User program failed with ValueError: ZIP does not support timestamps before 1980

Question

Hi i'm running into an issue in Azure ML-

I'm getting the error "User program failed with ValueError: ZIP does not support timestamps before 1980"
when it calls the

when i run the following part of the code as a part of a pipeline

from azureml.train.estimator import Estimator

script_folder = os.getcwd()
print("script folder : ",script_folder)
est = Estimator(source_directory=script_folder,
compute_target=compute_target,
environment_definition=env,
entry_script='train.py')

run = exp.submit(config=est)
print(run)

Not sure what is causing this but when i run the file independently of the pipeline it runs fine

Answer 1

There may be an issue with the timestamp of your file or parent folder. Try moving your file to another directory or creating a new workspace, if that doesn't help, please let me know so we can investigate further. The product team are aware of this issue and currently working on a resolution. Thanks.

Note: try adding the following to the parent script to create a new directory (outside of the current working directory) and copy the child script there:

# create child_run_directory
    os.makedirs('/tmp/child_dir', exist_ok=True)

    os.system('cp ./child.py /tmp/child_dir/child.py')



    # kick off child run

    child_config = ScriptRunConfig(

        source_directory="/tmp/child_dir",

        script='child.py',

        run_config=rc,

        ...

        )