I have one Windows 2008 R2 node where NPM sends a "Node is experiencing high packet loss" alert. Again, it's a Windows 2008 R2 machine hosted on vSphere ESXi 5.0 with a vmxnet adapter. NPM is 11.0.1. Several other nodes have the same configuration, Same NIC adapter, same host, same or greater utilization. This node provides a RemoteApp for users. Two other RemoteApp servers never say a word.
I've tried the usual 2008 R2 vmxnet adapter tweaks, RSS, tcp chimney offload, etc.. on this node as well as others to replicate the problem. No effect.
It just seems to send out random events that it's experiencing high packet loss. Nothing in reports or anything else in NPM seem to show that it does have any packet loss and a continuous ping throughout the day from the NPM server to the node doesn't drop a packet, and yet an alert shows up.
I have tried removing the node from NPM and then adding it back. Same thing. The only idea I have left is to simply build another VM doing the same thing, but that would create a brief outage for users.
Anyone have any thoughts?
Did you ever get this sorted, a recent upgrade on our systems to ESX 5.1 are showing exactly the same.
Hosts are reporting 50 to 70 percent packet loss, but continuous pings show nothing? What exactly is causing the packet loss metric?