> My bad. Looks like the thrift server is cycling through various AMs it > started when the thrift server was started. I think this is different from > either Hive 2.0.1 or LLAP.
This has been roughly been possible since hive-1.0, if you follow any of the Tez BI tuning guides over the last 4 releases [1]. The issue is more of a "User left Tableau open and went for lunch" capacity issue, so there is no direct connectivity from the JDBC connection -> Tez session. This is the default with LLAP and this specific tuning was off previously when the same Hive was used for ETL, Reporting and BI interchangeably. > Is there a setting I'm missing or is this expected behavior. There's probably some value in removing the round-robin through the pool and switching over to the "latest used first" to help a single user avoid confusion. The queue in TezSessionPoolManager has to be a dequeue, to allow that - so that a single user coming back from lunch has to only cold-start one AM before they get fast queries on LLAP. Cheers, Gopal [1] - https://docs.hortonworks.com/HDPDocuments/HDP2/HDP-2.4.3/bk_performance_tuning/content/section_set_up_interactive_queues_for_HDP_2-2.html