Multi-node lightweight Kubernetes setup

Synopsis

Kubernetes, also known as K8s (because there are 8 letters between K and s), is an open-source system for automating deployment, scaling, and management of containerized applications.

Creating Kubernetes playgrounds can be costly for students and hobbyists who might not have access to organization-grade resources. Also, developers might want a playground K8s environment to test out their configurations or implement some proof-of-concepts. This article is targeted towards this audience.

In this article, I wanted to give a brief overview of how to create a lightweight Kubernetes environment (having 2 nodes) on an average laptop (Windows OS, Intel i5 processor, and 8 GB RAM). We will also deploy a demo application across both these nodes.

There are many tools available to create lightweight Kubernetes clusters such as K3s, Minikube, Kind, MicroK8s, K0s, etc. Each of these has its pros, cons, and use-cases. In this article, I am going to use K3s.

The purpose of this environment would be to understand the architecture, play around with configurations and commands while learning Kubernetes. This environment cannot be used to deploy full-fledged applications because of limited resources.

Prerequisites

1. The first thing we need is a virtualization platform. In this article, I will be using VMware but any platform can be used. The most common ones are VMware and VirtualBox.
2. Alpine Linux ISO. Alpine is a popular lightweight and stripped-down Linux distribution.

Configuration

Installing Alpine Linux

I won’t be covering the steps to install Alpine Linux in this article but it's fairly easy to do and only takes up to 20–30 mins. Follow the official installation guide, VMware OR VirtualBox. Make sure that you install Alpine Linux in System Disk Mode so that our RAM usage is not very high.

I recommend allocating 2 GB RAM and 2 processors to the VM.

We need 2 VMs in our environment, one will act as the master node and the second one will act as a worker node. To create the 2nd VM, you can just clone the existing VM or set up a new one as per your need.

Installing k3s

k3s is open-source certified Kubernetes distribution built for resource-constrained environments. k3s is production-ready and fully conformant. Check out k3s GitHub.

Let’s proceed with the installation. First, we need a k3s server (Kubernetes cluster master).

Installing k3s on Alpine is fairly simple. Run the following command on the VM you wish to keep as cluster master.

curl -sfL https://get.k3s.io | sh -

This command will install k3s and start the k3s service. To verify the installation, run the following command.

rc-udpate

k3s should be listed in the list of services

k3 is present in the list of running services

That’s it for installing the k3s server, now the next step will be to set up the k3s agent (worker) node.

Before we proceed with the agent installation, we need 2 values.

1. K3S_URL i.e. k3s server URL: Check your VM’s IP address using the ifconfig command. Your k3s server URL will be https://<ip-address>:6443
2. K3S_TOKEN aka K3S_CLUSTER_SECRET: This can be obtained by running cat /var/lib/rancher/k3s/server/node-token command on your cluster master.

Now let's login to our 2nd VM to install k3s in worker mode.

Run the following command,

curl -sfL https://get.k3s.io | K3S_URL=<your server url> K3S_TOKEN=<your server token> sh -

Once the k3s is installed and the service is running, you can verify the installation by running the rc-update command on the worker node.

As we can see, here k3-agent service is running which means k3s is running in the agent mode.

We can now run the kubectl get nodes command on the cluster master.

2 nodes in Ready state

Deploying a demo application

Now we are ready to deploy our application on the cluster we created.

You can clone this repo which contains the demo app we are going to deploy, https://github.com/lakhinsu/k3s-demo-app

Before we deploy this app, we need to build and publish our docker image on a docker repository. We will not cover this here but you can check out this Docker documentation.

Now that our image is published, you can replace your-image placeholder in the k8s.deployment.yaml file.

We need to create a deployment on our Kubernetes environment. Refer to the k8s.deployment.yamlfile.

Next, we need a Kubernetes service for our application. Refer to the k8.service.yaml file.

Last but not the least, we need to create an Ingress object in Kubernetes to allow outside access to our application. Refer to the k8s.ingress.yaml file.

To create these objects in our Kubernetes cluster, run the following commands on the Kubernetes master node.

• kubectl apply -f k8s.deployment.yaml
• kubectl apply -f k8s.service.yaml
• kubectl apply -f k8s.ingress.yaml

Once our application is deployed, you can run this command to check our deployment object. kubectl describe deployments

we have 2 available replicas of our app which is as expected

We can also check status of the pods on which our app is running, use the command kubectl get pods -o wide

pods status output

In the above screenshot, we can see that our pods are running on our 2 nodes, this is as expected because of the podAntiAffinity we specified in the k8s.deployment.yaml .

Accessing our application

Now that our application is deployed, let’s try and access it.

We can access our application at http://<k3s-server-url>:80/ using any REST client.

Response from our application

Wrapping up

Finally, now that we have successfully created a k3s cluster and deployed our application, let’s see what our resource usage looks like.

Screenshot from my laptop for resource usage of both the VMs

We can see that resource usage is very low for both our VMs. If you see heavy resource usage then make sure that the Alpine Linux is installed in the System Disk mode.

References

Listing down the documentation and articles I have referred.

• https://wiki.alpinelinux.org/wiki/Installation#Installation_Overview
• https://wiki.alpinelinux.org/wiki/Install_Alpine_on_VMware_Workstation
• https://docs.docker.com/engine/reference/commandline/build/
• https://rancher.com/docs/k3s/latest/en/

Thanks and Regards