CKAD — Understanding Jobs

2 minute read

What Is a Job in Kubernetes?

A Job creates one or more Pods and ensures they run to completion. Unlike Deployments (which keep Pods running forever), Jobs are designed for batch processing — tasks that start, do their work, and exit.

This topic falls under the Application Design and Build domain (20% of the CKAD exam).

Real-World Use Cases

Use Case Example
Database migrations Apply schema changes before deploying a new app version
Data processing / ETL Transform and load data from one system to another
Batch emails Send weekly newsletter to all subscribers
Report generation Generate monthly billing reports
Test suites Run integration tests as part of CI/CD
Cleanup tasks Delete old temporary files or expired records

Jobs vs Deployments

Feature Job Deployment
Purpose Run to completion, then exit Run forever
Restart behavior Does not restart after success Restarts if Pod dies
Pod lifecycle Finite (has an end) Infinite (runs continuously)
Use case Batch processing, one-time tasks Web servers, APIs, long-running services
restartPolicy Never or OnFailure Always (default)

Creating a Basic Job

Imperative (Fast for Exams)

kubectl create job backup-job --image=busybox:1.36 -- /bin/sh -c "echo 'Backup completed at $(date)'"

This is the fastest way to create a Job during the CKAD exam. The -- separates kubectl arguments from the container command.

Declarative (YAML)

apiVersion: batch/v1
kind: Job
metadata:
  name: backup-job
spec:
  template:
    spec:
      containers:
      - name: backup
        image: busybox:1.36
        command: ["/bin/sh", "-c"]
        args: ["echo 'Backup started' && sleep 5 && echo 'Backup completed'"]
      restartPolicy: Never

kubectl apply -f backup-job.yaml

# Watch the Job progress
kubectl get jobs -w
# Expected:
# NAME         COMPLETIONS   DURATION   AGE
# backup-job   0/1           5s         5s
# backup-job   1/1           10s        10s

# Check the logs
kubectl logs job/backup-job
# Expected:
# Backup started
# Backup completed

Job YAML Structure

apiVersion: batch/v1
kind: Job
metadata:
  name: <job-name>
spec:                          ← Job spec
  completions: 1               ← How many successful Pods needed
  parallelism: 1               ← Max Pods running at once
  backoffLimit: 6              ← Max retries on failure
  activeDeadlineSeconds: 300   ← Total time limit
  ttlSecondsAfterFinished: 60  ← Auto-delete after completion
  template:                    ← Pod template (same as Deployments)
    spec:
      containers:
      - name: <container>
        image: <image>
        command: [...]
      restartPolicy: Never     ← Required: Never or OnFailure

Important: Jobs require restartPolicy: Never or restartPolicy: OnFailure. You cannot use Always (the default) because the Pod needs to be able to exit.

Generating Job YAML Quickly

On the CKAD exam, generate a YAML template and edit it:

# Generate YAML without creating the Job
kubectl create job my-job --image=busybox:1.36 --dry-run=client -o yaml -- /bin/sh -c "echo hello" > job.yaml

# Edit as needed
vi job.yaml

# Apply
kubectl apply -f job.yaml

Exam tip: Always use kubectl create job ... --dry-run=client -o yaml to generate a starting template. It saves time compared to writing YAML from scratch.

Next: Learn about running multiple Pods with Completions & Parallelism →

Updated:

You may also enjoy