kubejob

Run Kubernetes jobs and wait for their completion.

Motivation

Sometimes you just want to run a job in a Kubernetes cluster and wait for its completion while keeping an eye on the container output. This usecase frequently appears in deployment pipelines eg. for running a smoke test or a database migration before the actual deployment continues.

How to install

go get -v -u github.com/egym-playground/kubejob

How to use

Assuming that your $GOPATH/bin is part of your $PATH, you can call it directly, eg.:

kubejob -n test demo/two-containers.job.yaml

If you don't specify a job file kubejob reads it from stdin. This can be useful if you need to do some processing with the file before you hand it to kubejob, eg. for replacing an image tag:

sed s/IMAGE_TAG/v1.42/ my.job.yaml | kubejob -n test

Avoiding unintended restarts

In order to avoid unintended restarts you should set .spec.template.spec.restartPolicy=Never and .spec.backoffLimit=0, eg.:

apiVersion: batch/v1
kind: Job
metadata:
  name: demo
spec:
  backoffLimit: 0
  template:
    spec:
      restartPolicy: Never
      containers:
      - name: count1
        image: ubuntu
        command: ["sh", "-c", "for i in `seq 1 10`; do echo $i; done; false"]

Example output

2018/06/08 16:18:33 demo-7zbq9: Pending
2018/06/08 16:18:36 demo-7zbq9: Running
container1: 1
container1: 2
container2: 11
container2: 12
container1: 3
container2: 13
container1: 4
container2: 14
container1: 5
container2: 15
container1: 6
container2: 16
container1: 7
container2: 17
container1: 8
container2: 18
container1: 9
container2: 19
container1: 10
container2: 20
2018/06/08 16:18:46 demo-7zbq9: Succeeded
2018/06/08 16:18:50 Job completed successfully
2018/06/08 16:18:50 Deleting job