[
https://issues.apache.org/jira/browse/APEXCORE-474?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15453918#comment-15453918
]
ASF GitHub Bot commented on APEXCORE-474:
-----------------------------------------
GitHub user sandeshh opened a pull request:
https://github.com/apache/apex-core/pull/378
APEXCORE-474
https://issues.apache.org/jira/browse/APEXCORE-474
Creating a separate container for the last unifier is not required. It only
changes the default behavior, users can still create the separate unifiers.
@PramodSSImmaneni & @tweise please review.
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/sandeshh/apex-core APEXCORE-474
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/apex-core/pull/378.patch
To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #378
----
commit 3c5bc858967b0dfbebc5422aaaba93fbc7e7a3c8
Author: sandeshh <[email protected]>
Date: 2016-08-31T18:26:12Z
changes
----
> Default unifier placement during M*1 deployment
> -----------------------------------------------
>
> Key: APEXCORE-474
> URL: https://issues.apache.org/jira/browse/APEXCORE-474
> Project: Apache Apex Core
> Issue Type: Improvement
> Reporter: Sandesh
>
> During M*1 deployment, unifier was deployed in the separate container. But
> there is no advantage in doing that.
> It is better to make the unifier THREAD_LOCAL with the downstream operator.
> ( https://issues.apache.org/jira/browse/APEXCORE-482 )
> Note:
> Recently saw one Kafka ETL app, that had a total of 18 containers allocated,
> but out of that 5 containers were allocated for default unifiers. It also
> means, lots of time is spent in SerDe.
> Implementing this feature will improve the performance greatly.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
