Confusion with Node, Cluster, Active Data and Replica Data.

I am still confused with Node, Cluster, Active Data and Replica Data.

Let's say, I have 2 clusters.
1 cluster has 2 servers.
Each server has 1 Active Node and 1 Replica Node

This is what I am understanding:
Server1
(1)Active Node :Doc1, Doc2, Doc3
(2)Replica Node: Doc2, Doc4, Doc6

Server2
(1)Active Node :Doc4, Doc5, Doc6
(2)Replica Node: Doc1, Doc3, Doc5

1. Docs will be saved across Active nodes.
2. Replica Node in each server does not save the completely identical list of doc saved in Active Node in the same server. But all replica docs are saved across the all Replica Nodes

Do I understand correctly so far?

Now if I want to have a DR cluster, should I have a whole set of cluster like the original, then, XDCR data from the original cluster to the DR cluster? If doing so, do I find the same docs are stored in the same node just as in the original cluster?

1 Answer

« Back to question.

Hi there, in your definition you use server and node as if they are 2 separate containment. So I'll rename you node to vbucket. vbucket is a physical partition of a bucket. there are many (1024) vbuckets in a buckets and vbuckets can be a active or a replica.

2 clusters. 1 cluster has 2 servers. Each server has some Active vbuckets and some Replica vbuckets.

Server1
Active vbuckets on server1 :Doc1, Doc2, Doc3
Replica vbuckets on server1: Doc4, Doc5, Doc6

Server2
Active vbuckets on server2 :Doc4, Doc5, Doc6
Replica vbuckets on server2: Doc1, Doc2, Doc3

Answers:
- docs are written to active vbuckets.
- replica vbuckets in each server save an identical copy of the active vbucket.

Hope this helps.
-cihan