Federate zone control plane

With Kuma you can first start with just one zone control plane and then federate it to a multi-zone deployment. This way you can:

  • see your mesh deployment in one centralized place
  • connect multiple zones and introduce cross zone connectivity
  • manage policies that are pushed to all zones

Prerequisites

  • Completed quickstart to set up a zone control plane with demo application

Start a global control plane

Start a new Kubernetes cluster

Currently, it’s not possible to deploy global and zone control plane in the same Kubernetes cluster, therefore we need a new Kubernetes cluster.

You can skip this step if you already have a Kubernetes cluster running. It can be a cluster running locally or in a public cloud like AWS EKS, GCP GKE, etc.

  1. kind create cluster --name=mesh-global

By default, Kind does not support LoadBalancer therefore we need to configure it by following this guide. Note that all public cloud providers support load balancers out of the box.

Deploy a global control plane

  1. helm install --kube-context=kind-mesh-global --create-namespace --namespace kuma-system \
  2. --set controlPlane.mode=global \
  3. --set controlPlane.defaults.skipMeshCreation=true \
  4. kuma kuma/kuma

We skip default mesh creation as we will bring mesh from zone control plane in the next steps.

Sync endpoint

Find the external IP and port of the kuma-global-zone-sync service in the kuma-system namespace:

  1. kubectl get services -n kuma-system
  2. NAMESPACE     NAME                   TYPE           CLUSTER-IP      EXTERNAL-IP      PORT(S)                                                                  AGE
  3. kuma-system   kuma-global-zone-sync     LoadBalancer   10.105.9.10     35.226.196.103   5685:30685/TCP                                                           89s
  4. kuma-system   kuma-control-plane     ClusterIP      10.105.12.133   <none>           5681/TCP,443/TCP,5676/TCP,5677/TCP,5678/TCP,5679/TCP,5682/TCP,5653/UDP   90s

In this example the value is 35.226.196.103:5685. You pass this as the value of <global-kds-address> when you federate the zone control plane. If you see <Pending> you either need to wait until load balancer is provisioned or you need to make sure that your Kubernetes cluster support provisioning load balancers.

Copy resources from zone to global control plane

To federate zone control plane without any traffic interruption, we need to copy resources like secrets, meshes etc. First, we need to expose API server of zone control plane:

  1. kubectl --context=kind-mesh-zone port-forward svc/kuma-control-plane -n kuma-system 5681:5681

Then we export resources:

  1. export ZONE_USER_ADMIN_TOKEN=$(kubectl --context=kind-mesh-zone get secrets -n kuma-system admin-user-token -ojson | jq -r .data.value | base64 -d)
  2. kumactl config control-planes add \
  3.   --address http://localhost:5681 \
  4.   --headers "authorization=Bearer $ZONE_USER_ADMIN_TOKEN" \
  5.   --name "zone-cp" \
  6.   --overwrite  
  8. kumactl export --profile=federation --format=kubernetes > resources.yaml

And finally, we apply resources on global control plane

  1. kubectl apply --context=kind-mesh-global -f resources.yaml

Connect zone control plane to global control plane

Update Helm deployment of zone control plane to configure connection to the global control plane.

  1. helm upgrade --kube-context=mesh-zone --namespace kuma-system \
  2. --set controlPlane.mode=zone \
  3. --set controlPlane.zone=zone-1 \
  4. --set ingress.enabled=true \
  5. --set controlPlane.kdsGlobalAddress=grpcs://<global-kds-address>:5685 \
  6. --set controlPlane.tls.kdsZoneClient.skipVerify=true \
  7. kuma kuma/kuma

Verify federation

To verify federation, first port-forward the API service from the global control plane to port 15681 to avoid collision with previous port forward.

  1. kubectl --context=kind-mesh-global port-forward svc/kuma-control-plane -n kuma-system 15681:5681

And then navigate to 127.0.0.1:15681/gui to see the GUI.

You should eventually see

  • a zone in list of zones
  • policies including redis MeshTrafficPermission that we applied in the quickstart guide.
  • data plane proxies for the demo application that we installed in the quickstart guide.

Apply policy on global control plane

We can check policy synchronization from global control plane to zone control plane by applying a policy on global control plane:

  1. echo "apiVersion: kuma.io/v1alpha1
  2. kind: MeshCircuitBreaker
  3. metadata:
  4.   name: demo-app-to-redis
  5.   namespace: kuma-system
  6.   labels:
  7.     kuma.io/mesh: default
  8. spec:
  9.   targetRef:
  10.     kind: MeshService
  11.     name: demo-app_kuma-demo_svc_5000
  12.   to:
  13.   - targetRef:
  14.       kind: MeshService
  15.       name: redis_kuma-demo_svc_6379
  16.     default:
  17.       connectionLimits:
  18.         maxConnections: 2
  19.         maxPendingRequests: 8
  20.         maxRetries: 2
  21.         maxRequests: 2" | kubectl --context=kind-mesh-global apply -f -

If we execute the following command:

  1. kubectl get --context=kind-mesh-zone meshcircuitbreakers -A

The policy should be eventually available in zone control plane

  1. NAMESPACE     NAME                                                TARGETREF KIND   TARGETREF NAME
  2. kuma-system demo-app-to-redis-65xb45x2xfd5bf7f        MeshService      demo-app_kuma-demo_svc_5000
  3. kuma-system mesh-circuit-breaker-all-default          Mesh

Next steps

  • Read the multi-zone docs to learn more about this deployment model and cross-zone connectivity.