[ 
https://issues.apache.org/jira/browse/YARN-3142?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15511003#comment-15511003
 ] 

Wangda Tan commented on YARN-3142:
----------------------------------

Thanks [~varun_saxena],

Yes I understand that this will not have any meaningful performance gain, but 
we can optimize it later like adding fine-grained locks. This is first step to 
us.

Generally the patch looks very good, only a few comments:

*Lock is not requred:*

{code}
  public Map<String, ResourceRequest> getResourceRequests(
      SchedulerRequestKey schedulerKey) {
      // ...
  }
{code}

*Lock is required:*

{code}
  public Resource getResource(SchedulerRequestKey schedulerKey) {
    // ... 
  }
{code}
(Imaging someone replace the request in another thread before returning)

{code}
transferStateFromPreviousAppSchedulingInfo
{code}
You may need to synchronize placesBlacklistedByApp and call 
placesBlacklistedByApp.addAll(appInfo.getBlackList()) to make it consistent to 
other blacklist-related operations.

> Improve locks in AppSchedulingInfo
> ----------------------------------
>
>                 Key: YARN-3142
>                 URL: https://issues.apache.org/jira/browse/YARN-3142
>             Project: Hadoop YARN
>          Issue Type: Sub-task
>          Components: resourcemanager, scheduler
>            Reporter: Wangda Tan
>            Assignee: Varun Saxena
>         Attachments: YARN-3142.01.patch
>
>




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

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

Reply via email to