Providing Hive and HCatalog Libraries for the Sqoop Job
With the support for HCatalog added to Sqoop, any HCatalog job depends on a set of jar files being available
both on the Sqoop client host and where the Map/Reduce tasks run. To run HCatalog jobs, the environment
variable HADOOP_CLASSPATH
must be set up as shown below before launching the Sqoop HCatalog jobs:
HADOOP_CLASSPATH=$(hcat -classpath) export HADOOP_CLASSPATH
The necessary HCatalog dependencies will be copied to the distributed cache automatically by the Sqoop job.