Lately , it was seen that in our production instances, using the spark-shell the developers could grab access to the whole hive, every database and tables.
Even though we have integrated ranger with hive , using the beeline the ranger policies are honored , but spark is able to bypass the ranger and the users are able to read the data from hive tables.
Does anyone has an idea how to apply authorization for the developers accessing hive through spark?
Or integrate spark with ranger-hive ?