One cluster with different provionces indexes or different clusters with different provionces region

i use es store different region data,such beijing,shanghai,chongqing and other provinces.
should i use provionces name as index name ,and unique all_data
for example
cluster all_data with 9 nodes
elasticsearch.yml all_data

PUT beijing/_doc/1
PUT chongqing/_doc/1
PUT heilongjiang/_doc/1

or use different clusters with each province
cluster beijing with 3 nodes
elasticsearch.yml beijing
cluster chongqing with 3 nodes chongqing
cluster heilongjiang with 3 nodes heilongjiang

PUT /beijing:data/_doc/1
PUT /chongqing:data/_doc/1
PUT /heilongjiang:data/_doc/1

i think the first way is more easy,but cannot Fine-grained add more nodes to special provionces,but the second is more complex,more clusters.
which way should i go?any other suggestions?

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.