[
https://issues.apache.org/jira/browse/MAPREDUCE-6160?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14219049#comment-14219049
]
Rohith commented on MAPREDUCE-6160:
-----------------------------------
bq. Is there any reason why the job is not available in History? I see in your
above comment the Job got succeeded, Do you see any exception in in MR AM/JHS
log for the same job?
MRAppMaster was failed to write to HDFS because of safemode and safemode was
recovered later.
> Potential NullPointerException in MRClientProtocol interface implementation.
> ----------------------------------------------------------------------------
>
> Key: MAPREDUCE-6160
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-6160
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Reporter: Rohith
> Assignee: Rohith
> Attachments: MAPREDUCE-6160.patch, MAPREDUCE-6160.patch
>
>
> In the implementation of MRClientProtocol, many methods can throw
> NullPointerExceptions. Instead of NullPointerExceptions, better to throw
> IOException with proper message.
> In the HistoryClientService class and MRClientService class has
> #verifyAndGetJob() method that return job object as null.
> {code}
> getTaskReport(GetTaskReportRequest request) throws IOException;
> getTaskAttemptReport(GetTaskAttemptReportRequest request) throws IOException;
> getCounters(GetCountersRequest request) throws IOException;
> getTaskAttemptCompletionEvents(GetTaskAttemptCompletionEventsRequest request)
> throws IOException;
> getTaskReports(GetTaskReportsRequest request) throws IOException;
> getDiagnostics(GetDiagnosticsRequest request) throws IOException;
> {code}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
