[
https://issues.apache.org/jira/browse/AIRFLOW-3280?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Eran updated AIRFLOW-3280:
--------------------------
Description:
Running the following dag:
{code:java}
from airflow.exceptions import AirflowException
from airflow.models import DAG
from airflow.operators.dummy_operator import DummyOperator
from airflow.operators.python_operator import PythonOperator
from airflow.operators.subdag_operator import SubDagOperator
dag = DAG(dag_id="test_dag", start_date='2018-10-01', schedule_interval=None)
inner_dag = DAG(dag_id="test_dag.sub_dag_task", start_date='2018-10-01',
schedule_interval=None)
sub_dag_operator = SubDagOperator(task_id="sub_dag_task", subdag=inner_dag,
dag=dag)
def throw_exception():
raise AirflowException()
dummy_operator1 = DummyOperator(task_id="dummy1", dag=inner_dag)
python_operator = PythonOperator(task_id="python1", dag=inner_dag,
python_callable=throw_exception)
dummy_operator2 = DummyOperator(task_id="dummy2", dag=inner_dag,
trigger_rule="all_failed")
dummy_operator3 = DummyOperator(task_id="dummy3", dag=inner_dag)
dummy_operator4 = DummyOperator(task_id="dummy4", dag=inner_dag)
dummy_operator5 = DummyOperator(task_id="dummy5", dag=inner_dag,
trigger_rule="all_failed")
dummy_operator1 >> python_operator >> dummy_operator2
python_operator >> dummy_operator3
dummy_operator3 >> dummy_operator2
dummy_operator2 >> dummy_operator4
dummy_operator2 >> dummy_operator5
{code}
would result in a sub_dag_task as 'failed', even though the inner dag (when
clicking 'zoom into Sub DAG') is marked as 'success', only because there are a
few tasks marked as failed (which is ok, if its not the end of the DAG)
was:
Running the following dag:
{code:java}
from airflow.exceptions import AirflowException
from airflow.models import DAG
from airflow.operators.dummy_operator import DummyOperator
from airflow.operators.python_operator import PythonOperator
from airflow.operators.subdag_operator import SubDagOperator
dag = DAG(dag_id="test_dag", start_date='2018-10-01', schedule_interval=None)
inner_dag = DAG(dag_id="test_dag.sub_dag_task", start_date='2018-10-01',
schedule_interval=None)
sub_dag_operator = SubDagOperator(task_id="sub_dag_task", subdag=inner_dag,
dag=dag)
def throw_exception():
raise AirflowException()
dummy_operator1 = DummyOperator(task_id="dummy1", dag=inner_dag)
python_operator = PythonOperator(task_id="python1", dag=inner_dag,
python_callable=throw_exception)
dummy_operator2 = DummyOperator(task_id="dummy2", dag=inner_dag,
trigger_rule="all_failed")
dummy_operator3 = DummyOperator(task_id="dummy3", dag=inner_dag)
dummy_operator4 = DummyOperator(task_id="dummy4", dag=inner_dag)
dummy_operator5 = DummyOperator(task_id="dummy5", dag=inner_dag,
trigger_rule="all_failed")
dummy_operator1 >> python_operator >> dummy_operator2
python_operator >> dummy_operator3
dummy_operator3 >> dummy_operator2
dummy_operator2 >> dummy_operator4
dummy_operator2 >> dummy_operator5
{code}
would result in a sub_dag_task as 'failed', even though the inner dag (when
clicking 'zoom into
Sub DAG') is marked as 'success', only because there are a few tasks marked as
failed (which is ok, if its not the end of the DAG)
> SubDagOperator marked as failed, even if inner sub dag succeeds
> ---------------------------------------------------------------
>
> Key: AIRFLOW-3280
> URL: https://issues.apache.org/jira/browse/AIRFLOW-3280
> Project: Apache Airflow
> Issue Type: Bug
> Components: subdag
> Affects Versions: 1.9.0
> Reporter: Eran
> Priority: Major
>
> Running the following dag:
> {code:java}
> from airflow.exceptions import AirflowException
> from airflow.models import DAG
> from airflow.operators.dummy_operator import DummyOperator
> from airflow.operators.python_operator import PythonOperator
> from airflow.operators.subdag_operator import SubDagOperator
> dag = DAG(dag_id="test_dag", start_date='2018-10-01', schedule_interval=None)
> inner_dag = DAG(dag_id="test_dag.sub_dag_task", start_date='2018-10-01',
> schedule_interval=None)
> sub_dag_operator = SubDagOperator(task_id="sub_dag_task", subdag=inner_dag,
> dag=dag)
> def throw_exception():
> raise AirflowException()
> dummy_operator1 = DummyOperator(task_id="dummy1", dag=inner_dag)
> python_operator = PythonOperator(task_id="python1", dag=inner_dag,
> python_callable=throw_exception)
> dummy_operator2 = DummyOperator(task_id="dummy2", dag=inner_dag,
> trigger_rule="all_failed")
> dummy_operator3 = DummyOperator(task_id="dummy3", dag=inner_dag)
> dummy_operator4 = DummyOperator(task_id="dummy4", dag=inner_dag)
> dummy_operator5 = DummyOperator(task_id="dummy5", dag=inner_dag,
> trigger_rule="all_failed")
> dummy_operator1 >> python_operator >> dummy_operator2
> python_operator >> dummy_operator3
> dummy_operator3 >> dummy_operator2
> dummy_operator2 >> dummy_operator4
> dummy_operator2 >> dummy_operator5
> {code}
> would result in a sub_dag_task as 'failed', even though the inner dag (when
> clicking 'zoom into Sub DAG') is marked as 'success', only because there are
> a few tasks marked as failed (which is ok, if its not the end of the DAG)
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
