0

I tried to follow the instructions mentioned on this page...

https://towardsdatascience.com/serverless-distributed-data-pre-processing-using-dask-amazon-ecs-and-python-part-1-a6108c728cc4

And got 2 errors. One is related to IAM role and the other is related to dask.

1) I checked that sagemaker policy includes full access and still getting this exception:

An error occurred (AccessDeniedException) when calling the UpdateService operation: User: arn:aws:sts::513469704633:assumed-role/AmazonSageMaker-ExecutionRole-20180101T192504/SageMaker is not authorized to perform: ecs:UpdateService on resource: *

2) I do not understand what does cancelled error mean:

CancelledError: ('head-1-5-from-delayed-12929833853fcc20e3534186ebfd77f0', 0)

Python notebook screenshot attached. enter image description here

shantanuo
  • 31,689
  • 78
  • 245
  • 403

1 Answers1

0

My first guess is that the dask workers you've started with fargate don't have sufficient permissions to read your data. You may want to read through these docs:

http://docs.dask.org/en/latest/remote-data-services.html#amazon-s3

MRocklin
  • 55,641
  • 23
  • 163
  • 235