NIC Teaming - Difference between Adapter Fault Tolerance and Switch Fault Tolerance?
Regarding Intel NIC teaming, can someone please explain the differences between Adapter Fault Tolerance and Switch Fault Tolerance?
According to the Intel documentation it says Adapter Fault Tolerance should be used when you have two or more connections to the same switch while switch fault tolerance should be used when you have two connections to two different switches.
However in the "Active Standby NIC Teaming without vPC" lab exercise they used Adapter Fault Tolerance even though they had connections to two different switches. Why is that?
I tried both team types and noticed switch fault tolerance did not fail over when I shut down the primary port on N5k.
I'm assuming this has something to do with the fact that the servers are not directly connected to the N5Ks but I'm not sure how adapter fault tolerance could detect the failure but switch fault tolerance did not. The documentation doesn't really explain how the failover mechanism is different.
And finally, which teaming type is most commonly used in enterprise environments where you have two different switches?