Kubernetes Demos

Quick Start

./scripts/toolbox.sh

Pods

1. Interactive with cleanup on exit.

   Docker command.

   docker run -ti --rm ubuntu bash

   Equivalent kubernetes command.

   kubectl run -ti --rm --image ubuntu -- bash

   Look around and exit using exit or CTRL+D

   The pod will automatically be deleted because of the --rm. Handy for quick experimentation without forgetting to clean up after you.

2. All resources, including this pod, are reteivable through kubectl. In the background it simply makes an API call.

   kubectl <verb> <resource> [<resource name>]

   Run a new pod that logs the time every second. Notice the missing -ti --rm flags. This pod will just start and execute whatever you've told it to somewhere on the cluster.

   kubectl run --image ubuntu clock -- /bin/bash -c 'while true; do date; sleep 1; done'

   Go find your pod.

   kubectl get pods

   Different formats can be returned from the API using --output or -o. The different formats can be found in kubectl's --help system. It's very good, use it!

   Get some more information about your pod.

   kubectl get pod clock -o wide

   Get the YAML representation of your pod and save it to a file.

   kubectl get pod bash -o yaml > clock.yaml

3. These YAML manifests can be used as your source of truth when it comes to your workloads and their configration.

   Delete your pod.

   kubectl delete pod clock

   Bring it back using the manifest file.

   kubectl apply -f clock.yaml

   You can also delete any number of resources that are present in a file or folder. Delete the pod again using it's manifest file.

   kubectl delete -f clock.yaml

4. Nginx

   kubectl exec nginx -- curl localhost kubectl run -ti --rm --image busybox -- sh wget -q -O-

   kubectl apply -f sidecar.yaml kubectl get pod kubectl exec -ti sidecar -c busybox -- sh wget -q -O- localhost ip addr

   kubectl exec -ti sidecar -c caddy -- sh ip addr

Services

1. Pods get different IP addresses when they get created. To get a stable IP that you can use in the cluster you can make use of a Service.

   kubectl expose pod nginx --port 80 curl nginx-ip

   kubectl apply -f ./ kubectl get service

   curl red blue web

   node port load balancer

Namespaces

1. kubectl get ns

Deployments

1. kubectl create --help

   kubectl create deployment podinfo --image stefanprodan/podinfo:6.1.0 --port 9898 -o yaml --dry-run=client > deployment.yaml kubectl create service clusterip podinfo --tcp 9898 -o yaml --dry-run=client > service.yaml kubectl create ingress podinfo --rule podinfo.localhost/*=podinfo:9898 -o yaml --dry-run=client > ingress.yaml