Major Description
The rest of the Hadoop ecosystem (Hive, SparkSQL, Impala) can't handle uppercase or non-ascii characters in table and column names. Users who create these tables through the Kudu client may not recognize that the table won't be addressable by SparkSQL and Impala. Removing support would be a breaking change, but we can provide a flag on the master that will reject any CREATE/ALTER statements which introduce non-conforming names.
Additionally, we should fix the CREATE EXTERNAL TABLE Impala logic to display an error when the column/table name doesn't conform.