RHive not working with CDH4

Question

Has anyone tried to make RHive work with cdh4? Is it compatible with cdh4?

I have tried asking this question on their google group but no answers yet! I have installed R, RHadoop and all related packages on cdh4 but I am stuck at RHive.

Using cdh4 for all environment variables, rhive.connect() gives me the following error -

WARN conf.Configuration: fs.default.name is deprecated. Instead, use fs.defaultFS
Error in .jfindClass(as.character(class)) : class not found

Any ideas/suggestions?

Thanks, Vaibhav

Answer 1

It looks like RHive was initially incompatible with CDH4, but this was fixed in this pull request. The only difference seems to be that you need to add HADOOP_HOME/client jars with name "hdfs" into classpath.

My suggestion would be to make sure your version includes this fix and if not upgrade to a more recent version of RHive.

Answer 2

RHive is compatible with any Hadoop version before YARN.

I tested it with Hadoop 0.20.205.0, it worked fine