hadoop-yarn-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Wangda Tan (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (YARN-3141) Improve locks in SchedulerApplicationAttempt/FSAppAttempt/FiCaSchedulerApp
Date Fri, 16 Sep 2016 18:02:21 GMT

    [ https://issues.apache.org/jira/browse/YARN-3141?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15496949#comment-15496949

Wangda Tan commented on YARN-3141:

Thanks for confirmation,

Updated ver.5 patch addressed your comments. 

bq.  you dropped the lock instead of downgrading it before the log.
In this patch, I didn't do improvements like downgrading locks in a method. Beyond FSAttempt.allocate/unreserveInternal,
there're several similar methods in CS as well. I would prefer to have a separate JIRA to
do more detailed improvements. And I don't know how much performance benefit we can gain from
downgrading the lock. But it does make code becomes more complexity and harder to be understood.
Are you OK with doing such detailed improvements as a separate task? Let's try to make this
patch simply convert synchronized lock to R/W lock.


> Improve locks in SchedulerApplicationAttempt/FSAppAttempt/FiCaSchedulerApp
> --------------------------------------------------------------------------
>                 Key: YARN-3141
>                 URL: https://issues.apache.org/jira/browse/YARN-3141
>             Project: Hadoop YARN
>          Issue Type: Sub-task
>          Components: resourcemanager, scheduler
>            Reporter: Wangda Tan
>            Assignee: Wangda Tan
>         Attachments: YARN-3141.1.patch, YARN-3141.2.patch, YARN-3141.3.patch, YARN-3141.4.patch
> Enhance locks in SchedulerApplicationAttempt/FSAppAttempt/FiCaSchedulerApp, as mentioned
in YARN-3091, a possible solution is using read/write lock. Other fine-graind locks for specific
purposes / bugs should be addressed in separated tickets.

This message was sent by Atlassian JIRA

To unsubscribe, e-mail: yarn-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: yarn-issues-help@hadoop.apache.org

View raw message