mfamador / k8s-demo

Kubernetes demo

Create a local Kubernetes cluster, e.g. k3d with 3 worker nodes

k3d cluster create -p 8080:80@loadbalancer --agents 3

Start the local Kubernetes cluster if you've created it before

k3d cluster start

Set the kubernetes context

kubectl config use-context k3d-k3s-default

or use kubectx

kubectx k3d-k3s-default

Create the demo namespace

kubectl create namespace demo

or

kubectl create -f 00-namepace.yaml

Set demo namespace by default

kubectl config set-context --current --namespace=demo

or use kubens (comes with kubectx)

kubens demo

Monitor all resources created in demo namespace

watch kubectl get all --namespace demo

Create a deployment with 3 pods

kubectl create -f 01-deployment.yaml

Create a service to load balance across the previously created pods

kubectl create -f 02-service.yaml

Make a request to the newly created service

create a port-forward directly to the service port:

kubectl port-forward svc/hello 8888:8000

and execute in another terminal

curl localhost:8888

Create an ingress to expose the service to the internet (needs an Ingress controller - K3d has a bundled Traefik)

kubectl create -f 03-ingress.yaml

Access hello service through the previously created ingress

curl -H "host:echo.com" localhost:8080

Should get the response You've hit hello-<ID> version:1

Look at the logs

kubectl logs -f deploy/hello

Scale the hello deployment

kubectl scale deployment hello --replicas 5

Look at the logs

kubectl logs -f deploy/hello

or use stern to see the logs from all the pods

stern hello

Do a rolling upgrade

Edit the deployment directly on k8s

kubectl edit deploy hello

Set image to marcoamador/hello:1.0.1

 curl -H "host:echo.com" localhost:8080

The response must be now You've hit hello-<ID> version:1.0.1

Create a configmap

kubectl create -f 04-configmap.yaml

Use the previously created configmap in a deployment

kubectl create -f 05-deployment-configmap.yaml

Connect to the previously created pod and check env variables

kubectl exec -it hello-configmap-<ID> -- sh

# env

or

kubectl exec hello-configmap-<ID> -- env

Create a secret and a deployment that attaches it

kubectl create -f 06-secret.yaml

kubectl create -f 07-deployment-secret.yaml

and "see" the secret:

kubectl exec hello-secret-<ID> -- env | grep SECRET_

Create a job and a cronjob

kubectl create -f 08-job.yaml

kubectl create -f 09-cronjob.yaml

Look at the logs

kubectl logs -f countdown-<ID>

or

kubectl logs -f job/countdown

Create a daemonset, a service and ingress

kubectl create -f 10-daemonset.yaml

kubectl create -f 11-service-ds.yaml

kubectl create -f 12-ingress-ds.yaml

curl -H "host:echo-ds.com" localhost:8080

Create a deployment with a sidecar

kubectl create -f 13-deployment-sidecar.yaml

see the logs from main container and sidecar in terminals side by side

sidecar cache proxy

kubectl logs -f deploy/hello-sidecar --container vache-proxy

the container with our API to be cached

kubectl logs -f deploy/hello-sidecar --container hello

or use stern

stern hello-sidecar --container vache-proxy
stern hello-sidecar --container hello

both container in same log:

stern hello-sidecar 

make some requests (repeat them to see that the requests are cached for 15 secs):

curl -H "host:echo-sidecar.com" http://localhost:8080/something/1

curl -H "host:echo-sidecar.com" http://localhost:8080/something/2

Kubernetes dashboard

Install dashboard

kubectl apply -f https://raw.githubusercontent.com/kubernetes/dashboard/v2.0.0/aio/deploy/recommended.yaml

Create admin user

kubectl apply -f 14-dashboard-admin-user.yaml

Get user token

kubectl -n kubernetes-dashboard describe secret $(kubectl -n kubernetes-dashboard get secret | grep admin-user | awk '{print $1}')

Access dashboard

kubectl proxy

open http://localhost:8001/api/v1/namespaces/kubernetes-dashboard/services/https:kubernetes-dashboard:/proxy/

Cleanup

Delete all created resources

./delete-all.sh

Stop local Kubernetes cluster

k3d cluster stop

Delete the local Kubernetes cluster

k3d cluster delete