Install Primary-Remote on different networks

Follow this guide to install the Istio control plane on cluster1 (the primary cluster) and configure cluster2 (the remote cluster) to use the control plane in cluster1. Cluster cluster1 is on the network1 network, while cluster2 is on the network2 network. This means there is no direct connectivity between pods across cluster boundaries.

Before proceeding, be sure to complete the steps under before you begin.

If you are testing multicluster setup on kind you can use MetalLB to make use of EXTERNAL-IP for LoadBalancer services.

In this configuration, cluster cluster1 will observe the API Servers in both clusters for endpoints. In this way, the control plane will be able to provide service discovery for workloads in both clusters.

Service workloads across cluster boundaries communicate indirectly, via dedicated gateways for east-west traffic. The gateway in each cluster must be reachable from the other cluster.

Services in cluster2 will reach the control plane in cluster1 via the same east-west gateway.

Set the default network for cluster1

If the istio-system namespace is already created, we need to set the cluster’s network there:

$ kubectl --context="${CTX_CLUSTER1}" get namespace istio-system && \
  kubectl --context="${CTX_CLUSTER1}" label namespace istio-system topology.istio.io/network=network1

Configure cluster1 as a primary

Create the Istio configuration for cluster1:

$ cat <<EOF > cluster1.yaml
apiVersion: install.istio.io/v1alpha1
kind: IstioOperator
spec:
  values:
    global:
      meshID: mesh1
      multiCluster:
        clusterName: cluster1
      network: network1
      externalIstiod: true
EOF

Apply the configuration to cluster1:

$ istioctl install --context="${CTX_CLUSTER1}" -f cluster1.yaml

Notice that values.global.externalIstiod is set to true. This enables the control plane installed on cluster1 to also serve as an external control plane for other remote clusters. When this feature is enabled, istiod will attempt to acquire the leadership lock, and consequently manage, appropriately annotated remote clusters that will be attached to it (cluster2 in this case).

Install the east-west gateway in cluster1

Install a gateway in cluster1 that is dedicated to east-west traffic. By default, this gateway will be public on the Internet. Production systems may require additional access restrictions (e.g. via firewall rules) to prevent external attacks. Check with your cloud vendor to see what options are available.

Zip

$ @samples/multicluster/gen-eastwest-gateway.sh@ \
    --network network1 | \
    istioctl --context="${CTX_CLUSTER1}" install -y -f -

If the control-plane was installed with a revision, add the --revision rev flag to the gen-eastwest-gateway.sh command.

Wait for the east-west gateway to be assigned an external IP address:

$ kubectl --context="${CTX_CLUSTER1}" get svc istio-eastwestgateway -n istio-system
NAME                    TYPE           CLUSTER-IP    EXTERNAL-IP    PORT(S)   AGE
istio-eastwestgateway   LoadBalancer   10.80.6.124   34.75.71.237   ...       51s

Expose the control plane in cluster1

Before we can install on cluster2, we need to first expose the control plane in cluster1 so that services in cluster2 will be able to access service discovery:

Zip

$ kubectl apply --context="${CTX_CLUSTER1}" -n istio-system -f \
    @samples/multicluster/expose-istiod.yaml@

If the control-plane was installed with a revision rev, use the following command instead:

Zip

$ sed 's/{{.Revision}}/rev/g' @samples/multicluster/expose-istiod-rev.yaml.tmpl@ | kubectl apply --context="${CTX_CLUSTER1}" -n istio-system -f -

Set the control plane cluster for cluster2

We need identify the external control plane cluster that should manage cluster2 by annotating the istio-system namespace:

$ kubectl --context="${CTX_CLUSTER2}" create namespace istio-system
$ kubectl --context="${CTX_CLUSTER2}" annotate namespace istio-system topology.istio.io/controlPlaneClusters=cluster1

Setting the topology.istio.io/controlPlaneClusters namespace annotation to cluster1 instructs the istiod running in the same namespace (istio-system in this case) on cluster1 to manage cluster2 when it is attached as a remote cluster.

Set the default network for cluster2

Set the network for cluster2 by adding a label to the istio-system namespace:

$ kubectl --context="${CTX_CLUSTER2}" label namespace istio-system topology.istio.io/network=network2

Configure cluster2 as a remote

Save the address of cluster1’s east-west gateway.

$ export DISCOVERY_ADDRESS=$(kubectl \
    --context="${CTX_CLUSTER1}" \
    -n istio-system get svc istio-eastwestgateway \
    -o jsonpath='{.status.loadBalancer.ingress[0].ip}')

Now create a remote configuration on cluster2.

$ cat <<EOF > cluster2.yaml
apiVersion: install.istio.io/v1alpha1
kind: IstioOperator
spec:
  profile: remote
  values:
    istiodRemote:
      injectionPath: /inject/cluster/cluster2/net/network2
    global:
      remotePilotAddress: ${DISCOVERY_ADDRESS}
EOF

Here we’re configuring the location of the control plane using the injectionPath and remotePilotAddress parameters. Although convenient for demonstration, in a production environment it is recommended to instead configure the injectionURL parameter using properly signed DNS certs similar to the configuration shown in the external control plane instructions.

Apply the configuration to cluster2:

$ istioctl install --context="${CTX_CLUSTER2}" -f cluster2.yaml

Attach cluster2 as a remote cluster of cluster1

To attach the remote cluster to its control plane, we give the control plane in cluster1 access to the API Server in cluster2. This will do the following:

• Enables the control plane to authenticate connection requests from workloads running in cluster2. Without API Server access, the control plane will reject the requests.

• Enables discovery of service endpoints running in cluster2.

Because it has been included in the topology.istio.io/controlPlaneClusters namespace annotation, the control plane on cluster1 will also:

• Patch certs in the webhooks in cluster2.

• Start the namespace controller which writes configmaps in namespaces in cluster2.

To provide API Server access to cluster2, we generate a remote secret and apply it to cluster1:

$ istioctl create-remote-secret \
    --context="${CTX_CLUSTER2}" \
    --name=cluster2 | \
    kubectl apply -f - --context="${CTX_CLUSTER1}"

Install the east-west gateway in cluster2

As we did with cluster1 above, install a gateway in cluster2 that is dedicated to east-west traffic and expose user services.

Zip

$ @samples/multicluster/gen-eastwest-gateway.sh@ \
    --network network2 | \
    istioctl --context="${CTX_CLUSTER2}" install -y -f -

Wait for the east-west gateway to be assigned an external IP address:

$ kubectl --context="${CTX_CLUSTER2}" get svc istio-eastwestgateway -n istio-system
NAME                    TYPE           CLUSTER-IP    EXTERNAL-IP    PORT(S)   AGE
istio-eastwestgateway   LoadBalancer   10.0.12.121   34.122.91.98   ...       51s

Expose services in cluster1 and cluster2

Since the clusters are on separate networks, we also need to expose all user services (*.local) on the east-west gateway in both clusters. While these gateways are public on the Internet, services behind them can only be accessed by services with a trusted mTLS certificate and workload ID, just as if they were on the same network.

Zip

$ kubectl --context="${CTX_CLUSTER1}" apply -n istio-system -f \
    @samples/multicluster/expose-services.yaml@

Since cluster2 is installed with a remote profile, exposing services on the primary cluster will expose them on the east-west gateways of both clusters.

Congratulations! You successfully installed an Istio mesh across primary and remote clusters on different networks!

Next Steps

You can now verify the installation.

Cleanup

1. Uninstall Istio in cluster1:

   $ istioctl uninstall --context="${CTX_CLUSTER1}" -y --purge
   $ kubectl delete ns istio-system --context="${CTX_CLUSTER1}"

2. Uninstall Istio in cluster2:

   $ istioctl uninstall --context="${CTX_CLUSTER2}" -y --purge
   $ kubectl delete ns istio-system --context="${CTX_CLUSTER2}"