When using RHBASE with Cloudera's Hadoop, is it necessary to use zookeeper?

[dataquerent]

I had read somewhere that Cloudera's Hadoop installation was an acceptable choice for using RHadoop.
I installed Cloudera's Hadoop on Centos, but my only goal is to use RHadoop.
Cloudera recommended installation of Zookeeper; I installed zookeeper and tried to start up rhbase.
However, hbase thrift start results in errors related to zookeeper.

First question: Is anyone in this group currently successfully using rhbase with Cloudera Hadoop?

Second question: My understanding is that the preferred method is just to type <hbase thrift start> at the console, then type<R> and enter the R environment, at which point one should be able to use commands like <library(rhbase)> and <hb.init()>.  Is this the correct way to use rhbase?

Third question: I have noted that previous people who have asked about zookeeper have been directed to ask Cloudera.  I guess this means that no one in this discussion group currently requires zookeeper to use hbase. Is that correct?

Thanks

[David Champagne]

We've sucessfully tested rhbase on both CDH3 and CDH4.   You have t o start the thrift server in order to use rhbase, and "base thrift start is the normal command to start the thrift server.  Zookeeper is a HBase requirement.  You'll need to refer to the HBase documentation from Cloudera

[Kiran Bhakre]

Zookeeper is requirement for the HBase itself  so it is to be install while installation of HBase .

Regards 

Kiran

[Los]

I am using rhbase with CDH 4.  I would be happy to answer any questions you have about it.

To your second question, yes, that should work.

As others have mentioned, you do need a Zookeeper quorum to start the HBase Thrift server.  On the host you want to run the server, make sure your HBase configuration lists all the hosts in the quorum: /etc/hbase/conf/hbase-site.xml .

-Carlos