FYI my table is partition by Year + Month + Day. Total file size in HDFS is 10TB. Total records is 21 Billion records. We have 8 data nodes in HDP. I'm using HDP-3.0.1.0 with Ambari. Here's my current Hive config: Tez Container Size: 3072 MB HiveServer2 Heap Size: 4096 MB Memory: 819.2 MB Data per Reducer: 2042.9 MB They are mostly the default values. Do they make sense? Any suggestion on which to increase / decrease for optimum performance?
... View more
