hadoop-mapreduce-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Gera Shegalov (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (MAPREDUCE-6721) mapreduce.reduce.shuffle.memory.limit.percent=0.0 should be legal to enforce shuffle to disk
Date Thu, 23 Jun 2016 00:29:16 GMT

    [ https://issues.apache.org/jira/browse/MAPREDUCE-6721?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15345470#comment-15345470
] 

Gera Shegalov commented on MAPREDUCE-6721:
------------------------------------------

Thanks for a prompt commit [~ozawa]

> mapreduce.reduce.shuffle.memory.limit.percent=0.0 should be legal to enforce shuffle
to disk
> --------------------------------------------------------------------------------------------
>
>                 Key: MAPREDUCE-6721
>                 URL: https://issues.apache.org/jira/browse/MAPREDUCE-6721
>             Project: Hadoop Map/Reduce
>          Issue Type: Improvement
>          Components: mrv2, task
>    Affects Versions: 2.6.4
>            Reporter: Gera Shegalov
>            Assignee: Gera Shegalov
>             Fix For: 2.9.0
>
>         Attachments: MAPREDUCE-6721.001.patch, MAPREDUCE-6721.002.patch
>
>
> We are potentially hitting an in-memory-shuffle-related reservation starvation resembling
MAPREDUCE-6445. To work it around, we wanted to disable in memory shuffle via mapreduce.reduce.shuffle.memory.limit.percent=0.0
that turned out to be disallowed by the current logic. So we had to resort to another small
float value such as 0.0001. However, zero is more logical imo.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

---------------------------------------------------------------------
To unsubscribe, e-mail: mapreduce-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: mapreduce-issues-help@hadoop.apache.org


Mime
View raw message