Skip to main content

​
Concept and Usage

A Job creates one or more pods and is considered complete when all the pods inside it complete successfully. When you delete a job, the pods created by the job are deleted too. Use cases:
  • Running a batch job
  • Backup operations

​
Job YAML

job.yaml
apiVersion: batch/v1
kind: Job
metadata:
  name: backup-job
spec:
  completions: 3    # Run until 3 successful pod completions
  parallelism: 2    # Run 2 pods in parallel at a time
  template:         # Pod template
    spec:
      containers:
        - name: backup
          image: busybox
          command: ["echo", "Backup operation"]
      restartPolicy: Never
  • completions — The job runs until this many pods complete successfully.
  • parallelism — How many pods to run in parallel at a time.
  • restartPolicy: Never — If a pod fails, it will not be restarted. Instead, a new pod is created to replace it.
If some pods fail, the job will create new pods to replace the failed ones until the specified number of completions is achieved. The job keeps retrying until all pods have successfully completed or the job is deleted.

​
Commands

kubectl get jobs
kubectl delete job <job-name>

Build docs developers (and LLMs) love

Get started for freeTalk to us