1

Looking for some advice here:

I currently have 3 SQL servers in a single AG with a single network card on each server. During times of heavy network usage, we are sometimes seeing failovers to a secondary server. After some research, I see that the way round this is to have a dedicated NIC for the AG traffic on a private IP address range.

What I was wondering is, it is possible to drop and create existing endpoints/listeners with the updated IP addresses or would I have to create a new AG and move databases across to it?

1 Answer 1

3

During times of heavy network usage, we are sometimes seeing failovers to a secondary server. After some research, I see that the way round this is to have a dedicated NIC for the AG traffic on a private IP address range.

I'm skeptical that this will actually fix the issue. You'd basically need to collect network captures to prove the point. You'd also need a separate network if the one you were on was overloaded if you didn't QoS the current traffic.

What I was wondering is, it is possible to drop and create existing endpoints/listeners with the updated IP addresses or would I have to create a new AG and move databases across to it?

It's completely possible to alter/drop the endpoint(s) while everything is running. There is no need to remove the availability group already setup. However, there will be communications issues and failures while the endpoints are down/unavailable.

Again, I'm not sure this is going to actually solve your issue.

2
  • Thank for your reply. The reason I started looking at this, is it seemed to be best practice to have a dedicated network for AG traffic. To complicate things even further, the 3 servers are virtual and I'm struggling to find anything that says multiple networks would help in this type of environment. At the moment, I'm just at the research stage and not about to do anything without further investigation. Commented May 7, 2019 at 11:59
  • @AndrewLackenby , generally with the virtual servers multiple network interfaces won't help and may actually exacerbate the issue. It generally comes down to the host and the guest settings for network interfaces both at a Windows level and at the hypervisor level. You may have to play with the settings but this mostly manifests iteself as extreme latency on the network. Commented May 7, 2019 at 13:32

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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