[
https://issues.apache.org/jira/browse/MAPREDUCE-2749?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13101534#comment-13101534
]
Thomas Graves commented on MAPREDUCE-2749:
------------------------------------------
Thanks for the review, I'll update the patch for the config changes. I'll take
another look at the test to see if I can get away from using minicluster. The
issue I have with this is that to me it is a system/component test and not a
unit test - when nodemanager fails to start another component (RM) should do or
in this case not do something. When I was looking it really seemed like to
really test it you want all the regular init/start code in the nm so mocking
that didn't seem useful. We could perhaps mock the rm register routine which
is what the mini cluster already does for in process communication.
> [MR-279] NM registers with RM even before it starts various servers
> -------------------------------------------------------------------
>
> Key: MAPREDUCE-2749
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-2749
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Components: mrv2
> Affects Versions: 0.23.0
> Reporter: Vinod Kumar Vavilapalli
> Assignee: Thomas Graves
> Fix For: 0.23.0
>
> Attachments: MAPREDUCE-2749-v2.patch, MAPREDUCE-2749-v3.patch,
> MAPREDUCE-2749-v4.patch, MAPREDUCE-2749.patch
>
>
> In case NM eventually fails to start the ContainerManager server because of
> say a port clash, RM will have to wait for expiry to detect the NM crash.
> It is desirable to make NM register with RM only after it can start all of
> its components successfully.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
