Elasticsearch snapshot in cluster environment

I have implemented a log server using a Elasticsearch cluster which has 2 elasticsearch nodes.

Indexing naming on both servers are same. As an example for index logstash_1, on both elasticsearch nodes shows as logstash_1

So If I create a snapshot on one elasticsearch node index, will it create the full snapshot of that index ?. My exact requirement is illustrated as below.

A few pointers:

  • snapshot is for an index (regardless number/distribution of shards - indices would have their shard spread across the cluster)
  • you cannot create two indices of the same name in the same Elasticsearch cluster. Therefore I assume if you say you have two indices with exact same name on two nodes, you really have 2 clusters of 1 node - or else you really have only one index (on 2 nodes cluster each indices would have its primary and replica shards distributed evenly)
  • In case you have 2 clusters, if you want one snapshot on each cluster you will need to generate one snapshot for each cluster

A few API you can use:

  • http://node_name:9200/_cat/nodes -> this will tell you how many node your cluster has (assuming ES runs on 9200)
  • http://node_name:9200/_cat/indices?v -> this will give you list of indices on your cluster
  • https://node_name:9200/_cat/shards?v -> this will tell you where each shard is located and if replicas are assigned