Skip to content

Cluster Creation

We're nearly there. We installed all our prerequisites, and now we're ready to create our Kubernetes cluster. We will aim at an environment that mimics a real-world cluster as closely as possible - this implicates a few things:

  • we want more than one node - otherwise we won't see many of Kubernetes' strengths in action
  • we want proper connectivity from the outside

Let's get started with KinD's config.

KinD

KinD gets configured at the time of cluster creation by passing a configuration file. So, let's start by creating a new working directory for the remainder of this workshop to use.

mkdir k8s-workshop && cd k8s-workshop
vim kind-config.yaml

K8s?

K8s has become a common abbreviation for Kubernetes, as the word is quite long and hard to type. It is pronounced kates and you will see it throughout this workshop.

Similar phenomena include O11y (observability), or A11y (accessibility).

Configuration

We want to create the following configuration - feel free to expand the various tooltips for explanations of the different settings.

kind: Cluster
apiVersion: kind.x-k8s.io/v1alpha4
name: workshop-cluster
nodes: # (1)!
  - role: control-plane
    extraPortMappings: # (2)!
      - containerPort: 80
        hostPort: 80
        listenAddress: "127.0.0.1"
        protocol: TCP
      - containerPort: 443
        hostPort: 443
        listenAddress: "127.0.0.1"
        protocol: TCP
      - containerPort: 31000
        hostPort: 31000
        listenAddress: "127.0.0.1"
        protocol: TCP
    labels: # (3)!
        ingress-ready: "true"
  - role: worker
    extraPortMappings:
      - containerPort: 31000
        hostPort: 31001
        listenAddress: "127.0.0.1"
        protocol: TCP
  - role: worker
    extraPortMappings:
      - containerPort: 31000
        hostPort: 31002
        listenAddress: "127.0.0.1"
        protocol: TCP
  1. We want three nodes - one control plane member and two workers. This checks the more than one node box.

  2. We want to expose ports 80 and 443 on the host machine. KinD can do this by itself, unless we are running on Docker Desktop - so it's better to make sure 😉

    We also want to expose port 31000 on all nodes. We will need this extra port for one of the labs later in this workshop.

    This checks the proper connectivity box.

  3. We want to label our control plane node so that we can deploy an ingress controller on it later on.

Deployment

Once we've written our config file, we can create our cluster. This will take a few minutes if this is your first KinD cluster, so feel free to grab a coffee or a tea in the meantime.

kind create cluster --config kind-config.yaml

Deployment crashed?

Unfortunately, KinD runs into some edge cases due to Docker Desktop, OS networking stacks, and other factors here and there. I tried to keep the installation path as 'broad' as possible, but if your installation fails here, check out the troubleshooting section of the KinD documentation.

Verification

The cluster is installed, we did it! 🥳 Let's check if everything's alright:

kubectl config use-context kind-workshop-cluster
kubectl get nodes

Once entered, the output should look like this:

NAME                             STATUS   ROLES           AGE     VERSION
workshop-cluster-control-plane   Ready    control-plane   4m6s    v1.27.3
workshop-cluster-worker          Ready    <none>          3m42s   v1.27.3
workshop-cluster-worker2         Ready    <none>          3m47s   v1.27.3

✅ Our cluster is indeed running and reachable using kubectl, and all nodes display a status of Ready.

This is it, as far as setting things up is concerned. From now on, we will learn how to actually interact with a Kubernetes cluster, how to deploy workloads, and how to connect to them!

Last update: September 20, 2023
Created: September 14, 2023