Skip to content

AlgorithmError while replicating Sagemaker Preprocessing in lambda. #1378

@Olamyy

Description

@Olamyy

Describe the bug
I'm using the sagemaker client in boto3 to start a processing job in a lambda function. For some reason, I keep getting an AlgorithmError when I check the status of the processing job. To ensure it's not from my end, I used exactly the same set variables and script from the official tutorial available at https://github.com/awslabs/amazon-sagemaker-examples/blob/master/sagemaker_processing/scikit_learn_data_processing_and_model_evaluation/scikit_learn_data_processing_and_model_evaluation.ipynb.

To reproduce
Here's the relevant part of my lambda function

parameters = {
  "image_uri": "492215442770.dkr.ecr.eu-central-1.amazonaws.com/sagemaker-scikit-learn:0.20.0-cpu-py3",
  "role_name": "arn:aws:iam::AccountID:role/myUser",
  "arguments": [
    "--train-test-split-ratio",
    "0.2"
  ],
  "environment": {},
  "entrypoint": [
    "python3",
    "s3://processing-test/processing.py"
  ],
  "resources": {
    "ClusterConfig": {
      "InstanceType": "ml.m5.xlarge",
      "InstanceCount": 1,
      "VolumeSizeInGB": 1,
      "VolumeKmsKeyId": ""
    }
  },
  "tags": [],
  "inputs": [
    {
      "InputName": "test_input_name",
      "S3Input": {
        "S3Uri": "s3://sagemaker-sample-data-eu-central-1/processing/census/census-income.csv",
        "S3DataType": "S3Prefix",
        "S3InputMode": "File",
        "LocalPath": "/opt/ml/processing/input/"
      }
    }
  ],
  "outputs": [
    {
      "OutputName": "test_output_name",
      "S3Output": {
        "S3Uri": "s3://processing-test/output/",
        "S3UploadMode": "EndOfJob",
        "LocalPath": "/opt/ml/processing/output/"
      },
      "KmsKeyId": "alias/processing-kms-key"
    }
  ],
  "max_runtime": 86400,
  "network_config": {},
  "experiment_config": {},
  "job_name": "zflow-sagemaker-processing-test-pr-1-74-29-03-2020-06-24"
}

botocore_config = Config(
        retries=dict(
            max_attempts=10
        )
    )

sagemaker_client = boto3.client('sagemaker', config=botocore_config)

request = sagemaker_client.create_processing_job(
        **parameters
    )

Expected behavior
I expect the code to start a processing job.

Screenshots or logs
If applicable, add screenshots or logs to help explain your problem.
image

System information
A description of your system. Please provide:

  • Framework version: Sklearn Version 0.20.0
    I used the official tutorial so the config should pretty much be the same as that.

Additional context
Add any other context about the problem here.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions