Step-by-step instructions for performing operations with Kubernetes.

Edit This Page

Set up Cluster Federation with Kubefed

Kubernetes version 1.5 and above includes a new command line tool called kubefed to help you administrate your federated clusters. kubefed helps you to deploy a new Kubernetes cluster federation control plane, and to add clusters to or remove clusters from an existing federation control plane.

This guide explains how to administer a Kubernetes Cluster Federation using kubefed.

Note: kubefed is a beta feature in Kubernetes 1.6.


This guide assumes that you have a running Kubernetes cluster. Please see one of the getting started guides for installation instructions for your platform.

Getting kubefed

Download the client tarball corresponding to the latest release and extract the binaries in the tarball with the commands:

# Linux
curl -LO https://storage.googleapis.com/kubernetes-release/release/$(curl -s https://storage.googleapis.com/kubernetes-release/release/stable.txt)/kubernetes-client-linux-amd64.tar.gz
tar -xzvf kubernetes-client-linux-amd64.tar.gz

# OS X
curl -LO https://storage.googleapis.com/kubernetes-release/release/$(curl -s https://storage.googleapis.com/kubernetes-release/release/stable.txt)/kubernetes-client-darwin-amd64.tar.gz
tar -xzvf kubernetes-client-darwin-amd64.tar.gz

# Windows
curl -LO https://storage.googleapis.com/kubernetes-release/release/$(curl -s https://storage.googleapis.com/kubernetes-release/release/stable.txt)/kubernetes-client-windows-amd64.tar.gz
tar -xzvf kubernetes-client-windows-amd64.tar.gz

Note: The URLs in the curl commands above download the binaries for amd64. If you are on a different architecture, please use a URL appropriate for your architecture. You can find the list of available binaries on the release page.

Copy the extracted binaries to one of the directories in your $PATH and set the executable permission on those binaries.

sudo cp kubernetes/client/bin/kubefed /usr/local/bin
sudo chmod +x /usr/local/bin/kubefed
sudo cp kubernetes/client/bin/kubectl /usr/local/bin
sudo chmod +x /usr/local/bin/kubectl

Install with snap on Ubuntu

kubefed is available as a snap application.

  1. If you are on Ubuntu or one of other Linux distributions that support snap package manager, you can install with:

    sudo snap install kubefed --classic
  2. Run kubefed version to verify that the version you’ve installed is sufficiently up-to-date.

Choosing a host cluster.

You’ll need to choose one of your Kubernetes clusters to be the host cluster. The host cluster hosts the components that make up your federation control plane. Ensure that you have a kubeconfig entry in your local kubeconfig that corresponds to the host cluster. You can verify that you have the required kubeconfig entry by running:

kubectl config get-contexts

The output should contain an entry corresponding to your host cluster, similar to the following:

CURRENT   NAME                                          CLUSTER                                       AUTHINFO                                      NAMESPACE
*         gke_myproject_asia-east1-b_gce-asia-east1     gke_myproject_asia-east1-b_gce-asia-east1     gke_myproject_asia-east1-b_gce-asia-east1

You’ll need to provide the kubeconfig context (called name in the entry above) for your host cluster when you deploy your federation control plane.

Deploying a federation control plane

To deploy a federation control plane on your host cluster, run kubefed init command. When you use kubefed init, you must provide the following:

If your host cluster is running in a non-cloud environment or an environment that doesn’t support common cloud primitives such as load balancers, you might need additional flags. Please see the on-premises host clusters section below.

The following example command deploys a federation control plane with the name fellowship, a host cluster context rivendell, and the domain suffix example.com.:

kubefed init fellowship \
    --host-cluster-context=rivendell \
    --dns-provider="google-clouddns" \

The domain suffix specified in --dns-zone-name must be an existing domain that you control, and that is programmable by your DNS provider. It must also end with a trailing dot.

Once the federation control plane is initialized, query the namespaces:

kubectl get namespace --context=fellowship

If you do not see the default namespace listed (this is due to a bug). Create it yourself with the following command:

kubectl create namespace default --context=fellowship

The machines in your host cluster must have the appropriate permissions to program the DNS service that you are using. For example, if your cluster is running on Google Compute Engine, you must enable the Google Cloud DNS API for your project.

The machines in Google Container Engine (GKE) clusters are created without the Google Cloud DNS API scope by default. If you want to use a GKE cluster as a Federation host, you must create it using the gcloud command with the appropriate value in the --scopes field. You cannot modify a GKE cluster directly to add this scope, but you can create a new node pool for your cluster and delete the old one. Note that this will cause pods in the cluster to be rescheduled.

To add the new node pool, run:

scopes="$(gcloud container node-pools describe --cluster=gke-cluster default-pool --format='value[delimiter=","](config.oauthScopes)')"
gcloud container node-pools create new-np \
    --cluster=gke-cluster \

To delete the old node pool, run:

gcloud container node-pools delete default-pool --cluster gke-cluster

kubefed init sets up the federation control plane in the host cluster and also adds an entry for the federation API server in your local kubeconfig. Note that in the beta release in Kubernetes 1.6, kubefed init does not automatically set the current context to the newly deployed federation. You can set the current context manually by running:

kubectl config use-context fellowship

where fellowship is the name of your federation.

Basic and token authentication support

kubefed init by default only generates TLS certificates and keys to authenticate with the federation API server and writes them to your local kubeconfig file. If you wish to enable basic authentication or token authentication for debugging purposes, you can enable them by passing the --apiserver-enable-basic-auth flag or the --apiserver-enable-token-auth flag.

kubefed init fellowship \
    --host-cluster-context=rivendell \
    --dns-provider="google-clouddns" \
    --dns-zone-name="example.com." \
    --apiserver-enable-basic-auth=true \

Passing command line arguments to federation components

kubefed init bootstraps a federation control plane with default arguments to federation API server and federation controller manager. Some of these arguments are derived from kubefed init’s flags. However, you can override these command line arguments by passing them via the appropriate override flags.

You can override the federation API server arguments by passing them to --apiserver-arg-overrides and override the federation controller manager arguments by passing them to --controllermanager-arg-overrides.

kubefed init fellowship \
    --host-cluster-context=rivendell \
    --dns-provider="google-clouddns" \
    --dns-zone-name="example.com." \
    --apiserver-arg-overrides="--anonymous-auth=false,--v=4" \

Configuring a DNS provider

The Federated service controller programs a DNS provider to expose federated services via DNS names. Certain cloud providers automatically provide the configuration required to program the DNS provider if the host cluster’s cloud provider is same as the DNS provider. In all other cases, you have to provide the DNS provider configuration to your federation controller manager which will in-turn be passed to the federated service controller. You can provide this configuration to federation controller manager by storing it in a file and passing the file’s local filesystem path to kubefed init’s --dns-provider-config flag. For example, save the config below in $HOME/coredns-provider.conf.

etcd-endpoints = http://etcd-cluster.ns:2379
zones = example.com.

And then pass this file to kubefed init:

kubefed init fellowship \
    --host-cluster-context=rivendell \
    --dns-provider="coredns" \
    --dns-zone-name="example.com." \

On-premises host clusters

API server service type

kubefed init exposes the federation API server as a Kubernetes service on the host cluster. By default, this service is exposed as a load balanced service. Most on-premises and bare-metal environments, and some cloud environments lack support for load balanced services. kubefed init allows exposing the federation API server as a NodePort service on such environments. This can be accomplished by passing the --api-server-service-type=NodePort flag. You can also specify the preferred address to advertise the federation API server by passing the --api-server-advertise-address=<IP-address> flag. Otherwise, one of the host cluster’s node address is chosen as the default.

kubefed init fellowship \
    --host-cluster-context=rivendell \
    --dns-provider="google-clouddns" \
    --dns-zone-name="example.com." \
    --api-server-service-type="NodePort" \

Provisioning storage for etcd

Federation control plane stores its state in etcd. etcd data must be stored in a persistent storage volume to ensure correct operation across federation control plane restarts. On host clusters that support dynamic provisioning of storage volumes, kubefed init dynamically provisions a PersistentVolume and binds it to a PersistentVolumeClaim to store etcd data. If your host cluster doesn’t support dynamic provisioning, you can also statically provision a PersistentVolume. kubefed init creates a PersistentVolumeClaim that has the following configuration:

apiVersion: v1
kind: PersistentVolumeClaim
    volume.alpha.kubernetes.io/storage-class: "yes"
    app: federated-cluster
  name: fellowship-federation-apiserver-etcd-claim
  namespace: federation-system
  - ReadWriteOnce
      storage: 10Gi

To statically provision a PersistentVolume, you must ensure that the PersistentVolume that you create has the matching storage class, access mode and at least as much capacity as the requested PersistentVolumeClaim.

Alternatively, you can disable persistent storage completely by passing --etcd-persistent-storage=false to kubefed init. However, we do not recommended this because your federation control plane cannot survive restarts in this mode.

kubefed init fellowship \
    --host-cluster-context=rivendell \
    --dns-provider="google-clouddns" \
    --dns-zone-name="example.com." \

kubefed init still doesn’t support attaching an existing PersistentVolumeClaim to the federation control plane that it bootstraps. We are planning to support this in a future version of kubefed.

CoreDNS support

Federated services now support CoreDNS as one of the DNS providers. If you are running your clusters and federation in an environment that does not have access to cloud-based DNS providers, then you can run your own CoreDNS instance and publish the federated service DNS names to that server.

You can configure your federation to use CoreDNS, by passing appropriate values to kubefed init’s --dns-provider and --dns-provider-config flags.

kubefed init fellowship \
    --host-cluster-context=rivendell \
    --dns-provider="coredns" \
    --dns-zone-name="example.com." \

For more information see Setting up CoreDNS as DNS provider for Cluster Federation.

Adding a cluster to a federation

Once you’ve deployed a federation control plane, you’ll need to make that control plane aware of the clusters it should manage. You can add a cluster to your federation by using the kubefed join command.

To use kubefed join, you’ll need to provide the name of the cluster you want to add to the federation, and the --host-cluster-context for the federation control plane’s host cluster.

Note: The name that you provide to the join command is used as the joining cluster’s identity in federation. This name should adhere to the rules described in the identifiers doc. If the context corresponding to your joining cluster conforms to these rules then you can use the same name in the join command. Otherwise, you will have to choose a different name for your cluster’s identity. For more information, please see the naming rules and customization section below.

The following example command adds the cluster gondor to the federation running on host cluster rivendell:

kubefed join gondor --host-cluster-context=rivendell

Note: Kubernetes requires that you manually join clusters to a federation because the federation control plane manages only those clusters that it is responsible for managing. Adding a cluster tells the federation control plane that it is responsible for managing that cluster.

Naming rules and customization

The cluster name you supply to kubefed join must be a valid RFC 1035 label and are enumerated in the Identifiers doc.

Furthermore, federation control plane requires credentials of the joined clusters to operate on them. These credentials are obtained from the local kubeconfig. kubefed join uses the cluster name specified as the argument to look for the cluster’s context in the local kubeconfig. If it fails to find a matching context, it exits with an error.

This might cause issues in cases where context names for each cluster in the federation don’t follow RFC 1035 label naming rules. In such cases, you can specify a cluster name that conforms to the RFC 1035 label naming rules and specify the cluster context using the --cluster-context flag. For example, if context of the cluster you are joining is gondor_needs-no_king, then you can join the cluster by running:

kubefed join gondor --host-cluster-context=rivendell --cluster-context=gondor_needs-no_king

Secret name

Cluster credentials required by the federation control plane as described above are stored as a secret in the host cluster. The name of the secret is also derived from the cluster name.

However, the name of a secret object in Kubernetes should conform to the DNS subdomain name specification described in RFC 1123. If this isn’t the case, you can pass the secret name to kubefed join using the --secret-name flag. For example, if the cluster name is noldor and the secret name is 11kingdom, you can join the cluster by running:

kubefed join noldor --host-cluster-context=rivendell --secret-name=11kingdom

Note: If your cluster name does not conform to the DNS subdomain name specification, all you need to do is supply the secret name via the --secret-name flag. kubefed join automatically creates the secret for you.

kube-dns configuration

kube-dns configuration must be updated in each joining cluster to enable federated service discovery. If the joining Kubernetes cluster is version 1.5 or newer and your kubefed is version 1.6 or newer, then this configuration is automatically managed for you when the clusters are joined or unjoined using kubefed join or unjoin commands.

In all other cases, you must update kube-dns configuration manually as described in the Updating KubeDNS section of the admin guide.

Removing a cluster from a federation

To remove a cluster from a federation, run the kubefed unjoin command with the cluster name and the federation’s --host-cluster-context:

kubefed unjoin gondor --host-cluster-context=rivendell

Turning down the federation control plane

Proper cleanup of federation control plane is not fully implemented in this beta release of kubefed. However, for the time being, deleting the federation system namespace should remove all the resources except the persistent storage volume dynamically provisioned for the federation control plane’s etcd. You can delete the federation namespace by running the following command:

$ kubectl delete ns federation-system


Create an Issue Edit this Page