You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@aurora.apache.org by "Bill Farner (JIRA)" <ji...@apache.org> on 2015/03/21 01:27:38 UTC

[jira] [Commented] (AURORA-1216) Factor in failovers when calculating time waiting for running in scheduler updater

    [ https://issues.apache.org/jira/browse/AURORA-1216?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14372375#comment-14372375 ] 

Bill Farner commented on AURORA-1216:
-------------------------------------

Relevant code is here: https://github.com/apache/incubator-aurora/blob/48dc4c7ad28eeea3e7108c1c0b408f4010a27aa1/src/main/java/org/apache/aurora/scheduler/updater/InstanceUpdater.java#L75-L77

> Factor in failovers when calculating time waiting for running in scheduler updater
> ----------------------------------------------------------------------------------
>
>                 Key: AURORA-1216
>                 URL: https://issues.apache.org/jira/browse/AURORA-1216
>             Project: Aurora
>          Issue Type: Task
>          Components: Scheduler
>            Reporter: David McLaughlin
>
> If the scheduler fails over, any instances stuck in PENDING will most likely kill an entire update if minWaitingTimeForRunningMs is set to a time less than the time it takes for a new leader to be elected. It would be better for the updater to use max(newLeaderReadyTimestamp, instanceUpdateStartTimeMs) to avoid this. 



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