Mailing-List: contact hadoop-dev-help@lucene.apache.org; run by ezmlm
Precedence: bulk
Reply-To: hadoop-dev@lucene.apache.org
Message-ID: <19852644.1186010153187.JavaMail.jira@brutus>
Date: Wed, 1 Aug 2007 16:15:53 -0700 (PDT)
From: "Owen O'Malley (JIRA)" <jira@apache.org>
To: hadoop-dev@lucene.apache.org
Subject: [jira] Commented: (HADOOP-1437) Eclipse plugin for developing and
 executing MapReduce programs on Hadoop
In-Reply-To: <2758007.1180141036124.JavaMail.jira@brutus>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


    [ https://issues.apache.org/jira/browse/HADOOP-1437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12517109 ] 

Owen O'Malley commented on HADOOP-1437:
---------------------------------------

I actually did play with it and sent the feedback straight to Eugene. I guess I should post it here, too.

I tried it out and it looks pretty good.

1. It is unfortunate that you need to unzip into the eclipse installation directory rather than using the eclipse plugin management tools.

2. I could use it to define Mappers and Reducers.

3. When I tried to define a Launcher, it didn't find any Mappers or Reducers, even though I'd just made them with the wizard.

4. When I defined a Launcher by hand, I didn't find a way to tell the plugin about it, so it wouldn't let me run a map/reduce application.

5. I used "Connect to a MapReduce cluster" and the first time I didn't provide a login, and it didn't actually connect to anything. I could not figure out how to delete it. The login seems to be required for it to work, but not required by the form.

6. I don't understand why it wants a login/password to connect to the cluster. I never feel comfortable putting a real password into non-system software.

7. After I made a second connection to a cluster with a username, I could browse the dfs.

8. Clicking on "Transfer files to the HDFS" didn't do anything that I could see.

9. It might be nice to have the template parameterized by the types of the map output key/value types and the reduce output key/value types.

10. Most reduces have the form:

public void reduce(WritableComparable _key, Iterator values, OutputCollector output, Reporter reporter) throws IOException {
  KeyType key = (KeyType) _key;
  while (values.hasNext()) {
     ValueType value = (ValueType) values.next();
     // process value
  }
}

and it might be worth making the template provide more guidance to newbies.

I think it is great that IBM is creating this plugin.


> Eclipse plugin for developing and executing MapReduce programs on Hadoop
> ------------------------------------------------------------------------
>
>                 Key: HADOOP-1437
>                 URL: https://issues.apache.org/jira/browse/HADOOP-1437
>             Project: Hadoop
>          Issue Type: New Feature
>         Environment: Eclipse 3.2.0+, Java 1.5.0+
>            Reporter: Eugene Hung
>         Attachments: mrt-eclipse-1.0.4.zip
>
>
> An Eclipse plugin for developing and executing MapReduce programs on remote Hadoop servers.  Automatically provides templates for creating Map/Reduce classes, transparently bundles the classes into JAR files and sends them to a remote server for execution.  Allows the user to easily view status of Hadoop jobs and browse/upload/delete files from the Hadoop DFS within the Eclipse IDE.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.