Home / server / Questions / 685706
Accepted
Arvo Bowen
Asked: 2015-04-27 11:19:47 +0800 CST

Adding a virtual NIC to a 2008R2 server makes internet unreachable, why?

  • 772

I currently have a Windows Server 2008R2 box with a single physical NIC. For what I want to do I need two network adaptors in the box. To accomplish this, I added a legacy hardware device "Network Adaptor" and used the Microsoft Loopback Adaptor as the driver.

Now I have two network adaptors:

  1. NIC1 - Physical NIC 10.71.1.3 (network: 10.71.1.0/24)
  2. NIC2 - Virtual (Loopback) NIC 10.71.2.3 (network: 10.71.2.0/24)

My goal is to get NIC2 to piggyback off of NIC1 and be able to ping another server (10.71.2.2) on the same switch (there is only one switch involved in this scenario).

Currently I can ping a server (10.71.1.5) with no issues because it's using the physical NIC as its gateway. But I can't ping 10.71.2.2. I'm guessing because the driver I used (loopback) is creating an isolated network on NIC2.

Please Note: Normally one would just add a secondary IP to NIC1 (the physical NIC). If I did I could ping 10.71.2.2 with no issues (I have tried and it works). But in this case I need two network adaptors and not just a single one to make what I'm trying to do work.

Ideas?

Updated: 4/28/2015 enter image description here

networking

2 Answers

  1. You need to add the Routing and Remote Access role to the server and set it up for IP routing.

    RRAS will handle the routing from Nic2 to the Nic1 network, but the other servers will need to know the route back to Nic2.

    On the other servers you will need

    route add 10.71.2.0 mask 255.255.255.0 10.71.1.3

    This will let the other servers know to route return traffic via 10.71.1.3

    • 2
  2. Best Answer

    In the end I decided to go a different route (no pun intended). I actually have multiple physical NICs installed and was using them for other things.

    I since have learned something new about clustering. I have always in the past dedicated a NIC for the sole purpose of the cluster heartbeat. We my heartbeat always had a single point of failure. That never seemed to be an issue for me but the more I read up on it the more I realized that you don't have to dedicate a NIC just for the heartbeat. There is actually very little traffic that is used for the heartbeat itself and just incase that NIC goes down you should always have another NIC that is allowed to be utilized by the cluster and permitted to send/receive cluster data. This can be found by right clicking then going to "Properties" for each of the "Networks" from within the cluster config itself. Just make sure "Allow cluster network communication..." is ticked. You can allow clicks to if you wish but this is intended for things like VMs in the cluster etc and is not needed for the heartbeat.

    So I ended up with 2 NICs with different subnets and both are being used as a heartbeat and that way they each show up as segregated networks in the cluster.

    • 0