Home → NOTES → CKAD Essential Tips & Commands

CKAD Essential Tips & Commands

What's Available in the Exam

# These are typically pre-configured in the exam environment:
alias k=kubectl
export do="--dry-run=client -o yaml"
export now="--force --grace-period 0"

# Check what's available when you start:
alias
env | grep -i kubectl

Quick Setup (If Needed)

# Only set essential ones if not already there:
alias k=kubectl
export do="--dry-run=client -o yaml"
# Don't waste time on more - just use these two!

Essential Must-Know Commands (Use 'k' instead of 'kubectl')

# Quick pod creation
k run nginx --image=nginx $do > pod.yaml

# Quick deployment creation
k create deploy nginx --image=nginx $do > deploy.yaml

# Quick service exposure
k expose pod nginx --port=80 --target-port=80 $do > svc.yaml

# Quick configmap creation
k create cm myconfig --from-literal=key=value $do > cm.yaml

# Quick secret creation
k create secret generic mysecret --from-literal=password=123 $do > secret.yaml

# Quick job creation
k create job myjob --image=busybox -- /bin/sh -c "sleep 30" $do > job.yaml

# Quick cronjob creation
k create cj mycron --image=busybox --schedule="*/1 * * * *" -- /bin/sh -c "date" $do > cj.yaml

Exam Speed Tips

1. Always Use Dry-Run First

# Generate YAML then edit instead of writing from scratch
k run pod1 --image=nginx $do > pod.yaml
vim pod.yaml
k apply -f pod.yaml

2. Use Tab Completion

# The exam environment supports tab completion
k get po<TAB>    # Completes to 'pods'
k get pods nginx<TAB>  # Completes pod names
k -n kube-sys<TAB>     # Completes namespace names

3. Quick Resource Management

# Scale quickly
k scale deploy nginx --replicas=3

# Delete quickly  
k delete pod nginx $now

# Get all resources
k get all

# Wide output for more info
k get pods -o wide

# Immediate commands (no YAML needed)
k run test --image=nginx --rm -it --restart=Never -- /bin/sh

Common Exam Scenarios

Multi-Container Pods

# Remember: containers share network and storage
# Use localhost for inter-container communication
# Check container logs: k logs podname -c containername

Resource Limits

# Always remember the format:
resources:
  requests:
    memory: "64Mi"
    cpu: "250m"
  limits:
    memory: "128Mi"
    cpu: "500m"

Environment Variables

# From literal
env:
- name: ENV_VAR
  value: "myvalue"

# From ConfigMap
env:
- name: ENV_VAR
  valueFrom:
    configMapKeyRef:
      name: myconfigmap
      key: mykey

# From Secret
env:
- name: PASSWORD
  valueFrom:
    secretKeyRef:
      name: mysecret
      key: password

Volume Mounts

# Remember the pattern:
volumeMounts:
- name: myvolume
  mountPath: /data
volumes:
- name: myvolume
  emptyDir: {}
# OR
- name: myvolume
  configMap:
    name: myconfigmap

Troubleshooting Quick Checks

# Pod not starting? Check these in order:
k describe pod podname     # Check events
k logs podname            # Check application logs
k get events --sort-by=.metadata.creationTimestamp
k get pod podname -o yaml # Check full configuration

Quick Reference Patterns

Liveness/Readiness Probes

livenessProbe:
  httpGet:
    path: /healthz
    port: 8080
  initialDelaySeconds: 30
  periodSeconds: 10

readinessProbe:
  httpGet:
    path: /ready
    port: 8080
  initialDelaySeconds: 5
  periodSeconds: 5

Security Context

securityContext:
  runAsUser: 1000
  runAsGroup: 3000
  fsGroup: 2000
  capabilities:
    add: ["NET_ADMIN"]
    drop: ["ALL"]

Service Types

# ClusterIP (default) - internal only
# NodePort - external via node port
# LoadBalancer - external via cloud LB

Memory Tricks

Port Numbers

HTTP: 80
HTTPS: 443
SSH: 22
Common app ports: 8080, 3000, 5000

Resource Units

CPU: 1000m = 1 core
Memory: 1Gi = 1024Mi

Restart Policies

Always (default for Deployments)
OnFailure (good for Jobs)
Never (good for one-time tasks)

Exam Day Strategy (CRITICAL!)

First 2 minutes: Check if k alias exists, if not: alias k=kubectl
Set namespace immediately if specified: k config set-context --current --namespace=exam
Don't waste time on complex aliases - just use k and muscle memory
Read question completely first
Use dry-run to generate YAML: k run test --image=nginx $do > test.yaml
Edit YAML file instead of writing from scratch
Test with k apply -f file.yaml
Verify with k get and k describe
Move to next question if stuck (mark and return)

Time Management

2 hours for ~15-20 questions = 6-8 minutes per question max
Don't spend more than 5-7 minutes per question initially
Mark difficult questions and return later
Always verify your solution works before moving on
Use k get all to see what you created
Speed comes from muscle memory, not fancy aliases

What You MUST Memorize (No Time to Look Up)

# These commands must be automatic:
k run name --image=nginx $do > file.yaml
k create deploy name --image=nginx $do > file.yaml  
k expose pod name --port=80 $do > file.yaml
k create cm name --from-literal=key=value $do > file.yaml
k create secret generic name --from-literal=key=value $do > file.yaml
k create job name --image=busybox -- command $do > file.yaml
k create cj name --image=busybox --schedule="* * * * *" -- command $do > file.yaml

# Quick checks:
k get pods -o wide
k describe pod name
k logs name
k delete pod name --force --grace-period=0

Reality Check for Exam

What's Usually Pre-Configured

alias k=kubectl (but verify!)
Basic bash completion
Standard Linux utilities

What You Must Type/Remember

export do="--dry-run=client -o yaml" (if not set)
All kubectl commands and options
YAML structure and indentation
Resource field names and values

Don't Waste Time On

Complex aliases (you won't remember them under pressure)
Bashrc modifications
Multiple export variables

Bottom Line: Practice with just k alias and $do variable. That's all you realistically have time to set up.

Quick Debugging (Memorize These Patterns)

# Pod stuck in Pending?
k describe pod podname | grep -i events

# Pod crashing?
k logs podname --previous

# Service not working?
k get endpoints servicename

# Can't schedule?
k get nodes
k describe node nodename

# Quick health check sequence:
k get pods                    # Overview
k describe pod podname        # Detailed events
k logs podname               # Application logs  
k get events --sort-by=.metadata.creationTimestamp  # Recent events

Last updated: 2025-08-26 20:00 UTC