[
https://issues.apache.org/jira/browse/MAPREDUCE-64?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12804855#action_12804855
]
Chris Douglas commented on MAPREDUCE-64:
----------------------------------------
The failing test, {{TestTaskTrackerBlacklist}}, is not related. Filed
MAPREDUCE-1412.
bq. mapreduce.map.sort.record.percent and mapreduce.map.sort.spill.percent
should be mapreduce.map.sort.record.fraction and
mapreduce.map.sort.spill.fraction , respectively. They scale from 0.0 to 1.0,
not from 0.0 to 100.0 or to 100 .
There are a number of floats ranging from 0 to 1 in the configuration named
\*.percent, so this will remain consistent. Since there is a large
renaming/deprecation of variables is in 0.21 (MAPREDUCE-849), feel free to
pursue this point in a separate issue covering all instances.
> Map-side sort is hampered by io.sort.record.percent
> ---------------------------------------------------
>
> Key: MAPREDUCE-64
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-64
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Reporter: Arun C Murthy
> Assignee: Chris Douglas
> Attachments: M64-0.patch, M64-0i.png, M64-1.patch, M64-1i.png,
> M64-2.patch, M64-2i.png, M64-3.patch, M64-4.patch, M64-5.patch, M64-6.patch,
> M64-7.patch, M64-8.patch, M64-9.patch
>
>
> Currently io.sort.record.percent is a fairly obscure, per-job configurable,
> expert-level parameter which controls how much accounting space is available
> for records in the map-side sort buffer (io.sort.mb). Typically values for
> io.sort.mb (100) and io.sort.record.percent (0.05) imply that we can store
> ~350,000 records in the buffer before necessitating a sort/combine/spill.
> However for many applications which deal with small records e.g. the
> world-famous wordcount and it's family this implies we can only use 5-10% of
> io.sort.mb i.e. (5-10M) before we spill inspite of having _much_ more memory
> available in the sort-buffer. The word-count for e.g. results in ~12 spills
> (given hdfs block size of 64M). The presence of a combiner exacerbates the
> problem by piling serialization/deserialization of records too...
> Sure, jobs can configure io.sort.record.percent, but it's tedious and
> obscure; we really can do better by getting the framework to automagically
> pick it by using all available memory (upto io.sort.mb) for either the data
> or accounting.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
