Cisco Support Community
Showing results for 
Search instead for 
Did you mean: 

Welcome to Cisco Support Community. We would love to have your feedback.

For an introduction to the new site, click here. If you'd prefer to explore, try our test area to get started. And see here for current known issues.

New Member

RSTP topology change


Could someone please clarify my doubts about topology change mechanism in rapid spanning tree. I am monitoring the BPDUs in one of the segments of my example network. In case of a topolgy change I see a TC BPDU on that segment. After the TC BPDU I don't see any BPDUs whatsoever for the next 35 seconds. That is equal to the sum of max age of 20 seconds and 3 hello durations of 5 seconds each. The segment remains completely quiet for this duration. After 35 seconds situation comes back to normal and I see one hello message every hello time interval. I am unable to understand this behaviour of RSTP. According to STP I should see TC BPDUs for 35 seconds but in RSTP situation should stabilise immediately without waiting for 35 seconds. But my observation doesn't fit into either STP or RSTP. Does anyone have a clue about what might be going on?



Re: RSTP topology change

No clue as to what is happening. Please, give us more detail on what your network is (and particularly what is on this segment) and what the topology change is.

Thanks and regards,


New Member

Re: RSTP topology change

Thanks for the reply. My network consists of five S2800 switches. They are all connected in a straight line. You can name them as SWI,SW2,SW3,SW4,SW5, connected linerly. I have mirrored a port on SW1 to monitor the traffic on the segment between SW1 and SW2. I have configured the port priorities in increasing order from Sw1 to SW5. So in a steady state SW5 acts as the root bridge and I see hello BPDUs being sent from SW2 to SW1.

Now, I disconnect SW5 from the network and monitor the BPDUs on segment between SW1 and SW2. Since SW4 is the next highest priority it becomes the root and send TC flag set BPDUS. This I can see on my segment of SW1. But the problem is that immediately after TC BPDU there are no BPDUs seen on the segment for the next 35 seconds. This duration is equal to Max age(35) + 3*Hello Time(5). After this time I see regular hello BPDUs from SW2 to SW1. Why is there a period of silence for 35 seconds? Is there somehting about RSTP I don't know? Is it something to do about switch settings or port settings? I have no clue.

Re: RSTP topology change

Hi Sadaf,

Interesting. Well, first I don't know about the S2800. Is that a Cisco switch? I've been there for a while and I'm not following the whole range of product that is so diverse now ;-)

I can't explain what you are seeing in terms of correct RSTP behavior. I can make some few remarks though.

Not really useful for your problem, but configuring the port priority has no impact on your setup. What is important is to configure the bridge priority, and in order to have sw5 be the root, it has to have the (numerically) lowest priority. I'm just mentioning this because I wanted to confirm that SW4 is the best secondary root, followed by SW3 and so on.

Hello-time is normally 2 seconds by default. Did you tune it to 5 seconds? (IEEE does not allow this, Cisco again does;-).

As far as the real RSTP spec is concerned, there should be no topology change at all when the root bridge disappear in your network. This is because RSTP only generates topology changes when a port is transitioning from a discarding to a forwarding state. However, our implementation (in Cisco) is generating a TC when there is a role change (root port to designated port). This behavior was added to cover some cases that the spec consider illegal. Anyway, in case of Cisco's RSTP, you could indeed have a topology change.

Now, if your bridges were running RSTP, you should not have this long period of time with no BPDU. An RSTP bridge keeps sending periodic BPDUs on its designated ports, even when it does not receive any on its root port.

What surprises the most is that you receive single BPDU with TC set. If you had some switches running STP in the network (say SW3), this could explain that you did not receive any bpdu from the root for max-age (20sec): SW3 would discard SW4's bpdu until it has aged out SW1's information. But here, you see a first BPDU from SW4 claiming to be root and with TC set. I'm really confused;-)

Could you confirm that you there are Cisco switches. In which case, could you give me the IOS release they are running? Could you also check that everybody is running rapid-PVST or MST? When this is confirmed, I can give you some debug command to enable on the respective switches in order to try to understand what is happening. So far, it's beating me;-)



New Member

Re: RSTP topology change

Thanks Francois. That was indeed helpful. The switches I am using are not cisco switches....but they are running RSTP protocol.

When I had written about priority, I had meant bridge priority.

I have changed the hello time to 5 seconds.

After reading your comments I feel one of my switches is running STP though it is claiming to be RSTP. I'll observe the scenario keeping STP in mind. Thanks again for your help.