Index creation across 3 nodes in couchbase


We have 3 node couchbase cluster. We create an index by connecting to one of the 3 nodes via GUI console. We see the index is created in node1 not in other 2 nodes.
The doubt here is when application will try to query data it will have to go always to node1 to fetch data.
Does not couchbase create indexes in all the nodes. I know it creates in round robin fashion. But how the query will always fetch data from node1 in case node1 goes down.
basically can someone explain me how index creation across nodes work and along with it how querying will work across 3 nodes.

Check Load balancing, Index Replication and High Availability with N1QL

For 5.0.0

cc @deepkaran.salooja