2 Replies Latest reply on Dec 25, 2015 6:50 PM by Fred Weston

    vSwitch adapter info incorrect in ESXi 5.5

    Fred Weston Enthusiast

      I have just installed a new 3 node cluster with vCenter Server Appliance 5.5 build 2001466.   2 ESXi nodes are 5.5 build 206810 and one is 5.5 build 1746018.  The servers are Dell R620s.


      During testing, I have noticed that the network adapter information displayed in the vSphere Client is incorrect.


      Each host has two vSwitches and each vSwitch has 2 physical NICs bound.  vSwitch0 is for management traffic and vSwitch1 is for VM port groups.  I have configured NIC teaming on each vSwitch to use 1 active adapter and 1 stand by adapter.


      Here is a screenshot of the NIC configuration for one of the hosts:

      vswitch.png

      The management IP address for this host is 192.168.199.26 and the vSphere client indicates that vmnic0 is the active adapter and vmnic1 is in stand by.


      The problem I have encountered is that when I look at the mac address table in my switches to see which switch port is associated with the MAC address associated with 192.168.199.26, I see that MAC address associated with the port for vmnic1, not vmnic0.  So vSphere Client is telling me that vmnic1 is in stand by and vmnic0 is active,  but that is not correct based on what I see from the switches.


      I have a Cisco 3850 switch stack with two switches in it.  For this specific host, vmnic0 is connected to GigabitEthernet1/0/1 and vmnic1 is connected to GigabitEthernet2/0/1


      Here is a screenshot illustrating the MAC address for this host being associated with the wrong switch port:

      switch.png

      I have verified the vmnic to switch port mapping by looking at the CDP info in the vSphere Client, by looking at the CDP info in the switches as well as physically inspecting the cabling.


      I also did some troubleshooting by running a continuous ping to the hosts management address and unplugging the ethernet cable from vmnic0 (which vSphere Client shows as active) and I observed no lost pings.  Then I reconnected vmnic0 and disconnected vmnic1 and observed a few lost pings before they came back.  After doing that I checked the switches again and the MAC address had moved from vmnic1 to vmnic0, so fail over is working, but vSphere Client just isn't showing the correct active adapter.


      So my question is: why is vSphere Client showing vmnic0 as active and vmnic1 as stand by when vmnic1 is actually the adapter that is being actively used for management traffic, and what can I do to fix this?