[
https://issues.apache.org/jira/browse/CRUNCH-392?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Josh Wills updated CRUNCH-392:
------------------------------
Attachment: CRUNCH-392b.patch
There you go-- that should be much clearer. I took your pastebin code and
turned it into an integration test, and then made a one-line change to
MRExecutor to catch all of the exceptions, not just the IOExceptions, which
causes the test case to run successfully.
> PipelineExecution.getStatus() does not seem to work
> ---------------------------------------------------
>
> Key: CRUNCH-392
> URL: https://issues.apache.org/jira/browse/CRUNCH-392
> Project: Crunch
> Issue Type: Bug
> Components: Core
> Affects Versions: 0.8.2
> Reporter: Jason Gauci
> Assignee: Josh Wills
> Priority: Critical
> Attachments: CRUNCH-392.patch, CRUNCH-392b.patch
>
>
> I wanted a way to kill the underlying hadoop job when performing a control-C
> so I created a job run wrapper. I noticed when the job fails, the
> PipelineExecution.getStatus() still reports running and the binary hangs.
> Please see http://pastebin.com/KgHYKepS for a test case.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
