[ 
https://issues.apache.org/jira/browse/YARN-3437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14508742#comment-14508742
 ] 

Junping Du commented on YARN-3437:
----------------------------------

bq. Also, when you said "may be move out of collector.start()", did you mean 
moving the collector.start() call outside the synchronization block? If so, I'd 
be hesitant to do that. We just had a discussion on this in another JIRA.
I see. We don't want the other concurrent thread get a non-start collector. We 
can improve this later (in some other JIRA), e.g. attach an additional monitor 
object to specific collector (rather than sync on all collectors), and do 
something like: startIfNotStarted(). 

Another NIT is to replace Collections.synchronizedMap with ConcurrentHashMap, 
the later one has better performance and better safety (never throw 
ConcurrentModificationException) but just not keep order of element which we 
don't need in our case. Given we have YARN-3390 to continue refactor work on 
the same piece of code, we can continue the discussion there.

Latest patch LGTM. +1. Will go ahead to commit it within 24h if no 
object/further comments from others.

> convert load test driver to timeline service v.2
> ------------------------------------------------
>
>                 Key: YARN-3437
>                 URL: https://issues.apache.org/jira/browse/YARN-3437
>             Project: Hadoop YARN
>          Issue Type: Sub-task
>          Components: timelineserver
>            Reporter: Sangjin Lee
>            Assignee: Sangjin Lee
>         Attachments: YARN-3437.001.patch, YARN-3437.002.patch, 
> YARN-3437.003.patch, YARN-3437.004.patch
>
>
> This subtask covers the work for converting the proposed patch for the load 
> test driver (YARN-2556) to work with the timeline service v.2.



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

Reply via email to