[
https://issues.apache.org/jira/browse/AIRFLOW-3369?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16816323#comment-16816323
]
Andrew Harmon commented on AIRFLOW-3369:
----------------------------------------
so this appears to be an issue if you use
{code:python}
schedule_interval=timedelta(days=1)
{code}
but not if you use
{code:python}
schedule_interval='@daily'
{code}
Bug?
> Un-pausing a DAG with catchup =False creates an extra DAG run (1.10)
> --------------------------------------------------------------------
>
> Key: AIRFLOW-3369
> URL: https://issues.apache.org/jira/browse/AIRFLOW-3369
> Project: Apache Airflow
> Issue Type: Bug
> Affects Versions: 1.10.0
> Reporter: Andrew Harmon
> Priority: Major
> Attachments: image.png
>
>
> If you create a DAG with catchup=False, when it is un-paused, it creates 2
> dag runs. One for the most recent scheduled interval (expected) and one for
> the interval before that (unexpected).
> *Sample DAG*
> {code:java}
> from airflow import DAG
> from datetime import datetime
> from airflow.operators.dummy_operator import DummyOperator
> dag = DAG(
> dag_id='DummyTest',
> start_date=datetime(2018,1,1),
> catchup=False
> )
> do = DummyOperator(
> task_id='dummy_task',
> dag=dag
> )
> {code}
> *Result:*
> 2 DAG runs are created. 2018-11-18 and 108-11-17
> *Expected Result:*
> Only 1 DAG run should have been created (2018-11-18)
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
