I've just had a new article published on IBM Developerworks that looks at how to use Hadoop and Couchbase together. This enables you to take advantage of the huge map/reduce functionality on very large datasets within Hadoop and combine that with the querying and indexing functionality in Couchbase Server 2.0 to enable you to pull out specific details.
This effectively adds querying to Hadoop using the familiar map/reduce structure for both your big data processing at the front end, and your redcued data set once the information is stored into Couchbase.
You can read the full article: https://www.ibm.com/developerworks/opensource/library/ba-hadoop-couchbase/