Installation Overview

Prerequisites

Karmada kubectl plugin

kubectl-karmada is the Karmada command-line tool that lets you control the Karmada control plane, it presents as the kubectl plugin. For installation instructions see installing kubectl-karmada.

Karmadactl

karmadactl is also the command-line tool that lets you control the Karmada control plane. Compared with kubectl-karmada, it is a complete CLI tool exclusive to Karmada. For installation instructions see installing karmadactl.

Installation Overview - 图1note

Although the above two tools have different forms, the related commands and options are exactly the same. The following takes kubectl-karmada as an example. Replacing it with karmadactl also works fine.

In actual use, you can choose a CLI tool according to your needs.

Install Karmada by Karmada command-line tool

Install Karmada on your own cluster

Assume you have put your cluster’s kubeconfig file to $HOME/.kube/config or specify the path with KUBECONFIG environment variable. Otherwise, you should specify the configuration file by setting --kubeconfig flag to the following commands.

Note: The init command is available from v1.0. Running init command requires escalated privileges for it to store public configurations (certs, crds) for multiple users under default location /etc/karmada, you can override this location via flags --karmada-data and --karmada-pki. Refer to CLI for more details or usage information.

Run the following command to install:

  1. kubectl karmada init

It might take about 5 minutes and if everything goes well, you will see outputs similar to:

  1. I1121 19:33:10.270959 2127786 tlsbootstrap.go:61] [bootstrap-token] configured RBAC rules to allow certificate rotation for all agent client certificates in the member cluster
  2. I1121 19:33:10.275041 2127786 deploy.go:127] Initialize karmada bootstrap token
  3. I1121 19:33:10.281426 2127786 deploy.go:397] create karmada kube controller manager Deployment
  4. I1121 19:33:10.288232 2127786 idempotency.go:276] Service karmada-system/kube-controller-manager has been created or updated.
  5. ...
  6. ...
  7. ------------------------------------------------------------------------------------------------------
  8.  █████   ████   █████████   ███████████   ██████   ██████   █████████   ██████████     █████████
  9. ░░███   ███░   ███░░░░░███ ░░███░░░░░███ ░░██████ ██████   ███░░░░░███ ░░███░░░░███   ███░░░░░███
  10.  ░███  ███    ░███    ░███  ░███    ░███  ░███░█████░███  ░███    ░███  ░███   ░░███ ░███    ░███
  11.  ░███████     ░███████████  ░██████████   ░███░░███ ░███  ░███████████  ░███    ░███ ░███████████
  12.  ░███░░███    ░███░░░░░███  ░███░░░░░███  ░███ ░░░  ░███  ░███░░░░░███  ░███    ░███ ░███░░░░░███
  13.  ░███ ░░███   ░███    ░███  ░███    ░███  ░███      ░███  ░███    ░███  ░███    ███  ░███    ░███
  14.  █████ ░░████ █████   █████ █████   █████ █████     █████ █████   █████ ██████████   █████   █████
  15. ░░░░░   ░░░░ ░░░░░   ░░░░░ ░░░░░   ░░░░░ ░░░░░     ░░░░░ ░░░░░   ░░░░░ ░░░░░░░░░░   ░░░░░   ░░░░░
  16. ------------------------------------------------------------------------------------------------------
  17. Karmada is installed successfully.
  19. Register Kubernetes cluster to Karmada control plane.
  21. Register cluster with 'Push' mode
  23. Step 1: Use "kubectl karmada join" command to register the cluster to Karmada control plane. --cluster-kubeconfig is kubeconfig of the member cluster.
  24. (In karmada)~# MEMBER_CLUSTER_NAME=$(cat ~/.kube/config  | grep current-context | sed 's/: /\n/g'| sed '1d')
  25. (In karmada)~# kubectl karmada --kubeconfig /etc/karmada/karmada-apiserver.config  join ${MEMBER_CLUSTER_NAME} --cluster-kubeconfig=$HOME/.kube/config
  27. Step 2: Show members of karmada
  28. (In karmada)~# kubectl --kubeconfig /etc/karmada/karmada-apiserver.config get clusters
  31. Register cluster with 'Pull' mode
  33. Step 1: Use "kubectl karmada register" command to register the cluster to Karmada control plane. "--cluster-name" is set to cluster of current-context by default.
  34. (In member cluster)~# kubectl karmada register 172.18.0.3:32443 --token lm6cdu.lcm4wafod2jmjvty --discovery-token-ca-cert-hash sha256:9bf5aa53d2716fd9b5568c85db9461de6429ba50ef7ade217f55275d89e955e4
  36. Step 2: Show members of karmada
  37. (In karmada)~# kubectl --kubeconfig /etc/karmada/karmada-apiserver.config get clusters

The components of Karmada are installed in karmada-system namespace by default, you can get them by:

  1. kubectl get deployments -n karmada-system
  2. NAME                           READY   UP-TO-DATE   AVAILABLE   AGE
  3. karmada-aggregated-apiserver   1/1     1            1           102s
  4. karmada-apiserver              1/1     1            1           2m34s
  5. karmada-controller-manager     1/1     1            1           116s
  6. karmada-scheduler              1/1     1            1           119s
  7. karmada-webhook                1/1     1            1           113s
  8. kube-controller-manager        1/1     1            1           2m3s

And the karmada-etcd is installed as the StatefulSet, get it by:

  1. kubectl get statefulsets -n karmada-system
  2. NAME   READY   AGE
  3. etcd   1/1     28m

The configuration file of Karmada will be created to /etc/karmada/karmada-apiserver.config by default.

Offline installation

When installing Karmada, the kubectl karmada init will download the APIs(CRD) from the Karmada official release page (e.g. https://github.com/karmada-io/karmada/releases/tag/v0.10.1) and load images from the official registry by default.

If you want to install Karmada offline, maybe you have to specify the APIs tar file as well as the image.

Use --crds flag to specify the CRD file. e.g.

  1. kubectl karmada init --crds /$HOME/crds.tar.gz

The images of Karmada components could be specified, take karmada-controller-manager as an example:

  1. kubectl karmada init --karmada-controller-manager-image=example.registry.com/library/karmada-controller-manager:1.0

Deploy HA

Use --karmada-apiserver-replicas and --etcd-replicas flags to specify the number of the replicas (defaults to 1).

  1. kubectl karmada init --karmada-apiserver-replicas 3 --etcd-replicas 3

Install Karmada in Kind cluster

kind is a tool for running local Kubernetes clusters using Docker container “nodes”. It was primarily designed for testing Kubernetes itself, not for production.

Create a cluster named host by hack/create-cluster.sh:

  1. hack/create-cluster.sh host $HOME/.kube/host.config

Install Karmada v1.2.0 by command kubectl karmada init:

  1. kubectl karmada init --crds https://github.com/karmada-io/karmada/releases/download/v1.2.0/crds.tar.gz --kubeconfig=$HOME/.kube/host.config

Check installed components:

  1. kubectl get pods -n karmada-system --kubeconfig=$HOME/.kube/host.config
  2. NAME                                           READY   STATUS    RESTARTS   AGE
  3. etcd-0                                         1/1     Running   0          2m55s
  4. karmada-aggregated-apiserver-84b45bf9b-n5gnk   1/1     Running   0          109s
  5. karmada-apiserver-6dc4cf6964-cz4jh             1/1     Running   0          2m40s
  6. karmada-controller-manager-556cf896bc-79sxz    1/1     Running   0          2m3s
  7. karmada-scheduler-7b9d8b5764-6n48j             1/1     Running   0          2m6s
  8. karmada-webhook-7cf7986866-m75jw               1/1     Running   0          2m
  9. kube-controller-manager-85c789dcfc-k89f8       1/1     Running   0          2m10s

Install Karmada by Helm Chart Deployment

Please refer to installing by Helm.

Install Karmada by Karmada Operator

Please refer to installing by Karmada Operator

Install Karmada by binary

Please refer to installing by binary.

Install Karmada from source

Please refer to installing from source.

Install Karmada for development environment

If you want to try Karmada, we recommend that build a development environment by hack/local-up-karmada.sh which will do the following tasks for you:

  • Start a Kubernetes cluster by kind to run the Karmada control plane, aka. the host cluster.
  • Build Karmada control plane components based on a current codebase.
  • Deploy Karmada control plane components on the host cluster.
  • Create member clusters and join Karmada.

1. Clone Karmada repo to your machine:

  1. git clone https://github.com/karmada-io/karmada

or use your fork repo by replacing your GitHub ID:

  1. git clone https://github.com/<GitHub ID>/karmada

2. Change to the karmada directory:

  1. cd karmada

3. Deploy and run Karmada control plane:

run the following script:

  1. hack/local-up-karmada.sh

This script will do the following tasks for you:

  • Start a Kubernetes cluster to run the Karmada control plane, aka. the host cluster.
  • Build Karmada control plane components based on a current codebase.
  • Deploy Karmada control plane components on the host cluster.
  • Create member clusters and join Karmada.

If everything goes well, at the end of the script output, you will see similar messages as follows:

  1. Local Karmada is running.
  3. To start using your Karmada environment, run:
  4.   export KUBECONFIG="$HOME/.kube/karmada.config"
  5. Please use 'kubectl config use-context karmada-host/karmada-apiserver' to switch the host and control plane cluster.
  7. To manage your member clusters, run:
  8.   export KUBECONFIG="$HOME/.kube/members.config"
  9. Please use 'kubectl config use-context member1/member2/member3' to switch to the different member cluster.

4. Check registered cluster

  1. kubectl get clusters --kubeconfig=/$HOME/.kube/karmada.config

You will get similar output as follows:

  1. NAME      VERSION   MODE   READY   AGE
  2. member1   v1.23.4   Push   True    7m38s
  3. member2   v1.23.4   Push   True    7m35s
  4. member3   v1.23.4   Pull   True    7m27s

There are 3 clusters named member1, member2 and member3 have registered with Push or Pull mode.