A couple of days ago, I was experimenting with setting up ESXi 4.1 on a new HP BL465c blade (we have a number of 465c blades with ESX 4.0). There are four physical NIC's on on the virtual switch. The teaming was set to "link status only". I changed it to "beacon probing" and almost immediately I got the red diamond icon on the blade and a "lost network connectivity" alarm. Checking the events showed each of the four NIC's going down , then about 2 seconds later they all showed back up. However, the alarm stayed until I right clicked it and set it to green.
I'm assuming the "lost network connection" was an effect from the beacon probing starting, but shouldn't the alarm have cleared when the NIC's were back up? I waited several hours and even rebooted the host but the alarm indication didn't go away until I manually set it to green. I thought the alarm indications went away on their own when the issue was fixed.
The other thing I noticed was that prior to the change, when I looked at the network adapters on the VSwitch properties, it showed data under the "observed IP ranges" for each NIC. After the network "loss" and recovery, two of the NIC's show "none" for the observed IP ranged. I modified the swtich properties and moved three of the four nics to the "unused adapters" so that only one NIC was active and confirmed that it was working by pinging a VM on the host. I did this for each NIC and confirmed that all of them are working. Is the fact that it shows "none" for the observed networks an issue I should be concerned about?
Mike O.
