Kubernetes Cluster: Installation, Configuration, and Management

Kubernetes, or K8s, is an open-source container orchestration platform developed by Google. The core concept behind Kubernetes is that a user installs it on a server, or more likely a cluster, and deploys various workloads on it. Kubernetes addresses challenges related to container creation, scaling, namespaces, access rights, and more. The primary interaction with the cluster is through YAML configuration files.

This tutorial will guide you through creating and deploying a Kubernetes cluster locally.

Creating Virtual Machines

Copy link

We will set up the Kubernetes cluster on two virtual machines: one acting as the master node and the other as a worker node. While deploying a cluster with only two nodes is not practical for real-world use, it is sufficient for educational purposes. If you wish to create a Kubernetes cluster with more nodes, simply repeat the process for each additional node.

1. We will use Oracle's VirtualBox to create virtual machines, which you can download from this link. After installation, proceed to create the virtual machines.
2. For the operating system, we will use Ubuntu Server, which can be downloaded here. After downloading, open VirtualBox.
3. Click "Create" in VirtualBox to create a new virtual machine. The default settings are sufficient, but allocate 3 GB of RAM and 2 CPUs for the master node (which manages the Kubernetes cluster) and 2 GB of RAM for the worker node. Kubernetes requires a minimum of 2 CPUs for the master node. Create two virtual machines this way.
4. After creating the virtual machines, create a boot image with the Ubuntu Server distribution. Go to "Storage" and click "Choose/Create a Disk Image."
5. Click "Add" and select the Ubuntu Server distribution. Then, start both machines and install the operating system by selecting "Try or Install Ubuntu." During installation, create users for each system and choose the default settings.
6. After installation, shut down both virtual machines and go to their settings. In the "Network" section, change the connection type to "Bridged Adapter" for each system so that the virtual machines can communicate with each other over the network.

System Preparation

Copy link

Network Configuration

Copy link

Set the node names for the cluster. On the master node, execute the following command:

sudo hostnamectl set-hostname master.local

On the worker node, execute:

sudo hostnamectl set-hostname worker.local

If there are multiple worker nodes, assign each a unique name: worker1.local, worker2.local, and so on.

To ensure that nodes are accessible by name, modify the hosts file on each node. Add the following lines:

192.168.43.80     master.local master
192.168.43.77     worker.local worker

Here, 192.168.43.80 and 192.168.43.77 are the IP addresses of each node. To find the IP address, use the ip addr command:

ip addr

Locate the IP address next to inet. Open the hosts file and make the necessary edits:

sudo nano /etc/hosts

To verify that the VMs can communicate with each other, ping the nodes:

ping 192.168.43.80

If successful, you will receive a response similar to this:

PING 192.168.43.80 (192.168.43.80) 56(84) bytes of data.
64 bytes from 192.168.43.80: icmp_seq=1 ttl=64 time=0.054 ms

You can enforce pod-to-pod isolation by applying the patterns in Kubernetes Cluster Network Policies tutorial—defining ingress and egress rules, using podSelector and namespace selectors to restrict traffic between master and worker nodes.

Updating Packages and Installing Additional Utilities

Copy link

Next, install the necessary utilities and packages on each node. These steps should be applied to each node unless specified otherwise. Start by updating the package list and systems:

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

Then install the following packages:

sudo apt-get install curl apt-transport-https git iptables-persistent -y

Swap File

Copy link

Kubernetes will not start with an active swap file, so it needs to be disabled:

sudo swapoff -a

To prevent it from reactivating after a reboot, modify the fstab file:

sudo nano /etc/fstab

Comment out the line with #:

# /swap.img      none    swap    sw      0       0

Kernel Configuration

Copy link

Load additional kernel modules:

sudo nano /etc/modules-load.d/k8s.conf

Add the following two lines to k8s.conf:

br_netfilter
overlay

Now, load the modules into the kernel:

sudo modprobe br_netfilter
sudo modprobe overlay

Verify the modules are loaded successfully:

sudo lsmod | egrep "br_netfilter|overlay"

You should see output similar to this:

overlay               147456  0
br_netfilter           28672  0
bridge                299008  1 br_netfilter

Create a configuration file to process traffic through the bridge in netfilter:

sudo nano /etc/sysctl.d/k8s.conf

Add the following two lines:

net.bridge.bridge-nf-call-ip6tables = 1
net.bridge.bridge-nf-call-iptables = 1

Apply the settings:

sudo sysctl --system

Docker Installation

Copy link

Run the following command to install Docker:

sudo apt-get install docker docker.io -y

For more details on installing Docker on Ubuntu, refer to the official guide.

After installation, enable Docker to start on boot and restart the service:

sudo systemctl enable docker
sudo systemctl restart docker

Kubernetes Installation

Copy link

Add the GPG key:

sudo curl -s https://packages.cloud.google.com/apt/doc/apt-key.gpg | sudo apt-key add -

Next, create a repository configuration file:

sudo nano /etc/apt/sources.list.d/kubernetes.list

Add the following entry:

deb https://apt.kubernetes.io/ kubernetes-xenial main

Update the apt-get package list:

sudo apt-get update

Install the following packages:

sudo apt-get install kubelet kubeadm kubectl

Installation is now complete. Verify the Kubernetes client version:

sudo kubectl version --client 

The output should be similar to this:

Client Version: version.Info{Major:"1", Minor:"24", GitVersion:"v1.24.2"}

Cluster Configuration

Copy link

Master Node

Copy link

Run the following command for the initial setup and preparation of the master node:

sudo kubeadm init --pod-network-cidr=10.244.0.0/16

The --pod-network-cidr flag specifies the internal subnet address, with 10.244.0.0/16 being the default value.

The process will take a few minutes. Upon completion, you will see the following message:

Then you can join any number of worker nodes by running the following on each as root:
kubeadm join 192.168.43.80:6443 --token f7sihu.wmgzwxkvbr8500al \
--discovery-token-ca-cert-hash sha256:6746f66b2197ef496192c9e240b31275747734cf74057e04409c33b1ad280321

Save this command to connect the worker nodes to the master node.

Create the KUBECONFIG environment variable:

export KUBECONFIG=/etc/kubernetes/admin.conf

Install the Container Network Interface (CNI):

kubectl apply -f https://raw.githubusercontent.com/coreos/flannel/master/Documentation/kube-flannel.yml

Worker Node

Copy link

On the worker node, run the kubeadm join command obtained during the master node setup. After this, on the master node, enter:

sudo kubectl get nodes

The output should be:

NAME                 STATUS      ROLES                        AGE    VERSION
master.local          Ready      control-plane,master          10m    v1.24.2
worker.local          Ready      <none>                        79s    v1.24.2

The cluster is now deployed and ready for operation.

Conclusion

Copy link

Setting up a Kubernetes cluster involves several steps, from creating and configuring virtual machines to installing and configuring the necessary software components. This tutorial provided a step-by-step guide to deploying a basic Kubernetes cluster on a local environment. While this setup is suitable for educational purposes, real-world deployments typically involve more nodes and more complex configurations. Kubernetes provides powerful tools for managing containerized applications, making it a valuable skill for modern IT professionals. By following this guide, you've taken the first steps in mastering Kubernetes and its ecosystem.