The esxcli network nic down/up commands fail to restart a NIC (2002233)

Details

• Using the esxcli network nic down and esxcli network nic up commands fails to restart a physical network adapter that is connected to an HP Virtual Connect Flex-10 Ethernet Module and is in a non-standard speed mode or a physical network adapter that resides in a Cisco UCS sever.
• Using the esxcli network nic down and esxcli network nic up commands fails to restart a physical network adapter that is connected to an Emulex OneConnect 10Gb UCNA and is in a non-standard speed mode or a physical network adapter that resides in a IBM Blade Center Server.

Another command that does not work with a HP Flex-10 or Cisco UCS network adapter is the esxcli network nic set -n vmnicX -a for autonegotiating the speed of an adapter with sequence number X.

This issue has been observed on  IBM Blade Servers utilizing the Emulex OneConnect 10Gb UCNA as well.

Errors that will be reported in the vmkernel.log file:

2013-10-23T05:41:42.588Z cpu36:10220)Uplink: 8603: Setting link down on physical adapter vmnic7.

2013-10-23T05:41:42.588Z cpu36:10220)Uplink: 8634: Requested link speed/duplex 0x9c4/0x2 is not valid.

Solution

Use these commands only with the network cards running auto-negotiation/10Mb/100Mb/1000Mb/10000Mb speeds. Do not use these commands on a  HP Flex-10, Cisco UCS  / IBM Blade Center Server, Emulex OneConnect 10Gb UCNA network adapter with non-standard speed. These commands do not support HP Flex-10, Cisco UCS, Emulex OneConnect 10Gb UCNA and other cards with non-standard speeds.

• To change the link state of the physical interface to down:
  
  esxcli network nic down -n vmnicX 
  
• To change the link state of the physical interface to up:
  
  esxcli network nic up -n vmnicX
  
• To auto-negotiate the speed of an adapter:
  
  esxcli network nic set -n vmnicX -a 

Source KB:-
http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=2002233&src=vmw_so_vex_ragga_1012