HA deployment via NSX

maguphe

We are planning to build HA Orion Server Deployment (NPM, NCM, and SAM) across two data centers (~40ms latency).

The two data centers will be connected via stretched NSX, so both data centers will be on a single subnet.

With the above environment, would we be able to deploy HA deployment (single-subnet method)?

As for the SQL database, we will be using the cluster SQL build-out, but would we need another cluster of backup SQL server in the secondary DC?

Would we also require a load balancer to assign the VIP?

maguphe

It's odd to answer my own question, but I was able to get some answers from a SolarWinds sales engineer. 

It seems like 200ms may be the ideal cut-off line for RTT between a poller and a DB, so I should be good with about 40ms RTT between the two data centers.

I believe the remote pollers can be as far as 500ms away from the main poller FYI.

As for the SQL database clustering, we've decided to deploy a single SQL server from each data center and try to cluster them across the NSX environment.

And, we wouldn't need a load balancer as the VIP will be assigned directly to both the main pollers, and the failover will occur automagically once the main poller loses connectivity.