CKAD Exercises - Part 2

August 2, 2021 10-minute read

Pod Design

Labels, Selectors, Annotations

Exercise 1

List all pods with all their labels.

Show Solution

kubectl get pods --show-labels

Exercise 2

Create two pods with image nginx. Add the label env=blue to one of them and and env=green to the other. List all pods that have the app label.

Show Solution

kubectl run nginx-blue --image=nginx --restart=Never -l env=blue 
kubectl run nginx-green --image=nginx --restart=Never -l env=green
kubectl get pods -l env --show-labels

Exercise 3

Assuming there is a pod with label env=green, change the label to env=red.

Show Solution

kubectl label pod nginx-blue env=red --overwrite=true

Exercise 4

Assuming there is a pod with label env=blue, remove the value blue from this label.

Show Solution

kubectl label pod nginx-blue env= --overwrite=true

Exercise 5

Assuming there is a pod labeled with env=blue, remove this label from the pod.

Show Solution

kubectl label pod nginx-blue env-

Exercise 6

Add the label gpu=hardware to one of your cluster nodes. Create a pod with image busybox that will automatically be assigned to run on the previously labeled node.

TIP: Use nodeSelector in the pod specification to assign the pod to a node.

Show Solution

kubectl get nodes

kubectl label node <one-of-your-nodes-here> gpu=hardware

kubectl run gpu-app --image=busybox --restart=Never --dry-run=client -o yaml > pod.yaml 

kubectl create -f pod.yaml

Add a podSelector to pod.yaml:

apiVersion: v1
kind: Pod
metadata:
  creationTimestamp: null
  labels:
    run: gpu-app
  name: gpu-app
spec:
  containers:
  - image: busybox
    name: gpu-app
    resources: {}
  nodeSelector: 
    gpu: hardware
  dnsPolicy: ClusterFirst
  restartPolicy: Never
status: {}

Please note that a nodeSelector value is automatically interpreted as a label.

Annotations

Exercise 7

Annotate a pod with team=devops.

Show Solution

kubectl annotate pod nginx team=devops

Exercise 8

Find a pods annotations.

Show Solution

kubectl describe pod nginx
# Please find annotations in the "Annotations" field.

Exercise 9

Remove a pods team=devops annotation.

Show Solution

kubectl annotate pod nginx team-

Pod Design

Deployments

Exercise 10

Create a deployment with image nginx.

Show Solution

kubectl create deploy nginx-webapp --image=nginx:1.20

Exercise 11

Get the replicaset that was created with a deployment.

Show Solution

kubectl get replicaset -l app=nginx-webapp

Exercise 12

Scale an existing deployment to 3 replicas.

Show Solution

kubectl scale deploy nginx-webapp --replicas=3

Exercise 13

Find all pods of a deployment.

Show Solution

# get the labels from a deployment
kubectl get deploy --show-labels

# use the `app` label to find all deployments
kubectl get pod -l app=nginx-webapp

Exercise 14

Get a deployments rollout status.

Show Solution

kubectl rollout status deployment nginx-webapp

Exercise 15

Change the image version of a deployment. Confirm the image version for the deployment was set correctly. Confirm the pods are started with the new image.

Show Solution

kubectl set image deploy/nginx-webapp nginx=nginx:1.19
kubectl get deployment nginx-webapp -o wide # alternatively: kubectl describe deploy nginx-webapp | grep Image
# Confirm that column "IMAGES" contains "nginx:1.19" 
kubectl get pods -l app=nginx-webapp -o yaml | grep "image: nginx:1.19"
# If, e.g. there are 2 pods within the deployment, confirm that the image was listed 2 times.

One alternative to using the “kubectl set image” command is to use the “edit” command and edit the image inside a text editor instead: kubectl edit deploy nginx-webapp.

Exercise 16

Show the revision history for a deployment.

Show Solution

kubectl rollout history deploy nginx-webapp

Exercise 17

Verify that the rollout of the new image version was successful.

Show Solution

kubectl rollout history deploy nginx-webapp
kubectl rollout status deployment nginx-webapp
# Make sure that you see a positive confirmation: deployment "nginx-webapp" successfully rolled out.
# If there the deployment is still ongoing or there is a problem, you will usually see a message that makes this clear, 
# such as the following: 
# Waiting for deployment "nginx-webapp" rollout to finish: 1 of 3 updated replicas are available…
# or:
# error: deployment "nginx-webapp" exceeded its progress deadline

Exercise 18

Rollback a deployment to the previous version.

Show Solution

kubectl rollout undo deployment nginx-webapp

Exercise 19

Set the image of a deployment to a wrong value. Confirm that the deployment is in an error state.

Show Solution

Use the set image command to change the image:

kubectl set image deploy/nginx-webapp nginx=nginx:does-not-exist

TIP: Alternatively, you can also use kubectl edit deployment nginx-webapp to edit the image in an editor like vim.

Confirm that the rollout is still in a pending state (the corresponding pods should be in status ErrImagePull):

kubectl rollout status deploy nginx-webapp
# Waiting for deployment "nginx-webapp" rollout to finish: 1 old replicas are pending termination...

kubectl get pods -l app=nginx-webapp
# NAME                            READY   STATUS         RESTARTS   AGE
# nginx-webapp-7d85f4f584-v9sqc   0/1     ErrImagePull   0          7s

Exercise 20

Suppose there are 10 revisions of a deployment. Rollback all revisions up to revision 1.

Show Solution

kubectl rollout undo deployment/nginx-webapp --to-revision=1

NOTE: If you leave away parameter --to-revision then the undo command will only undo the latest revision.

Exercise 21

Suppose there are 10 revisions of a deployment. Show the history for revision 4.

Show Solution

kubectl rollout history deploy nginx-webapp --revision=4
# Output:
# deployment.apps/nginx-webapp with revision #4
# Pod Template:
#   Labels:        app=nginx-webapp
#   Containers:
#    nginx:
#     Image:       nginx:1.19
#     Port:        <none>
#     Host Port:   <none>
#     Environment: <none>
#     Mounts:      <none>
#   Volumes:       <none>

NOTE: Observe how the image is part of the output!

Exercise 22

Pause the rollout of a deployment.

Show Solution

kubectl rollout pause deploy nginx-webapp

Exercise 23

Resume the rollout of a deployment.

Show Solution

kubectl rollout resume deploy nginx-webapp

Exercise 24

Create an autoscaling rule for a deployment. Set the autoscaling CPU target to 70%. Use 2 as the minimum and 3 as the maximum number of replicas.

Show Solution

kubectl autoscale deploy webapp --cpu-percent=70 --min=2 --max=3

Make sure a HorizontalPodAutoscaler was successfully created.

kubectl autoscale deploy webapp --cpu-percent=70 --min=2 --max=3 
# Output:
# NAME           REFERENCE                 TARGETS         MINPODS   MAXPODS   REPLICAS   AGE
# nginx-webapp   Deployment/nginx-webapp   <unknown>/70%   2         3         0          7s

Verify that two pods have been created:

kubectl get pods -l app=nginx-webapp
# Output:
# NAME                          READY   STATUS    RESTARTS   AGE
# nginx-webapp-ccf759f6-czv7s   1/1     Running   0          24m
# nginx-webapp-ccf759f6-xr6r2   1/1     Running   0          24m

Exercise 25

Remove autoscaling for a deployment.

Show Solution

Delete the HorizontalPodAutoscaler that was created when the corresponding autoscaling rule has been created.

kubectl delete hpa nginx-webapp
# Output:
# horizontalpodautoscaler.autoscaling "nginx-webapp" deleted

Exercise 26

Delete a deployment.

Show Solution

kubectl delete deployment nginx-webapp

NOTE: Observe how the pods are being deleted with the deployment as well!

Jobs

Exercise 27

Create a job named envprinter with image busybox that prints all environment variables and then quits.

TIP: To print all environment variables, use the linux command env.

Show Solution

kubectl create job envprinter --image=busybox -- env
# Output: 
# job.batch/envprinter created

Exercise 28

Check the status of a job named envprinter.

Show Solution

kubectl get job envprinter
# Output:
# NAME         COMPLETIONS   DURATION   AGE
# envprinter   1/1           5s         1m14s

kubectl describe job envprinter
# Output: 
# ...
# Type    Reason            Age    From            Message
# ----    ------            ----   ----            -------
# Normal  SuccessfulCreate  4m40s  job-controller  Created pod: envprinter-rn272
# Normal  Completed         4m37s  job-controller  Job completed

TIP: You can add parameter -w to watch for changes (such as the completion status).

Exercise 29

Print the logs for a job with the name envprinter.

Show Solution

kubectl logs job/envprinter
# Output:
# PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin
# HOSTNAME=envprinter-crzbq
# KUBERNETES_PORT=tcp://10.245.0.1:443
# KUBERNETES_PORT_443_TCP=tcp://10.245.0.1:443
# KUBERNETES_PORT_443_TCP_PROTO=tcp
# KUBERNETES_PORT_443_TCP_PORT=443
# KUBERNETES_PORT_443_TCP_ADDR=10.245.0.1
# KUBERNETES_SERVICE_HOST=10.245.0.1
# KUBERNETES_SERVICE_PORT=443
# KUBERNETES_SERVICE_PORT_HTTPS=443
# HOME=/root

TIP: You can add parameter -f to follow the logs. In this case the logs will be streamed.

Exercise 30

Delete the job with the name envprinter.

Show Solution

kubectl delete job envprinter
# Output:
# job.batch "envprinter" deleted

Exercise 31

Create a job named greeter with image busybox that executes sh -c 'echo Hi; sleep 3; echo Bye;' and then exists. Let Kubernetes run it 10 times in a row (sequentially). Watch the execution progress. Delete the job after all executions completed.

Show Solution

kubectl create job greeter --image=busybox --dry-run=client -o yaml -- sh -c 'echo Hi; sleep 3; echo Bye;' > greeter.yaml

The above command will create the YAML file greeter.yaml. Open this file in a text editor like vim and set attribute spec.parallelism to value 10 like this:

apiVersion: batch/v1
kind: Job
metadata:
  creationTimestamp: null
  name: greeter
spec:
  parallelism: 10 # <--- Add this line to the YAML file!
  template:
    metadata:
      creationTimestamp: null
    spec:
      containers:
      - command:
        - sh
        - -c
        - echo Hi; sleep 3; echo Bye;
        image: busybox
        name: greeter
        resources: {}
      restartPolicy: Never
status: {}

Create the job using the spec from greeter.yaml:

kubectl create -f greeter.yaml 
# Output:
# job.batch/greeter created

Watch the progress:

kubectl get job/greeter -w
# Output:
# NAME      COMPLETIONS   DURATION   AGE
# greeter   0/10          3s         3s
# greeter   1/10          6s         6s
# greeter   2/10          11s        11s
# greeter   3/10          16s        16s
# greeter   4/10          21s        21s
# greeter   5/10          26s        26s
# greeter   6/10          31s        31s
# greeter   7/10          37s        37s
# greeter   8/10          42s        42s
# greeter   9/10          47s        47s
# greeter   10/10         52s        52s

IMPORTANT: Observe how we used parameter -w to watch for status changes. This command will then block and update the status regularly.

Delete the job:

kubectl delete job greeter
# Output:
# valuejob.batch "greeter" deleted

Exercise 32

Get all pods that belong to a job named greeter.

Show Solution

kubectl get pods -l job-name=greeter
# Output:
# NAME            READY   STATUS      RESTARTS   AGE
# greeter-4zkm7   0/1     Completed   0          9m8s
# greeter-5hzsv   0/1     Completed   0          9m14s
# greeter-cj22v   0/1     Completed   0          8m56s
# greeter-fq9lk   0/1     Completed   0          9m2s
# greeter-h7xkr   0/1     Completed   0          8m50s
# greeter-jt67x   0/1     Completed   0          9m26s
# greeter-kg7mj   0/1     Completed   0          9m46s
# greeter-qrzv6   0/1     Completed   0          9m20s
# greeter-wm9xx   0/1     Completed   0          9m32s
# greeter-wpxr8   0/1     Completed   0          9m38s

Exercise 33

Create a job named greeter with image busybox that executes sh -c 'echo Hi; sleep 3; echo Bye;' and then exists. Let Kubernetes run it 10 times in parallel. Watch the execution progress. Delete the job after all executions completed.

Show Solution

kubectl create job greeter --image=busybox --dry-run=client -o yaml -- sh -c 'echo Hi; sleep 3; echo Bye;' > greeter.yaml

The above command will create the YAML file greeter.yaml. Open this file in a text editor like vim and set attribute spec.completions to value 10 like this:

apiVersion: batch/v1
kind: Job
metadata:
  creationTimestamp: null
  name: greeter
spec:
  parallelism: 10 # <--- Add this line to the YAML file!
  template:
    metadata:
      creationTimestamp: null
    spec:
      containers:
      - command:
        - sh
        - -c
        - echo Hi; sleep 3; echo Bye;
        image: busybox
        name: greeter
        resources: {}
      restartPolicy: Never
status: {}

Create the job using the spec from greeter.yaml:

kubectl create -f greeter.yaml 
# Output:
# job.batch/greeter created

Watch the progress:

kubectl get job/greeter -w
# Output:
# NAME      COMPLETIONS   DURATION   AGE
# greeter   0/1 of 10     2s         2s
# greeter   1/1 of 10     7s         7s
# greeter   2/1 of 10     8s         8s
# greeter   3/1 of 10     8s         8s
# greeter   4/1 of 10     8s         8s
# greeter   5/1 of 10     8s         8s
# greeter   6/1 of 10     8s         8s
# greeter   7/1 of 10     8s         8s
# greeter   8/1 of 10     12s        12s
# greeter   9/1 of 10     12s        12s
# greeter   10/1 of 10    12s        12s

IMPORTANT: Observe how we used parameter -w to watch for status changes. This command will then block and update the status regularly.

Delete the job:

kubectl delete job greeter
# Output:
# valuejob.batch "greeter" deleted

CronJobs

Exercise 34

Create a cron job named time-printer that runs each minute to print the current date/time and then exit. Verify that the job actually runs each minute. Delete the job when you’re done.

TIP: You can use the unix command date to print the current date/time.

TIP: The cron expression for a 1 minute schedule is */1 * * * *.

Show Solution

Create the job:

kubectl create cronjob time-printer --image=busybox --schedule="*/1 * * * *" -- sh -c "date"
# Output:
# cronjob.batch/time-printer created

Watch new pods being created each minute:

kubectl get pods -w
# Output:
# NAME                            READY   STATUS        RESTARTS   AGE
# time-printer-1621548660-7vbgg   0/1     Completed     0          3m8s
# time-printer-1621548720-lgdk4   0/1     Completed     0          2m8s
# time-printer-1621548780-zg6mk   0/1     Completed     0          67s
# time-printer-1621548840-5t76k   0/1     Completed     0          7s
# time-printer-1621548660-7vbgg   0/1     Terminating   0          3m14s
# time-printer-1621548660-7vbgg   0/1     Terminating   0          3m14s
# time-printer-1621548900-wqgb6   0/1     Pending       0          0s
# time-printer-1621548900-wqgb6   0/1     Pending       0          0s
# ...

NOTE: Observe how we used parameter -w to watch the list of pods. This will block and report status updates.

Delete the job:

kubectl delete cronjob time-printer
# Output:
# cronjob.batch "time-printer" deleted

IMPORTANT: Observe how all pods that were created for the cron job are also being deleted with the cron job!

Table of Contents

Pod Design

Labels, Selectors, Annotations

Exercise 1

Exercise 2

Exercise 3

Exercise 4

Exercise 5

Exercise 6

Annotations

Exercise 7

Exercise 8

Exercise 9

Pod Design

Deployments

Exercise 10

Exercise 11

Exercise 12

Exercise 13

Exercise 14

Exercise 15

Exercise 16

Exercise 17

Exercise 18

Exercise 19

Exercise 20

Exercise 21

Exercise 22

Exercise 23

Exercise 24

Exercise 25

Exercise 26

Jobs

Exercise 27

Exercise 28

Exercise 29

Exercise 30

Exercise 31

Exercise 32

Exercise 33

CronJobs

Exercise 34