SageMaker Bring Your Own Container - on local mode - connection to S3

[idanmoradarthas]

Describe the feature you'd like
During implementation of SageMaker BYOC with local mode with Python SDK, we notice that it connects to S3 (to Sagemaker object's - default bucket), places all of the artifacts in there, and then downloading it again into the local container. We thought that local mode means that no external connections are done, but that's not the case in our implementation.

How would this feature be used? Please describe.
We want local mode to be able to be completely local. So if we want to, we can indicate to it to fully work without the need to connect to web at all.

Additional context
We used this code as a reference: https://github.com/aws-samples/amazon-sagemaker-local-mode/blob/main/scikit_learn_bring_your_own_container_local_processing/scikit_learn_bring_your_own_container_local_processing.py

[alar0330]

Seems to be related: #3031

[mufaddal-rohawala]

Fix released. Should be fixed in release >= v2.89.0

[clausagerskov]

@mufaddal-rohawala this is not fixed

[clausagerskov]

this bug fix listed in the release was to change a required "s3:ListAllMyBuckets" to a required "s3:ListBucket".
that does not at all address the issue here, which is fully local development

[clausagerskov]

@shreyapandit

[fahran-wallace]

I'm experiencing the same I think, using sagemaker[local]==2.222.0.

A code snippet extracted from this example: https://github.com/aws-samples/amazon-sagemaker-local-mode/blob/main/xgboost_script_mode_local_training_and_serving/xgboost_script_mode_local_training_and_serving.py

from sagemaker import TrainingInput
from sagemaker.xgboost import XGBoost, XGBoostModel
from sagemaker.local import LocalSession

DUMMY_IAM_ROLE = 'arn:aws:iam::111111111111:role/service-role/AmazonSageMaker-ExecutionRole-20200101T000001'
LOCAL_SESSION = LocalSession()
LOCAL_SESSION.config = {'local': {'local_code': True}}  # Ensure full code locality, see: https://sagemaker.readthedocs.io/en/stable/overview.html#local-mode
FRAMEWORK_VERSION = "1.7-1"

def main():
    xgb_inference_model = XGBoostModel(
        model_data="./tests/resources/models/fake/model.json",
        role=DUMMY_IAM_ROLE,
        entry_point="inference.py",
        source_dir="./src",
        framework_version=FRAMEWORK_VERSION,
        sagemaker_session=LOCAL_SESSION
    )

    print('Deploying endpoint in local mode')
    predictor = xgb_inference_model.deploy(
        initial_instance_count=1,
        instance_type="local",
    )

def test_inference_endpoint():
    main()

Expected behaviour: Model repacked locally
Actual behaviour: Failure when creating an S3 bucket during the deploy job. I...really do not want my code being uploaded to random S3 buckets!

operation_name = 'CreateBucket'
api_params = {'Bucket': 'sagemaker-eu-west-1-xxxxxxxxxxxx', 'CreateBucketConfiguration': {'LocationConstraint': 'eu-west-1'}}```

[clausagerskov]

so is this fixed then?