[ https://issues.apache.org/jira/browse/MAPREDUCE-1131?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Aaron Kimball updated MAPREDUCE-1131:
-------------------------------------

    Attachment: MAPREDUCE-1131.patch

This patch adds a parameter ({{mapreduce.task.profile.enable.fetch}}) which defaults to true; if set true, profiler outputs are downloaded by the client. If false, the client does not attempt to fetch the output files. Users of other profilers can disable local fetches so that profiling does not interfere with job success reporting.

This adds a testcase which runs a job on a MiniMR cluster with profiling enabled, but an empty {{agentlib}} string. This causes no output to be generated by a profiler. With fetching disabled, this still allows the job to succeed.

> Using profilers other than hprof can cause JobClient to report job failure
> --------------------------------------------------------------------------
>
>                 Key: MAPREDUCE-1131
>                 URL: https://issues.apache.org/jira/browse/MAPREDUCE-1131
>             Project: Hadoop Map/Reduce
>          Issue Type: Bug
>          Components: client
>            Reporter: Aaron Kimball
>            Assignee: Aaron Kimball
>         Attachments: MAPREDUCE-1131.patch
>
>
> If task profiling is enabled, the JobClient will download the {{profile.out}} file created by the tasks under profile. If this causes an IOException, the job is reported as a failure to the client, even though all the tasks themselves may complete successfully. The expected result files are assumed to be generated by hprof. Using the profiling system with other profilers will cause job failure.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.