I had faced an issue where , the mangement network connectivity on the server lost , as soon a addition vnics were added to the UCS service profile ,
please see the mac address for the below nics
the nic order mac address are
vmnic0 : 03
vmnic1 : 00
vmnic2 : 01
vmnic3 : 02
but the UCS service profile would have the mac address for the corresponding nic as different , See the attached snippet for clarifications
the first Vnic in UCS has : 03 which is same as of our nic
the second Vnic on the UCS has different mac as that of our esxi host . hence the order was completely changed ,
Imagine if you have allowed Vlan 145 on the both the nics 0 and 1 for the management traffic and choose nic vnic 0 and 1 on the virtual switch end , the host will not respond ,
cisco has identified the bug , please see the relevant bug id bug/CSCuv34051
this issue can be fixed by the following method
- Place the ESXi/ESX host into maintenance mode.
- Remove all VMNICs from the virtual switches except
vmnic0
, which is used for management. Make note of which VMNICs are assigned to which vSwitches. - Shut down the ESXi/ESX host.
- Remove/disable all NICs except
vmnic0
. On-board NICs can be disabled in the BIOS. NICs installed in PCI slots should be physically removed. Make note of the slot each one came from. - Boot the server. At boot time, all removed/disabled VMNICs are removed from the
esx.conf
file. - Shut down the ESXi/ESX host.
- Add/enable all required NICs.Enable the on-board NICs in the BIOS, and re-insert the PCI bus NICs into the slots they came from.
- Boot the ESXi/ESX host. All VMNICs are assigned per segment, bus, slot, and function ID, and are ordered correctly.
- Assign VMNICs back to the virtual switches they were assigned to before.
- Exit maintenance mode on the ESXi/ESX host.
else it can also be done at UCS service profile using cisco locacli commands
https://tools.cisco.com/bugsearch/bug/CSCuv34051
Hope this helps to figure out the issue
Thanks and regards,
Nithyanathan R