[
https://issues.apache.org/jira/browse/MAPREDUCE-3479?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Tom White updated MAPREDUCE-3479:
---------------------------------
Attachment: MAPREDUCE-3479.patch
The failing test was indicating an underlying problem with the first patch:
passing the Cluster object to the Job causes the configuration objects to get
out of sync (which is what MAPREDUCE-1788 was addressing) so the YARNRunner (or
LocalJobRunner) would use the configuration from the Cluster, which would not
reflect any changes made to it during job submission (e.g. setting distributed
cache properties). So instead this patch updates the Cluster object in the
JobClient to be the one used for job submission.
All tests pass for me now.
> JobClient#getJob cannot find local jobs
> ---------------------------------------
>
> Key: MAPREDUCE-3479
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-3479
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Components: client
> Affects Versions: 0.23.0
> Reporter: Tom White
> Assignee: Tom White
> Attachments: MAPREDUCE-3479.patch, MAPREDUCE-3479.patch
>
>
> The problem is that JobClient#submitJob doesn't pass the Cluster object to
> Job for the submission process, which means that two Cluster objects and two
> LocalJobRunner objects are created. LocalJobRunner keeps an instance map of
> job IDs to Jobs, and when JobClient#getJob is called the LocalJobRunner with
> the unpopulated map is used which results in the job not being found.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
