Have a little question about NIC Teaming (NFT) in CUCM 5.x and above.
Main interest of mine: can I connect 2 (two) NICs of, for example,
HP-based MCS to separate physical access layer switches ?
I've attached diagram to illustrate the case.
The following prerequisites have been accomplished:
- 'set network failover enable' is entered via CUOS's CLI
- the output of 'show network failover' shows there are 3 (three) network interfaces
- mac address for all of them (BOND0, ETH-0, ETH-1) is similar: 0011.2233.4455
All is working fine.
Then if something goes wrong with Sw-A, cable from Sw-A to CUCM or ETH-0 itself,
CUOS activate its ETH-1. Before this activation occurs (i.e. before failure occurs)
will mac 0011.2233.4455 exist in the mac-address-table of Sw-B ?
Is it a stack? in which case, build a port channel with the ports that connect to you CUCM
let me add to this, If you would build NIC pairing Active/Avtive, yes port channel is the way to go.
If active/passive is used, you do not need any additional config on the switches, as the MAC address of you NIC pair will only be learned on one switchport at the same time anyway
No, access layer switches don't make a stack, they are separate devices.
You mean while ETH-1 is in STANDBY mode, no mac address will exist on the switchport where ETH-1 connects to ?
Is this true until CUCM activate ETH-1 if failure occurs on ETH-0 link ?