[ http://issues.apache.org/jira/browse/HADOOP-59?page=comments#action_12368188 ]
Doug Cutting commented on HADOOP-59: ------------------------------------ Michel, I think that -D should define Hadoop configuration properties, not JVM system properties. And I don't think Hadoop configurations should by default include all of the JVM's system properties. Finally, please include good user-level javadoc on all public & protected items, attach Apache's license at the top of each file, try to bundle things into a single patch file, etc. These kinds of things make it much easier for me to commit a patch. Ideally all that I need to do is read the patch to see that it looks reasonable, apply it with 'patch -p 0 < patchFile', run unit tests and commit. Including new unit tests is also a good idea. If I need to clean up javadoc, licenses, formatting, etc. before I can commit then I will be less inclined to process a contribution promptly. Doug > support generic command-line options > ------------------------------------ > > Key: HADOOP-59 > URL: http://issues.apache.org/jira/browse/HADOOP-59 > Project: Hadoop > Type: Improvement > Components: conf > Versions: 0.1 > Reporter: Doug Cutting > Assignee: Michel Tourn > Fix For: 0.1 > Attachments: Tool.java, ToolBase.java, bashfile.patch, toolbase.patch > > Hadoop commands should all support some common options. For example, it > should be possible to specify the namenode, datanode, and, for that matter, > any config option, in a generic way. > This could be implemented with code like: > public interface Tool extends Configurable { > void run(String[] args) throws Exception; > } > public class ToolBase implements Tool extends Configured { > public final void main(String[] args) throws Exception { > Configuration conf = new Configuration(); > ... parse config options from args into conf ... > this.configure(conf); > this.run(); > } > } > public MyTool extends ExcecutableBase { > public static void main(String[] args) throws Exception { > new MyTool().main(args); > } > } > The general command line syntax could be: > bin/hadoop [generalOptions] command [commandOptions] > Where generalOptions are things that ToolBase handles, and only the > commandOptions are passed to Tool.run(). The most important generalOption > would be '-D', which would define name/value pairs that are set in the > configuration. This alone would permit folks to set the namenode, datanode, > etc. -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://issues.apache.org/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira
