You are viewing documentation for a previous version of Anthos GKE on-prem. View the latest documentation.

Creating a Service and an Ingress

This page shows how to create a Kubernetes Ingress object in an GKE on-prem user cluster. An Ingress object must be associated with one or more Service objects, each of which is associated with a set of Pods.

SSH into your admin workstation

SSH into your admin workstation:

ssh -i ~/.ssh/vsphere_workstation ubuntu@[IP_ADDRESS]

where [IP_ADDRESS] is the IP address of your admin workstation.

Do all of the remaining steps in this topic on your admin workstation.

Creating a Deployment

Here's a manifest for a Deployment:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: hello-deployment
spec:
  selector:
    matchLabels:
      greeting: hello
  replicas: 3
  template:
    metadata:
      labels:
        greeting: hello
    spec:
      containers:
      - name: hello-world
        image: "gcr.io/google-samples/hello-app:2.0"
        env:
        - name: "PORT"
          value: "50000"
      - name: hello-kubernetes
        image: "gcr.io/google-samples/node-hello:1.0"
        env:
        - name: "PORT"
          value: "8080"

For the purpose of this exercise, these are the important points to understand about the Deployment manifest:

  • Each Pod that belongs to the Deployment has the greeting: hello label.

  • Each Pod has two containers.

  • The env fields specify that the hello-app containers listen on TCP port 50000, and the node-hello containers listen on TCP port 8080. For hello-app, you can see the effect of the PORT environment variable by looking at the source code.

Copy the manifest to a file named hello-deployment.yaml, and create the Deployment:

kubectl apply --kubeconfig [USER_CLUSTER_KUBECONFIG] -f hello-deployment.yaml

where [USER_CLUSTER_KUBECONFIG] is the kubeconfig file for your user cluster.

Exposing your Deployment with a Service

To provide a stable way for clients to send requests to the Pods of your Deployment, create a Service.

Here's a manifest for a Service that exposes your Deployment to clients inside and outside of your cluster:

apiVersion: v1
kind: Service
metadata:
  name: hello-service
spec:
  type: NodePort
  selector:
    greeting: hello
  ports:
  - name: world-port
    protocol: TCP
    port: 60000
    targetPort: 50000
  - name: kubernetes-port
    protocol: TCP
    port: 60001
    targetPort: 8080

Copy the manifest to a file named hello-service.yaml, and create the Service:

kubectl apply --kubeconfig [USER_CLUSTER_KUBECONFIG] -f hello-service.yaml

View the Service:

kubectl --kubeconfig [USER_CLUSTER_KUBECONFIG] get service hello-service --output yaml

The output shows the nodePort values that have been assigned to the Service:

apiVersion: v1
kind: Service
metadata:
  ...
  name: hello-service
  namespace: default
  ...
spec:
  clusterIP: 10.105.252.237
  externalTrafficPolicy: Cluster
  ports:
  - name: world-port
    nodePort: 31807
    port: 60000
    protocol: TCP
    targetPort: 50000
  - name: kubernetes-port
    nodePort: 30734
    port: 60001
    protocol: TCP
    targetPort: 8080
  selector:
    greeting: hello
  sessionAffinity: None
  type: NodePort
status:
  loadBalancer:
    ingress:
    - {}

In the preceding output, the ports field is an array of ServicePort objects: one named world-port and one named kubernetes-port.

These are the ways a client can call the Service using world-port:

  • A client running on one of the cluster nodes can send a request to the clusterIP on port. In this example, 10.105.252.237:60000. The request is forwarded to a member Pod on targetPort. In this example, [POD_IP_ADDRESS]:50000.

  • A client can send a request to the IP address of any cluster node on nodePort. In this example, [NODE_IP_ADDRESS]:31807. The request is forwarded to a member Pod on targetPort. In this example, [POD_IP_ADDRESS]:50000.

These are the ways a client can call the Service using kubernetes-port:

  • A client running on one of the cluster nodes can send a request to the clusterIP on port. In this example, 10.105.252.237:60001. The request is forwarded to a member Pod on targetPort. In this example, [POD_IP_ADDRESS]:8080.

  • A client can send a request to the IP address of any cluster node on nodePort. In this example, [NODE_IP_ADDRESS]:30734. The request is forwarded to a member Pod on targetPort. In this example, [POD_IP_ADDRESS]:8080.

Creating an Ingress

Here's a manifest for an Ingress:

apiVersion: extensions/v1beta1
kind: Ingress
metadata:
  name: my-ingress
spec:
  rules:
  - http:
      paths:
      - path: /greet-the-world
        backend:
          serviceName: hello-service
          servicePort: 60000
      - path: /greet-kubernetes
        backend:
          serviceName: hello-service
          servicePort: 60001

Copy the manifest to a file named my-ingress.yaml, and create the Ingress:

kubectl apply --kubeconfig [USER_CLUSTER_KUBECONFIG] -f my-ingress.yaml

As part of creating your user cluster, you specified a virtual IP address (VIP) for ingress by providing a value for usercluster.vips.ingressvip in your configuration file.

When a client sends a request to your user cluster ingress VIP, the request is routed to your F5 BIG-IP load balancer. The load balancer forwards the request to an ingress Service running in your user cluster. The ingress Service is configured to forward the request to different backends depending on the path in the request URL.

It is important to understand that there are two different Services related to the steps in this topic:

  • Your Service named hello-service. This is a Service that you created to expose the Pods of your hello-deployment Deployment.

  • The ingress Service that runs in the gke-system namespace of your user cluster. This Service is part of your cluster infrastructure.

The /greet-the-world path

In your Ingress manifest, you can see a rule that says the path /greet-the-world is associated with serviceName: hello-service and servicePort: 60000. Recall that 60000 is the port value in the world-port field of your hello-service Service.

- name: world-port
    nodePort: 31807
    port: 60000
    protocol: TCP
    targetPort: 50000

The ingress Service chooses a cluster node and forwards the request to the node on nodePort. In this example, [NODE_IP_ADDRESS]:31807. The iptables rules on the node forward the request to a member Pod on port 50000. The container listening on port 50000 displays a Hello World! message.

The /greet-kubernetes path

In your Ingress manifest, you can see a rule that says the path /greet-kubernetes is associated with serviceName: hello-service and servicePort: 60001. Recall that 60001 is the port value in the kubernetes-port field of your hello-service Service.

- name: kubernetes-port
    nodePort: 30734
    port: 60001
    protocol: TCP
    targetPort: 8080

The ingress Service chooses a cluster node and forwards the request to the node on nodePort. In this example, [NODE_IP_ADDRESS]:30734. The iptables rules on the node forward the request to a member Pod on port 8080. The container listening on port 8080 displays a Hello Kubernetes! message.

Test the Ingress:

Test the Ingress using the /greet-the-world path:

curl [USER_CLUSTER_INGRESS_VIP]/greet-the-world

The output shows a Hello, world! message:

Hello, world!
Version: 2.0.0
Hostname: ...

Test the Ingress using the /greet-kubernetes path:

curl [USER_CLUSTER_INGRESS_VIP]/greet-kubernetes

The output shows a Hello, Kubernetes! message:

Hello Kubernetes!

Cleaning up

Delete your Ingress:

kubectl --kubeconfig [USER_CLUSTER_KUBECONFIG] delete ingress my-ingress

Delete your Service:

kubectl --kubeconfig [USER_CLUSTER_KUBECONFIG] delete service hello-service

Delete your Deployment:

kubectl --kubeconfig [USER_CLUSTER_KUBECONFIG] delete deployment hello-deployment

What's next

Overview of advanced installation topics

Creating dashboards to monitor cluster health