...
- When using the HBase Input or Output steps from within a Pentaho MapReduce job you must have the HBase jar on the HADOOP_CLASSPATH on each node running a TaskTracker.
- When using CDH3u1 and above the Hive JDBC driver fails retrieving the last row. You must replace the Hive JDBC driver in Kettle's libext/JDBC directory and PRD's lib/jdbc with this driver hive-jdbc-0.7.0-pentaho-SNAPSHOT.jar.