Learning Kubernetes with some kindness

I really enjoy being able to get familiar with some thing by locally installing it and monkeying with it a bit. Things that run in Kubernetes (k8s) and all the complications that come with it are no different. Having a convenient way to spin up and tear down a cluster easily makes learning about Cloud Native Computing Foundation (CNCF) projects much more fun. Here’s a short post that demonstrates how to create a four node k8s cluster.

Until recently, I had used vagrant to handle the cluster creation (see https://github.com/galexrt/k8s-vagrant-multi-node for an example of how that works). Although that works, it’s pretty resource intensive and as I was building out how infrastructure would play together it just became to much for my laptop to handle.

I’ve tried minikube a bit, but it just wasn’t the same experience (and not possible in some cases) when trying to segregate workloads.

A coworker did a demo with kind (Kubernetes IN Docker), https://kind.sigs.k8s.io/, and I was instantly sold on it. It uses docker containers as its k8s nodes, so it’s much easier on the system resources. Since it was made to test kubernetes, the clusters that it creates are multinode and compatible with everything I’ve tried to do with them so far. And it’s fast!

Requirements for running kind are minimal, you need to have a newer version of GoLang installed and, of course, docker. Installation is pretty straightforward and described pretty well here: https://kind.sigs.k8s.io/docs/user/quick-start#installation

Example of getting up and running with a 4 node ( 1 master / 3 worker ) cluster.

❯ kind create cluster --config cluster.yaml
Creating cluster "kind" ...
 ✓ Ensuring node image (kindest/node:v1.21.1) 
 ✓ Preparing nodes
 ✓ Writing configuration
 ✓ Starting control-plane 
 ✓ Installing CNI 
 ✓ Installing StorageClass 
 ✓ Joining worker nodes 
Set kubectl context to "kind-kind"
You can now use your cluster with:

kubectl cluster-info --context kind-kind

Thanks for using kind! 

❯ k get nodes
NAME                 STATUS   ROLES                  AGE     VERSION
kind-control-plane   Ready    control-plane,master   7m5s    v1.21.1
kind-worker          Ready    <none>                 6m32s   v1.21.1
kind-worker2         Ready    <none>                 6m32s   v1.21.1
kind-worker3         Ready    <none>                 6m32s   v1.21.1

This is the config I’ll be using in the next few posts. Here are the guts of the cluster.yaml file:

# a cluster with a control-plane node and 3 workers
kind: Cluster
apiVersion: kind.x-k8s.io/v1alpha4
nodes:
- role: control-plane
  kubeadmConfigPatches:
  - |
    kind: InitConfiguration
    nodeRegistration:
      kubeletExtraArgs:
        node-labels: "ingress-ready=true"
  extraPortMappings:
  - containerPort: 80
    hostPort: 80
    protocol: TCP
  - containerPort: 443
    hostPort: 443
    protocol: TCP
- role: worker
- role: worker
- role: worker
containerdConfigPatches:
- |-
  [plugins."io.containerd.grpc.v1.cri".registry.mirrors."localhost:5000"]
    endpoint = ["http://kind-registry:5000"]

When you’re done with it or goof it up so bad it needs to be rebuilt, getting rid of it is as easy as this

❯ kind delete cluster
Deleting cluster "kind" ...

Now you know how to create a local k8s cluster the kind way.