Showing results for 
Search instead for 
Did you mean: 

4 nodes HA setup possible ?

Hi Gurus,
                     Herewith I come across one more query regarding HA setup in windows / solaris platform.

Is it possible to make HA setup with 4 nodes ? This is how my setup should looks like

Server A (in primary ) --- Server A1 (in secondary) (This is the back for my server A)

Server B (in primary) - server B1 (in secondary) (this is the back for my server B)


Here i am using replication network to replicate the data between the servers. Is this possible within single cluster setup.  I heard like we are supporting upto 32 nodes, but the possibility for back up server is only 1. Is that so ?

If so , how can  i acheive the above setup. I come across with an idea like creating 2 clusters (each for 2 pair of setup). Will that work fine . ???

any comments on this is much appreciated.



2 Replies

Hi Yes you can have upto 32

Yes you can have upto 32 nodes in the same cluster, note that the nodes will need the same o/s.

If I understood your scenario, you are planning on having two servers at one site A and two nodes at site B, with replication between the sites. If you have the private network infrastucutre (for heartbeats) you could configure this as one cluster or two clusters.

Look up "Replicated Data cluster" in the VCS Users guide for an overview.


From a Windows SFW HA point

From a Windows SFW HA point of view:

Single cluster with replication between nodes = Replicated Data Cluster (RDC)
Multipl clusters with replication between cluster nodes = Global Cluster Option (GCO)

Within a single cluster setup, you will require using a synchronous replication so that all nodes have access to exact same data at all times.
This is to ensure that a local failover can occur without any delays or difference in data between the replicated volumes / disks.
(Synchronous replicaion may cause delays in data write acknowledgements to the apps if the replication is slow so needs a good/fast network)
Usually, RDC is created if you have 1 node using 1 storage array and 1 node using another within a very close physical location. (fast link)

In a global cluster, you actually setup minimum 2 separate clusters. 1 primary and 1 dr for example.
Primary cluster will have multple nodes for local failovers from 1 storage.
DR cluster will have replicated data from the primary cluster. (Usually Asynchronous using slower WAN links for replication)

Your scenario can be setup using either of two methods depending on what you think it is capable of.

Cluster 1 = Server A & Server A1

Cluster 1 = Server A
Cluster 2 = Server A1