CKA Specialization Unit 2 || Workloads, Deployments, Helm and Scheduling

Lab Environment:

  • Control Plane: k8sm
  • Worker Nodes: k8sw1, k8sw2
  • Kubernetes: v1.35
  • Lab Namespace: cka-unit2

Task 1: Explore Kubernetes Workloads
Objective: Identify Deployments, ReplicaSets, StatefulSets, DaemonSets, Jobs and CronJobs.

kubectl get all -A
kubectl get deployments -A
kubectl get replicasets -A
kubectl get statefulsets -A
kubectl get daemonsets -A
kubectl get jobs,cronjobs -A
kubectl explain deployment
kubectl explain deployment.spec

Exercise: Explain the difference between Deployment, StatefulSet, DaemonSet and Job.

Task 2: Create and Manage Deployments
Objective: Deploy Nginx and manage application replicas.

kubectl create namespace cka-unit2
kubectl create deployment nginx –image=nginx:stable –replicas=3 -n cka-unit2
kubectl get deployments -n cka-unit2
kubectl get pods -n cka-unit2 -o wide
kubectl get rs -n cka-unit2
kubectl describe deployment nginx -n cka-unit2
kubectl scale deployment nginx –replicas=5 -n cka-unit2
kubectl get pods -n cka-unit2

Expected: Five Nginx Pods running under one Deployment.
Exercise: Scale the Deployment down to two replicas.

Task 3: Rolling Updates and Rollbacks
Objective: Update application images and restore previous versions.

kubectl set image deployment/nginx nginx=nginx:alpine -n cka-unit2
kubectl rollout status deployment/nginx -n cka-unit2
kubectl rollout history deployment/nginx -n cka-unit2
kubectl rollout undo deployment/nginx -n cka-unit2
kubectl rollout restart deployment/nginx -n cka-unit2
kubectl get deployment nginx -n cka-unit2 -o jsonpath='{.spec.template.spec.containers[*].image}’

Exercise: Deploy an invalid image, identify the failure and roll back.

Task 4: Deployment Strategies and Resource Management
Objective: Configure CPU, memory and rolling-update parameters.

kubectl set resources deployment nginx –requests=cpu=100m,memory=64Mi –limits=cpu=500m,memory=256Mi -n cka-unit2
kubectl describe deployment nginx -n cka-unit2
kubectl top pods -n cka-unit2
kubectl patch deployment nginx -n cka-unit2 –type=strategic -p ‘{“spec”:{“strategy”:{“type”:”RollingUpdate”,”rollingUpdate”:{“maxSurge”:1,”maxUnavailable”:0}}}}’
kubectl get deployment nginx -n cka-unit2 -o yaml

Exercise: Explain maxSurge, maxUnavailable, requests and limits.

Task 5: ConfigMaps and Secrets
Objective: Manage application configuration and sensitive values.

kubectl create configmap app-config –from-literal=APP_MODE=production -n cka-unit2
kubectl create secret generic app-secret –from-literal=DB_PASSWORD=test123 -n cka-unit2
kubectl get configmaps -n cka-unit2
kubectl get secrets -n cka-unit2
kubectl describe configmap app-config -n cka-unit2
kubectl describe secret app-secret -n cka-unit2
kubectl set env deployment/nginx –from=configmap/app-config -n cka-unit2
kubectl exec -n cka-unit2 deployment/nginx — printenv APP_MODE

Exercise: Create a ConfigMap containing two environment variables and verify them inside a Pod.

Task 6: Helm Package Manager
Objective: Install Kubernetes applications using Helm.

helm version
helm list -A
helm repo add bitnami https://charts.bitnami.com/bitnami
helm repo update
helm search repo nginx
helm show chart bitnami/nginx
helm show values bitnami/nginx
helm install webapp bitnami/nginx -n cka-unit2
helm list -n cka-unit2
helm status webapp -n cka-unit2
helm get manifest webapp -n cka-unit2
kubectl get pods,svc -n cka-unit2

Exercise: Identify the Helm release name, chart version and application version.
Note: Helm chart availability and container-image access may vary.

Task 7: Helm Upgrades and Rollbacks
Objective: Upgrade, inspect and restore Helm releases.

helm show values bitnami/nginx > values.yaml
helm get values webapp -n cka-unit2
helm upgrade webapp bitnami/nginx -n cka-unit2 –set service.type=ClusterIP
helm history webapp -n cka-unit2
helm rollback webapp 1 -n cka-unit2
helm status webapp -n cka-unit2
helm template webapp bitnami/nginx -n cka-unit2 > rendered.yaml
helm uninstall webapp -n cka-unit2

Exercise: Upgrade a Helm release and roll back to revision 1.

Task 8: Kubernetes Scheduling
Objective: Understand how the scheduler selects nodes.

kubectl get nodes -o wide
kubectl get pods -A -o wide
kubectl describe node k8sw1
kubectl describe node k8sw2
kubectl get nodes –show-labels
kubectl describe nodes | grep -A 8 ‘Allocated resources’
kubectl get pods -n kube-system -l component=kube-scheduler
kubectl logs -n kube-system kube-scheduler-k8sm –tail=50
kubectl get events -A –sort-by=.metadata.creationTimestamp

Exercise: Identify which worker nodes host the Nginx replicas and inspect their scheduling events.

Task 9: NodeSelector and Node Affinity
Objective: Schedule workloads onto specific nodes.

kubectl label node k8sw1 disktype=ssd
kubectl get nodes –show-labels
kubectl run node-test –image=nginx:stable -n cka-unit2 –dry-run=client -o yaml > node-test.yaml

Edit node-test.yaml and add under spec:
nodeSelector:
disktype: ssd

kubectl apply -f node-test.yaml
kubectl get pod node-test -n cka-unit2 -o wide
kubectl describe pod node-test -n cka-unit2

Exercise: Configure required node affinity for k8sw1 and compare it with nodeSelector.

Task 10: Taints and Tolerations
Objective: Restrict workload scheduling on worker nodes.

kubectl taint nodes k8sw2 dedicated=testing:NoSchedule
kubectl describe node k8sw2 | grep Taints
kubectl get nodes -o wide

Create toleration.yaml:
apiVersion: v1
kind: Pod
metadata:
name: toleration-test
namespace: cka-unit2
spec:
nodeSelector:
kubernetes.io/hostname: k8sw2
tolerations:

  • key: dedicated
    operator: Equal
    value: testing
    effect: NoSchedule
    containers:
  • name: nginx
    image: nginx:stable

kubectl apply -f toleration.yaml
kubectl get pod toleration-test -n cka-unit2 -o wide
kubectl taint nodes k8sw2 dedicated=testing:NoSchedule-

Exercise: Explain NoSchedule, PreferNoSchedule and NoExecute.

Task 11: Pod Affinity and Anti-Affinity
Objective: Control Pod placement relative to other Pods.

kubectl get pods -n cka-unit2 -o wide
kubectl get nodes -L kubernetes.io/hostname
kubectl edit deployment nginx -n cka-unit2

Add under spec.template.spec:
affinity:
podAntiAffinity:
requiredDuringSchedulingIgnoredDuringExecution:
– labelSelector:
matchLabels:
app: nginx
topologyKey: kubernetes.io/hostname

kubectl get pods -n cka-unit2 -o wide
kubectl describe deployment nginx -n cka-unit2
kubectl get events -n cka-unit2 –sort-by=.metadata.creationTimestamp

Exercise: Verify that matching replicas run on different eligible nodes. Extra replicas may remain Pending if there are insufficient nodes.

Task 12: Cordon, Drain and Uncordon
Objective: Perform Kubernetes node maintenance.

kubectl get nodes
kubectl cordon k8sw1
kubectl get nodes
kubectl drain k8sw1 –ignore-daemonsets –delete-emptydir-data
kubectl get pods -A -o wide
kubectl uncordon k8sw1
kubectl get nodes

Exercise: Explain the differences between cordon, drain and uncordon.
Warning: Run drain only on a disposable lab worker, as it can disrupt running applications.

Task 13: Troubleshoot Pending Pods
Objective: Identify scheduling failures.

kubectl get pods -A –field-selector=status.phase=Pending
kubectl describe pod node-test -n cka-unit2
kubectl get events -n cka-unit2 –sort-by=.metadata.creationTimestamp
kubectl describe nodes
kubectl get nodes –show-labels
kubectl describe nodes | grep -i taints
kubectl top nodes

Exercise: Create a Pod with a nonexistent nodeSelector label, diagnose FailedScheduling and correct the configuration.

Task 14: Workload Troubleshooting
Objective: Investigate unhealthy Pods and application failures.

kubectl get pods -A –field-selector=status.phase!=Running,status.phase!=Succeeded
kubectl get deployments -A
kubectl get replicasets -A
kubectl describe pods -n cka-unit2
kubectl logs deployment/nginx -n cka-unit2
kubectl rollout status deployment/nginx -n cka-unit2
kubectl get events -n cka-unit2 –sort-by=.metadata.creationTimestamp
kubectl top pods -n cka-unit2

Exercise: Simulate ImagePullBackOff, inspect Pod events and restore a working image.

Task 15: Cleanup and Cluster Verification
Objective: Remove lab resources and restore normal scheduling.

kubectl delete namespace cka-unit2
kubectl label node k8sw1 disktype-
kubectl taint nodes k8sw2 dedicated=testing:NoSchedule-
kubectl uncordon k8sw1
kubectl uncordon k8sw2
kubectl get nodes
kubectl get pods -n kube-system
kubectl get namespaces
kubectl get pods -A -o wide

Exercise: Verify that all three nodes are Ready, worker nodes are schedulable and the lab namespace has been removed.

Leave a Reply

Your email address will not be published. Required fields are marked *