I have this problem in my environment and no one seems to understand or have a solution.
When I ping my hosts, it has package lost on the management network, to the point that I can't log ssh. And the host are constantly disconnecting and reconnecting from vCenter.
HA keeps tryng to fail over the VMs.
I have four ESXi hosts em 6.5 version (update 2), DRS and HA enabled.
My vSwitch has two physical adapters, both active and configured with PortID (or route based on originating virtual port). I have NSX and vxlan,
All using the same vSwitch, physical adapters and nework. I have no package lost on the vxlan.
My first suspicius is that I had misconfiguration on the physical network switch. So I ask my network guy to check, there's none.
I tryed to disable one physical adapters, on one of the hosts, and the package loss immediately stops (in that specific host, not in the whole cluster). I put the two adapters active again and problem comes back.
Then I try to reboot one of the hosts and now everything work just fine on that host. No package loss, no disconnects.
I been trying to investigate the other tree host, that ares stil problematic but until now I have no clue.
There's anyone whos seen something like this?
Update firmware and driver
You have said that one physical NIC has been disabled and problem is gone.
Physical NIC is a single device not VMKernel adapter or physical port.