Skip to content

local mode permission denied when file is written from running job to mounted volume #2647

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
jmahlik opened this issue Sep 21, 2021 · 1 comment · Fixed by #3017
Closed

Comments

@jmahlik
Copy link
Contributor

jmahlik commented Sep 21, 2021

Describe the bug
Permission denied in local mode when files are created by the running job. This occurs on the ProcessingOutput volumes when attempting to clean up. The job doesn't exit cleanly.

Fix is here:

#3017

To reproduce
Write a file to a mounted docker volume as a non-root user using local mode.

Expected behavior
Try except the error.

Screenshots or logs

---> 76     shutil.rmtree(source)
     77     return final_uri
...
PermissionError: [Errno 13] Permission denied: 'myfile2.txt'

System information
A description of your system. Please provide:

  • SageMaker Python SDK version: latest
  • Framework name (eg. PyTorch) or algorithm (eg. KMeans): local mode
  • Framework version: n/a
  • Python version: python 3.9
  • CPU or GPU: cpu
  • Custom Docker image (Y/N): Y

Additional context
Add any other context about the problem here.

~local mode

@david-waterworth
Copy link

I'm still experiancing with with sagemaker==2.224.2

INFO:sagemaker.local.entities:Pipeline step 'XXX' FAILED. Failure message is: PermissionError: [Errno 13] Permission denied: '/tmp/tmpq_a0b8qh/output/...'

I added some instrumentation and the problem isn't deleting the file, the permission denied comes from the earlier line when the file is uploaded to s3

sagemaker_session.upload_data(source, bucket, path)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging a pull request may close this issue.

3 participants