Class HadoopFlow is the Apache Hadoop specific implementation of a {@link Flow}.
HadoopFlow must be created through a {@link HadoopFlowConnector} instance.
If classpath paths are provided on the {@link FlowDef}, the Hadoop distributed cache mechanism will be used to augment the remote classpath.
Any path elements that are relative will be uploaded to HDFS, and the HDFS URI will be used on the JobConf. Note all paths are added as "files" to the JobConf, not archives, so they aren't needlessly uncompressed cluster side.
@see HadoopFlowConnector