You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@airflow.apache.org by "t oo (Jira)" <ji...@apache.org> on 2019/12/25 22:13:00 UTC
[jira] [Updated] (AIRFLOW-5006) Need task_failures audit trail for
failed run in metastore db even after running 'clear' results in next
DagRun successful
[ https://issues.apache.org/jira/browse/AIRFLOW-5006?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
t oo updated AIRFLOW-5006:
--------------------------
Description:
if a dagrun fails (and so some rows go to task_fail table in metastore) but then after running 'clear' from cli to restart the dagrun (and its successful run) the rows from metastore's task_fail table get deleted. is this expected? or is there some way i can keep history of all failures in metastore?
even history of old successful runs are lost when doing clear for same exec date
was:if a dagrun fails (and so some rows go to task_fail table in metastore) but then after running 'clear' from cli to restart the dagrun (and its successful run) the rows from metastore's task_fail table get deleted. is this expected? or is there some way i can keep history of all failures in metastore?
> Need task_failures audit trail for failed run in metastore db even after running 'clear' results in next DagRun successful
> --------------------------------------------------------------------------------------------------------------------------
>
> Key: AIRFLOW-5006
> URL: https://issues.apache.org/jira/browse/AIRFLOW-5006
> Project: Apache Airflow
> Issue Type: Improvement
> Components: database
> Affects Versions: 1.10.3, 1.10.5
> Environment: mysql RDS metastore, localexecutor
> Reporter: t oo
> Priority: Major
>
> if a dagrun fails (and so some rows go to task_fail table in metastore) but then after running 'clear' from cli to restart the dagrun (and its successful run) the rows from metastore's task_fail table get deleted. is this expected? or is there some way i can keep history of all failures in metastore?
>
> even history of old successful runs are lost when doing clear for same exec date
--
This message was sent by Atlassian Jira
(v8.3.4#803005)