Cluster Installation

The installation process will be carried out on four virtual machines (VMs) with identical parameters.

• control-plane: test-master1, test-master2, test-master3
• worker nodes: test-worker1

The installation will be performed using the k3sup tool directly from your local machine.

Prerequisites

Running Linux virtual servers with SSH access. It is best to create them using the repository according to the VM creation documentation.

known_hosts

Before starting the installation process, it is advisable to log in to each VM at least once to avoid being prompted to accept the host key.

Tools

• k3sup
• kubectl

Installation

Before starting the installation, create the .kube directory locally.

k3sup options

Selected k3sup options

Option	Description
--ip	IP address of the host on which k3sup will perform operations
--tls-san	IP address (and DNS name) that will serve as the shared cluster IP in kube-vip
--cluster	Option used on the host where the cluster is being initialized
--server-ip	Option to specify the IP address of the node where the cluster was initialized, used when joining additional nodes
--server	Option used to add a host to the control-plane

Cluster initialization

test-master1

k3sup install \
--ip 192.168.3.17 \
--tls-san 192.168.3.22 \
--tls-san testk3s-ha-cluster.lan \
--k3s-extra-args "--disable servicelb" \
--cluster \
--k3s-channel latest \
--local-path $HOME/.kube/config \
--user cloud-user \
--merge

This command initializes the Kubernetes cluster on the host specified by the --ip option.

.kube/config

If the following error appears during cluster initialization:

Error: open /home/cloud-user/.kube/config: no such file or director

It means that k3sup was unable to configure cluster access for you. Log in to the newly created node and copy the contents of /etc/rancher/k3s/k3s.yaml to your local ~/.kube/config. You only need to change the line server: https://127.0.0.1:6443 and replace it with the IP address of the host that was just configured.

Once the command finishes, the first cluster node is up and running.
The installation ends with the following messages:

Merging config into file: /home/cloud-user/.kube/config
Saving file to: /home/cloud-user/.kube/config

# Test your cluster with:
export KUBECONFIG=/home/cloud-user/.kube/config
kubectl config use-context default
kubectl get node -o wide

Verifying that Kubernetes is working:

🖥️ kubectl get nodes

NAME           STATUS   ROLES                AGE   VERSION
test-master1   Ready    control-plane,etcd   12m   v1.36.1+k3s1

Installing kube-vip

Important

This step is required before proceeding with k3s installation on the remaining cluster nodes.

All commands are executed on test-master1.

sudo su -
mkdir -p /var/lib/rancher/k3s/server/manifests/
curl https://kube-vip.io/manifests/rbac.yaml > /var/lib/rancher/k3s/server/manifests/kube-vip-rbac.yaml
kubectl apply -f https://kube-vip.io/manifests/rbac.yaml
export VIP=192.168.3.22
export INTERFACE=eth0
KVVERSION=$(curl -sL https://api.github.com/repos/kube-vip/kube-vip/releases | jq -r ".[0].name")
alias kube-vip="ctr image pull ghcr.io/kube-vip/kube-vip:$KVVERSION; ctr run --rm --net-host ghcr.io/kube-vip/kube-vip:$KVVERSION vip /kube-vip"
kube-vip manifest daemonset --services --inCluster --arp --interface eth0 | kubectl apply -f -

After running these commands, the kube-vip pod should start:

root@test-master1:~# kubectl get pods -n kube-system | grep kube-vip
kube-vip-ds-xr8f5                         1/1     Running     0             11m

Joining control-plane nodes

Add additional hosts by changing the IP address in the --ip option:

k3sup join \
--ip 192.168.3.xx \
--server-ip 192.168.3.17 \
--server \
--k3s-channel latest \
--user cloud-user

Joining worker nodes

When joining worker nodes, omit the --server option:

k3sup join \
--ip 192.168.3.xx \
--server-ip 192.168.3.17 \
--k3s-channel latest \
--user cloud-user

Summary

After completing all operations, we have a ready 4-node Kubernetes cluster:

🖥️ kubectl get nodes

NAME           STATUS   ROLES                AGE    VERSION
test-master1   Ready    control-plane,etcd   95m    v1.36.1+k3s1
test-master2   Ready    control-plane,etcd   64m    v1.36.1+k3s1
test-master3   Ready    control-plane,etcd   4m3s   v1.36.1+k3s1
test-worker1   Ready    <none>               34s    v1.36.1+k3s1

kube-vip DaemonSet running on all control-plane (master) nodes:

🖥️ kubectl get pods -n kube-system -o wide | grep kube-vip

kube-vip-ds-4ww75                         1/1     Running     0              70m     192.168.3.18   test-master2   <none>           <none>
kube-vip-ds-b99sh                         1/1     Running     0              6m24s   192.168.3.20   test-worker1   <none>           <none>
kube-vip-ds-xglx4                         1/1     Running     0              9m52s   192.168.3.19   test-master3   <none>           <none>
kube-vip-ds-xr8f5                         1/1     Running     0              87m     192.168.3.17   test-master1   <none>           <none>