[
https://issues.apache.org/jira/browse/MAPREDUCE-4049?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13250275#comment-13250275
]
Mariappan Asokan commented on MAPREDUCE-4049:
---------------------------------------------
Hi Avner,
I worked on MAPREDUCE-2454(to make sort pluggable in Hadoop) and posted a
patch on top of trunk version 1221902 a while back. The patch was created on
top of the trunk since ReduceTask.java was already refactored nicely and I was
advised to work on the trunk version.
Please take a look at the patch file mapreduce-2454.patch posted in
MAPREDUCE-2454. If you want, I can post a patch on top of the latest trunk.
The patch decoupled the merge from shuffle by creating ShuffleRunner and
ShuffleCallback interfaces. The MergeManager implements the ShuffleCallback
and the shuffle itself implements ShuffleRunner interface.
Since you are making shuffle as pluggable, I notice some overlapping changes.
If I can be of any assistance to reduce the conflict between our patches,
please let me know. Meanwhile, I will go over the details of your patch and
get back. Do you have a patch created on top of trunk?
Also, I would like to hear opinions from other developers who have shown
interest in this Jira.
> 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: 0.23.1, 1.0.1
> Reporter: Avner BenHanoch
> Labels: merge, plugin, rdma, shuffle
> Attachments: HADOOP-1.0.2.patch, HADOOP-1.0.x.patch, Hadoop Shuffle
> Consumer Plugin TLD.rtf, Hadoop Shuffle Provider Plugin TLD.rtf,
> MAPREDUCE-4049-branch-1.0.2.patch, mapred-site.xml, mapred.diff, src.tgz,
> test.diff
>
>
> 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:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
