I'm trying to follow the steps in the Spark 2.0.0 Developer Preview, but am running into an error whenever I try to create a SQLContext.
org.apache.spark.SparkException: Unable to create database default as failed to create its directory maprfs:///hive/warehouse
The path /hive/warehouse is the correct location of where we store our hive tables, but I'm unclear why Spark is trying to recreate that directory. Of course, I don't have permission to create it so the job fails. Note that this happens even if I'm making a test DataFrame; it's not specific to accessing Hive functionality.
I've been at this for a couple days without much progress so I appreciate any tips. Thanks for your help!