Create a Service and an Ingress

Version 1.13 

• Latest
• Version 1.16
• Earlier versions (unsupported)

This document shows how to create a Kubernetes Ingress object in a user, hybrid, or standalone cluster for Google Distributed Cloud. An Ingress is associated with one or more Services, each of which is associated with a set of Pods.

Before you begin

Get an SSH connection from your cluster to your admin workstation.

Create 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 CLUSTER_KUBECONFIG -f hello-deployment.yaml

Replace CLUSTER_KUBECONFIG with the name of the kubeconfig file for your cluster.

Expose 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 your cluster:

apiVersion: v1
kind: Service
metadata:
  name: hello-service
spec:
  type: ClusterIP
  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 CLUSTER_KUBECONFIG -f hello-service.yaml

Replace CLUSTER_KUBECONFIG with the name of the kubeconfig file for your cluster.

View the Service:

kubectl --kubeconfig CLUSTER_KUBECONFIG get service hello-service --output yaml

The output shows the value of clusterIP that has been given to the Service. For example:

apiVersion: v1
kind: Service
metadata:
  annotations:
    ...
spec:
  clusterIP: 10.96.14.249
  clusterIPs:
  - 10.96.14.249
  ipFamilies:
  - IPv4
  ipFamilyPolicy: SingleStack
  ports:
  - name: world-port
    port: 60000
    protocol: TCP
    targetPort: 50000
  - name: kubernetes-port
    port: 60001
    protocol: TCP
    targetPort: 8080
  selector:
    greeting: hello
  sessionAffinity: None
  type: ClusterIP
status:
  loadBalancer: {}

In the preceding output, the ports field is an array of ServicePort objects: one named world-port and one named kubernetes-port. For more information about the Service fields, see ServiceSpec in the Kubernetes documentation.

These are the ways a client can call the Service:

• Using world-port: A client running on one of the cluster nodes sends a request to the clusterIP on port. In this example, 10.96.14.249:60000. The request is forwarded to a member Pod on targetPort. In this example, POD_IP_ADDRESS:50000.

• Using kubernetes-port: A client running on one of the cluster nodes sends a request to the clusterIP on port. In this example, 10.96.14.249:60001. The request is forwarded to a member Pod on targetPort. In this example, POD_IP_ADDRESS:8080.

Ingress components

These are some of the cluster components related to ingress:

Note: The Istio components that are preinstalled on Google Distributed Cloud support ingress only. These components are not intended to provide general service mesh capabilities. For information on setting up a fully functional service mesh, see Setup in the Istio documentation or Set up a multi-cluster mesh outside Google Cloud in the Cloud Service Mesh documentation.

• The istio-ingress Deployment. This is the ingress proxy. The ingress proxy forwards traffic to internal Services according to rules specified in an Ingress object.

• The istio-ingress Service. This Service exposes the istio-ingress Deployment.

• The istiod Deployment. This is the ingress controller. The ingress controller watches the creation of Ingress objects and configures the ingress proxy accordingly.

All of these Istio in-cluster components are installed in the gke-system namespace. This namespace doesn't conflict with a full Istio/Cloud Service Mesh installation.

Create an Ingress

Here's a manifest for an Ingress:

apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
  name: my-ingress
spec:
  rules:
  - http:
      paths:
      - path: /greet-the-world
        pathType: Exact
        backend:
          service:
            name: hello-service
            port:
              number: 60000
      - path: /greet-kubernetes
        pathType: Exact
        backend:
          service:
            name: hello-service
            port:
              number: 60001

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

kubectl apply --kubeconfig CLUSTER_KUBECONFIG -f my-ingress.yaml

When you create a user cluster, you specify a value for loadbalancer.ingressVIP in the cluster configuration file. This IP address is configured on the cluster load balancer. When you create an Ingress, the Ingress is given this same VIP as its external IP address.

When a client sends a request to your user cluster ingress VIP, the request is routed to your load balancer. The load balancer uses the istio-ingress Service to forward the request to the ingress proxy, which runs in your user cluster. The ingress proxy is configured to forward the request to different backends depending on the path in the request URL.

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 section of your hello-service Service.

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

The ingress Service forwards the request to clusterIP:50000. The request then goes to one of the member Pods of the hello-service Service. The container, in that Pod, 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 section of your hello-service Service.

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

The ingress Service forwards the request to clusterIP: 8080. The request then goes to one of the member Pods of the hello-service Service. The container, in that Pod, listening on port 8080 displays a Hello Kubernetes! message.

Test the Ingress

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

curl CLUSTER_INGRESS_VIP/greet-the-world

Replace CLUSTER_INGRESS_VIP with the external IP address of the Ingress.

The output shows a Hello, world! message:

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

Test the Ingress using the /greet-kubernetes path:

curl CLUSTER_INGRESS_VIP/greet-kubernetes

The output shows a Hello, Kubernetes! message:

Hello Kubernetes!

Disable bundled Ingress

The Ingress capability bundled with Google Distributed Cloud supports ingress functionality only. You may choose to integrate with Istio or Cloud Service Mesh. These products offer additional benefits of a fully functional service mesh, such as mutual TLS (mTLS), ability to manage authentication between services, and workload observability. If you integrate with Istio or Cloud Service Mesh, we recommend that you disable the bundled Ingress capability.

You can enable or disable bundled Ingress with the spec.clusterNetwork.bundledIngress field in your cluster configuration file. This field is available to version 1.13.0 clusters and higher only. The bundledIngress field defaults to true and isn't present in the generated cluster configuration file. This field is mutable and can be changed when you create or update a version 1.13.0 or higher cluster. You can also specify this field when you upgrade a cluster to version 1.13.0 or higher.

The following sample cluster configuration file shows how you configure your cluster to disable the bundled Ingress capability:

apiVersion: v1
kind: Namespace
metadata:
  name: cluster-hybrid-basic
---
apiVersion: baremetal.cluster.gke.io/v1
kind: Cluster
metadata:
  name: hybrid-basic
  namespace: cluster-hybrid-basic
spec:
  type: hybrid
  profile: default
  anthosBareMetalVersion: 1.13.0
  gkeConnect:
    projectID: project-fleet
  controlPlane:
    nodePoolSpec:
      nodes:
      - address: 10.200.0.2
  clusterNetwork:
    bundledIngress: false
    pods:
      cidrBlocks:
      - 192.168.0.0/16
    services:
      cidrBlocks:
      - 172.26.232.0/24
...

Set up HTTPS for Ingress

If you want to accept HTTPS requests from your clients, the ingress proxy must have a certificate so it can prove its identity to your clients. This proxy must also have a private key to complete the HTTPS handshake.

The following example uses these entities:

• Ingress proxy: Participates in the HTTPS handshake, and then forwards packets to member Pods of the hello-service Service.

• Domain for the hello-service Service: altostrat.com in Example Org

Follow these steps:

1. Create a root certificate and private key. This example uses a root certificate authority of root.ca.example.com in Root CA Example Org.

   openssl req -x509 -sha256 -nodes -days 365 -newkey rsa:2048 -subj \
       '/O=Root CA Example Inc./CN=root.ca.example.com' -keyout root-ca.key \
       -out root-ca.crt

2. Create a certificate signing request:

    openssl req -out server.csr -newkey rsa:2048 -nodes -keyout server.key -subj \
        "/CN=altostrat.com/O=Example Org"

3. Create a serving certificate for the ingress proxy.

   openssl x509 -req -days 365 -CA root-ca.crt -CAkey root-ca.key -set_serial 0 \
       -in server.csr -out server.crt

   You have now created the following certificates and keys:

   • root-ca.crt: Certificate for the root CA
   • root-ca.key: Private key for the root CA
   • server.crt: Serving certificate for the ingress proxy
   • server.key: Private key for the ingress proxy

4. Create a Kubernetes Secret that holds the serving certificate and key.

   kubectl create secret tls example-server-creds --key=server.key --cert=server.crt \
       --namespace gke-system

   The resulting Secret is named example-server-creds.

   Note: TLS secrets must be defined in the gke-system namespace.

Create a Deployment and Service

If you created a Deployment and a Service in the HTTP portion of this guide, leave those in place. If you did not, create them now, following the steps described for HTTP.

Create an Ingress

If you previously created an Ingress in the HTTP portion, delete that Ingress before proceeding.

Delete the Ingress:

kubectl --kubeconfig CLUSTER_KUBECONFIG delete ingress my-ingress

To handle traffic for the Service that you created previously, create a new Ingres that has a tls section. This will enable HTTPS between clients and the ingress proxy.

Here's a manifest for an Ingress:

apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
  name: my-ingress-2
spec:
  tls:
  - hosts:
    - altostrat.com
    secretName: example-server-creds
  rules:
  - host: altostrat.com
    http:
      paths:
      - path: /greet-the-world
        pathType: Exact
        backend:
          service:
            name: hello-service
            port:
              number: 60000
      - path: /greet-kubernetes
        pathType: Exact
        backend:
          service:
            name: hello-service
            port:
              number: 60001

Save the manifest in a file named my-ingress-2.yaml, and create the Ingress:

kubectl apply --kubeconfig CLUSTER_KUBECONFIG -f my-ingress-2.yaml

Confirm by testing.

• Test the /greet-the-world path:

  curl -v --resolve altostrat.com:443:CLUSTER_INGRESS_VIP\
      https://altostrat.com/greet-the-world \
      --cacert root-ca.crt

  Output:

  Hello, world!
  Version: 2.0.0
  Hostname: hello-deployment-5ff7f68854-wqzp7
  

• Test the /greet-kubernetes path:

  curl -v --resolve altostrat.com:443:CLUSTER_INGRESS_VIP \
      https://altostrat.com/greet-kubernetes --cacert root-ca.crt

  Output:

  Hello Kubernetes!
  

Create a LoadBalancer Service

The LoadBalancer type is an extension of the NodePort type. So a Service of type LoadBalancer has a cluster IP address and one or more nodePort values. By default, Kubernetes allocates node ports to LoadBalancer Services. These allocations can quickly exhaust available node ports from the 2,768 allotted to your cluster. To save node ports, disable load balancer node port allocation by setting the allocateLoadBalancerNodePorts field to false in the LoadBalancer Service spec. This setting prevents Kubernetes from allocating node ports to LoadBalancer Services. For more information, see Disabling load balancer NodePort allocation in the Kubernetes documentation.

Here's a manifest to create a Service that doesn't use any node ports:

apiVersion: v1
kind: Service
metadata:
  name: service-does-not-use-nodeports
spec:
  selector:
    app: my-app
  type: LoadBalancer
  ports:
  - port: 8000
  # Set allocateLoadBalancerNodePorts to false
  allocateLoadBalancerNodePorts: false

Cleaning up

Delete your Ingress:

kubectl --kubeconfig CLUSTER_KUBECONFIG delete ingress INGRESS_NAME

Replace INGRESS_NAME with the name of the Ingress, such as my-ingress or my-ingress-2.

Delete your Service:

kubectl --kubeconfig CLUSTER_KUBECONFIG delete service hello-service

Delete your Deployment:

kubectl --kubeconfig CLUSTER_KUBECONFIG delete deployment hello-deployment

Delete your LoadBalancer Service:

kubectl --kubeconfig CLUSTER_KUBECONFIG delete service service-does-not-use-nodeports