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

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

For an introduction to the new site, click here. And see here for current known issues.

New Member

question on STP converge delay

Hi !

I would like to know if I increase the value of the hello timer from default value to 10 sec. What is append on the covergence delay ?

I think this delay will increase too but I'm not sure.

some one can confirm ?

4 ACCEPTED SOLUTIONS

Accepted Solutions
Hall of Fame Super Bronze

Re: question on STP converge delay

It's not going to wait 10 seconds before sending the first hello BDPU during uplink, it will send the hello BDPU right away. It shouldn't affect convergence.

However, you must make sure this command is entered at the root switch so you must modify the priority on the switch, do not leave it as default.

HTH,

__

Edison.

Re: question on STP converge delay

Hi Christian,

As Edison says, the convergence time will not change, because it is determined by the Max Age Timer and the Forward Delay Timer.

However, given a big enough network, 10 sec hello time will increase the possibility for a blocked port to not hear a hello within the Max Age Timer and

1. the currently blocked port (let's say Fa1/0) will go into the listening state.

2. when Fa1/0 goes into listening state, the current root port (let's say Fa1/24) is immediately blocked to prevent loops.

3. this will also block traffic going through Fa1/24.

4. finally Fa1/0 may hear a hello again, and it will go back to blocking beacuse its root path cost is higher than for Fa1/24.

5. Fa1/24 goes from blocking into listening state and after 2 times the Forward Delay time will go into forwarding again.

And this situation can repeat many times.

So, what I want to say here that in specific circumstances increasing the hello time may get the network unstable by forcing the network to reconverge again and again, and the traffic path will be repeatedly interrupted.

Cheers:

Istvan

Re: question on STP converge delay

Hi Christian,

Above I made a mistake in my analyis:

Fa1/24 will not go into blocking state, but sometimes a bridging loop may form.

Fa1/0 will make an assumption that its link to the neighbor switch is interrupted if no BPDUs are received for the Max Age time.

Thanks:

Istvan

Re: question on STP converge delay

Increasing the hello timer to 10 seconds is risky unless you also increase the max age timer. Think of them like any other hello timer and hold timer pair. By default, the hold timer (max age) is 10 times the hello timer to allow for several propagation hops.

If you do increase the max age timer, then that will increase the convergence time by the same amount in many failure situations, specifically the ones where the failure is not detected directly by something closer to the root.

Just out of interest ... why do you want to do this? It is unusual that 2 seconds is too long. Do you have some problem that maybe we need to examine the diagnosis more closely?

Kevin Dorrell

Luxembourg

5 REPLIES
Hall of Fame Super Bronze

Re: question on STP converge delay

It's not going to wait 10 seconds before sending the first hello BDPU during uplink, it will send the hello BDPU right away. It shouldn't affect convergence.

However, you must make sure this command is entered at the root switch so you must modify the priority on the switch, do not leave it as default.

HTH,

__

Edison.

Re: question on STP converge delay

Hi Christian,

As Edison says, the convergence time will not change, because it is determined by the Max Age Timer and the Forward Delay Timer.

However, given a big enough network, 10 sec hello time will increase the possibility for a blocked port to not hear a hello within the Max Age Timer and

1. the currently blocked port (let's say Fa1/0) will go into the listening state.

2. when Fa1/0 goes into listening state, the current root port (let's say Fa1/24) is immediately blocked to prevent loops.

3. this will also block traffic going through Fa1/24.

4. finally Fa1/0 may hear a hello again, and it will go back to blocking beacuse its root path cost is higher than for Fa1/24.

5. Fa1/24 goes from blocking into listening state and after 2 times the Forward Delay time will go into forwarding again.

And this situation can repeat many times.

So, what I want to say here that in specific circumstances increasing the hello time may get the network unstable by forcing the network to reconverge again and again, and the traffic path will be repeatedly interrupted.

Cheers:

Istvan

Re: question on STP converge delay

Hi Christian,

Above I made a mistake in my analyis:

Fa1/24 will not go into blocking state, but sometimes a bridging loop may form.

Fa1/0 will make an assumption that its link to the neighbor switch is interrupted if no BPDUs are received for the Max Age time.

Thanks:

Istvan

Re: question on STP converge delay

Increasing the hello timer to 10 seconds is risky unless you also increase the max age timer. Think of them like any other hello timer and hold timer pair. By default, the hold timer (max age) is 10 times the hello timer to allow for several propagation hops.

If you do increase the max age timer, then that will increase the convergence time by the same amount in many failure situations, specifically the ones where the failure is not detected directly by something closer to the root.

Just out of interest ... why do you want to do this? It is unusual that 2 seconds is too long. Do you have some problem that maybe we need to examine the diagnosis more closely?

Kevin Dorrell

Luxembourg

New Member

Re: question on STP converge delay

I'm currently suddying for my CCNP certification

in my book on exercice section on STP topic the author aks the question, but in the answer that part of the question was forgot.

thanks a lot for your help !

259
Views
0
Helpful
5
Replies