Query service errors when one node is down

6.5.0-1423

-Setup 2 data nodes and 1 query/index

-Bucket with primary index and data loaded

-Disable autofailover

-Shutdown one data node

-Query: "select * from bucket limit 1000"

-30s to run

-Error: "Error performing bulk get operation - cause: {1 errors, starting with dial tcp 10.0.253.113:11210: getsockopt: connection refused}"

Expectation: 

a) faster detection of node down

b) partial results

c) getfromreplica