Major Description
Hi All,
I am using Apache Drill 1.8.0 on AWS EMR and joining two hive tables. Below is sample query. This working fine in Drill CLI but giving below error after running few minutes. If i try simple select query (select t1.col from hive.table t1) it works fine in both Drill CLI and UI. Only problem with join query.
If i cancel the join query from background, it displays results in UI. This is very strange situation.
Drill has been installed on a AWS node which has 32 GB RAM and 80 GB storage. I didn't specify memory separately to Drill. I am trying to join two tables have rows 4607818 & 14273378 respectively. Please find attached drillbit.log file as well.
Only my confusion here is Drill CLI is working fine with this data and showing output then why UI doesn't display output and throwing error. While running query from UI, if i cancel it from background then UI immediately display result.
I believe i am missing some configuration (which is holding o/p in buffer and after cancellation display result) here and need your help.
Join Query
--------------
select t1.col FROM hive.table1 as t1 join hive.table2 as t2 on t1.col = t2.col limit 1000;
Error
-------
Query Failed: An Error Occurred org.apache.drill.common.exceptions.UserRemoteException: SYSTEM ERROR: RpcException: Data not accepted downstream. Fragment 1:4 [Error Id: 0b5ed2db-3653-4e3a-9c92-d0a6cd69b66e on ip-172-31-16-222.us-west-2.compute.internal:31010]