[
https://issues.apache.org/jira/browse/TUSCANY-2471?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12612633#action_12612633
]
Jean-Sebastien Delfino commented on TUSCANY-2471:
-------------------------------------------------
That looks like a good starting point for experimenting with Hadoop.
To help others in the project try it, could you please provide a Maven pom.xml
for that code?
I guess the Maven pom should build the few classes you have in your patch, the
Wordcount jar, and reference the required dependencies from Hadoop. The hadoop
JARs do not seem to be available in the Apache Maven repos yet (unless I missed
them) so I suggest the following:
1. ask the Hadoop project (on their dev list) if they already have their JARS
in a public repos and if not if they could please publish them
2. in the meantime install the Hadoop JARs manually in your local Maven repos
and write in this JIRA or better in a README file the instructions to do it
(for others who will want to try your code)
Thanks!
> Two small test functions that submit a MR job without the console, and the
> start of a Java Component that submits a MR job.
> ---------------------------------------------------------------------------------------------------------------------------
>
> Key: TUSCANY-2471
> URL: https://issues.apache.org/jira/browse/TUSCANY-2471
> Project: Tuscany
> Issue Type: New Feature
> Environment: Mac OS 10.5.2
> Reporter: Chris Trezzo
> Priority: Minor
> Attachments: patch-JIRA-2471, wordcount.jar
>
>
> The Test class submits a MR job using the runJar API. (This is basically
> doing the same thing as the hadoop shell script)
> The Test2 class submits a MR job without calling the main or run method in
> org.apache.hadoop.examples.WordCount, but the job is still submitted using a
> JAR file.
> The services package includes incomplete code for a java SCA component that
> will do the same thing as Test2 through Tuscany.
> In order for these methods to work, Hadoop must be running (I have it running
> in pseudo-distributed mode). Also the Hadoop library and the Hadoop/conf
> directory must be included in the class path.
> Attached is the patch and wordcount.jar
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
