Kubernetes - Jobs 👔

Description 👀

Jobs are a way to run a pod to completion. A job creates one or more pods and ensures that a specified number of them successfully terminate. As pods successfully complete, the job tracks the successful completions. When a specified number of successful completions is reached, the job itself is complete. Deleting a job will clean up the pods it created.

Examples of `jobs` 🧑‍💼

performing a one-time task such as running a script or setting up a work directory for a deployment
computations for a finite amount of data, such as analyzing a large data set or rendering a frames of a movie
and many other use cases where you want to run a pod to completion and then stop.

Basic `Commands` 📝

get jobs
```
  kubectl get jobs
```
output job logs
```
  kubectl logs <job-pod-name>
```

Creating a `Job` 🛠️

By default Kubernetes will restart a pod if it fails. This is not a desirable action for a job.

spec section of a pod that will not be reset

  spec:    
      # default is Always
      # restartPolicy: Always
      # set to Never to run a pod to completion
      restartPolicy: Never

sample job definition:

  apiVersion: batch/v1
  kind: Job
  metadata:
      name: math-add-job
  spec:
      # to run multiple of the same job at the same time set completions to a number greater than 1
      # by default the pods are made one after the other
      completions: 3
      template:
          spec:
              containers:
                - name: math-add
                  image: ubuntu
                  command: ["expr", "77", "+", "5"]
                  restartPolicy: Never

to run jobs in parallel set spec.parallelism to a number greater than 1

  apiVersion: batch/v1
  kind: Job
  metadata:
      name: math-add-job
  spec:
      completions: 3
      # to run jobs in parallel set parallelism to a number greater than 1
      parallelism: 3
      template:
          spec:
              containers:
                - name: math-add
                  image: ubuntu
                  command: ["expr", "77", "+", "5"]
                  restartPolicy: Never

Examples 🧩

sample job definition

  apiVersion: batch/v1
  kind: Job
  metadata:
      name: throw-dice-job
  spec:
      completions: 1
      template:
          metadata:
              name: throw-dice-pod
          spec:
              containers:
              - image: kodekloud/throw-dice
                name: throw-dice
              restartPolicy: Never

updated job definition to increase spec.backoffLimit from default of 6 to 10

  apiVersion: batch/v1
  kind: Job
  metadata:
      name: throw-dice-job
  spec:
      completions: 1
      template:
          metadata:
              name: throw-dice-pod
          spec: 
              containers:
              - image: kodekloud/throw-dice
                name: throw-dice
              restartPolicy: Never
      backoffLimit: 10

updated job to run as many times required for three completions

  apiVersion: batch/v1
  kind: Job
  metadata:
      name: throw-dice-job
  spec:
      completions: 3
      template:
          metadata:
              name: throw-dice-pod
          spec: 
              containers:
              -  image: kodekloud/throw-dice
                  name: throw-dice
              restartPolicy: Never 

sample output

  🚀  kubernetes-zero-to-pro ➜ kc get pods
  NAME                   READY   STATUS      RESTARTS   AGE
  throw-dice-job-2hf2r   0/1     Completed   0          119s
  throw-dice-job-4zt5w   0/1     Completed   0          96s
  throw-dice-job-8mljc   0/1     Completed   0          2m23s
  throw-dice-job-k4m2g   0/1     Error       0          2m13s
  throw-dice-job-lwfpl   0/1     Error       0          2m17s
  throw-dice-job-qbmzl   0/1     Error       0          110s
  throw-dice-job-rjz54   0/1     Error       0          114s
  throw-dice-pod         0/1     Error       0          38m

notice the time it took to run the job and the number of pods that were created

optimize the job to run three jobs in parallel

  apiVersion: batch/v1
  kind: Job
  metadata:
      name: throw-dice-job
  spec:
      completions: 3
      parallelism: 3
      template:
          metadata:
              name: throw-dice-pod
          spec: 
              containers:
              -  image: kodekloud/throw-dice
                  name: throw-dice
              restartPolicy: Never 