[
https://issues.apache.org/jira/browse/HADOOP-3923?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12646709#action_12646709
]
Tsz Wo (Nicholas), SZE commented on HADOOP-3923:
------------------------------------------------
Hi Lohit, are you saying that we should keep TaskGraphServlet in
JobTrackerStatusHttpServer? Without TaskGraphServlet,
JobTrackerStatusHttpServer is quite empty and may not deserve to have a class
for that. If we keep TaskGraphServlet in JobTrackerStatusHttpServer and put
future servlet classes as static classes in JobTrackerStatusHttpServer, it will
make JobTrackerStatusHttpServer unnecessary messy. So it might be better to
remove this not-so-useful class. No?
> Deprecate org.apache.hadoop.mapred.StatusHttpServer
> ---------------------------------------------------
>
> Key: HADOOP-3923
> URL: https://issues.apache.org/jira/browse/HADOOP-3923
> Project: Hadoop Core
> Issue Type: Improvement
> Components: mapred
> Reporter: Tsz Wo (Nicholas), SZE
> Assignee: Tsz Wo (Nicholas), SZE
> Priority: Minor
> Attachments: 3923_20081111.patch, 3923_20081111b.patch
>
>
> StatusHttpServer is extending org.apache.hadoop.http.HttpServer. The only
> additional feature in StatusHttpServer is TaskGraphServlet. StatusHttpServer
> is used by TaskTracker and JobTracker. It turns that only JobTracker uses
> TaskGraphServlet but not TaskTracker.
> So, we should deprecate StatusHttpServer, let TaskTracker and JobTracker use
> HttpServer directly, and add TaskGraphServlet to JobTracker.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
