airflow-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "George Leslie-Waksman (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (AIRFLOW-326) Unclean scheduler restarts result in zombie Task Instances
Date Wed, 13 Jul 2016 16:42:20 GMT

    [ https://issues.apache.org/jira/browse/AIRFLOW-326?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15375323#comment-15375323
] 

George Leslie-Waksman commented on AIRFLOW-326:
-----------------------------------------------

Because of our deploy environment (heroku like), we occasionally have to restart everything
to get changes through. We're investigating alternatives, but for now we occasionally have
no option other than restarting the scheduler (and all of the workers).

> Unclean scheduler restarts result in zombie Task Instances
> ----------------------------------------------------------
>
>                 Key: AIRFLOW-326
>                 URL: https://issues.apache.org/jira/browse/AIRFLOW-326
>             Project: Apache Airflow
>          Issue Type: Bug
>          Components: celery
>    Affects Versions: Airflow 1.7.1.3
>            Reporter: George Leslie-Waksman
>
> When a task is running and the scheduler and task both die before the task completes
execution, the task is not identified as failed. The task appears to be in a "Running" state
even though it has long since stopped running.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message