The goal of this package is to allow convenient processing on a Hadoop cluster of large data sets. It is based on rmr2 but should be easier to use and more abstracted from the underlying mapreduce computational model.
Please visit the RHadoop wiki for details or go directly to the tutorial for a gentle introduction.