Beginner’s Guide : Simple hello Kubernetes (all-in-one) on a single VPS

5 min readNov 23, 2022

https://www.leverege.com/iot-ebook/from-docker-to-kubernetes

Kubernetes, also known as K8s, is an open-source system for automating deployment, scaling, and management of containerized applications. In this tutorial I’m using kubeadm to bootstrap a single-node Kubernetes cluster and then taint / force the control plane’s scheduler to schedule our pod’s deployment on control plane node. If u want to try kubernetes on your local PC, you can try using minikube.

VPS Requirements

https://kubernetes.io/docs/setup/production-environment/tools/kubeadm/install-kubeadm/

You can use https://idcloudhost.com as your VPS provider. Their VPS’s cheap (estimated IDR 87,000 / month or $6 / month) option meets our minimum requirements. For 2 CPU cores, 2GB RAM, and 20GB storage, you just need to pay IDR 120/hour or $0.0083/hour.
The payment method is pay as you go , means pay per hour. You can stop the VPS anytime to reduce the charge and if you stop the VPS, you will be charged for storage service only. You also can delete the VPS service that you just built, so you won’t be charged at all.

In this tutorial, I’m using Ubuntu 18.04 as my VPS’s operating system

Objectives

We will setup a single-node Kubernetes cluster on our VPS, deploy a simple nodejs “Hello, World!” app to our Kubernetes, and finally be able to access it via our VPS’s public IP

Setup Kubernetes and Hello World App

Here are all the complete steps to set up a single-node Kubernetes cluster and make it accessible from outside of our VPS network. Please don’t leave any steps, and follow them step by step.

Disable Swap

We MUST disable swap in order for the kubelet to work properly.
https://discuss.kubernetes.io/t/swap-off-why-is-it-necessary/6879/4

First disable swap

sudo swapoff -a

And then to disable swap on startup in /etc/fstab

sudo sed -i '/ swap / s/^\\(.*\\)$/#\\1/g' /etc/fstab

Install docker

If you already install docker via install -y docker.io uninstall it first.

And then, follow these tutorial https://docs.docker.com/engine/install/ubuntu/

Install cri-dockerd

Reference https://kubernetes.io/docs/setup/production-environment/container-runtimes/#docker

Complete all the steps by following the tutorial on the link below.

Install kubelet, kubeadm, and kubectl

Follow all the steps (in the link above) until you get this output at the final steps

Init kubeadm with calico network provider and cri-dockerd cri socket

First run

sudo su

then

kubeadm init --pod-network-cidr=192.168.0.0/16 --cri-socket=unix:///var/run/cri-dockerd.sock

After completely initiated kubeadm, then run the followings

mkdir -p $HOME/.kube

sudo cp -i /etc/kubernetes/admin.conf $HOME/.kube/config

sudo chown $(id -u):$(id -g) $HOME/.kube/config

Install calico

https://projectcalico.docs.tigera.io/getting-started/kubernetes/self-managed-onprem/onpremises

Our Kubernetes is going to be ready

Run kubectl get pods -A --watch to watch our pods status in a realtime

Wait for a minute until all Pod’s STATUS is Running and the Ready value is 1/1. After a minute, let’s check whether our Kubernetes is ready or not by running the following command:

kubectl get pods -A

Try run kubectl get nodes . If the STATUS value is Ready, then our Kubernetes is ready to be used.

Deploy simple nodejs “Hello, World!” application to our Kubernetes

Run the following command to deploy simple “Hello, World!” application that runs inside the Kubernetes pod. reinhardjs/node-hello:1.0.0 is the docker image that I pushed to the docker hub.

kubectl create deployment hello-node --image=reinhardjs/node-hello:1.0.0

You can set whatever you want for the pod’s name, in this case, I’m using hello-node as my pod’s name. Remember, it’s not Kubernetes node, but pod. We are using our VPS, as a single node cluster that will run the Kubernetes cluster (control-plane node) and run all of our pods on top of it. Instead of setting up multiple VPS to implement multi node (expensive), we just use a single Kubernetes cluster (all-in-one).

Then try to run kubectl get pods. You will get your pod status always (stuck) in pending status. Our application will never be deployed, because of this, try run kubectl describe pods

As the docs said:

By default, your cluster will not schedule Pods
on the control plane nodes for security reasons
https://kubernetes.io/docs/setup/production-environment/tools/kubeadm/create-cluster-kubeadm/#control-plane-node-isolation

So, to be able to use the Kubernetes cluster as our pod’s container, we need to taint the control-plane node

kubectl taint nodes — all node-role.kubernetes.io/control-plane-

After that, our Kubernetes will automatically try to deploy our node “Hello, World!” application

Wait until our application is running on top of created pods

In order to be able to access the pod from outside VPS’s network (remote), we need to expose our hello-node pod. We are using Kubernetes’s load balancer to expose our pod.

The port 3000 below is the running port of our nodejs hello application

kubectl expose deployment hello-node --type=LoadBalancer --port=3000

Now, run the command below to get the list of running services

kubectl get services

And voila! We can access the simple “Hello, World!” application via our VPS’s public IP with the port of 30472, created by Kubernetes’s load balancers randomly.