Need GP tool for executing mapreduce jobs using HCatalog

[climbage]

Justification

• Oozie has far too complicated of a set up for simply executing a a jar file
• HCatalog is now part of Hive which comes standard with most Hadoop distributions
• Simple rest interface

% curl -s -d user.name=ctdean \
   -d jar=wordcount.jar \
   -d class=org.myorg.WordCount \
   -d libjars=transform.jar \
    -d arg=wordcount/input \
    -d arg=wordcount/output \
    'http://localhost:50111/templeton/v1/mapreduce/jar'

See https://cwiki.apache.org/confluence/display/Hive/WebHCat+Reference+MapReduceJar.