CKA (Other exams) study plan

Install a three-node cluster with kubeadm on a local VM set, writing down every command in a personal runbook

Explain the role of each control-plane component (kube-apiserver, etcd, kube-scheduler, kube-controller-manager) and where each runs as a static pod

Back up and restore etcd using etcdctl snapshot save and snapshot restore, then verify the API server returns healthy after restore

Inspect and modify kubelet and static pod manifests by hand, then confirm the change takes effect after a kubelet restart

Perform a control-plane upgrade and a worker upgrade with kubeadm upgrade, including drain and uncordon

Rotate cluster certificates and identify expiry dates with kubeadm certs check-expiration

Configure RBAC Roles, ClusterRoles, RoleBindings and ServiceAccounts to grant least-privilege access, then prove denial with kubectl auth can-i

Add a second worker node by generating and using a join token, and remove a node cleanly

Create Deployments, DaemonSets, Jobs and CronJobs from scratch and from generated YAML without a browser

Apply nodeSelector, node affinity, pod affinity and anti-affinity, and taints with tolerations to force scheduling outcomes

Configure resource requests and limits and observe what happens to a pod that exceeds its memory limit

Perform rolling updates and rollbacks with kubectl rollout and record the revision history

Expose workloads with ClusterIP, NodePort, LoadBalancer and headless Services and explain when each is appropriate

Write NetworkPolicies that allow traffic from one labeled namespace only, and verify the block with a test pod

Configure an Ingress resource and confirm path-based routing to two backends

Trace a DNS lookup from inside a pod with nslookup and explain the CoreDNS service records involved

Create PersistentVolumes, PersistentVolumeClaims and StorageClasses and bind a claim to a specific volume

Mount a PVC into a Deployment and prove data survives pod deletion

Change a Deployment's access mode requirement and diagnose why the claim stays Pending

Use kubectl describe and events to debug a failed mount

Diagnose a Pending pod, a CrashLoopBackOff pod and an ImagePullBackOff pod using describe, logs --previous and events

Fix a broken kubelet on a worker node and get the node back to Ready

Repair a Service with a selector that matches no pods and a Service with the wrong targetPort

Read node conditions and correlate them with pod eviction behavior

Complete a full timed practice run of install and cluster maintenance tasks using only the Kubernetes documentation site

Practice switching context and namespace quickly with kubectl config use-context and --namespace to avoid wasted time

Rebuild a broken control-plane component from its static pod manifest under time pressure

Audit your own errors and list the three task types that cost you the most time

Run two full timed mock exams and score each against the domain weights

Drill the weakest task types from week 7 until they can be completed without documentation

Verify speed on imperative kubectl create and run commands for pods, deployments, services and configmaps

Finalize a one-page command cheat sheet you can review in the last hour before the exam