Not able to pass environment variable from tfjob to pod

[ghost]

Here is how the tfjob looks like:

apiVersion: "kubeflow.org/v1"
kind: "TFJob"
metadata:
name: mlmodel-train-recipe-textcls
namespace: kubeflow
spec:
cleanPodPolicy: All
tfReplicaSpecs:
Worker:
replicas: 1
restartPolicy: Never
template:
spec:
containers:
- name: tensorflow
image: gcr.io/*****/mlmodel-train-recipe-textcls
command:
- "python"
- "services/bert_textcls/execute_job.py"
env:
- name: JOB_ID
value: "xyz"

Only for reference, I have kept a JOB_ID value as xyz but in the actual pipeline, I am using cloud builder and replacing the job_id on the fly. The tfjob worker gets created but when the value of os.getenv("JOB_ID") is always None.

Is it a bug in the latest tf-operator?

There is one in reference - https://www.kubeflow.org/docs/components/training/tftraining/

I am using kubeflow 1.3 and version 1.1.0 of tfoperator installed using kustomization scripts for kubeflow overlays - https://github.com/kubeflow/tf-operator/tree/v1.1.0/manifests

[gaocegege]

I cannot reproduce it, could you please format the code to show if there is a problem with indentation.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.