[
https://issues.apache.org/jira/browse/MAPREDUCE-6765?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Haibo Chen updated MAPREDUCE-6765:
----------------------------------
Attachment: mapreduce6765.001.patch
The patch is trivial. In cases where the resource requested by container is
not supported, do not scheduler container requests, since the job will soon be
killed.
There is no test since the fix is trivial, and ScheduledRequests is an inner
class whose ScheduledRequests.addMap or ScheduledRequests.addReduce are hard to
spy on.
> MR should not schedule container requests in cases where reducer or mapper
> containers demand resource larger than the maximum supported
> ---------------------------------------------------------------------------------------------------------------------------------------
>
> Key: MAPREDUCE-6765
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-6765
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Components: mr-am
> Affects Versions: 2.7.2
> Reporter: Haibo Chen
> Assignee: Haibo Chen
> Priority: Minor
> Fix For: 2.9.0
>
> Attachments: mapreduce6765.001.patch
>
>
> When mapper or reducer containers request resource larger than the
> maxResourceRequest in the cluster, job is to be killed. In such cases, it is
> unnecessary to still schedule container requests.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
