Hi,
We have a couple of DELL R910 servers and have had problems with random but very infrequent network connectivity loss.
In the most recent case, a DELL R910 running Windows 2008 R2 Enterprise showed up in our monitoring software as offline.
This machine connects to the production network via teamed NIC cards, each NIC connects to a separate Cisco 2960 switch (which is stacked using stackwise to appear as one switch) using the Broadcomm Teaming and Loadbalancing software.
A third onboard NIC is then connected to a separate LAN/Switch (2960) not load balanced. This network is used for large data transfers only.
We could not ping the server, nor did VNC/RDP/UNC access to the machine respond. In this case we turned to the remote access card and got access to the desktop of the server.
Everything appeared normal (power on, Windows responsive). So we tried a couple of ping tests. The server could not contact any hosts on the production network, nor on the secondary network. All NIC cards were reporting that they were connected as normal (link lights OK, etc..). Other hosts could not ping the server via its address on the production network or on the secondary network.
There were no errors in the machines event log (except messages relating to loss of DC connectivity, etc..)
We tried to disable and re-enable the NIC cards through Windows, but the process hung in the disabling state and nothing changed. As a last resort, I unplugged the cables physically and plugged them back in, but this did not make a difference.
Naturally a reboot fixed this issue, however these servers are running Oracle so misc. loss of connectivity to the server is a bit of an issue (to put it mildly).
Has anyone seen anything like this before?