - { name: dsr-livy-8, port: 30008, nodePort: 30008 }
    - { name: dsr-livy-9, port: 30009, nodePort: 30009 }

    - { name: dsr-spark-0, port: 30010, nodePort: 30010 }
    - { name: dsr-spark-1, port: 30011, nodePort: 30011 }
    - { name: dsr-spark-2, port: 30012, nodePort: 30012 }
    - { name: dsr-spark-3, port: 30013, nodePort: 30013 }
    - { name: dsr-spark-4, port: 30014, nodePort: 30014 }
    - { name: dsr-spark-5, port: 30015, nodePort: 30015 }
    - { name: dsr-spark-6, port: 30016, nodePort: 30016 }
    - { name: dsr-spark-7, port: 30017, nodePort: 30017 }
    - { name: dsr-spark-8, port: 30018, nodePort: 30018 }
    - { name: dsr-spark-9, port: 30019, nodePort: 30019 }

The example exposes the Livy and Spark ports through NodePort, using the same port numbers inside and outside the container.

It exposes the MapR Data Science Refinery UI port through a LoadBalancer service. To access the UI, use the following URL with port 32000, as defined in the mapping:

https://<k8s-worker-ip>:32000

<k8s-worker-ip> is the public IP address of any Kubernetes worker node. The worker node proxies the port into the MapR Data Science Refinery service.