Experiencing odd network issues with Hyper-V guest virtual servers is an old issue, but it is one I still occasionally see. It especially crops up on HPe Proliant Servers. What typically happens is that your Hyper-V guest virtual servers seem to have random networking issues. Some of the virtual servers on the same Hyper-V host may not experience the network issue at all. From my experience, the issue is typically triggered under a heavy network load on a virtual server, such as a large file transfer between two servers. Rebooting the Hyper-V host will resolve the issue temporarily, and the Hyper-V guest servers will operate normally for some time.
The cause is an issue with an outdated Broadcom NetExtreme network adapter driver. The issue is supposed to be resolved by updating your network adapter driver to a version greater than 16.8. However, in my experience, the driver update does not always resolve the issue. If you update the driver and the issue persists, or if you don’t want to update your network driver for some reason, you can alternatively disable VMQ (VM Queueing) on the Hyper-V host network adapter that is used by your Hyper-V virtual switch.
Please note that if your network adapters are 1G, then disabling VMQ will have no effect since VMQ is disabled on the Hyper-V virtual switch for virtual machines that are using 1G network adapters. Microsoft has a Knowledge Base article describing the symptoms and resolution that can be found here: https://support.microsoft.com/en-us/help/2986895/virtual-machines-lose-network-connectivity-when-you-use-broadcom-netxt.
The article states that this applies to only Hyper-V versions 2012 and 2012 R2, but I have seen this issue on 2008 R2 as well. It may also be an issue in Windows 2016 and 2019.
If you’re experiencing any network issues with your virtual servers and need assistance with troubleshooting, do not hesitate to contact us at any time.
