hadoop-common-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Nigel Daley (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HADOOP-4688) multiple spills/reducers test case for mergeParts() needed
Date Thu, 20 Nov 2008 00:07:44 GMT

    [ https://issues.apache.org/jira/browse/HADOOP-4688?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12649220#action_12649220
] 

Nigel Daley commented on HADOOP-4688:
-------------------------------------

+1 to such a test!

> multiple spills/reducers test case for mergeParts() needed
> ----------------------------------------------------------
>
>                 Key: HADOOP-4688
>                 URL: https://issues.apache.org/jira/browse/HADOOP-4688
>             Project: Hadoop Core
>          Issue Type: Test
>          Components: mapred
>    Affects Versions: 0.19.1
>            Reporter: Yuri Pradkin
>
> The first  patch version (incorrect) for http://issues.apache.org/jira/browse/HADOOP-4614
exposed a problem that existing tests do not exercise some code paths of MapOutputBuffer::mergeParts().
 In particular, we need a test that would validate output when numSpills >1 and partitions
>1.
> This might be a test that has more than 1 reducer and forces multiple first level (map)
spills by setting *io.sort.mb* (and possibly *io.sort.spill.percent*) to small values.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message