[
https://issues.apache.org/jira/browse/MAPREDUCE-6449?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14904749#comment-14904749
]
Anubhav Dhoot commented on MAPREDUCE-6449:
------------------------------------------
You are mostly right. The objective is to distinguish a local exception versus
a remote exception which happens because of incorrect reuse of
YarnRuntimeException in MR. We do not want to wrap these exception, but use a
MR specific exception type to distinguish between the 2 cases. This is clear in
the case of MRApps.taskType throwing this exception and its callers catching
this exception such as HsWebServices and AMWebServices as you pointed out. If
we are simply using it as an unchecked exception and not expecting to catch it,
we can create a MR specific RuntimeException.
> MR Code should not throw and catch YarnRuntimeException to communicate
> internal exceptions
> ------------------------------------------------------------------------------------------
>
> Key: MAPREDUCE-6449
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-6449
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Reporter: Anubhav Dhoot
> Assignee: Neelesh Srinivas Salian
>
> In discussion of MAPREDUCE-6439 we discussed how throwing and catching
> YarnRuntimeException in MR code is incorrect and we should instead use some
> MR specific exception.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
