Using Jobs

The purpose of the Job resource in Kubernetes is to run tasks that can complete, which makes them not ideal for long-running applications, but great for batch jobs or similar tasks that can benefit from parallelism.

Here's what a Job spec YAML looks like:

job-1.yaml

apiVersion: batch/v1
kind: Job
metadata:
  name: runner
spec:
  template:
    spec:
      containers:
      - name: run-job
        image: node:lts-jessie
        command: ["node", "job.js"]
      restartPolicy: Never
  backoffLimit: 4

This Job will start a single Pod, and run a command, node job.js, until it completes, at which point the Pod will shut down. In this and the future examples, we assume that the container image used has a file, job.js, that runs...