Shared control plane (multi-network)

10 minute read

Follow this guide to configure a multicluster mesh using a shared control plane with gateways to connect network-isolated clusters. Istio’s location-aware service routing feature is used to route requests to different endpoints, depending on the location of the request source.

By following the instructions in this guide, you will setup a two-cluster mesh as shown in the following diagram:

Shared Istio control plane topology spanning multiple Kubernetes clusters using gateways

The primary cluster, cluster1, runs the full set of Istio control plane components while cluster2 only runs Istio Citadel, Sidecar Injector, and Ingress gateway. No VPN connectivity nor direct network access between workloads in different clusters is required.

Prerequisites

Two or more Kubernetes clusters with versions: 1.13, 1.14, 1.15.
Authority to deploy the Istio control plane
Two Kubernetes clusters (referred to as cluster1 and cluster2).
The Kubernetes API server of cluster2 MUST be accessible from cluster1 in order to run this configuration.

You can use the kubectl command to access both the cluster1 and cluster2 clusters with the --context flag, for example kubectl get pods --context cluster1. Use the following command to list your contexts:

$ kubectl config get-contexts
CURRENT   NAME       CLUSTER    AUTHINFO       NAMESPACE
*         cluster1   cluster1   user@foo.com   default
          cluster2   cluster2   user@foo.com   default

Store the context names of your clusters in environment variables:
```
$ export CTX_CLUSTER1=$(kubectl config view -o jsonpath='{.contexts[0].name}')
$ export CTX_CLUSTER2=$(kubectl config view -o jsonpath='{.contexts[1].name}')
$ echo CTX_CLUSTER1 = ${CTX_CLUSTER1}, CTX_CLUSTER2 = ${CTX_CLUSTER2}
CTX_CLUSTER1 = cluster1, CTX_CLUSTER2 = cluster2
```
If you have more than two clusters in the context list and you want to configure your mesh using clusters other than the first two, you will need to manually set the environment variables to the appropriate context names.

Setup the multicluster mesh

In this configuration you install Istio with mutual TLS enabled for both the control plane and application pods. For the shared root CA, you create a cacerts secret on both cluster1 and cluster2 clusters using the same Istio certificate from the Istio samples directory.

The instructions, below, also set up cluster2 with a selector-less service and an endpoint for istio-pilot.istio-system that has the address of cluster1 Istio ingress gateway. This will be used to access pilot on cluster1 securely using the ingress gateway without mutual TLS termination.

Setup cluster 1 (primary)

Deploy Istio to cluster1:

$ kubectl create --context=$CTX_CLUSTER1 ns istio-system
$ kubectl create --context=$CTX_CLUSTER1 secret generic cacerts -n istio-system --from-file=samples/certs/ca-cert.pem --from-file=samples/certs/ca-key.pem --from-file=samples/certs/root-cert.pem --from-file=samples/certs/cert-chain.pem
$ istioctl manifest apply --context=$CTX_CLUSTER1 \
  -f install/kubernetes/operator/examples/multicluster/values-istio-multicluster-primary.yaml

Wait for the Istio pods on cluster1 to become ready:

$ kubectl get pods --context=$CTX_CLUSTER1 -n istio-system
NAME                                      READY   STATUS    RESTARTS   AGE
istio-citadel-55d8b59798-6hnx4            1/1     Running   0          83s
istio-galley-c74b77787-lrtr5              2/2     Running   0          82s
istio-ingressgateway-684f5df677-shzhm     1/1     Running   0          83s
istio-pilot-5495bc8885-2rgmf              2/2     Running   0          82s
istio-policy-69cdf5db4c-x4sct             2/2     Running   2          83s
istio-sidecar-injector-5749cf7cfc-pgd95   1/1     Running   0          82s
istio-telemetry-646db5ddbd-gvp6l          2/2     Running   1          83s
prometheus-685585888b-4tvf7               1/1     Running   0          83s

Create an ingress gateway to access service(s) in cluster2:
```
$ kubectl apply --context=$CTX_CLUSTER1 -f - <<EOF
apiVersion: networking.istio.io/v1alpha3
kind: Gateway
metadata:
  name: cluster-aware-gateway
  namespace: istio-system
spec:
  selector:
    istio: ingressgateway
  servers:
  - port:
      number: 443
      name: tls
      protocol: TLS
    tls:
      mode: AUTO_PASSTHROUGH
    hosts:
    - "*.local"
EOF
```
This Gateway configures 443 port to pass incoming traffic through to the target service specified in a request’s SNI header, for SNI values of the local top-level domain (i.e., the Kubernetes DNS domain). Mutual TLS connections will be used all the way from the source to the destination sidecar.
Although applied to cluster1, this Gateway instance will also affect cluster2 because both clusters communicate with the same Pilot.
Determine the ingress IP and port for cluster1.
1. Set the current context of kubectl to CTX_CLUSTER1
```
$ export ORIGINAL_CONTEXT=$(kubectl config current-context)
$ kubectl config use-context $CTX_CLUSTER1
```
2. Follow the instructions in Determining the ingress IP and ports, to set the INGRESS_HOST and SECURE_INGRESS_PORT environment variables.
3. Restore the previous kubectl context:
```
$ kubectl config use-context $ORIGINAL_CONTEXT
$ unset ORIGINAL_CONTEXT
```
4. Print the values of INGRESS_HOST and SECURE_INGRESS_PORT:
```
$ echo The ingress gateway of cluster1: address=$INGRESS_HOST, port=$SECURE_INGRESS_PORT
```
Update the gateway address in the mesh network configuration. Edit the istio ConfigMap:
```
$ kubectl edit cm -n istio-system --context=$CTX_CLUSTER1 istio
```
Update the gateway’s address and port of network1 to reflect the cluster1 ingress host and port, respectively, then save and quit. Note that the address appears in two places, the second under values.yaml:.
Once saved, Pilot will automatically read the updated network configuration.

Setup cluster 2

Export the cluster1 gateway address:
```
$ export LOCAL_GW_ADDR=$(kubectl get --context=$CTX_CLUSTER1 svc --selector=app=istio-ingressgateway \
    -n istio-system -o jsonpath='{.items[0].status.loadBalancer.ingress[0].ip}') && echo ${LOCAL_GW_ADDR}
```
This command sets the value to the gateway’s public IP and displays it.
The command fails if the load balancer configuration doesn’t include an IP address. The implementation of DNS name support is pending.

Deploy Istio to cluster2:

$ kubectl create --context=$CTX_CLUSTER2 ns istio-system
$ kubectl create --context=$CTX_CLUSTER2 secret generic cacerts -n istio-system --from-file=samples/certs/ca-cert.pem --from-file=samples/certs/ca-key.pem --from-file=samples/certs/root-cert.pem --from-file=samples/certs/cert-chain.pem
$ CLUSTER_NAME=$(kubectl --context=$CTX_CLUSTER2 config view --minify=true -o jsonpath='{.clusters[].name}')
$ istioctl manifest apply --context=$CTX_CLUSTER2 \
  --set profile=remote \
  --set values.global.mtls.enabled=true \
  --set values.gateways.enabled=true \
  --set values.security.selfSigned=false \
  --set values.global.controlPlaneSecurityEnabled=true \
  --set values.global.createRemoteSvcEndpoints=true \
  --set values.global.remotePilotCreateSvcEndpoint=true \
  --set values.global.remotePilotAddress=${LOCAL_GW_ADDR} \
  --set values.global.remotePolicyAddress=${LOCAL_GW_ADDR} \
  --set values.global.remoteTelemetryAddress=${LOCAL_GW_ADDR} \
  --set values.gateways.istio-ingressgateway.env.ISTIO_META_NETWORK="network2" \
  --set values.global.network="network2" \
  --set values.global.multiCluster.clusterName=${CLUSTER_NAME} \
  --set autoInjection.enabled=true

Wait for the Istio pods on cluster2, except for istio-ingressgateway, to become ready:

$ kubectl get pods --context=$CTX_CLUSTER2 -n istio-system -l istio!=ingressgateway
NAME                                      READY   STATUS    RESTARTS   AGE
istio-citadel-55d8b59798-nlk2z            1/1     Running   0          26s
istio-sidecar-injector-5749cf7cfc-s6r7p   1/1     Running   0          25s

Determine the ingress IP and port for cluster2.
1. Set the current context of kubectl to CTX_CLUSTER2
```
$ export ORIGINAL_CONTEXT=$(kubectl config current-context)
$ kubectl config use-context $CTX_CLUSTER2
```
2. Follow the instructions in Determining the ingress IP and ports, to set the INGRESS_HOST and SECURE_INGRESS_PORT environment variables.
3. Restore the previous kubectl context:
```
$ kubectl config use-context $ORIGINAL_CONTEXT
$ unset ORIGINAL_CONTEXT
```
4. Print the values of INGRESS_HOST and SECURE_INGRESS_PORT:
```
$ echo The ingress gateway of cluster2: address=$INGRESS_HOST, port=$SECURE_INGRESS_PORT
```
Update the gateway address in the mesh network configuration. Edit the istio ConfigMap:
```
$ kubectl edit cm -n istio-system --context=$CTX_CLUSTER1 istio
```
Update the gateway’s address and port of network2 to reflect the cluster2 ingress host and port, respectively, then save and quit. Note that the address appears in two places, the second under values.yaml:.
Once saved, Pilot will automatically read the updated network configuration.

Prepare environment variables for building the n2-k8s-config file for the service account istio-reader-service-account:

$ CLUSTER_NAME=$(kubectl --context=$CTX_CLUSTER2 config view --minify=true -o jsonpath='{.clusters[].name}')
$ SERVER=$(kubectl --context=$CTX_CLUSTER2 config view --minify=true -o jsonpath='{.clusters[].cluster.server}')
$ SECRET_NAME=$(kubectl --context=$CTX_CLUSTER2 get sa istio-reader-service-account -n istio-system -o jsonpath='{.secrets[].name}')
$ CA_DATA=$(kubectl get --context=$CTX_CLUSTER2 secret ${SECRET_NAME} -n istio-system -o jsonpath="{.data['ca\.crt']}")
$ TOKEN=$(kubectl get --context=$CTX_CLUSTER2 secret ${SECRET_NAME} -n istio-system -o jsonpath="{.data['token']}" | base64 --decode)

Create the n2-k8s-config file in the working directory:

$ cat <<EOF > n2-k8s-config
apiVersion: v1
kind: Config
clusters:
  - cluster:
      certificate-authority-data: ${CA_DATA}
      server: ${SERVER}
    name: ${CLUSTER_NAME}
contexts:
  - context:
      cluster: ${CLUSTER_NAME}
      user: ${CLUSTER_NAME}
    name: ${CLUSTER_NAME}
current-context: ${CLUSTER_NAME}
users:
  - name: ${CLUSTER_NAME}
    user:
      token: ${TOKEN}
EOF

Start watching cluster 2

Execute the following commands to add and label the secret of the cluster2 Kubernetes. After executing these commands Istio Pilot on cluster1 will begin watching cluster2 for services and instances, just as it does for cluster1.
```
$ kubectl create --context=$CTX_CLUSTER1 secret generic n2-k8s-secret --from-file n2-k8s-config -n istio-system
$ kubectl label --context=$CTX_CLUSTER1 secret n2-k8s-secret istio/multiCluster=true -n istio-system
```

Wait for istio-ingressgateway to become ready:

$ kubectl get pods --context=$CTX_CLUSTER2 -n istio-system -l istio=ingressgateway
NAME                                    READY     STATUS    RESTARTS   AGE
istio-ingressgateway-5c667f4f84-bscff   1/1       Running   0          16m

Now that you have your cluster1 and cluster2 clusters set up, you can deploy an example service.

Deploy example service

As shown in the diagram, above, deploy two instances of the helloworld service, one on cluster1 and one on cluster2. The difference between the two instances is the version of their helloworld image.

Deploy helloworld v2 in cluster 2

Create a sample namespace with a sidecar auto-injection label:

$ kubectl create --context=$CTX_CLUSTER2 ns sample
$ kubectl label --context=$CTX_CLUSTER2 namespace sample istio-injection=enabled

Deploy helloworld v2:

$ kubectl create --context=$CTX_CLUSTER2 -f @samples/helloworld/helloworld.yaml@ -l app=helloworld -n sample
$ kubectl create --context=$CTX_CLUSTER2 -f @samples/helloworld/helloworld.yaml@ -l version=v2 -n sample

Confirm helloworld v2 is running:

$ kubectl get po --context=$CTX_CLUSTER2 -n sample
NAME                             READY     STATUS    RESTARTS   AGE
helloworld-v2-7dd57c44c4-f56gq   2/2       Running   0          35s

Deploy helloworld v1 in cluster 1

Create a sample namespace with a sidecar auto-injection label:

$ kubectl create --context=$CTX_CLUSTER1 ns sample
$ kubectl label --context=$CTX_CLUSTER1 namespace sample istio-injection=enabled

Deploy helloworld v1:

$ kubectl create --context=$CTX_CLUSTER1 -f @samples/helloworld/helloworld.yaml@ -l app=helloworld -n sample
$ kubectl create --context=$CTX_CLUSTER1 -f @samples/helloworld/helloworld.yaml@ -l version=v1 -n sample

Confirm helloworld v1 is running:

$ kubectl get po --context=$CTX_CLUSTER1 -n sample
NAME                            READY     STATUS    RESTARTS   AGE
helloworld-v1-d4557d97b-pv2hr   2/2       Running   0          40s

Cross-cluster routing in action

To demonstrate how traffic to the helloworld service is distributed across the two clusters, call the helloworld service from another in-mesh sleep service.

Deploy the sleep service in both clusters:

$ kubectl apply --context=$CTX_CLUSTER1 -f @samples/sleep/sleep.yaml@ -n sample
$ kubectl apply --context=$CTX_CLUSTER2 -f @samples/sleep/sleep.yaml@ -n sample

Wait for the sleep service to start in each cluster:

$ kubectl get po --context=$CTX_CLUSTER1 -n sample -l app=sleep
sleep-754684654f-n6bzf           2/2     Running   0          5s

$ kubectl get po --context=$CTX_CLUSTER2 -n sample -l app=sleep
sleep-754684654f-dzl9j           2/2     Running   0          5s

Call the helloworld.sample service several times from cluster1 :

$ kubectl exec --context=$CTX_CLUSTER1 -it -n sample -c sleep $(kubectl get pod --context=$CTX_CLUSTER1 -n sample -l app=sleep -o jsonpath='{.items[0].metadata.name}') -- curl helloworld.sample:5000/hello

Call the helloworld.sample service several times from cluster2 :

$ kubectl exec --context=$CTX_CLUSTER2 -it -n sample -c sleep $(kubectl get pod --context=$CTX_CLUSTER2 -n sample -l app=sleep -o jsonpath='{.items[0].metadata.name}') -- curl helloworld.sample:5000/hello

If set up correctly, the traffic to the helloworld.sample service will be distributed between instances on cluster1 and cluster2 resulting in responses with either v1 or v2 in the body:

Hello version: v2, instance: helloworld-v2-758dd55874-6x4t8
Hello version: v1, instance: helloworld-v1-86f77cd7bd-cpxhv

You can also verify the IP addresses used to access the endpoints by printing the log of the sleep’s istio-proxy container.

$ kubectl logs --context=$CTX_CLUSTER1 -n sample $(kubectl get pod --context=$CTX_CLUSTER1 -n sample -l app=sleep -o jsonpath='{.items[0].metadata.name}') istio-proxy
[2018-11-25T12:37:52.077Z] "GET /hello HTTP/1.1" 200 - 0 60 190 189 "-" "curl/7.60.0" "6e096efe-f550-4dfa-8c8c-ba164baf4679" "helloworld.sample:5000" "192.23.120.32:15443" outbound|5000||helloworld.sample.svc.cluster.local - 10.20.194.146:5000 10.10.0.89:59496 -
[2018-11-25T12:38:06.745Z] "GET /hello HTTP/1.1" 200 - 0 60 171 170 "-" "curl/7.60.0" "6f93c9cc-d32a-4878-b56a-086a740045d2" "helloworld.sample:5000" "10.10.0.90:5000" outbound|5000||helloworld.sample.svc.cluster.local - 10.20.194.146:5000 10.10.0.89:59646 -

In cluster1, the gateway IP of cluster2 (192.23.120.32:15443) is logged when v2 was called and the instance IP in cluster1 (10.10.0.90:5000) is logged when v1 was called.

$ kubectl logs --context=$CTX_CLUSTER2 -n sample $(kubectl get pod --context=$CTX_CLUSTER2 -n sample -l app=sleep -o jsonpath='{.items[0].metadata.name}') istio-proxy
[2019-05-25T08:06:11.468Z] "GET /hello HTTP/1.1" 200 - "-" 0 60 177 176 "-" "curl/7.60.0" "58cfb92b-b217-4602-af67-7de8f63543d8" "helloworld.sample:5000" "192.168.1.246:15443" outbound|5000||helloworld.sample.svc.cluster.local - 10.107.117.235:5000 10.32.0.10:36840 -
[2019-05-25T08:06:12.834Z] "GET /hello HTTP/1.1" 200 - "-" 0 60 181 180 "-" "curl/7.60.0" "ce480b56-fafd-468b-9996-9fea5257cb1e" "helloworld.sample:5000" "10.32.0.9:5000" outbound|5000||helloworld.sample.svc.cluster.local - 10.107.117.235:5000 10.32.0.10:36886 -

In cluster2, the gateway IP of cluster1 (192.168.1.246:15443) is logged when v1 was called and the gateway IP in cluster2 (10.32.0.9:5000) is logged when v2 was called.

Cleanup

Execute the following commands to clean up the example services and the Istio components.

Cleanup the cluster2 cluster:

$ istioctl manifest generate --context=$CTX_CLUSTER2 \
  --set profile=remote \
  --set values.global.mtls.enabled=true \
  --set values.gateways.enabled=true \
  --set values.security.selfSigned=false \
  --set values.global.controlPlaneSecurityEnabled=true \
  --set values.global.createRemoteSvcEndpoints=true \
  --set values.global.remotePilotCreateSvcEndpoint=true \
  --set values.global.remotePilotAddress=${LOCAL_GW_ADDR} \
  --set values.global.remotePolicyAddress=${LOCAL_GW_ADDR} \
  --set values.global.remoteTelemetryAddress=${LOCAL_GW_ADDR} \
  --set values.gateways.istio-ingressgateway.env.ISTIO_META_NETWORK="network2" \
  --set values.global.network="network2" \
  --set autoInjection.enabled=true | kubectl --context=$CTX_CLUSTER2 delete -f -
$ kubectl delete --context=$CTX_CLUSTER2 ns sample
$ unset CTX_CLUSTER2 CLUSTER_NAME SERVER SECRET_NAME CA_DATA TOKEN INGRESS_HOST SECURE_INGRESS_PORT INGRESS_PORT LOCAL_GW_ADDR

Cleanup the cluster1 cluster:

$ istioctl manifest generate --context=$CTX_CLUSTER1 \
  -f install/kubernetes/operator/examples/multicluster/values-istio-multicluster-primary.yaml | kubectl --context=$CTX_CLUSTER1 delete -f -
$ kubectl delete --context=$CTX_CLUSTER1 ns sample
$ unset CTX_CLUSTER1
$ rm n2-k8s-config