Adding connections to UCS Fabric Interconnects

gcook0001 · ‎10-03-2018

We are looking at adding additional connections between our UCS 6248UPs and our catalyst CS3850s.

Currently we have the following

FI-A to primary cs3850

FI-B to secondary cs3850

We would like to add

FI-A to secondary cs3850

FI-B to primary cs3850

From what I understand the configurations are as follows

FI-A and FI-B

add an additional Uplink port on each

cs3850

current connection

interface TenGigabitEthernet1/1/3

description "TRUNK to UCS"

switchport trunk allowed vlan 1,2,3

switchport mode trunk

spanning-tree portfast trunk

new connection

interface TenGigabitEthernet1/1/4

description "TRUNK to UCS"

switchport trunk allowed vlan 1,2,3

switchport mode trunk

spanning-tree portfast trunk

spanning-tree is configured on the switches as follows

primary

spanning-tree mst configuration
name Region1
revision 10
instance 1 vlan 1
instance 2 vlan 2
instance 3 vlan 3
!
spanning-tree mst 0-3 priority 0

secondary

spanning-tree mst configuration
name Region1
revision 10
instance 1 vlan 1
instance 2 vlan 2
instance 3 vlan 3
!
spanning-tree mst 0-3 priority 8192

From what I understand this should work. If there is something I am missing please let me know.

I have attached a diagram for clarrification

Thanks Gordon

Kirk J · ‎10-03-2018

Greetings.

Had you considered putting the 3850's in a stackwise config?

You could cut down on the trunk interfaces between the two switches, and could also do port-channels from Each FI to a pair of ports (one on each stack member) which would have faster lacp port fail-over redundancy reaction, and better lacp load balancing.

If you are sticking to the dual physical/logical switch pair, then what you have looks fine.

The port-fast is correct for FI switching in end-host mode.

The MST related should have no bearing on the non-STP participating FI uplinks in port-fast mode.

Thanks,

Kirk...

gcook0001 · ‎10-04-2018

Thanks for the reply. I appreciate the feedback.

We originally had these setup in a stack but had issues. These switches don't support a rolling upgrade and we did not realize this. The first firmware upgrade we did caused havoc. Since both switches booted at the same time all the servers shutdown since they didn't detect a network connection. It took us two days to recover.