deleteNamespacedJob with propagationPolicy as 'Foreground' does not delete pods created by the job

[rnarayan-perforce]

Describe the bug
deleteNamespacedJob with propagationPolicy as 'Foreground' does not delete pods created by the job

await this.kubernetesBatchApi.deleteNamespacedJob({
      name: "job-name",
      namespace: "namespace",
      body: {},
      gracePeriodSeconds: 60,
      propagationPolicy: 'Foreground'
});

Client Version
e.g. 1.1.2, 1.4.0 & 1.3.0

Server Version
e.g. 1.32

To Reproduce
Call deleteNamespacedJob with a job name and use the following commands on two different terminals to list the jobs and the pod:

watch kubectl get jobs -n namespace

watch kubectl get pods -n namespace

The job gets deleted but the pod continues to run.

Expected behavior
Since we are using Foreground propagationPolicy, the pods should get deleted first and then the job. In this case the pod does not get deleted at all.

Example Code

await this.kubernetesBatchApi.deleteNamespacedJob({
      name: "job-name",
      namespace: "namespace",
      body: {},
      gracePeriodSeconds: 60,
      propagationPolicy: 'Foreground'
});

Environment (please complete the following information):

• OS: LINUX
• Node.js version 22
• Cloud runtime: AWS EKS

[mstruebing]

Hey @rnarayan-perforce I just tried it with latest main and with 1.4.0 and it seems to work just fine:

This is the code I've used:

import * as k8s from '../../../dist/index.js';

const kc = new k8s.KubeConfig();
kc.loadFromDefault();

const k8sApi = kc.makeApiClient(k8s.BatchV1Api);

const namespace = 'my-namespace';

const res = await k8sApi.deleteNamespacedJob({
    name: 'long-running-job',
    namespace,
    gracePeriodSeconds: 60,
    propagationPolicy: 'Foreground',
});

console.log('res:', res);

Are there any more information we need to reproduce this?
You could create an integration test with the failing behavior similar to: https://github.com/kubernetes-client/javascript/blob/main/src/test/integration/patchNamespace.ts

Screen.Recording.2025-11-19.at.09.38.32.mov

[rnarayan-perforce]

Thank you for quick reply. I will try to share some better steps to reproduce this.

[mstruebing]

Does it work if you do it manually? I assume it could also be some kind of misconfiguration on your cluster.

[rnarayan-perforce]

I tried doing the same with kubectl command and it works fine:

kubectl delete jobs <job-name> -n <namespace>

[mstruebing]

You did not pass the foreground as well as the grace period as far as I can see like in the initial example.
Can you create an example that fails in KinD with the kuberente-client/javascript?

[rnarayan-perforce]

My colleague, @smahajan0228 found the issue in my code snippet. It was the empty body (body: {}) parameter that I was passing. After removing it the deletion works as expected.

So I changed my code snippet from this:

await this.kubernetesBatchApi.deleteNamespacedJob({
      name: "job-name",
      namespace: "namespace",
      body: {},
      gracePeriodSeconds: 60,
      propagationPolicy: 'Foreground'
});

to:

await this.kubernetesBatchApi.deleteNamespacedJob({
      name: "job-name",
      namespace: "namespace",
      gracePeriodSeconds: 60,
      propagationPolicy: 'Foreground'
});