SKLearn Processing Container - Error: "WARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager."

0

Hey all,

I am trying to run the script below in the writefile titled "vw_aws_a_bijlageprofile.py". This code has worked for me using other data sources, but now I am getting the following error message from the CloudWatch Logs:

"***2022-08-24T20:09:19.708-05:00

WARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv***"

Any idea how I get around this error?

Full code below.

Thank you in advance!!!!

%%writefile vw_aws_a_bijlageprofile.py

import os
import sys
import subprocess
def install(package):
    subprocess.check_call([sys.executable, "-q", "-m", "pip", "install", package])
install('awswrangler')
install('tqdm')
install('pandas')
install('botocore')
install('ruamel.yaml')
install('pandas-profiling')
import awswrangler as wr
import pandas as pd
import numpy as np
import datetime as dt
from dateutil.relativedelta import relativedelta
from string import Template
import gc
import boto3

from pandas_profiling import ProfileReport

client = boto3.client('s3')
session = boto3.Session(region_name="eu-west-2")


def run_profile():



    query = """
    SELECT  * FROM "intl-euro-archmcc-database"."vw_aws_a_bijlage"
    ;
    """
                                        #swich table name above
        
    tableforprofile = wr.athena.read_sql_query(query,
                                            database="intl-euro-archmcc-database",
                                            boto3_session=session,
                                            ctas_approach=False,
                                            workgroup='DataScientists')
    print("read in the table queried above")

    print("got rid of missing and added a new index")

    profile_tblforprofile = ProfileReport(tableforprofile, 
                                  title="Pandas Profiling Report", 
                                  minimal=True)

    print("Generated table profile")
                                      
    return profile_tblforprofile


if __name__ == '__main__':

    profile_tblforprofile = run_profile()
    
    print("Generated outputs")

    output_path_tblforprofile = ('/opt/ml/processing/output/profile_vw_aws_a_bijlage.html')
                                    #switch profile name above
    print(output_path_tblforprofile)
    
    profile_tblforprofile.to_file(output_path_tblforprofile)

import sagemaker
from sagemaker.processing import ProcessingInput, ProcessingOutput

session = boto3.Session(region_name="eu-west-2")

bucket = 'intl-euro-uk-datascientist-prod'

prefix = 'Mark'

sm_session = sagemaker.Session(boto_session=session, default_bucket=bucket)
sm_session.upload_data(path='vw_aws_a_bijlageprofile.py',
                                bucket=bucket,
                                key_prefix=f'{prefix}/source')
import boto3
#import sagemaker
from sagemaker import get_execution_role
from sagemaker.sklearn.processing import SKLearnProcessor

region = boto3.session.Session().region_name


S3_ROOT_PATH = "s3://{}/{}".format(bucket, prefix)

role = get_execution_role()
sklearn_processor = SKLearnProcessor(framework_version='0.20.0',
                                     role=role,
                                     sagemaker_session=sm_session,
                                     instance_type='ml.m5.24xlarge',
                                     instance_count=1)
sklearn_processor.run(code='s3://{}/{}/source/vw_aws_a_bijlageprofile.py'.format(bucket, prefix),
                      inputs=[],
                      outputs=[ProcessingOutput(output_name='output',
                                                source='/opt/ml/processing/output',
                                                destination='s3://intl-euro-uk-datascientist-prod/Mark/IODataProfiles/')])
1 個回答
0

This is not an error. This is just a warning message that something can go wrong, e. g. on your local machine or on the on-premise server, where you usually setup a virtual env.

In contrast, Amazon SageMaker is a managed cloud service that is designed to work with containers and to run your code as root without a virtual env. There are no different package managers, there's only one package manager, and no conflicting behaviour can occur in this scenario.

So, this warning is expected in the SageMaker environment and you can just safely ignore this log message.

profile pictureAWS
Ivan
已回答 2 年前
  • huh, that is really interesting. Maybe that isn't the problem then. I just re-ran the above and got the same error message, maybe I didn't include enough details though: sagemaker-sklearn-container 1.0 requires jinja2==2.10.2, but you have jinja2 3.1.2 which is incompatible. sagemaker-sklearn-container 1.0 requires MarkupSafe==1.1.1, but you have markupsafe 2.1.1 which is incompatible. sagemaker-sklearn-container 1.0 requires numpy==1.19.5, but you have numpy 1.21.6 which is incompatible. sagemaker-sklearn-container 1.0 requires pandas==0.25.*, but you have pandas 1.3.5 which is incompatible.

  • I can send you the cloudwatch processing jobs error log in a CSV or image file but even when I screenshot the error log, which is a 74 KB PNG file, I cannot add it here.

  • Hey @Ivan - can I send over the error log please?

您尚未登入。 登入 去張貼答案。

一個好的回答可以清楚地回答問題並提供建設性的意見回饋,同時有助於提問者的專業成長。

回答問題指南