Lab 06 - Deployment lifecycle

On this lab you’ll create a new deployment and work with revisions.

On this lab

• Create your first Deployment
• Update deployment
• Rollback deployment

Create your first Deployment

Create a file named mydep-v1.yml and add the following content.

apiVersion: apps/v1
kind: Deployment
metadata:
  name: mydep
spec:
  replicas: 3
  selector:
    matchLabels:
      app: myapp
  template:
    metadata:
      labels:
        app: myapp
        version: v1
    spec:
      containers:
        - name: myapp
          image: nginx:1.17
          ports:
            - containerPort: 80

Let’s create this deployment on your cluster.

kubectl apply -f mydep-v1.yml

Check if everything is running properly. For that you need to check deployments, replica sets and pods.

kubectl get deployments

kubectl get rs

kubectl get pods --show-labels

After last command you must see an output similar with this.

NAME                     READY   STATUS    RESTARTS   AGE   LABELS
mydep-74c8d46695-jmrmm   1/1     Running   0          7s    app=myapp,pod-template-hash=74c8d46695,version=v1
mydep-74c8d46695-m8rgh   1/1     Running   0          7s    app=myapp,pod-template-hash=74c8d46695,version=v1
mydep-74c8d46695-stfk9   1/1     Running   0          7s    app=myapp,pod-template-hash=74c8d46695,version=v1

Pay attention on label version=v1 on all pods.

You have used nginx:1.17 as image for your container. Let’s use a newer version.

Edit mydep-v1.yml file and change the value of property image to nginx:1.18.

Run the apply command again to make this change on your deployment.

kubectl apply -f mydep-v1.yml

Let’s check the new pods. If you are fast enough you may see new pods to be created and older pods to be terminated.

kubectl get pods --show-labels

You may notice that new pods are being created (STATUSwith value ContainerCreating) and older pods are being terminated (STATUSwith value Terminating)

Take the name of one of the pods to check if the image was changed. Use that name on next command.

kubectl describe pod <POD_NAME>

Scroll up on the output and check the image used.

Containers:
  myapp:
    Container ID:   docker://c06e9537476ef35ed845cfac276cac1cbf2c5fe6a777376f7ea0300bc453c312
    Image:          nginx:1.18

You’re done with your first deployment.

Update deployment

Create 3 new files named mydep-v2.yml, mydep-v3.yml and mydep-v4.yml.

Use the same content that you used on file mydep-v1.yml changing only line 14 with following content:

• mydep-v2.yml file: version: v2
• mydep-v3.yml file: version: v3
• mydep-v4.yml file: version: v4

Now let’s apply each deployment and after the update of each one, check the labels on your pods.

Starting on file mydep-v2.yml.

kubectl apply -f mydep-v2.yml
kubectl get pods --show-labels

You must see version=v2on pod’s labels list.

Next file mydep-v3.yml.

kubectl apply -f mydep-v3.yml
kubectl get pods --show-labels

You must see version=v3on pod’s labels list.

Finally file mydep-v4.yml.

kubectl apply -f mydep-v4.yml
kubectl get pods --show-labels

You must see version=v4on pod’s labels list.

With this process you confirm how it’s easy to make updates on pods that are the workloads of your application.

Rollback deployment

Finally, let’s check how can we use the revisions that where created automatically to revert to previous versions.

First, let’s rollback to previous version.

kubectl rollout undo deploy mydep

To check if was properly applied, check pods label again.

kubectl get pods --show-labels

You must see version=v3on pod’s labels list.

Now let’s rollback to initial version.

kubectl rollout undo deploy mydep --to-revision=1

You must see version=v1on pod’s labels list.

And you’re done working with deployments. Let’s cleanup your cluster.

kubectl delete deployment mydep

Next you’ll see how you can work with this deployments as a unique object.

Next Lab: Managing services >

Return home >