Description
When the datasource is kerberized, the behavior of reloading kerberos token is different between "Cube Build" and "Load Hive Table".
I summarized the detail of this behavior in the attached image.
I think BeelineHiveClient#init shuold be implementad like following.
http://appcrawler.com/wordpress/2015/06/18/examples-of-connecting-to-kerberos-hive-in-jdbc/