hadoop-yarn-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sunil G (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (YARN-3136) getTransferredContainers can be a bottleneck during AM registration
Date Mon, 30 Mar 2015 18:12:02 GMT

    [ https://issues.apache.org/jira/browse/YARN-3136?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14387114#comment-14387114

Sunil G commented on YARN-3136:

Hi [~jlowe] [~jianhe]
Bug type IS2_INCONSISTENT_SYNC (click for details) 
In class org.apache.hadoop.yarn.server.resourcemanager.scheduler.AbstractYarnScheduler
Field org.apache.hadoop.yarn.server.resourcemanager.scheduler.AbstractYarnScheduler.applications
Synchronized 90% of the time
Unsynchronized access at AbstractYarnScheduler.java:[line 138]
Unsynchronized access at AbstractYarnScheduler.java:[line 165]
Unsynchronized access at AbstractYarnScheduler.java:[line 233]

As "applications" is now a concurrent version, I feel we do not need a lock. Kindly share
your opinion.

test case failure is not related.

> getTransferredContainers can be a bottleneck during AM registration
> -------------------------------------------------------------------
>                 Key: YARN-3136
>                 URL: https://issues.apache.org/jira/browse/YARN-3136
>             Project: Hadoop YARN
>          Issue Type: Sub-task
>          Components: scheduler
>    Affects Versions: 2.6.0
>            Reporter: Jason Lowe
>            Assignee: Sunil G
>         Attachments: 0001-YARN-3136.patch, 0002-YARN-3136.patch, 0003-YARN-3136.patch,
0004-YARN-3136.patch, 0005-YARN-3136.patch, 0006-YARN-3136.patch, 0007-YARN-3136.patch, 0008-YARN-3136.patch,
> While examining RM stack traces on a busy cluster I noticed a pattern of AMs stuck waiting
for the scheduler lock trying to call getTransferredContainers.  The scheduler lock is highly
contended, especially on a large cluster with many nodes heartbeating, and it would be nice
if we could find a way to eliminate the need to grab this lock during this call.  We've already
done similar work during AM allocate calls to make sure they don't needlessly grab the scheduler
lock, and it would be good to do so here as well, if possible.

This message was sent by Atlassian JIRA

View raw message