AWS Sagemaker - Access Denied

3
votes

I'm working through some of the example Sagemaker notebooks, and I receive the following Access Denied error when trying to run the linear_time_series_forecast example:

ValueError: Error training linear-learner-2017-12-21-15-29-34-676: Failed Reason: ClientError: Data download failed:AccessDenied (403): Access Denied

I can manually download and upload from my S3 bucket using the AWS command line interface, but the Jupyter notebook fails.

Can someone please help me with this error?

1
amazon-web-servicesamazon-s3amazon-sagemaker
Are you running your Jupyter notebook on Sagemaker's Notebook Instance, or are you running the notebook locally on your computer? - spg
@spg I'm running the notebook through Sagemaker's notebook instance. - mmmgood

1 Answers

2
votes

Thanks for using Amazon SageMaker!

Looks like this question was also answered on the AWS Forums: https://forums.aws.amazon.com/thread.jspa?threadID=270054&tstart=0

The IAM Role referenced by

role = get_execution_role()

needs to have a policy attached to it that grants S3:GetObject permission on the S3 bucket holding your training data.