BPDU guard - weird situation

Question

Hi guys,
This morning unpleasant surprise happened to me. One of critical ports was err-disabled because of BPDU guard (device B). This wouldn't be surprise if this port (on Device B) wasn't configured as L3 port (I agree that BPDU filter shouldn't be enabled at all here, this is legacy config), and other end have BPDU filter enabled (Device A). Here is port config:
Device A:
interface GigabitEthernet4/0/24
 switchport access vlan 10
 switchport trunk encapsulation dot1q
 switchport mode access
 switchport nonegotiate
 logging event trunk-status
 spanning-tree bpdufilter enable
Device B:
interface GigabitEthernet2/45
 no switchport
 ip address 10.0.0.1 255.255.252.0
 ip helper-address 172.16.249.5
 logging event link-status
 logging event trunk-status
 spanning-tree portfast
 spanning-tree bpduguard enable
Log from Device B indicating that it was err-disabled:
Apr 20 20:08:52.336 CETS: %SPANTREE-2-BLOCK_BPDUGUARD: Received BPDU on port Gi2/45 with BPDU Guard enabled. Disabling port.
Apr 20 20:08:52.336 CETS: %PM-4-ERR_DISABLE: bpduguard error detected on Gi2/45, putting Gi2/45 in err-disable state
Log form Device A indicating that BPDU never sent from this port:
DeviceA#show spanning-tree vlan 10 detail
 Port 186 (GigabitEthernet4/0/24) of VLAN0010 is designated forwarding
   Port path cost 4, Port priority 128, Port Identifier 128.186.
   Designated root has priority 28740, address 001a.6da4.f000
   Designated bridge has priority 28740, address 001a.6da4.f000
   Designated port id is 128.186, designated path cost 0
   Timers: message age 0, forward delay 0, hold 0
   Number of transitions to forwarding state: 1
   Link type is point-to-point by default
   Bpdu filter is enabled
   BPDU: sent 0, received 0
Did anyone had ever similar experience? By all logical explanations, this should never happen
Thanks

Answer 1

On the other hand, most SOHO switches do not implement Spanning Tree. If you are concerned about users installing switches, you need to take other precautions as well.
You can stop the users using a switch to fan out a port, by configuring port security and only allowing one MAC address on the port.
The BPDU guard will give you some protection against certain malicious user practices, even if the rogue switch does not do Spanning Tree. For example, the user who plug in a SOHO switch, and then plugs two other ports of that SOHO switch back-to-back with a cross-cable. In this case, your Catalyst will see its own BPDUs circulating round the loop, and will close the port down. (If the SOHO switch is not doing Spanning Tree, then it will pass the BPDUs through transparently.) This is why you should not have bdpu-guard and bpdu-filter on the same port.
Kevin Dorrell
Luxembourg