kosteev edited a comment on issue #18229: URL: https://github.com/apache/airflow/issues/18229#issuecomment-920679992
I could reproduce it in Airflow 2.1.2. Code of DAG: ``` default_args = {'wait_for_downstream': True, 'provide_context': False} command = 'sleep 1' with DAG( 'test_downstream', start_date=datetime.datetime.now() - datetime.timedelta(hours=1), schedule_interval='*/1 * * * *', default_args=default_args, max_active_runs=100, dagrun_timeout=datetime.timedelta(minutes=60), catchup=True) as dag: start_task = BashOperator( task_id='start_task', bash_command=command, retries=3, trigger_rule='all_success', dag=dag) end_task = BashOperator( task_id='end_task', bash_command=command, retries=3, trigger_rule='all_success', dag=dag) for target_date in range(0, 2): bash_cmd_j000 = BashOperator( task_id='bash_cmd_j000_' + str(target_date), bash_command=command, retries=3, trigger_rule='all_success', dag=dag) bash_cmd_j001 = BashOperator( task_id='bash_cmd_j001_' + str(target_date), bash_command=command, retries=3, trigger_rule='all_success', dag=dag) bash_cmd_j002 = BashOperator( task_id='bash_cmd_j002_' + str(target_date), bash_command=command, retries=3, trigger_rule='all_success', dag=dag) bash_cmd_j003 = BashOperator( task_id='bash_cmd_j003_' + str(target_date), bash_command=command, retries=3, trigger_rule='all_success', dag=dag) start_task >> bash_cmd_j000 >> bash_cmd_j001 >> bash_cmd_j002 >> bash_cmd_j003 >> end_task ``` After some time some tasks are scheduled and marked as failed (and up_for_retry) with log message: ``` [2021-09-16 07:56:46,290] {taskinstance.py:887} INFO - Dependencies not met for <TaskInstance: test_downstream.bash_cmd_j001_0 2021-09-16T06:58:00+00:00 [queued]>, dependency 'Previous Dagrun State' FAILED: The tasks downstream of the previous task instance <TaskInstance: test_downstream.bash_cmd_j001_0 2021-09-16 06:57:00+00:00 [success]> haven't completed (and wait_for_downstream is True). [2021-09-16 07:56:46,560] {local_task_job.py:96} INFO - Task is not able to be run ``` -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-unsubscr...@airflow.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org