[
https://issues.apache.org/jira/browse/MAPREDUCE-7166?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Lei Li updated MAPREDUCE-7166:
------------------------------
Target Version/s: 3.3.0
Description: When node lost, all mappers on this node will be
rescheduled, but it's unnecessary for succeeded map only task because map only
job has no reducer, and map output has been written to HDFS when it become
succeeded, so we can ignore this node lost event. (was: map only job have no
reducers, so when a mapper task is succeeded, it's output have been written to
HDFS. )
> map-only job should ignore node lost event when task is already succeeded
> -------------------------------------------------------------------------
>
> Key: MAPREDUCE-7166
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-7166
> Project: Hadoop Map/Reduce
> Issue Type: Improvement
> Components: mrv2
> Affects Versions: 2.7.5, 3.0.0, 3.1.1, 2.8.5, 2.7.7, 2.9.2
> Reporter: Zhaohui Xin
> Assignee: Lei Li
> Priority: Major
> Labels: newbie
> Attachments: MAPREDUCE-7166.001.patch, MAPREDUCE-7166.002.patch
>
>
> When node lost, all mappers on this node will be rescheduled, but it's
> unnecessary for succeeded map only task because map only job has no reducer,
> and map output has been written to HDFS when it become succeeded, so we can
> ignore this node lost event.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
