Spanning tree question!!!!!

masterk

Hello all

I have a question thats bugging me.

Say you have switches A,B and C connected in a triangle.

Some documentation says that if Switch A saw its own BPDU come back to it, then it would detect a loop and block the port.

But some documentation says that Switch A sends a BPDU to Switch B and C, they accept it, modify the cost and send it on, therefore not broadcasting it around the loop, so switch A would never see its own BPDU to know there is a loop?

Do you see what i mean?

danielno8

If the loop was caused by a hub, then the switch would see it's own BPDU come back to it.

masterk

Think i am getting Loopguard mixed up with STP, anyone else agree?

masterk

danielno8 wrote: »

If the loop was caused by a hub, then the switch would see it's own BPDU come back to it.

Thanks yes i agree, and that is essentially the same as plugging one port on a switch into another of its own ports.

So if you had one switch by itself, it is the Root Bridge, then connected one of its ports to another, it will block one port.

So my question is which function blocks one of those ports? As i thought the first rule of STP is that the root bridge has all of its ports as a designated port?

Is there a special function that stops a switch looping back to itself?

ehnde

masterk wrote: »

As i thought the first rule of STP is that the root bridge has all of its ports as a designated port?

Is there a special function that stops a switch looping back to itself?

You are correct, the root switch doesn't do any blocking, the link is blocked on the other end.

masterk

ehnde wrote: »

You are correct, the root switch doesn't do any blocking, the link is blocked on the other end.

Ah indeed, but if you take a quick look at my previous post, i was talking about one single switch by itself. It is still running spanning tree and it is its own root bridge.

So when i connect fa0/1 to fa0/2 via a crossover lead it blocks one port, and that breaks spanning tree rule No. 1 of not having any ports blocked on the root bridge?

See what i am getting at? I wanted to know what function it is that lets a switch detect its own BPDU's and react accordingly, as from my understanding STP prevents loops by design, it doesnt detect them.

alan2308

masterk wrote: »

See what i am getting at? I wanted to know what function it is that lets a switch detect its own BPDU's and react accordingly, as from my understanding STP prevents loops by design, it doesnt detect them.

It see's a frame, it looks at the frame. That's it. And it does detect a loop before preventing it. It blocks a port because it sees two paths back to the root bridge, so that is detecting it.

Did you expect the switch to allow the loop?

ehnde

I know your question is theoretical, but this behavior is important. Try connecting two switch ports on one switch (crossover cable) and give it the no spanning-tree command. Hilarity ensues Oh...if you have taken some or all of the fans out of your switches you might not want to do this. It's likely to overheat and die a permanent death.