configuring trunk with Cisco Nexus 7000

Yes, I have vPC on Po interfaces. Though the working ports can work with/out vPC. I'm still wondering why F5 say LACP with Cisco should be set PASSIVE/short/MAC address. It's working with active/long/IP address.

The reasons for using a vPC are simply to avoid a site outage due to faiulure of >1 device... With a vPC yu can have EITHER switch or Viprion up & running.

FWIW it works fine when configured correctly. All my Viprions are vPC connected to dual Nexus 7010's using vPC's.

Hey Elias... You do have 1 trunk configured only on each Viprion don't you?

H

OK. So my nexus config (Which works) is

interface port-channel202
  description MAN: Viprion-1 20Gb channel
  switchport
  switchport mode trunk
  switchport trunk native vlan 1000
  switchport trunk allowed vlan 105,108,1024-1255,2005-2006,2015-2016
  switchport trunk allowed vlan add 4080,4086-4087
  vpc 202

And then on the interfaces

interface Ethernet4/18
  description MAN: Viprion-1 10Gb int 1.1
  switchport
  switchport mode trunk
  switchport trunk native vlan 1000
  switchport trunk allowed vlan 105,108,1024-1255,2005-2006,2015-2016
  switchport trunk allowed vlan add 4080,4086-4087
  channel-group 202 mode active
  no shutdown

The other interface is of course connected to Nexus 2... That config is identical.

Then the Viprion is configured with a single 'trunk' (F5 Port-Channel).

When you do a 'sh vpc ' is it Status==Up and Consistency==Success?

well...well... well...

I wanted to provide feedback for the resolution of my problem. Hopefully, it will help someone down the line.

If you have ever resolved a problem by accident, I know I have. I found the problem to be channel-group id on port channel interface membership. The LTM channel-group number must be the same on both Nexus 7k. I was following the ASA port-group id style. I discovered by accident when one LTM was taken off, then I connected sinlge LTM to both nexus 7k, everything came up. Then I said dawn it. I brought the second LTM back and criss-crossed each everything worked fine. Solution: Each LTM must have identical channel-group number on each Nexus 7k. For ASA, this is not the case when you criss-cross connect.

So you have both channel groups set for 202?

LTM's don't have channel-group id's... LTM's use a NAME to refer to a trunk (BigIP trunk == Cisco port-channel/channel-group).

Yeah, there's probably a cisco note somewhere that says the port-channel id has to equal the vPC ID...

TBH it wouldn't occur to be to use different one... Do ASA's even do vPC's? They barely do port-channels...

H

One thing to note it's important that the F5 is set up for MIST (multi instance spanning tree), we got hit where a 10.4 device was still running single instance spanning tree (under network->spanning tree options). This was an artifact from when the box was running 9.x (which didn't support MIST), it was still set to single instance. When we added a newly created vlan to a trunk that terminates on a Nexus VPC (running MIST by default) went blocking, listening, leanrning, forwarding for all the vlans on the trunk causing a blip.

Phil

I would recommend letting your switch run span tree and leaving it disabled on the F5, have you considered that or am I missing something?