0

The problem: I am unable to get traffic between a Hyper-V host and it's VM to utilize an internal virtual switch while an external virtual switch is also connected.

Currently running a single Win2012 R2 server with Hyper-V installed as a test environment. I have one Hyper-V VM, VM1, which needs fast access to a file share located on the Hyper-V host VM1 is hosted on. The Hyper-V host has two Hyper-V virtual switches configured. One is an external virtual switch and is tied to the physical 1GB network adapter on the Hyper-V host. The other is an internal virtual switch and as such is not tied to the Hyper-V hosts physical network adaptor. VM1 has two virtual adaptors configured on it; the external and internal adaptors as outlined above. The external virtual switch's subnet is 192.132.4.x with both the host and VM1 having a unique IP address on that subnet. The internal virtual switch's subnet is 10.10.0.x with both the host and VM1 having a unique IP address on that subnet.

Hyper-V Host:
Internal IP: 10.10.0.10
External IP: 192.132.4.10

VM1:
Internal IP: 10.10.0.20
External IP: 192.132.4.20

The problem occurs like this. Both the internal and external virtual adaptors are enabled. Logged into VM1, I copy a 2GB file from the Hyper-V host's share at \\10.10.0.10\share to VM1's C:. I have Task Manager open and am on the Performance tab where I can see which adaptors the traffic is using based on utilization. When I transfer the 2GB file the external adaptor is clearly being used to transfer the file and the internal adaptor shows no traffic. What I want to see is the traffic traversing the internal adaptor since the \\10.10.0.10 IP does not exist on the 192.132.4.x subnet which the external adaptor's virtual switch is connected to. Now, I disable the external virtual adaptor on VM1 and repeat accessing the same share at \\10.10.0.10\share; the file transfers to VM1 using the internal virtual adaptor this time.

I don't understand how the external adaptor can be used to access a path destined for the 10.10.0.x subnet in the first place; that adaptor can not access that subnet.

Aside from Performance monitor clearly showing the traffic usage the internal vswitch is not capped by the physical 1GB adaptor and really flies speed-wise. This is the reason for this setup; 100-120MB's vs 600MB's. This is a developmental environment and I am aware this would not be good for a production environment.

3
  • It seems like you're either mis-judging the situation, or your 'internal' switch is not really 'internal'. double/triple check your virtual switches and their associations with physical NICs. the VM host's 'internal' network should not be associated with a physical NIC. I'd even suggest deleting/recreating your virtual switches.
    – goofology
    Apr 7, 2015 at 5:41
  • Virtual switches are all as indicated. Nice thing about Hyper-V, it's dead simple without VMM's fabric involved. This is actually the third from scratch environment created to eliminate any quirks related to setup; so vswitch recreation to the extreme.
    – MIke
    Apr 7, 2015 at 23:25
  • for what it's worth, i have several VMs configured exactly like you describe, and I haven't ever had this issue. All paths work as designed. not much help, but lack of ability to recreate the issue suggests there may be a quirk in your setup. I'd offer to connect to you by TeamViewer but not sure if that is allowed or if you'd trust a total stranger.
    – goofology
    Apr 8, 2015 at 1:26

0

You must log in to answer this question.

Browse other questions tagged .