“It is also recommended for an MSSQL Always-On Availability Group to have listeners pre-configured on both the source and target IP ranges too”
My question is how to do this?
Using the Windows Failover Cluster management console (WFCMC), adding a “Client Access Point” to the Role, the IP address you are permitted to add is restricted to the subnets currently known to the cluster.
I’ve also used WFCMC to give the cluster core services their own additional IP addresses for use when in DR.
Temporarily adding another IP address from the DR or ISOLATED subnet to the NIC, does allow an IP address from those subnets to be defined to the cluster and role. Can anyone provide documentation on whether this is the correct approach?
We’ve discovered that when such a cluster is rebooted in production, the role brings all three IP addresses online, but worse still, it registers all of them with DNS. Clients receive all three addresses when resolving for the SQL always on listener name: production IP, Isolation IP and DR IP. This causes problems!
Manually taking those IP addresses offline using WFCMC does not de-register them from DNS, they need to be removed from DNS manually.
What is the correct way to prepare a SQL always-on availability group to fail over to a DR site using a different IP subnet, in such a way that production is not negatively affected?
Mark F.
