Certified Kubernetes Administrator (CKA) Practice Exam - Questions
1. Create a pod called `nginx-secure` that uses the `nginx` image and runs in the `restricted`
namespace.
2. Create a deployment called `redis-deploy` with 2 replicas using the `redis:6.0` image and expose
it via ClusterIP service.
3. Set a resource limit of 200m CPU and 256Mi memory for a container in pod `nginx-limited` using
the `nginx` image.
4. Create a pod `nginx-probe` with readiness and liveness HTTP probes on port 80.
5. Display the node on which the pod `nginx-secure` is running.
6. Create a persistent volume called `pv-data` of size 1Gi using hostPath `/mnt/data`.
7. Create a PersistentVolumeClaim for `pv-data` and mount it in a pod running busybox.
8. Create a service account `appuser` and use it in a pod called `frontend`.
9. Taint a node `worker1` to only allow pods with the toleration
`key=dedicated,value=frontend:NoSchedule`.
10. Get all pods using more than 100Mi of memory.
11. Create a Job that runs `echo Hello CKA` and exits.
12. Patch deployment `web` to use image `nginx:1.18`.
13. Configure a network policy to allow ingress to pod label app=web only from pods with label
access=true.
14. Create an Ingress to expose `web-svc` service on `/web` path.
15. Upgrade a cluster node's kubelet configuration to use a custom `--cgroup-driver=systemd` flag.
16. Create a pod that mounts a ConfigMap with key `ENV=prod` as an environment variable.
17. Enable audit logging in the Kubernetes API server.
18. Backup etcd data to `/backup/etcd-snapshot.db`.
19. Restore etcd from a snapshot.
20. Create a Horizontal Pod Autoscaler for `web-deploy` between 1 and 5 replicas at 80% CPU
target.
21. Create a CronJob that runs every 5 minutes and echoes `Hi`.
22. Configure RBAC to allow `appuser` to list pods in `dev` namespace.
23. View API resources available in the cluster.
24. Show the control plane component status.
25. Force delete a pod stuck in terminating state.
26. Create a pod `curlpod` and use it to test service `web-svc` on port 80.
27. Create a pod with init containers that sleeps 10s before starting main container.
28. Create a deployment with anti-affinity rules to spread pods across nodes.
29. Create a pod that writes to a file every 30 seconds and test log rotation.
30. Debug why a pod `nginx-pod` is not starting.
Certified Kubernetes Administrator (CKA) Practice Exam - Answers
1. Create a pod called `nginx-secure` that uses the `nginx` image and runs in the `restricted`
namespace.
Answer:
kubectl create ns restricted
kubectl run nginx-secure --image=nginx -n restricted
2. Create a deployment called `redis-deploy` with 2 replicas using the `redis:6.0` image and expose
it via ClusterIP service.
Answer:
kubectl create deployment redis-deploy --image=redis:6.0 --replicas=2
kubectl expose deployment redis-deploy --port=6379 --target-port=6379 --type=ClusterIP
3. Set a resource limit of 200m CPU and 256Mi memory for a container in pod `nginx-limited` using
the `nginx` image.
Answer:
Create a YAML file with `resources.limits` section and apply using kubectl apply -f file.yaml
4. Create a pod `nginx-probe` with readiness and liveness HTTP probes on port 80.
Answer:
Define the pod spec with `readinessProbe` and `livenessProbe` using HTTP GET on port 80.
5. Display the node on which the pod `nginx-secure` is running.
Answer:
kubectl get pod nginx-secure -o wide -n restricted
6. Create a persistent volume called `pv-data` of size 1Gi using hostPath `/mnt/data`.
Answer:
Define a PersistentVolume YAML with capacity 1Gi and hostPath `/mnt/data`, then apply it.
7. Create a PersistentVolumeClaim for `pv-data` and mount it in a pod running busybox.
Answer:
Create a PVC and a pod that mounts the claim, using a shared volumeMounts path.
8. Create a service account `appuser` and use it in a pod called `frontend`.
Answer:
kubectl create sa appuser
Add `serviceAccountName: appuser` in the pod spec.
9. Taint a node `worker1` to only allow pods with the toleration
`key=dedicated,value=frontend:NoSchedule`.
Answer:
kubectl taint nodes worker1 dedicated=frontend:NoSchedule
10. Get all pods using more than 100Mi of memory.
Answer:
kubectl top pod --all-namespaces | awk '$4 > 100'
11. Create a Job that runs `echo Hello CKA` and exits.
Answer:
Define a Job YAML that runs busybox with command `echo Hello CKA`.
12. Patch deployment `web` to use image `nginx:1.18`.
Answer:
kubectl set image deployment/web nginx=nginx:1.18
13. Configure a network policy to allow ingress to pod label app=web only from pods with label
access=true.
Answer:
Define a NetworkPolicy with podSelector app=web and from: podSelector matchLabels access=true.
14. Create an Ingress to expose `web-svc` service on `/web` path.
Answer:
Define an Ingress resource mapping `/web` to `web-svc`.
15. Upgrade a cluster node's kubelet configuration to use a custom `--cgroup-driver=systemd` flag.
Answer:
Edit `/var/lib/kubelet/config.yaml` and restart kubelet with systemctl.
16. Create a pod that mounts a ConfigMap with key `ENV=prod` as an environment variable.
Answer:
kubectl create configmap env-cm --from-literal=ENV=prod
Use `envFrom` or `env` in pod spec.
17. Enable audit logging in the Kubernetes API server.
Answer:
Modify kube-apiserver manifest to include `--audit-log-path` and `--audit-policy-file`.
18. Backup etcd data to `/backup/etcd-snapshot.db`.
Answer:
ETCDCTL_API=3 etcdctl snapshot save /backup/etcd-snapshot.db --endpoints <endpoint> --cacert
--cert --key
19. Restore etcd from a snapshot.
Answer:
Stop etcd, use `etcdctl snapshot restore`, then point etcd to the restored data dir.
20. Create a Horizontal Pod Autoscaler for `web-deploy` between 1 and 5 replicas at 80% CPU
target.
Answer:
kubectl autoscale deployment web-deploy --min=1 --max=5 --cpu-percent=80
21. Create a CronJob that runs every 5 minutes and echoes `Hi`.
Answer:
kubectl create cronjob hi-job --schedule="*/5 * * * *" --image=busybox -- echo Hi
22. Configure RBAC to allow `appuser` to list pods in `dev` namespace.
Answer:
Create Role with list pods and bind it to `appuser` using RoleBinding.
23. View API resources available in the cluster.
Answer:
kubectl api-resources
24. Show the control plane component status.
Answer:
kubectl get componentstatus
25. Force delete a pod stuck in terminating state.
Answer:
kubectl delete pod <pod-name> --grace-period=0 --force
26. Create a pod `curlpod` and use it to test service `web-svc` on port 80.
Answer:
kubectl run curlpod --image=busybox -it --restart=Never -- curl web-svc:80
27. Create a pod with init containers that sleeps 10s before starting main container.
Answer:
Define initContainers section with sleep 10, then normal containers start.
28. Create a deployment with anti-affinity rules to spread pods across nodes.
Answer:
Use `podAntiAffinity` with `requiredDuringSchedulingIgnoredDuringExecution`.
29. Create a pod that writes to a file every 30 seconds and test log rotation.
Answer:
Pod uses busybox and writes to file in a loop. Set up logrotate manually.
30. Debug why a pod `nginx-pod` is not starting.
Answer:
kubectl describe pod nginx-pod
kubectl logs nginx-pod
