- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi,
I am running a 4 node Hyper-V R2 cluster with exactly the same hardware on each node using the latest NIC Intel drivers (15.8.76.0).
There are two * Intel ET Quad port server adapters in each server.
I created a two port VMLB team (one NIC from each adapter).
These are connected to seperate switches to provide fault tolerance in the event of adapter or switch failure.
First problem is that the team status shows one of the NICs in the team as active and the other as disabled.
What could cause this issue?
Second issue is when I enable VMDQ (virtual machine queues) on the Intel adapter and then start a VM up on the cluster the node where the VM is running blue screens. The VM then fails over to another node in the cluster and tries to startup. Again the node bluescreens etc etc. If I disable virtual machine queues on the Intel adapter then the bluescreens stop and the cluster is stable.
I purchased the Intel ET Quad adapters specifically to be able to use VMDQ but at present this causes bluescreens (latest drivers).
Any ideas?
Link Copied
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
For the VMLB team, you must connect all the ports in the team to the same switch. Give that a try, and you should see the VMLB team ports load balance like you would expect.
Question on enabling virtual machine queues. Are the adapter ports part of a team when you enable the virtual machine queues? I would like to get more information about your hardware and how the adapters and virtual switch are configured. I will contact you through a private message so we can work through this.
Mark H
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
"For the VMLB team, you must connect all the ports in the team to the same switch. Give that a try, and you should see the VMLB team ports load balance like you would expect."
Are you sure about this? I'm pretty confident we have a Hyper-V Cluster with ET Quad Port Adapters running, where the ports (teamed and in VMLB mode) are connected to two different switches - as is needed for fault tolerance on the switch level.
EDIT: I just double checked - all four adapters in team are active.
EDIT 2: To OP did you try enabling VMQ on both adapters, before creating the team?
- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page