SQL Mirroring Network Question (Private IP or IP on same Subnet)

My DBA is has mirroing set up on 3 SQL database servers.  The database continue to fail over.  We are trying to set up a dedicated NIC to pass the mirroring traffic.  The DBA is stating we need another NIC and IP on the same subnet as the primary NIC/IP and the IP needs to be registered in DNS so his mirroring solution can use FQDN to route the traffic.  My understanding is the NIC/IP should be on a separte private network and not registered in DNS.  Which way is the best practice?

His way:
NIC 1:
NIC 2: (register in DNS or modify the host file)

My way:
NIC 1:
NIC 2: (No register in DNS)
Who is Participating?

[Product update] Infrastructure Analysis Tool is now available with Business Accounts.Learn More

I wear a lot of hats...

"The solutions and answers provided on Experts Exchange have been extremely helpful to me over the last few years. I wear a lot of hats - Developer, Database Administrator, Help Desk, etc., so I know a lot of things but not a lot about one thing. Experts Exchange gives me answers from people who do know a lot about one thing, in a easy to use platform." -Todd S.

Ryan McCauleyEnterprise Analytics ManagerCommented:
It really shouldn't matter - what matters is the network congestion level of the network switches the traffic is passing through. If the switches are busy, it doesn't matter if you're using a dedicated subnet or not - it won't be able to route the traffic without dropping packets. If the congestion point really is your network card (which I'd argue isn't likely unless your server is really, really busy), then adding another NIC will work no matter how  you IP it.

I'm not sure if this is the reasoning being used here, but there's a perception that just putting things on a "dedicated subnet" will improve performance, but I haven't seen that to be the case - if your equipment still goes through the same switches and network equipment, it will face the same congestion. An exception might be if your original subnet requires the packets to leave the switch for some reason and your switches are smart enough to realize that your new "dedicated subnet" exists only on the single switch and can keep packets local. Another exception would be if your switches can be set up with VLAN isolation on them - in that case, then a dedicated VLAN would definitely improve stability, since traffic can be isolated and prioritized appropriately and is it's mingling with regular user traffic. I'd recommend that all of your heartbeat/verification operations (everything extremely time-sensitive) be done on a dedicated VLAN, since it keeps this traffic out of the way.

That said, I don't think the IP scheme you choose matters as much as what's changing when you add that second NIC and why its thought that additional port will improve performance. It may do that, but I think understanding why matters a lot more than the IP address you give the new port.

Experts Exchange Solution brought to you by

Your issues matter to us.

Facing a tech roadblock? Get the help and guidance you need from experienced professionals who care. Ask your question anytime, anywhere, with no hassle.

Start your 7-day free trial
Conrad_BelAuthor Commented:
Thanks for the comment.  A follow up question and statement:

We would use a dedicated VLAN (we call it the service LAN) with isolated traffic.  

Question:  My DBA seems to think it has to have the new second IP registered in DNS for mirroring.  But if it's private on private LAN then it can't register in DNS.  Can you set up mirroing to use the Private IP without DNS entry?

I have set up cluster but not mirroring.  With a cluster you do not register DNS on the heartbeat NIC.
Ryan McCauleyEnterprise Analytics ManagerCommented:
Sure - just provide it the remote IP and you'll be fine without DNS. DNS only exists to let you resolve a friendly name into an IP address - all the network communication with the devices is done via IP anyways and DNS just makes it easy for the actual network users.

Just a note - When you set up the NIC on the private VLAN you'll give it an IP and subnet mask but no default gateway (that's important) - if you provide a default gateway so that two different NICs on different VLANs have one, the server will not have any idea how to reliably route traffic off that subnet and you may lose network connectivity to the server, intermittently or totally. Your second NIC will be able to communicate with any IP on the same subnet (determined by the network mask you use) without a default gateway.
Conrad_BelAuthor Commented:
Understood.  No DFGW and both of the new NICS will have hard coded IP's.

Thanks for your assistance and advice.  Much appreciated.
It's more than this solution.Get answers and train to solve all your tech problems - anytime, anywhere.Try it for free Edge Out The Competitionfor your dream job with proven skills and certifications.Get started today Stand Outas the employee with proven skills.Start learning today for free Move Your Career Forwardwith certification training in the latest technologies.Start your trial today
Microsoft SQL Server 2008

From novice to tech pro — start learning today.