Hive + druid (all or most all columns) come back as null

Hi all, I’ve managed to get the my Kafka–> druid working (hive/druid setup on 3 nodes using HDP), and have confirmed my druid queries are now yielding correct results.

I have Hive setup, and (create external tables.etc), doing a select count(*) om my druid datasource yields the correct numbers.

however when ever I try to look at actual columns, the results are nearly always null…

I suspect…maybe the issue is the meta-database?

when I setup my cluster using HDP, I believe I chose derby (embedded) as the metadata store…which I suspect might be the cause?

So my questions are:

When Hive connects to druid, does it also connect to ‘druid’ meta-database? I’m guessing yes…

Is there anyway to ‘migrate’ from derby-embedded to mysql?

i.e. seeing

hive.druid.metadata.uri: jdbc:mysql://localhost:3306/druid

My question is where does the derby db live? same location as overlord? or it replicated to multiple druid nodes?

I’m assuming derby should never be used in a clustered setting.