• export DOCKER_ID_USER="username"

• docker pull $DOCKER_ID_USER/hello_node
• docker run -d -p 8080:8080 $DOCKER_ID_USER/hello_node
• docker ps to see what's running (node the CONTAINER ID)
• open http://localhost:8080/
• docker stop <CONTAINER ID>

• docker login

• docker build -t $DOCKER_ID_USER/hello_node .
• docker push $DOCKER_ID_USER/hello_node

• kubectl run hello-node --image=index.docker.io/$DOCKER_ID_USER/hello_node --port=8080
• Check the deployment is there with kubectl get deployments
• Check the pods are up with kubectl get pods
• Check it's running properly before exposure with kubectl exec -ti hello-node-7ddff5778f-nt2g6 curl localhost:8080
• Expose with kubectl expose deployment/hello-node --type="NodePort" --port=8080
• Check with kubectl get services
• Get a full description with kubectl describe services/hello-node
• Pull the NodePort into an environment variable for easier use with export NODE_PORT=$(kubectl get services/hello-node -o go-template='{{(index .spec.ports 0).nodePort}}') && echo NODE_PORT=$NODE_PORT
• Open the app in a web browser: http://<IP_OF_ANY_NODE>:<NODE_PORT>/
• Clean up with kubectl delete service,deployment hello-node

• kubectl apply -f kubernetes.yaml (apply is impdepotent so use instead of create which is just for creating). Using 1 file so the service is created first which causes the container's env vars to contain service info plus scheduling should spread them between hosts.

The interactive tutorials on kubernetes.io are pretty good.