Hyper-V Guest Network connectivity not restored after reboot

Have you tried creating a new vSwitch and moving it to that switch to see if the problem persist? It could be an issue with that vSwitch. 

First: Updates on servers running automatically is bad. Really bad. Please don't do that.

Second: Is the host set up with Broadcom Gigabit network ports? Make sure to disable Virtual Machine Queues on all of the ports in the network team the virtual switch resides on.

Third: Check to make sure ALL ports in the team are functioning.

Fourth: If no team (bad), then make sure that the port the virtual switch is bound to is functioning correctly on the switch side and the network adapter side. Make sure speed negotiation is correct and the patch cable isn't bad.

Simple fix: Remove the vSwitch from the VM, delete the vSwitch, recreate the vSwitch, then bind it to the VM again.

Are VLANs involved?

We have updates scheduled for that particular server on the last sunday of the month. We review and determine what updates will go in. then we let it go at midnight.

There is VLans involved. The switch has a 10GB trunk to it.

No Teaming

It only seems to effect a 2/10 vms on that switch, all ten on same vlan

That means that the server has extra ports?

I suggest eliminating that single point of failure by setting up a team and binding the virtual switch to it. No switch programming is required.

Plug in one of those ports, set up the switch trunk and VLAN(s), and bind a new vSwitch to that port. Flip the VM's vNIC over to the new vSwitch and see if they setting down.

No it doesn't,

This is a very small deployment. There are two dl380 G9s with a dual Ten GB cards installed. The site has a Cisco Switch which has 2 ten gig ports. Each server has a ten gig trunk to it. We have aVeeam server also which has direct 10Gb links between itself and each of the DL380's


Ideally we would have some more switching capability however its not my call to make that purchase. Maybe I have been lucky but the chance of  switchport/nic failing Vs the cost of a decent TenG switch does not lean in favor of the cost of the switch with tis particular client. I could revert the server to teamed 1GB as there is plenty of ports for that. and given the amount of users probably wouldn't have too much of an impact. Be just a shame to leave a 10Gb port empty!

With that setup we'd do the following:
 * Team the two 10GbE ports with Switch Embedded Teaming
 ** vNIC for Backup (if needed)
 ** QoS the setup if needed to keep backup bandwidth in line
 * Team two 1GbE ports for server Management

If there is a high bandwidth file server then 10GbE would be the vSwitch to use. All others are fine on a 1GbE pair of ports teamed and dedicated to that virtual switch.

I have two very thorough EE articles on all things Hyper-V:
Some Hyper-V Hardware and Software Best Practices
Practical Hyper-V Performance Expectations

Philip, 

Sorry for the late response, it caught be again last weekend and thought i would come back to this issue. I really dont see that this is a connectivity issue between the hyper V host and my network. Whenever this issue happens it only occurs on the server that had just rebooted. I have 10 servers running on this host and all others are using the same virtual switch happily. Really thinking just to move to ESXi. this is really inconvenient. 

Some fringe ideas: RMM on the box or in-guest? Atero?

Client A/V on the host or in-guest? Is there a firewall component? Is it blocking on startup? Going into some form of limp mode?

We have a lot of servers managed on Hyper-V and don't have this issue. So, what could it be? Something non-standard that's running in-guest is where I'm leaning. What's running in the VM that shouldn't be? Client A/V or "security" software is my best guess. The RMM software agent would be next.

our deployments are very simple. we use Sophos av. the only RMM we have is we use cloudradar

Remove both from the Exchange VM and make sure everything is working as expected.

No A/V on the host either. That's bad news.

Once the two are removed reboot and see if it behaves.

This issue popped up again last night and it made me want to revisit this Thread. Surely someone else has faced this 

Just to re-cap, We have no Antivirus on the physical host

We only have one tenGB connection to the physical server for VLAN trunks that the guests connect to. We have another 1GB interface for management

I don't see how the physical connections to the server are causing this for the reason that of the 10 guests hosted on this host, only one of them failed to reconnect, 9 guests connected fine and services working, However one didn't and the fix is to open the network settings on the within hyper V manager and disconnect the guest from the virtual switch, apply change then reconnect it. After that network traffic resumes to the guest.

Regards