Kubernetes Cluster API: a step by stap guide
In this post I will show you how you can install and manage a Kubernetes cluster with Cluster API.
What is ClusterAPI?
The Cluster API project was started by the Kubernetes Special Interest Group (SIG) Cluster Lifecycle and automates cluster lifecycle management for platform operators using Kubernetes-style APIs and conventions. Currently, more than 100 Kubernetes installers and distributions exist, each with a unique default configurations. Acknowledging this ClusterAPI concentrated on community based providers that bootstrap different kinds of Kubernetes clusters and architectures.
The most commonly used providers are cloud based solutions like AWS, Azure or GCP, or private cloud providers (CloudStack, OpenStack, OpenNebula), but you can also use virtualization engines like vSphere, Proxmox, Nautanix, KubVirt, Virtink, Harwester. The most intrestingsolutions are the hardware based providers like Metal3.
Create ClusterAPI Clusters with Kind
For this demo I will use ClusterAPI’s kind provider to create a cluster. The only dependencies are kubectl and kind. The we will install ClusterAPI’s client.
curl -L https://github.com/kubernetes-sigs/cluster-api/releases/download/v1.0.4/clusterctl-linux-amd64 -o clusterctl
chmod +x ./clusterctl
sudo mv ./clusterctl /usr/local/bin/clusterctl
clusterctl version
# OR
brew install kind
brew install kubernetes-cli
brew install kubectx
brew install clusterctl
To manage the new clusters we need a management kubernetes cluster where we will install the ClusterAPI controller and it’s CRDs. For this demo I will use a kind cluster for this porpose too.
Fo this the management kind cluster need to be communicat with the underlying docker engine on the host. To do that we need to mount the docker engine socket into the kind cluster’s vm.
cat > cluster-mgmt.yaml <<EOF
kind: Cluster
apiVersion: kind.x-k8s.io/v1alpha4
nodes:
- role: control-plane
extraMounts:
- hostPath: /var/lib/docker
containerPath: /var/lib/docker
- hostPath: /var/run/docker.sock
containerPath: /var/run/docker.sock
EOF
kind create cluster --config cluster-mgmt.yaml --name mgmt
docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
414fd857c0fb kindest/node:v1.32.2 "/usr/local/bin/entr…" About a minute ago Up 58 seconds 127.0.0.1:43483->6443/tcp mgmt-control-plane
1815dce2b51f nginx:latest "/docker-entrypoint.…" 2 weeks ago Up 49 minutes 0.0.0.0:80->80/tcp, 0.0.0.0:443->443/tcp kubedash-proxy
Install the ClusterAPI operator and CRDs:
export CLUSTER_TOPOLOGY=true
clusterctl init --infrastructure docker
kubectl get pod -n capd-system
kubectl get pod -n capi-kubeadm-bootstrap-system
kubectl get pod -n capi-kubeadm-control-plane-system
kubectl get crd | grep cluster
Once the management cluster is ready, you can create your first workload cluster.
clusterctl generate cluster capi-quickstart --flavor development \
--kubernetes-version v1.31.0 \
--control-plane-machine-count=1 \
--worker-machine-count=0 \
> capi-quickstart.yaml
kubectl apply -f capi-quickstart.yaml
kubectl get cluster
clusterctl describe cluster capi-quickstart
kubectl get kubeadmcontrolplane
NAME CLUSTERCLASS PHASE AGE VERSION
capi-quickstart quick-start Provisioned 5m47s v1.31.0
Deploy CNI
# First, get the kubeconfig
clusterctl get kubeconfig capi-quickstart > kubeconfig.capi-quickstart.yaml
# Install Calico
kubectl --kubeconfig=./kubeconfig.capi-quickstart.yaml apply \
-f https://raw.githubusercontent.com/projectcalico/calico/v3.27.0/manifests/calico.yaml
Upgrade cluster
As I sead earlier ClusterApi is a Kubernetes cluster lifecycle manager so you can use it to upgrade your cluster:
kubectl get kubeadmcontrolplane,machinedeployments
kubectl patch cluster capi-quickstart --type json \
--patch '[{"op": "replace", "path": "/spec/topology/version", "value": "v1.32.2"}]'
