Virtual Machines lose access to network

Question

We have continual problems with losing network connectivity to the virtual machines hosted on a Hyper-V host (the host is running Hyper-V Server 2012 R2), it goes down 2 – 3 times per week.
Situation:
All 3 Windows 2008 R2 virtual machines on the host lose connectivity to the outside world at the same random time (they are the only 3 on the host).
When this occurs the host can still ping the 3 virtual machines and the host is reachable.
When this occurs the 3 virtual machines can ping the host and each other, but nothing else outside of the host (request timed out followed by destination host unreachable)
Nothing in the hardware or Windows event logs on the host indicates there is a problem. No indications of issues in the VM event logs either.
Restarting the virtual machines does not fix the issue
Brand new HP hardware, Broadcom nic drivers were upated early on
VMQ is disabled on the VM's
We are not using nic teaming or clustering
Restarting the host temporarily resolves the problem (during the restart the vms are just placed in a saved state).
The host is running Hyper-V Server 2012 R2
The switch ports that the 2 network cables are plugged into (1 for RDP access to host and 1 for the VM traffic) show no errors on the switch.
This sounds like an issue with the Virtual switch but I can't find anything in the logs that gives me a hint.

Answer 1

Thank you for this post.  I have had similar issues with the following configuration
5 Node Fail Over Cluster - 5 HP DL580-G7 Servers
Two NIC Teams Configured 
4 1 GBPS NIC's - Two Native to the integrated 4 Port on the mother board and Two in an expansion card 
- The Integrated NICS are HP NC 375 I (Two ports are 1 GPBS and Two are 10 GBPS)
- The expansion NICS are HP 332T adapters
All four of the 1 GBPS NICS are teamed - this is the team the Virtual Switch for all the VM's is using
Both of the 10 GBPS NICS are teamed
The Load Balancing for both teams is set to switch independent and dynamic
Only two of the hosts are experiencing issues with the Guest VM's losing connection
The OS is Server 2012R2 Data Center Full Install
All recommended hot fixes and updates for a fail over cluster have been applied
All HP Firmware Updates have been applied (Current Through June 2014)
Setting the VMQ to disabled appears to have resolved this issue for me, I will continue to monitor until I am comfortable moving production Guests back to the affected Hyper-V Hosts.
I noted while attempting to set the VMQ to disable that only the 332T Adapters were set to true in the first place
TMB