[
https://issues.apache.org/jira/browse/MAPREDUCE-1912?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12909165#action_12909165
]
Vinay Kumar Thota commented on MAPREDUCE-1912:
----------------------------------------------
Current approach is bit confusing for me as a user. So, I have one suggestion
for using the Rumen commands as 'mrfoldtrace' and 'mrgentrace' options in the
top level itself instead of using -fold and -generate-trace options under
Rumen. In my suggestion, the output is in the following way.
If user enter the following command, the output for specific to Rumen should be
in the following way.
${HADOOP_HOME}/bin/hadoop jar build/hadoop-mapred-tools-0.22.0-SNAPSHOT.jar
Output is
mrfoldtrace : Command for fold/scale a mapreduce job trace.
mrgentrace: Command for generating the mapreduce job trace with MapReduce job
history logs.
> [Rumen] Add a driver for Rumen tool
> ------------------------------------
>
> Key: MAPREDUCE-1912
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-1912
> Project: Hadoop Map/Reduce
> Issue Type: Improvement
> Components: tools/rumen
> Affects Versions: 0.22.0
> Reporter: Amar Kamat
> Assignee: Amar Kamat
> Fix For: 0.22.0
>
> Attachments: mapreduce-1912-v1.1.patch, mapreduce-1912-v1.7.4.patch,
> mapreduce-1912-v1.7.7.patch
>
>
> Rumen, as a tool, has 2 entry points :
> - Trace builder
> - Folder
> It would be nice to have a single driver program and have 'trace-builder' and
> 'folder' as its options.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
