Phase 1: Local Cluster + kubectl Fluency

Time: week 1. Goal: a running cluster on your Mac and muscle memory with kubectl.

Install a local cluster

Recommendation: kind (Kubernetes-in-Docker, lightweight, disposable), or just enable Kubernetes in Docker Desktop / OrbStack since you already run Docker.

brew install kind kubectl k9s
kind create cluster --name k8dev
kubectl cluster-info

Install k9s too - it's a TUI for the cluster and you'll thank me later. You learn resource relationships just by navigating it.

The kubectl verbs that matter

These cover 95% of day-to-day usage:

kubectl get pods                  # list resources
kubectl describe pod <name>       # details + events (your #1 debugging tool)
kubectl logs <pod>                # container logs
kubectl exec -it <pod> -- sh      # shell into a container
kubectl apply -f file.yaml        # declare desired state
kubectl delete -f file.yaml       # remove it
kubectl port-forward svc/x 8080:80  # tunnel to a service
kubectl explain deployment.spec   # offline field docs, underrated

Kubeconfig and contexts

kubectl decides which cluster to talk to via ~/.kube/config. This matters the moment you have more than one cluster (local + work).

kubectl config get-contexts
kubectl config use-context kind-k8dev

Exercise

Full playbook with the kind config, expected output, and verify checklist: Lab 1. Use the lab - it creates the cluster with the port mappings that Lab 4 needs.

Run an nginx pod imperatively, poke at it, then delete it:

kubectl run web --image=nginx
kubectl get pods -w
kubectl port-forward pod/web 8080:80   # open http://localhost:8080
kubectl logs web
kubectl exec -it web -- sh
kubectl delete pod web

Notice the pod does not come back after deletion. That's the lesson leading into Phase 2: bare pods have no controller watching them. Nothing declared that this pod should exist, so nothing recreates it.

Checkpoint

• You can create and destroy a kind cluster without looking anything up.
• You can get, describe, log, and exec against a pod from memory.
• You know what file kubectl reads to find your cluster.

Next: Phase 2: Pods, Deployments, Services