The old changing of the MAC address causing NIC/IP issues has raised it’s ugly head again…this time during a planned migration of one of our VCSA from one vCenter to another vCenter. Ran into an issue where the source VM running on an ESXi 5.5 Update 3a host, had a MAC address that wasn’t compatible with the destination host running ESXi 6.0.

BAD_MAC

Somewhere along the lines during the creation of this VM (and others in this particular Cluster) the assigned MAC address conflicted with the reserved MAC ranges from VMware. There is a workaround to this as mentioned in this post but It was too late for the VCSA and upon reboot I could see that it couldn’t find eth0 and had created a new eth1 interface that didn’t have any IP config. The result was all vCenter services struggled to start and eventually timed out rendering the appliance pretty much dead in the water.

BAD_MAC2

To fix the issue, firstly you need to not down the current MAC address assigned to the VM.

There is an additional eth interface picked up by the appliance that needs to be removed and an adjustment made to the initial eth0 config…After boot, wait for the services to time out (can take 20-30 minutes) and then ALT-F1 into the console. Login using the root account and enable the shell.

BAD_MAC4

  • cd /etc/udev/rules.d/
  • Modify 70-persistent-net.rules and change the MAC address to the value recorded for eth0.
  • Comment out or remove the line corresponding to the eth1 interface.
  • Save and close the file and Reboot the Appliance.

BAD_MAC5

All things being equal you should have a working VSCA again.

References:

https://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=2012451