logo

kobevino

Kubernetes Cluster Setup using Kubespray

Thu, Oct 9, 2025

If you're a frontend engineer, you should learn Kubernetes. As your applications grow, they need to be separated and deployed independently. Eventually, this leads to adopting a micro-frontend architecture, which fits perfectly with Kubernetes.

I’ve been trying to create a Kubernetes cluster using Vagrant and VirtualBox on my local machine for the past few days. However, it was quite difficult to run the cluster locally due to limited RAM.

So I've decided to create the cluster using EC2 instances, with Kubespray controller running locally.

First, let's create a master node and worker nodes.

Create a Master node and Worker nodes.

ec2-instance

You should create the master node using an EC2 instance of t3.medium or higher. For the worker nodes, it's recommnded to use at least t3.small instances.

Master Node - CPU: 2, RAM: 4 | Wokerer Node - CPU: 2, RAM: 2

inbound-rule

You need to open ports 6443 and 10250 in your EC2 security group.

  • 6443is used by the Kubernetes API server on the master node.
  • 10250 is used by Kubelet on each node, allowing the master to communicate with the workers.

These ports must be open so that the Kubespray playbook can successfully set up the cluster.

Create a KubeSpray controller on my computer.

$ mkdir kubespray
$ cd kubespray
$ git clone https://github.com/kubernetes-sigs/kubespray .
# Ensure you have installed Docker then
$ docker run --rm -it --name kubespray --mount type=bind,source="$(pwd)"/inventory/sample,dst=/inventory \
  --mount type=bind,source="${HOME}"/.ssh/id_rsa,dst=/root/.ssh/id_rsa \
  quay.io/kubespray/kubespray:v2.28.0 bash

You should create an ssh key on your local machine before running the Kubespray conainer.

Run the Kubespray playbooks

# Inside the container
root@bbaa7247daad:/kubespray# vi /inventory/inventory.ini
master ansible_host=<Pulbic IP> ip=<Private IP> etcd_member_name=etcd1 ansible_user=ubuntu 
 
[etcd:children]
kube_control_plane
 
[kube_node]
worker1 ansible_host=<Pulbic IP> ip=<Private IP> ansible_user=ubuntu 
worker2 ansible_host=<Pulbic IP> ip=<Private IP> ansible_user=ubuntu

You should modify the inventory.ini file as shown above.

# Copy the local public key
$ cat ~/.ssh/id_rsa.pub
$ ssh -i "your-key-pair.pem" ubuntu@<master-public-ip>
ubuntu@<master-public-ip>:~$ echo "Paste your public key" >> ~/.ssh/authorized_keys

You should add public key to each instance to authorize ssh access. Please do the same on the worker nodes.

# Inside the container
root@bbaa7247daad:/kubespray# ansible-playbook -i /inventory/inventory.ini --private-key /root/.ssh/id_rsa --become cluster.yml

ansible-playbook

If the playbook finishes without any errors, as shown above, your Kubernetes cluster has been successfully set up.

Check Master Node

$ ssh -i "your-key-pair.pem" ubuntu@<master-public-ip>
ubuntu@<master-public-ip>:~$ sudo kubectl version
Client Version: v1.32.5
Kustomize Version: v5.5.0
Server Version: v1.32.5
ubuntu@<master-public-ip>:~$ sudo kubectl get nodes
NAME      STATUS   ROLES           AGE   VERSION
master    Ready    control-plane   12m   v1.32.5
worker1   Ready    <none>          11m   v1.32.5
worker2   Ready    <none>          11m   v1.32.5

Boom! Success.

ubuntu@<master-public-ip>:~$ mkdir -p $HOME/.kube
ubuntu@<master-public-ip>:~$ sudo cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
ubuntu@<master-public-ip>:~$ sudo chown $(id -u):$(id -g) $HOME/.kube/config

If you want to use kubectl without sudo, follow the steps above.

ubuntu@<master-public-ip>:~$ kubectl get pod -n kube-system
NAME                                       READY   STATUS    RESTARTS   AGE
calico-kube-controllers-588d6df6c9-g5wx8   1/1     Running   0          13m
calico-node-2vrhd                          1/1     Running   0          14m
calico-node-l5bsl                          1/1     Running   0          14m
calico-node-pw95t                          1/1     Running   0          14m
coredns-5c54f84c97-8dr66                   1/1     Running   0          13m
coredns-5c54f84c97-spr5f                   1/1     Running   0          13m
dns-autoscaler-56cb45595c-prs5h            1/1     Running   0          13m
kube-apiserver-master                      1/1     Running   0          15m
kube-controller-manager-master             1/1     Running   1          15m
kube-proxy-2cpzg                           1/1     Running   0          14m
kube-proxy-bzpqx                           1/1     Running   0          14m
kube-proxy-h5zl7                           1/1     Running   0          14m
kube-scheduler-master                      1/1     Running   1          15m
nginx-proxy-worker1                        1/1     Running   0          14m
nginx-proxy-worker2                        1/1     Running   0          14m
nodelocaldns-g8kqw                         1/1     Running   0          13m
nodelocaldns-gwlfx                         1/1     Running   0          13m
nodelocaldns-j4jdm                         1/1     Running   0          13m
ubuntu@<master-public-ip>:~$ kubectl run apache --image httpd
ubuntu@<master-public-ip>:~$ kubectl run nginx --image nginx
ubuntu@<master-public-ip>:~$ kubectl get po -o wide
NAME     READY   STATUS    RESTARTS   AGE    IP               NODE      NOMINATED NODE   READINESS GATES
apache   1/1     Running   0          111s   10.233.125.2     worker2   <none>           <none>
nginx    1/1     Running   0          103s   10.233.105.131   worker1   <none>           <none>

Wrap Up

Creating the cluster was challenging, but I learned a lot about how Kubernetes works and how its components interact. Now, let's practice k8s.