Skip to main content

Deploy Kubernetes On-Premises From Zero

Resource Preparation​

NameIPRoleRAMCPUOS
k8s-master-1192.168.1.111Control Plane32Ubuntu 22.04 LTS
k8s-master-2192.168.1.112Worker32Ubuntu 22.04 LTS
k8s-master-3192.168.1.113Worker32Ubuntu 22.04 LTS

Implementation Steps​

1) Update system​

sudo apt update -y && sudo apt upgrade -y

2) Add hosts (above all 3 virtual machines)​

Open file host to config ip address

vim /etc/hosts/

Add those configs into file

192.168.1.111 k8s-master-1
192.168.1.112 k8s-master-2
192.168.1.113 k8s-master-3

Saved and check

ping <ip-addr>

3) Create a new user​

We should create/switch to another user (avoid using root) to install kubernetes

Create devops user and add to sudo group

adduser devops
usermod -aG sudo devops

Switch to devops

su devops
cd /home/devops

4) Turn off swap​

Because Kubernetes require actual RAM to ensure about performance and stable

# turn off swap forever
sudo sed -i '/swap/s/^/#/' /etc/fstab

5) Download and add module into kernel​

System will download overlay & br_netfilter package manually, this is two modules which require for k8s implement

  • overlay: Support overlay filesystem, it's necessary for containerd to manage container image.

  • br_netfilter: Allow filter network package on bridge network, it's very important for Kubernetes to handle traffic between container.

sudo modprobe overlay
sudo modprobe br_netfilter

To make system automatically install two modules above after restart, we need modify config file by command below

sudo tee /etc/modules-load.d/containerd.conf <<EOF
overlay
br_netfilter
EOF

6) Config network​

Those config to make sure:

  • Kubernetes can manage traffic between containers and Pods.
  • Firewall rules can apply to traffic via bridge network.
  • The system can routing package between nodes in cluster.
echo "net.bridge.bridge-nf-call-ip6tables = 1" | sudo tee -a /etc/sysctl.d/kubernetes.conf
echo "net.bridge.bridge-nf-call-iptables = 1" | sudo tee -a /etc/sysctl.d/kubernetes.conf
echo "net.ipv4.ip_forward = 1" | sudo tee -a /etc/sysctl.d/kubernetes.conf

Apply sysctl config

sudo sysctl --system

7) Install docker​

sudo apt install -y curl gnupg2 software-properties-common apt-transport-https ca-certificates
sudo curl -fsSL https://download.docker.com/linux/ubuntu/gpg | sudo gpg --dearmour -o /etc/apt/trusted.gpg.d/docker.gpg
sudo add-apt-repository "deb [arch=amd64] https://download.docker.com/linux/ubuntu $(lsb_release -cs) stable"

8) Install & config containerd​

This is container runtime which is use by Kubernetes for manage container

sudo apt install -y containerd.io

After installed, we need config containerd use Systemd to manage resource group (cgroup) => This make resource management become consistent.

containerd config default | sudo tee /etc/containerd/config.toml >/dev/null 2>&1
sudo sed -i 's/SystemdCgroup = false/SystemdCgroup = true/g' /etc/containerd/config.toml

Restart containerd

sudo systemctl restart containerd
sudo systemctl enable containerd

9) Install Kubernetes package​

Config to download from legit source

echo "deb [signed-by=/etc/apt/keyrings/kubernetes-apt-keyring.gpg] https://pkgs.k8s.io/core:/stable:/v1.30/deb/ /" | sudo tee /etc/apt/sources.list.d/kubernetes.list
curl -fsSL https://pkgs.k8s.io/core:/stable:/v1.30/deb/Release.key | sudo gpg --dearmor -o /etc/apt/keyrings/kubernetes-apt-keyring.gpg

Install packages (kubelet, kubeadm, kubectl)

sudo apt install -y kubelet kubeadm kubectl

For avoid auto upgrade version => Ensure package versions are consistent.

sudo apt-mark hold kubelet kubeadm kubectl

Deploy model (1 master - 2 worker)​

1master2workers

Standard model

Init Kubernetes on master node​

sudo kubeadm init
mkdir -p $HOME/.kube
sudo cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
sudo chown $(id -u):$(id -g) $HOME/.kube/config
kubectl apply -f https://raw.githubusercontent.com/projectcalico/calico/v3.25.0/manifests/calico.yaml

After the initial setup, it'll have command follow format below - use it to add worker

sudo kubeadm join <ip-master-node>:6443 --token <your_token> --discovery-token-ca-cert-hash <your_sha>

Add worker node​

Copy command from master node and paste it to machine where role determine is worker role (maybe you need to run it with sudo)

Verify​

Run kubectl get node -o wide in master node to check worker added

verify-workers

Example output

Conclusion​

Just one of over 35 ways to install Kubernetes. In the future, to add more workers to your system, simply run

sudo kubeadm join <ip-master-node>:6443 --token <your_token> --discovery-token-ca-cert-hash <your_sha>

In the next post of series, I will show you how to use cloud for Kubernetes Happy Coding!