VTP over PRP Channel

[BrianSekleckiGE]

IE4000 (or Stratix 5400) user question:

 Has anyone successfully made VTP version 3 (or v2) work over a PRP interface?

  Debug sw-vlan * seems to show the VTP library transmitting VTP on the PR1 meta-interface instead of the underlying physical interfaces (D'OH!) 

  That wont work, obviously.  I'm filing with a ticket with Rockwell now, hopefully they'll forward it to the Cisco TAC before 2023 >:}

~BAS

[Albert Mitchell]

BAS,

can you provide more details on the issue?   i want to make sure i understand the defintion of "meta-interface' properly.

VTP transmissions should be looking at the PRP interface as a logical EtherChannel interface.  at least that's the expected behavior.

ALBERT

[BrianSekleckiGE]

Yes, in example

 Gi1/1  - PRP Channel 1, Lan_A

 Gi1/2 - PRP Channel 2, LAN_B

 Gi1/3 - Standard STP Link

 Gi1/4 - Standard STP Link

 PRP1 - Meta/Logical Interface

With: debug sw-vlan vtp *

Upon update, I see:

  ... subset #1 [Gi1/3]

  ....subset #1 [Gi1/4]

  ... subset #1 [PRP1]

(But not Gi1/1 or Gi1/2)

But since VTP is bi-directional in V3, shouldn't the Sw-VLAN utilize the underlying physical interfaces for PRP channel 1?

[Albert Mitchell]

Brian,

Question:  with PRP, should VTP use individual physical interfaces in channle instead of the logical interface (eg: meta-interface)????

I don't think so. 

Reason:  VTP doesn't know its a PRP channel.  to the VTP logic, it only see's the logical interface, not the physical interfaces.  VTP logic is unaware of the fact that its a PRP logical channel instead of normal EtherChannel.   

for Normal EtherChannels where mulitple physicals have the same source and destination device, VTP messages only need to be sent over 1 of the physical links.  and it doesn't matter which.   and VTP doesn't care which physical interface is used to transmit the msg.  so, the VTP logic uses the logical interface instead of physical.   someother logic determine which physical interface is used (eg: load balance scheme).  VTP doesn't care.

[BrianSekleckiGE]

Thank you for your insight.

Consider in a dual-Redbox (Dual HSRP L3 Switch) topology:

https://literature.rockwellautomation.com/idc/groups/literature/documents/td/enet-td021_-en-p.pdf

Consider a standard ring topology with a PRP and Ring or Star topology of access switches.

The Redbox will have a LAN_A and LAN_B connection to one port on access switch.

In which case, control protocols like CDP or VTP need to know that there are two different switches/peer devices attached to physical interfaces.

~BAS

[BrianSekleckiGE]

[BrianSekleckiGE]

After a restful weekend hiking in The Vecors, I'm say to sorry to add clarification:

My problem report is not a matter of semantics; in my case VTP isn't functioning.

 My "IES"  (LAN_A and LAN_B switches depicted above) never receive any VTP traffic from the RedBoxes (which also happen to be VTP master).  

In DEBUG, I observe VTP frames transmitting outbound on the PRP1 interface on RedBox, but the same DEBUG on the IES never see the VTP frame never arrive/receive.