t oo created AIRFLOW-6335:
-----------------------------
Summary: dag_processor_manager timeout logs should be ERROR not
INFO/WARN
Key: AIRFLOW-6335
URL: https://issues.apache.org/jira/browse/AIRFLOW-6335
Project: Apache Airflow
Issue Type: Improvement
Components: DAG, scheduler
Affects Versions: 1.10.6
Reporter: t oo
I triggered a large DAG with 18k tasks, after 30mins still nothing started
going to scheduled/queued/running state.
All airflow processes (scheduler, worker.etc) were up. I then went searching
through the logs for ERROR, nothing came up! After a lot of digging i found
below in dag_processor_manager.log:
||File Path||PID||Runtime||# DAGs||# Errors||Last Runtime||Last Run||
|/home/ec2-user/airflow/dags/redact1.py|31366|261.22s|0|-1|361.32s|2019-12-24T17:10:44|
|/home/ec2-user/airflow/dags/redact2.py| | |1|0|1.00s|2019-12-24T17:16:29|
[2019-12-24 17:40:48,739] \{dag_processing.py:1324} INFO - Processor for
/home/ec2-user/airflow/dags/redact1.py with PID 17307 started at
2019-12-24T17:34:47.417660+00:00 has timed out, killing it.
[2019-12-24 17:40:49,696] \{dag_processing.py:1324} INFO - Processor for
/home/ec2-user/airflow/dags/redact1.py with PID 17307 started at
2019-12-24T17:34:47.417660+00:00 has timed out, killing it.
[2019-12-24 17:40:49,697] \{dag_processing.py:1191} WARNING - Processor for
/home/ec2-user/airflow/dags/redact1.py exited with return code -9.
Solution:
Change from INFO to ERROR:
[https://github.com/apache/airflow/blob/1.10.6/airflow/utils/dag_processing.py#L1321]
Change from WARN to ERROR:
[https://github.com/apache/airflow/blob/1.10.6/airflow/utils/dag_processing.py#L1189]
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
