Another STP point of confusion

I've seen a lot of example diagrams like this one:

And what they show is that the link with the highest port number on Switch B ends up being blocked. But what I can't tell is if gi2/24 is blocked because it's higher than gi2/23, or is gi2/24 blocked because it's connected to gi2/11 which is higher than gi2/10?

In other words, if I had this configuration:

Then would gi2/24 still be blocked, or would gi2/23 be blocked?

Find more posts tagged with

Comments

sschwieterman

I am also studying for my ICND2 exam currently. Am taking the exam this Thursday so it would help if others could confirm, but...

My understanding is that the blocked port is based on the higher port of the root switch, not the non-root switch. In your example, the first picture port g2/24 would be blocked and the second picture port 2/23 would be blocked.

sucanushie

Because the top switch is the Root Bridge None of it's ports will be blocking.

So because they are on the same switch and would have the same Bridge ID the lowest port would win and become the root port and G 2/24 would be blocking.

doctorlexus

sschwieterman wrote: »

I am also studying for my ICND2 exam currently. Am taking the exam this Thursday so it would help if others could confirm, but...

My understanding is that the blocked port is based on the higher port of the root switch, not the non-root switch. In your example, the first picture port g2/24 would be blocked and the second picture port 2/23 would be blocked.

Thanks for the reply. This is the way Boson explains it, but I've seen other explanations, which leads me to believe a lot of people, even those trying to instruct it, don't really understand how it works.

I'm leaning toward your explanation as being correct.

sucanushie

Your question was Then would gi2/24 still be blocked, or would gi2/23 be blocked?

Good luck on your exam

networker050184

The information used to make the decision is received in the BPDU. So when you look at it that way it is apparent it's the remote port ID.

aftereffector

sucanushie wrote: »

Because the top switch is the Root Bridge None of it's ports will be blocking.

So because they are on the same switch and would have the same Bridge ID the lowest port would win and become the root port and G 2/24 would be blocking.

This is the correct answer.

doctorlexus

networker050184 wrote: »

The information used to make the decision is received in the BPDU. So when you look at it that way it is apparent it's the remote port ID.

This makes sense.

doctorlexus

aftereffector wrote: »

This is the correct answer.

No, sucanushie's post had nothing to do with my question. Never was there any debate that none of the root bridge's ports would be blocked.

OctalDump

Try it out in PacketTracer, or on real gear. At some point you are going to have to use real gear, so this is a great opportunity to see how things work.

sschwieterman

So I ran this in NetSim to test it out like Octal suggested.

First scenario: I had port 5 on Switch 1 connected to port 15 on Switch 2, and port 20 on Switch 1 connected to port 10 on Switch 2.

Second scenario: I had port 5 on Switch 1 connected to port 5 on Switch 2, and port 15 on Switch 1 connected to port 15 on Switch 2.

In both scenarios port 15 on Switch 2 was blocked, so it does seem the highest port on the non-root bridge becomes the blocked port (assuming link speed is the same between both connections).

Does this sound accurate to the more seasoned CCNAs on here?

Sy Kosys

sschwieterman wrote: »

...so it does seem the highest port on the non-root bridge becomes the blocked port (assuming link speed is the same between both connections).

There are two equal-cost paths between the directly-connected switches. When the root port is being elected, the path/port costs are evaluated to determine which one will be determined as the shortest path to the root bridge.

In the case of multiple equal-cost paths to the root bridge, further evaluations are made and the tie-breaker is the port number on the non-root bridge. The lower interface number is chosen as the root port, and the higher interface number is put into a blocking state.

Thus, taking us pack to the original post, Switch A is the root bridge and Switch B goes through the aforementioned election process to determine which port gets blocked to prevent an STP loop. In all scenarios, the higher interface number (when dealing with multiple equal-cost links connecting to the root bridge) gets the axe.

sschwieterman

Sy Kosys wrote: »

There are two equal-cost paths between the directly-connected switches. When the root port is being elected, the path/port costs are evaluated to determine which one will be determined as the shortest path to the root bridge.

In the case of multiple equal-cost paths to the root bridge, further evaluations are made and the tie-breaker is the port number on the non-root bridge. The lower interface number is chosen as the root port, and the higher interface number is put into a blocking state.

Thus, taking us pack to the original post, Switch A is the root bridge and Switch B goes through the aforementioned election process to determine which port gets blocked to prevent an STP loop. In all scenarios, the higher interface number (when dealing with multiple equal-cost links connecting to the root bridge) gets the axe.

Thank you for the confirmation!

doctorlexus

Well, this has been a bit of a rollercoaster. Good to know how it actually works.

Launchpad

Sy Kosys wrote: »

There are two equal-cost paths between the directly-connected switches. When the root port is being elected, the path/port costs are evaluated to determine which one will be determined as the shortest path to the root bridge.

In the case of multiple equal-cost paths to the root bridge, further evaluations are made and the tie-breaker is the port number on the non-root bridge. The lower interface number is chosen as the root port, and the higher interface number is put into a blocking state.

Thus, taking us pack to the original post, Switch A is the root bridge and Switch B goes through the aforementioned election process to determine which port gets blocked to prevent an STP loop. In all scenarios, the higher interface number (when dealing with multiple equal-cost links connecting to the root bridge) gets the axe.

I just reviewed an ICND2 CBTnuggets video and Keith Barker explicitly said that with all else being equal it's the advertising lowest port, not the receiving lowest port, that determines which port will be forwarding.

This would suggest that gi2/23 is blocking in the second diagram above, despite the fact that it has a lower port number than the other non-root interface of gi2/24, because that link has the higher of the two interfaces on the other (root) end.

Sy Kosys

Launchpad wrote: »

I just reviewed an ICND2 CBTnuggets video and Keith Barker explicitly said that with all else being equal it's the advertising lowest port, not the receiving lowest port, that determines which port will be forwarding.

Correct, because all switches in the topology are sending (and receiving) BPDUs.

Launchpad wrote: »

This would suggest that gi2/23 is blocking in the second diagram above, despite the fact that it has a lower port number than the other non-root interface of gi2/24, because that link has the higher of the two interfaces on the other (root) end.

Negative. The final tiebreaker is the lowest sender port ID which determines the root port. The root bridge does not initiate any tie-breaking process, as it's already the root.

When all other things have been calculated as equal, the tiebreaker conditions, in order, are as follows:
- Lowest root bridge ID
- Lowest root path cost to the root bridge
- Lowest sender bridge ID
- Lowest sender port ID

Since we have a straightforward two-switch configuration with Switch A known as the root bridge, Switch B has already met the first 3 of the 4 tiebreakers. Regardless of which switch port(s) connect to B from the perspective of A, in our scenario Switch B will ALWAYS select it's lowest sending port ID as the root port, and place all others in a blocking state.

GDaines

Sy Kosys wrote: »

The final tiebreaker is the lowest sender port ID which determines the root port. The root bridge does not initiate any tie-breaking process, as it's already the root.

When all other things have been calculated as equal, the tiebreaker conditions, in order, are as follows:
- Lowest root bridge ID
- Lowest root path cost to the root bridge
- Lowest sender bridge ID
- Lowest sender port ID

Since we have a straightforward two-switch configuration with Switch A known as the root bridge, Switch B has already met the first 3 of the 4 tiebreakers. Regardless of which switch port(s) connect to B from the perspective of A, in our scenario Switch B will ALWAYS select it's lowest sending port ID as the root port, and place all others in a blocking state.

Well I get it now. The Root Bridge (switch) never blocks ports, so it's always from the perspective of all the other devices communicating with the root bridge. So in the second diagram from the given example, given that the path costs are identical, Switch B will use the lowest sender port ID to determine the Root Port, thus placing GI2/23 in forwarding mode, and GI2/24 in blocking mode.

And just to confirm my grasp, if the connection via GI2/23 was 2 hops through a 100Mbps switch (2x19=38 cost?), but GI2/24 was 3 hops through only gigabit switches (3x4=12 cost), then port GI2/24 would be the designated Root Port for SwitchB?

james43026

networker050184 wrote: »

The information used to make the decision is received in the BPDU. So when you look at it that way it is apparent it's the remote port ID.

I second this. Also the root bridge would never have any ports blocked, as someone else already stated. The switch making the STP decision, will look at the BPDU's that it has received, and will check the port priority that it received on both interfaces, if they are the same, then it will go with the lowest port number that it received, the corresponding link that received it will become the root port / designated port.

Sy Kosys

GDaines wrote: »

And just to confirm my grasp, if the connection via GI2/23 was 2 hops through a 100Mbps switch (2x19=38 cost?), but GI2/24 was 3 hops through only gigabit switches (3x4=12 cost), then port GI2/24 would be the designated Root Port for SwitchB?

Correct. And as a point of clarification, the Switch B port numbers in this scenario would not matter, as there is no longer an equal-cost path to the root bridge. Tiebreaker #2 would apply in this case.

Launchpad

Sy Kosys wrote: »

Since we have a straightforward two-switch configuration with Switch A known as the root bridge, Switch B has already met the first 3 of the 4 tiebreakers. Regardless of which switch port(s) connect to B from the perspective of A, in our scenario Switch B will ALWAYS select it's lowest sending port ID as the root port, and place all others in a blocking state.

Are you saying that the advertised lowest port (in the BPDUs coming from Switch A) is NOT used in the tiebreaking process for Switch B to determine which port is forwarding and which is blocking? And that only Switch B's own port numbers are relevant?

It was my understanding that the final tiebreaker for the non-root switch is the lowest physical port number received in the BPDUs from the neighboring switch. In the OP's second scenario, since Switch A's lowest advertised port connects to Gi 2/24 on Switch B, Switch B should be forwarding on that port and blocking on Gi 2/23.

doctorlexus

Launchpad wrote: »

It was my understanding that the final tiebreaker for the non-root switch is the lowest physical port number received in the BPDUs from the neighboring switch. In the OP's second scenario, since Switch A's lowest advertised port connects to Gi 2/24 on Switch B, Switch B should be forwarding on that port and blocking on Gi 2/23.

That was how I understood it at first from descriptions I've read. But according to sschwieterman's lab test, Gi 2/24 is blocked in both scenarios. I can't argue with experimental evidence.

GDaines

GDaines wrote: »

In the second diagram from the given example, given that the path costs are identical, Switch B will use the lowest sender port ID to determine the Root Port, thus placing GI2/23 in forwarding mode, and GI2/24 in blocking mode.

I got it wrong! GI2/24 becomes the Root Port in diagram2 as it's connected to the lower port number on the Root Bridge.

Sy Kosys

And I am revisiting this as well. My apologies for the erroneous information, I was wrong for the second scenario and put the config to the home lab and sonofagun it was not what I thought.

I mistook "lowest interface of sender" to mean that of the nonroot switch.

Important thing is, this triggered a conversation which has us all engaged and now we all have definitive proof.

Sorry guys, will double-check my work next time!

doctorlexus