01-12-2011 11:09 AM - edited 03-06-2019 02:57 PM
Say switch A and C are connected via switch B (which has spanning-tree disabled) and a loop is created on B. What does switch A do when it receives inferior BPDU's from itself?
Solved! Go to Solution.
01-13-2011 01:00 PM
Dear friends,
Please allow me to join the discussion.
First, the words "superior" and "inferior" are relative: anytime we speak about superior and inferior BPDUs, we have to say what is that particular BPDU superior/inferior to. There is no superior or inferior BPDU per se, rather, of each two BPDUs sent by two different ports (on the same switch or on different switches), one is superior to another (and vice versa, the second is inferior to the first). The design of STP makes sure that each two BPDUs can be compared and decided which of them is superior/inferior to the other (in other words, the BPDUs create a total ordering).
Second, per the diagram, a BPDU sent by the Switch A may loop on the Switch B and get back to the Switch A. If the BPDU is received by another port than the one that sent the BPDU, normal election of a Designated/Backup port would ensue. However, if the BPDU is received by the same port that originally sent it, the STP will declared the port as Broken (Looped) and will put it into Blocking state. It will not be err-disabled.
If you decide to try this out experimentally, you may find that the port on Switch A eventually gets err-disabled; however, that will not be caused by the STP but by an independent mechanism: each port on Catalyst switches emits so-called LOOP frames. We have been in extensive discussions with Giuseppe Larosa about its purpose and each one of us still seems to be somehow biased about it Nevertheless, if a port receives a LOOP frame it has sent itself, it will be err-disabled immediately.
So in this topology, the STP would merely block the port, however, the LOOP mechanism would sooner or later shut it down completely.
Best regards,
Peter
01-12-2011 11:55 AM
Please confirm if this is how the topology looks as per your example
SWITCH-A---------------SWITCH-B------------------SWITCH-C
Or, SwitchA and Switch C are directly connected to each other as well
SwitchA -------------Switch-C
\ /
\ /
\ /
\ /
Switch-B(spanning tree disabled)
01-12-2011 12:04 PM
Yeah only except A & C don't connect directly to each other, only through switch B which has STP disabled
01-12-2011 03:12 PM
If A and C are not connected to each other, the loop never completes
The BPDU sent out by the switch A is received on int 1 of switchB and goes out int 2 on switchB to switchC
The Switch C sends its BPDU out through int 2 on Switch B and it reaches switch A via int 1.
A switch does not send the traffic back the interface it receives on, therefore once the BPDU leaves the the switch it does not see it back unless there is an alternate path.
01-12-2011 06:38 PM
Thanks Gurpreet - I undstand that though... but my question was what if a loop was created on switch B? (ports 9 & 10 on switch B inadvertently connected), and B has spanning tree turned off. BPDU from switch A reaches B on 1/1 and gets looped on ports 9 & 10 on switch B before going back out 1/1 and hitting switch A again. What happens then?
01-12-2011 03:14 PM
A switch can not send superior and inferior BPDU at the same time
The superiority or the inferiority of the BPDU depends upon the combination of two factors
1. Priority of the switch
2. Mac addres of the switch
You may want to refer to the following link
http://www.cisco.com/image/gif/paws/10556/spanning_tree1.swf
01-13-2011 12:47 PM
Thank you for attaching the diagram
As per the above diagram, i believe the switch should loop back error message and put the interface into error disabled state.
01-13-2011 01:00 PM
Dear friends,
Please allow me to join the discussion.
First, the words "superior" and "inferior" are relative: anytime we speak about superior and inferior BPDUs, we have to say what is that particular BPDU superior/inferior to. There is no superior or inferior BPDU per se, rather, of each two BPDUs sent by two different ports (on the same switch or on different switches), one is superior to another (and vice versa, the second is inferior to the first). The design of STP makes sure that each two BPDUs can be compared and decided which of them is superior/inferior to the other (in other words, the BPDUs create a total ordering).
Second, per the diagram, a BPDU sent by the Switch A may loop on the Switch B and get back to the Switch A. If the BPDU is received by another port than the one that sent the BPDU, normal election of a Designated/Backup port would ensue. However, if the BPDU is received by the same port that originally sent it, the STP will declared the port as Broken (Looped) and will put it into Blocking state. It will not be err-disabled.
If you decide to try this out experimentally, you may find that the port on Switch A eventually gets err-disabled; however, that will not be caused by the STP but by an independent mechanism: each port on Catalyst switches emits so-called LOOP frames. We have been in extensive discussions with Giuseppe Larosa about its purpose and each one of us still seems to be somehow biased about it Nevertheless, if a port receives a LOOP frame it has sent itself, it will be err-disabled immediately.
So in this topology, the STP would merely block the port, however, the LOOP mechanism would sooner or later shut it down completely.
Best regards,
Peter
01-13-2011 01:50 PM
Awsome!!!!
Thanks for the detailed explanation
01-14-2011 04:01 AM
Thanks Peter! great explanation.
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide