The "Using Hive with HBase", linked here, contains the following instructions:
To allow Hive scripts to use HBase, add the following statements to the top of each script. Replace the <component_version> strings with current version numbers for CDH, Guava and the Hive HBase handler....
ADD JAR /usr/lib/hive/lib/zookeeper.jar;
ADD JAR /usr/lib/hive/lib/hive-hbase-handler-<Hive-HBase-Handler_version>-cdh<CDH_version>.jar
ADD JAR /usr/lib/hive/lib/guava-<Guava_version>.jar;
For example,
ADD JAR /usr/lib/hive/lib/zookeeper.jar;
ADD JAR /usr/lib/hive/lib/hive-hbase-handler-0.10.0-cdh4.4.0.jar
ADD JAR /usr/lib/hive/lib/guava-11.0.2.jar;
However, when issuing a INSERT OVERWRITE TABLE hbase_table ... command in hive, I received the following error:
Error: java.lang.ClassNotFoundException: org.apache.hadoop.hbase.mapreduce.TableInputFormatBase
This was fixed by adding the following jar in addition to the above jars:
ADD JAR /usr/lib/hive/lib/hbase.jar
And I think you should also point out the fact that people using Hive with HBase don't need to issue the ADD JAR command at the begining of every single script; all they need to do is edit the hive-site.xml file on each node they execute Hive from and include the following property:
<property>
<name>hive.aux.jars.path</name>
<value>file:///usr/lib/hive/lib/zookeeper.jar,file:///usr/lib/hive/lib/hive-hbase-handler-0.10.0-cdh4.5.0.jar,file:///usr/lib/hive/lib/guava-11.0.2.jar,file:///usr/lib/hive/lib/hbase.jar</value>
</property>
