Shreyas Joshi created AIRFLOW-736:
-------------------------------------
Summary: sub dag status still "running" even if all task instances
are complete
Key: AIRFLOW-736
URL: https://issues.apache.org/jira/browse/AIRFLOW-736
Project: Apache Airflow
Issue Type: Bug
Components: scheduler
Affects Versions: Airflow 1.7.1.3
Reporter: Shreyas Joshi
Sometimes, a sub dag is still running, even though all the task instances in it
are successful. The log is an endless list of waiting for all the tasks to
finish like:
{noformat}
[2017-01-05 00:46:26,031] {jobs.py:997} INFO - [backfill progress] | waiting: 1
| succeeded: 6 | kicked_off: 7 | failed: 0 | skipped: 4 | deadlocked: 0
[2017-01-05 00:46:31,033] {jobs.py:997} INFO - [backfill progress] | waiting: 1
| succeeded: 6 | kicked_off: 7 | failed: 0 | skipped: 4 | deadlocked: 0
{noformat}
with this line hidden in there:
{noformat}
2017-01-05 00:10:21,037] {jobs.py:965} ERROR - The airflow run command failed
at reporting an error. This should not occur in normal circumstances. Task
state is 'running',reported state is 'success'. TI is <TaskInstance:
something.something_child 2017-01-03 23:00:00 [running]>
{noformat}
There are no dependencies on past runs. We are using the CeleryExecutor with
three worker nodes. AIRFLOW-396 seems very similar to this.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
