1

In my company we have chosen to implement Multi-site SQL Server Failover cluster. Basically Data center A has 2 nodes (Active/Passive) and Data Center B geographically located acts as the DR. I believe there will be a SAN replication from Data Center A Data Center B.

If in future we want to implement a Multi node AlwaysOn and availability group HADR solution. By that i mean (Data center A has 2 active node and 2xSANs, and DBs are part of a group. Data center B has an Active node and its own SAN which act as a DR site). Can you guide me on how to implement this process or articles that explain this in details.

How can I migrate from a geo-clustreing to an High Availability Disaster Recovery? and what should I consider?

1

After playing with Distributed Availability Groups between geographically dispersed datacenters, my advice is to have an Availability Group in your primary datacenter for local HA, but use log shipping to the secondary datacenter.

This gives you maximum reliability without the pains of a DAG. Failover between datacenters is slower and requires manual intervention, but given that failing between datacenters means that something has gone terribly wrong™ then a little bit of manual intervention is probably acceptable. - jonathan-fite

1

The best practice for your scenario is: 2 nodes in Location A with their subnet say Network_A 192.168.1.x and one node in location B (DR) with subnet Network_B 192.168.2.x.

The weight in Location A is more so you don't need to add disk or folder to your cluster. Your failover Cluster will take 2 IPs one from Network_A and the other from Network_B and your Listener too. The Network_B IPs appears in your cluster offline unless you Failover the DR site, in normal cases you can use SQL database in your DR as read-only.

One important thing: if you have SAN replication don't make your Databases as part of this plan nor part of RDM (raw mapped drive).

Not the answer you're looking for? Browse other questions tagged or ask your own question.