1: don't treat trivial clusters as clusters - Alibaba uses this config
2: report synchronous failures immidiately and accurately

fix #2642

For context, Alibaba reports from cluster nodes as 2 nodes:

aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa {proxy endpoint} myself,master - 0 0 0 connected 0-16383
bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb {proxy endpoint} slave aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa 0 0 0 connected

which confuses things somewhat; this isn't a valid/normal cluster configuration, so: ignore it!

Other relevant metadata:

from config:

from info:

# Cluster
cluster_enabled:1
databases:256
nodecount:2