[jira] [Commented] (AIRFLOW-96) s3_conn_id of s3KeySensor cannot be defined using an environment variable

["Chris Riccomini (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/AIRFLOW-96?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15278584#comment-15278584 ] 

Chris Riccomini commented on AIRFLOW-96:
----------------------------------------

It sounds like the S3 key sensor needs to be updated to pay attention to environment variables as well. I think S3KeySensor should use the S3hook, which extends BaseHook. BaseHook has get_connection, which honors the environment variable.

Would you like to submit a PR for this?

> s3_conn_id of s3KeySensor cannot be defined using an environment variable
> -------------------------------------------------------------------------
>
>                 Key: AIRFLOW-96
>                 URL: https://issues.apache.org/jira/browse/AIRFLOW-96
>             Project: Apache Airflow
>          Issue Type: Bug
>          Components: operators
>    Affects Versions: Airflow 1.6.2
>         Environment:     Python Version: 2.7.11
>     Operating System: OS X El Capitan 10.11.4
>            Reporter: dmtran
>            Priority: Minor
>
> According to https://pythonhosted.org/airflow/concepts.html#connections, Airflow has the ability to reference connections via environment variables from the operating system. The environment variable needs to be prefixed with AIRFLOW_CONN_ to be considered a connection.
> This doesn't work with an S3KeySensor, the following exception is raised:
> {noformat}
> [2016-05-10 17:01:37,101] {models.py:1041} ERROR - conn_id doesn't exist in the repository
> Traceback (most recent call last):
>   File "/usr/local/lib/python2.7/site-packages/airflow/models.py", line 1000, in run
>     result = task_copy.execute(context=context)
>   File "/usr/local/lib/python2.7/site-packages/airflow/operators/python_operator.py", line 65, in execute
>     return_value = self.python_callable(*self.op_args, **self.op_kwargs)
>   File "/Users/dmtran/git_repos/coordinator/airflow/dags/test-s3.py", line 24, in check_key_in_s3
>     s3_conn_id='S3_CONNECTION')
>   File "/usr/local/lib/python2.7/site-packages/airflow/utils.py", line 461, in wrapper
>     result = func(*args, **kwargs)
>   File "/usr/local/lib/python2.7/site-packages/airflow/operators/sensors.py", line 332, in __init__
>     raise AirflowException("conn_id doesn't exist in the repository")
> AirflowException: conn_id doesn't exist in the repository
> {noformat}
> You can reproduce this issue by triggering a DAGrun of the following DAG:
> {code}
> from airflow.hooks.base_hook import CONN_ENV_PREFIX
> from airflow.operators import *
> from airflow.models import DAG
> from datetime import datetime
> import os
> args = {
>     'owner': 'airflow',
>     'start_date': datetime(2016, 5, 10, 7)
> }
> dag = DAG(dag_id='test-s3',
>     default_args=args,
>     schedule_interval=None)
> def check_key_in_s3(**context):
>     os.environ[CONN_ENV_PREFIX + 'S3_CONNECTION'] = '{ "aws_access_key_id": "dummyAccessKey",  "aws_secret_access_key": "dummySecretKey" }'
>     sensor = S3KeySensor(
>         task_id='s3keysensor',
>         bucket_name='dummy_bucket',
>         bucket_key='dummy_key',
>         dag=dag,
>         s3_conn_id='S3_CONNECTION')
>     sensor.execute(context)
> check_s3_key_operator = PythonOperator(
>     task_id='check_key_in_s3',
>     python_callable=check_key_in_s3,
>     provide_context=True,
>     dag=dag)
> {code}
> The exception is raised because of the following lines in method __init__ of class S3KeySensor:
> {code}
>         db = session.query(DB).filter(DB.conn_id == s3_conn_id).first()
>         if not db:
>             raise AirflowException("conn_id doesn't exist in the repository")
> {code}



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)