Hi,
We're having a very odd and elusive networking problem with a Hyper-V 2012 R2 server.
Shortly are upgrading all our servers to Server 2012 R2 and migrating to Hyper-V 2012 R2 users started to notice that Outlook would disconnect and redirected documents would become inaccessible randomly at varying times throughout the day.
After some troubleshooting we managed to isolate the following behaviour.
Pings to the VM's drop off first followed shortly by pings to the Hyper-V host.
The packets are dropped for a period of around 2 minutes intermittently around 1-2 times per day.
The timing seems completely random.
Neither the VM's nor Hyper-V host register any loss of network connection.
VM's can not talk to each other over the vSwitch whilst the problem is occurring.
Normal service resumes after usually around 2 minutes or so.
Nothing seems to be logged in the event log on any of the machines to do with network issues. The only trace we can find is EventID 2070 been logged on the exchange servers application log to say that the DC is unavailable. And on the domain controller/file server we can see LDAP session time outs and SMB handles been scavenged.
The server is running on a ML350P Gen8 (Broadcom NIC) - All firmware and drivers have been updated.
All servers are up to date with all the latest updates.
We have also tried a second Intel based NIC but this has made no difference as did swapping all network cables and the switch.
There are currently 4 VM's connected to a single vSwitch which is linked to a dedicated port on the Intel NIC - management is not allowed through this interface but is done through the second interface on the Intel NIC.
Has anyone come across this before? It is causing a lot of problems with the client and we are very close to rebuilding the Hyper-V host with an earlier version of Windows server.
Thanks,
Richard.
.