I would appreciate if someone can provide answers to the following -
1) Is impalad process multi-threaded? From the functionality it seems to me
(heartbeat thread talks with statestored, thread(s) for planner, thread(s)
for coordinator, thread(s) from executor )
2) Does coordinator wait for each fragment to return the whole data before
performing aggregation/sorting/topn operations? Or it starts fetching data
as generated by the fragments running on another node or coordinator node
starts pulling data as it is generated by fragments on another node? Or it
is dependent on the operators created by the planner in the query plan?
Thanks in advance.