[
https://issues.apache.org/jira/browse/HADOOP-1432?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12502742
]
Tahir Hashmi commented on HADOOP-1432:
--------------------------------------
News here. Thanks to pushing, shoving and helping from Devaraj, I've got
Grizzly to serve map outputs. In terms of high level changes, StatusHttpServer
now has a new constructor in addition to the earlier one. The new constructor
accepts a class that implements Grizzly's TCP Adapter. Alert! this makes the
constructor dependent on Grizzly but then constructors are disposable ;-). If
this constructor is called, StatusHttpServer will start Jetty as well as
Grizzly. By default, Grizzly will be started on a random port between 10k and
64k -1. In case the port is already taken, port rolling has been provided for
and it happens independent of jetty port rolling. The mapOutputPort is a new
piece of information which is ultimately communicated to Reduce tasks.
Unit tests pass. RandomWriter on a single node works. You need to apply the
attached patch and unpack the attached tarball in src/lib to make things work.
Comments invited.
> Use Grizzly for Fetching Map Output in Shuffle
> ----------------------------------------------
>
> Key: HADOOP-1432
> URL: https://issues.apache.org/jira/browse/HADOOP-1432
> Project: Hadoop
> Issue Type: Improvement
> Components: mapred
> Reporter: Tahir Hashmi
> Assignee: Tahir Hashmi
> Attachments: 1432.patch, grizzly.tgz
>
>
> As mentioned in HADOOP-1273 and references therefrom, Jetty 6 still doesn't
> seem to be stable enough for use in Hadoop. Instead, we've decided to
> consider the usage of Grizzly Framework [https://grizzly.dev.java.net/] for
> NIO based communication.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
