[jira] [Commented] (AIRFLOW-6389) add config for 'allow_multi_scheduler_instances' default True

["t oo (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/AIRFLOW-6389?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17045895#comment-17045895 ] 

t oo commented on AIRFLOW-6389:
-------------------------------

t
https://jira.apache.org/jira/browse/AIRFLOW-6389 any suggestions on which .py file to change to implement this? could jobs table help?

6 replies

ash  22 hours ago
That is more than a single file change

ash  22 hours ago
It's my next two months worth of work!

t  22 hours ago
I'm not saying support HA, just prevent 2nd scheduler pid starting

ash  21 hours ago
Oh to detect running pid that should be in the cli/commands/ folder

ash  21 hours ago
I don't think there's a need for a config option

ash  21 hours ago
If it's configured to write a pid, and that already exists and is a live process it should refuse to run

> add config for 'allow_multi_scheduler_instances' default True
> -------------------------------------------------------------
>
>                 Key: AIRFLOW-6389
>                 URL: https://issues.apache.org/jira/browse/AIRFLOW-6389
>             Project: Apache Airflow
>          Issue Type: New Feature
>          Components: scheduler
>    Affects Versions: 1.10.6
>            Reporter: t oo
>            Priority: Minor
>
> right now common deployment pattern with blue/green build is:
> 1. on EC2 1, start scheduler
> 2. Assign 'final' DNS to EC2 1
> 3. create EC2 2
> 4. start scheduler on EC2 2
> 5.  Assign 'final' DNS to EC2 2
> 6. Teardown EC2 1
> Issue is that since the metastore db (ie mysql) is shared to both EC2s there is a period of time between point 4 and 6 above where there are multiple schedulers running. To avoid this proposing config for 'allow_multi_scheduler_instances' that when set to False, the startup of scheduler will detect that another scheduler is running then exit (ie not startup) with WARNING message
> 7. We have cron/systemd setup to keep retrying to to start the scheduler pid, so as soon as point 6 completes scheduler should successfully launch on EC2 1



--
This message was sent by Atlassian Jira
(v8.3.4#803005)