hadoop-mapreduce-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Avner BenHanoch (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (MAPREDUCE-4049) plugin for generic shuffle service
Date Wed, 29 Aug 2012 07:57:11 GMT

    [ https://issues.apache.org/jira/browse/MAPREDUCE-4049?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13443886#comment-13443886

Avner BenHanoch commented on MAPREDUCE-4049:

I wish that you can handle this limitation also through this JIRA itself so that both shuffle
consuming and providing will be pluggable and other provider implementation also can work.
its just my wish
I am intending to handle it anyway.  I also need it.  I just said that the consumer side is
standalone and worth commit as is.
In order to commit the fix for the TaskAttemptImpl.java limitation, I just need to know the
desired behavior.  I see two possible ways:
 * I can  call _serviceData.put(...)_ for any auxiliary service
 * We can create new entry in XML, so we'll have 2 sets of AuxiliaryServices, one that gets
APPLICATION_INIT events and one that does not get them.
*If someone tells me the desired way, I will submit it very soon!*

So, won't be there any new interface ShuffleProviderPlugin in mrv2 as it is already pluggable
by adding as part of auxiliary services?
right.  It is not required in mrv2. Yarn arch already provides this functionality thru auxiliary

On a separate note,
*Can someone please tell me how to see javac warnings of Unit Tests?*  My patch is currently
rejected because of such a warning.  I want to make sure this warning is cleaned before I
resubmit the patch.  I am using "mvn test ..." for compiling and running tests.  However,
this hides javac output.  Do you know of a way to see it.
> plugin for generic shuffle service
> ----------------------------------
>                 Key: MAPREDUCE-4049
>                 URL: https://issues.apache.org/jira/browse/MAPREDUCE-4049
>             Project: Hadoop Map/Reduce
>          Issue Type: Improvement
>          Components: performance, task, tasktracker
>    Affects Versions: 1.0.3, 1.1.0, 2.0.0-alpha, 3.0.0
>            Reporter: Avner BenHanoch
>              Labels: merge, plugin, rdma, shuffle
>         Attachments: HADOOP-1.x.y.patch, Hadoop Shuffle Consumer Plugin TLD.rtf, Hadoop
Shuffle Provider Plugin TLD.rtf, mapred-site.xml, mapreduce-4049.patch
> Support generic shuffle service as set of two plugins: ShuffleProvider & ShuffleConsumer.
> This will satisfy the following needs:
> # Better shuffle and merge performance. For example: we are working on shuffle plugin
that performs shuffle over RDMA in fast networks (10gE, 40gE, or Infiniband) instead of using
the current HTTP shuffle. Based on the fast RDMA shuffle, the plugin can also utilize a suitable
merge approach during the intermediate merges. Hence, getting much better performance.
> # Satisfy MAPREDUCE-3060 - generic shuffle service for avoiding hidden dependency of
NodeManager with a specific version of mapreduce shuffle (currently targeted to 0.24.0).
> References:
> # Hadoop Acceleration through Network Levitated Merging, by Prof. Weikuan Yu from Auburn
University with others, [http://pasl.eng.auburn.edu/pubs/sc11-netlev.pdf]
> # I am attaching 2 documents with suggested Top Level Design for both plugins (currently,
based on 1.0 branch)

This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

View raw message