Connection to different nodes of a cluster


I wanted to understand about multiples nodes of a cluster. We got different servers like Data, FTS, Index, Query for 1 cluster. Therefore, considering the best practice, should we configure connections separately for Create/Update and Query.

We are using .Net Core Technology stack. Currently I have configured Query server for all CRUD operations that includes FTS operations too. So, what I wanted to know is :

  1. For create and update scenarios should we configure and make use of only Data server/nodes exclusively in the connection and use in while Create/Update document?
  2. For Query scenarios( N1QL and FTS) , should we configure only Query specfic server/nodes connections separately and use it for retrieval scenarios?

Hi @nnaavveen,

Short version:

When using the .NET SDK, it doesn’t matter which nodes your specify. The SDK will discover them all.

Long version:

When using the .NET SDK, you specify a list of one or more Servers in the cluster. The SDK will connect to the first node it finds in that list. From that point on, Couchbase will give the SDK a complete map of every server in the cluster. As a developer, you don’t really need to know which servers have which services.

However, it’s a good idea to list as many nodes as possible in your configuration, just in case one of the nodes becomes inaccessible.