Replicating data between an Elastic Cluster Nodes

[A.Hani]

Hi
New to Elastic here and would like to get help with some inquiries.
Scenario:

1- Let's assume that I have a cluster of (2) master nodes, is it possible to configure each node to hold its own data? In other words, Node N1 only has the data that it's receiving and this data does not get replicated to N2 and vice versa.

2- If I got a one-node Elasticsearch cluster up and running with the default configurations and containing data, would it be possible to reconfigure this node for another node to join the cluster without touching the data it holds, or will I have to redo the installation from scratch?

[Christian_Dahlqvist]

@A.Hani

1- Let's assume that I have a cluster of (2) master nodes, is it possible to configure each node to hold its own data? In other words, Node N1 only has the data that it's receiving and this data does not get replicated to N2 and vice versa.

That is what will happen if they are set up as two separate clusters. If they form a single cluster that is not possible. Also note that you need 3 master eligible nodes in order to achieve a highly available cluster that can continue operating if one node goes down.

@A.Hani

2- If I got a one-node Elasticsearch cluster up and running with the default configurations and containing data, would it be possible to reconfigure this node for another node to join the cluster without touching the data it holds, or will I have to redo the installation from scratch?

If you add a node to the cluster data will be distributed across the cluster. You can to some extent control this through shard allocation filtering but I would recommend against this for this type of small cluster.

Why do you feel the need to control where the data resides?

[A.Hani]

Thank you for the prompt reply. It's just that I'm having a situation where I got a production node up and running and there is a need to add another node and make it join the cluster but I do not have enough resources (disk storage) for that second node to hold the data from the First node.

Regarding My second question : [quote="A.Hani, post:1, topic:322737"]
2- If I got a one-node Elasticsearch cluster up and running with the default configurations (Bootstrapped) and containing data, would it be possible to reconfigure this node for another node to join the cluster without touching the data it holds, or will I have to redo the installation from scratch?
[/quote]