Server Fault is a question and answer site for system and network administrators. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I have a PowerEdge server that has 1 management NIC and 4 other NICs. I have the management NIC set up so that I can access iDRAC and it works great. Now I'd like to set up Windows 2008 R2 Hyper-V. Our network has two kinds of IP addresses. One is the 10.10.x.x kind that you can't get to from outside the network. The other is a public kind, like 128.210.x.x.

I thought I should set up one NIC with a 10.10.x.x IP address and the other with as 128.210.x.x IP address so that the Hyper-V parent partition can use the 10.10.x.x one and be relatively safe from outside attack. Then the child partitions/guest OS's running on top of Hyper-V could use the public NIC, so that people could get at them from anywhere (some of these are web servers).

When I try to set them up this way, Windows complains about having multiple default gateways. If I give both NICs public IPs then things work -- that is, both parent and child partitions can get out to the internet. Could someone who knows vastly more about networking than me explain what I should be doing here? The basic idea is to have one VM running a file server, accessible on a 10.10.x.x IP or via WINS like \FILESERVER and then to have a bunch of other VMs that are separate webservers. Thanks.

share|improve this question

Your problem stems from a misunderstanding about default gateways. There is only one total, not one per NIC. Setting additional default gateways is for redundancy only.

You only have one default gateway, but you still need to set route for the other interfaces for non-default situations. If you don't do this, all traffic leaving your server will attempt to go out of the interface that has the default gateway on it, since that is the only gateway. It's pretty obvious why this won't work well for you.

So, in conclusion, if you define your default gateway for the 10.x.x.x interface, you need to add static routes for the 128.210.x.x interfaces which define the next hop for traffic leaving those NICs.

share|improve this answer
So I would set up the first NIC with a 10.10.x.y IP and a default gateway of 10.10.x.254, and then the next NIC as 128.210.x.y and just leave the default gateway blank in the settings dialog? And then set up a route (using something like ) to route 128.210.x.y to...what? The router at 128.210.x.254? – user21715 Sep 1 '10 at 3:27
So maybe I should enumerate an example. Suppose we have the following: DRAC:, gateway Then NIC1 (Windows Server 2008 R2):, gateway Finally, Hyper-V VMs: Webserver1 128.210.x.77 Webserver2 128.210.x.81 Fileserver So those VMs should use NIC2. – user21715 Sep 1 '10 at 3:55
@smorfnip - ya, pretty much. You can leave the DRAC out of these examples, it's superflous since it is disjoint from the host OS. – MDMarra Sep 1 '10 at 3:58

It's possible that windows is throwing a fit about shared NICs. The recommended setup with HyperV is to dedicate one NIC to the host OS and one NIC for the use of HyperV guests (minimum). I believe the idea is that no NICs are shared between the host and VMs. So long as you have at least one that shows up only in the host's networking, and another that is only assigned to one of the HyperV networks, anything you do with the remaining NICs is fine.

Though you could certainly set up multiple IP ranges on the same physical NIC for the guests to use.

share|improve this answer
I believe that he does have 1 NIC for management and the others for guests. Unless I am misunderstanding. – MDMarra Sep 1 '10 at 2:27
If iDRAC works the way I believe it does, its NIC is not accessible to the OS at all (but only to its controller card), so I'd probably use one for the host, and the other 3 for guest use. – Kara Marfia Sep 1 '10 at 2:32
@Kara - Right, He then goes on to say that the 10.x.x.x interface is for the parent partition, which is the equivalent of the management interface, correct? I don't use Hyper-V, but that was my understanding. – MDMarra Sep 1 '10 at 2:44
@MarkM I understand 'parent' to refer to the host OS. implies that the NIC it uses is unavailable to anything else. the 10.x.x.x is just the internal-only address space, likely NAT'ed to the external 128.210.x.x addresses, but I thought I'd save questions about firewalls until after we nailed down the first issue. /grin – Kara Marfia Sep 1 '10 at 2:48
@Kara - I guess we'll wait for the OP to update. I read it as, he has the DRAC working, he has a NIC on an internal 10.x.x.x address and some other NICs on 128.210.x.x. IT just sounded like he was trying to put a default gateway on each NIC, which would obviously cause issues in a multi-homed situation. – MDMarra Sep 1 '10 at 2:54

Honestly, I think that you should be looking at this differently. You need to some networking paths for Hyper-V itself and some for the VMs. These should go through different NICs.

If you're setting up a cluster, or iSCSI, or something of that nature, you want to reserve NICs for those things. I don't know what "iDRAC" is, so I'll assume it's something that eats an entire management NIC. If not, you might want to share Hyper-V management with iDRAC and use that NIC for both.

Take the remaining NICs and bind Virtual Networks to them. Uncheck the "allow management OS to share this network adapter" box. You don't need that unless NICs are scarce.

Now your Hyper-V server has only 10.x.x.x addresses, through the management interfaces. Your VMs are bound to separate NICs and they can get whatever IP addresses are appropriate. The VMs aren't even connected to Hyper-V, except through the external switch.

share|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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