You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@hdt.apache.org by "Mirko Kaempf (JIRA)" <ji...@apache.org> on 2013/04/06 16:45:16 UTC
[jira] [Commented] (HDT-20) Excessive number of jars added to
MapReduce project classpath
[ https://issues.apache.org/jira/browse/HDT-20?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13624424#comment-13624424 ]
Mirko Kaempf commented on HDT-20:
---------------------------------
I suggest the following:
- Install the Remote System Explorer
- connect to a Cluster-Node, e.g. in my case a node with the CDH4.2 (based on the parcel installation)
- beside the incubator-hdt projects I have a Java-project "hadoop-developement template" to test different things around hadoop. This is my developement target.
- within this project I create a linked-folder (a linked folder within Eclipse) which points to the location of the jar Files within the cluster.
- those jar-files can be added to the projects class path.
I think we need a connector, which either connects directly to the classpath of a certain cluster node (we assume all nodes are equally installed) or
we maintain profiles for Hadoop Installations. The second part would lead to a massive amount of work. So an direct "JAR importer / synchronizer" which grabs a
"jar-profile from a cluster node can solve a lot of problems. In this situation the developer has exactly that JARs which are available in his cluster.
Beside this we should maintain a selection of relevant profile, to work against, even id you are not connected directly with a certain cluster.
(will create a new Issue as "Improvement")
> Excessive number of jars added to MapReduce project classpath
> -------------------------------------------------------------
>
> Key: HDT-20
> URL: https://issues.apache.org/jira/browse/HDT-20
> Project: Hadoop Development Tools
> Issue Type: Bug
> Reporter: Adam Berry
> Priority: Minor
> Original Estimate: 0h
> Remaining Estimate: 0h
>
> After all of the hadoop-core dependencies were added, all of those jars are now added to the bundle classpath, only really hadoop-core needs to be.
--
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