readme.md

Pods

Start a pod (imperative way):

kubectl run nginx --image nginx:1.16.0

Create a pod (declarative way):

kubectl apply -f 2.deployments/pod.yaml

Lets get the pod ip:

kubectl get po -o wide

In a separate terminal, get a shell inside the cluster and lets contact the above web apps:

kubectl run -it alpine --image=alpine:3.19 --restart=Never /bin/sh

Inside the pod shell run the following commands

apk add curl

curl http://<pod-ip>:3000

curl http://<pod-ip>:3000/callforpapers

Replicaset

Create the replicaset:

kubectl apply -f 2.deployments/replicaset.yaml

Delete one of the pods and see what happens:

kubectl delete pod <pod-name>

Deployments

Create the deployment, rollout strategy:

kubectl apply -f 2.deployments/deployment-rollout.yaml

Check the new deployment, the replicaset created underneath and the pods:

kubectl get deploy

kubectl get rs

kubectl get po

How it works:

Rollout strategy

Apply a new image so we can see the rollout in action

kubectl set image deployment.v1.apps/nginx-rollout nginx=nginx:1.16.1

Watch how the pods rollout progresses

kubectl get po -w

See the switch of the replicasets:

kubectl get rs

Apply an image that doesn't exist and see how the pods rollout is blocked

kubectl set image deployment.v1.apps/nginx-rollout nginx=nginx:1.161

Recreate strategy

Create another deployment with recreate strategy:

kubectl apply -f 2.deployments/deployment-recreate.yaml

Check how the pods are being affected in a separate window:

kubectl get po -w

Apply a new image so we can see the recreate in action

kubectl set image deployment.v1.apps/nginx-recreate nginx=nginx:1.16.2

Apply an image that doesn't exist and see how the pods recreate is done whatever the consequences

kubectl set image deployment.v1.apps/nginx-recreate nginx=nginx:1.161

What about canary deployments or A/B testing?