k8s external ip configuration

@jacobtomlinson made a comment here on externalizing services. I would suggest that the default be a NodePort for a two reasons:

• Ingress/LoadBalancers can be tricky to setup/configure – especially for beginners
  • note: on clouds (GKE at least) ingress is automatically setup
• Testing may be challenging (though not impossible)

NodePorts are builtin to k8s, we can easily test, and the setup should work out of the box without additional configuration

When using a NodePort I believe the pod/service is accessible by any nodeIP: https://kubernetes.io/docs/tutorials/services/source-ip/#source-ip-for-services-with-type-nodeport

For configuring a NodePort I would recommend looking at: https://kubernetes.io/docs/tasks/access-application-cluster/service-access-application-cluster/

or the following:

kind: Service 
apiVersion: v1 
metadata:
  name: SERVICENAME
spec:
  type: NodePort
  selector:
    app: APPNAME
  ports:
    # nodePort - a static port assigned on each the node
    # port - port exposed internally in the cluster (assigned on creation)
    # targetPort - the container port to send requests to
    - nodePort: 30163
      port: 8080 

cc @jcrist

@jacobtomlinson do you have thoughts here ?

[jcrist]

Lots of helm charts I see seem to either use ClusterIP or LoadBalancer as a default. I feel like LoadBalancer makes more sense in this case, especially given they tend to “just work” with cloud providers.

I went with LoadBalancer as the default for this reason, but hopefully have things done in a way that users could use NodePort if needed.

[jcrist]

After some conversations with dev-ops people, I’m moving to just having the user specify the scheduler proxy address as well.

This was done in #71.