Question on adding spanning tree capability for redundancy

spanning tree close the loop

Hello,
I have a question about adding spanning tree capability for redundancy. I understand basically how it works, but I'd like to save money by only buying one managed switch, and combine it with some unmanaged switches.

Here's my idea:
If my layout is like this: a Host, connected by a single link to M1 (managed switch w/RSTP), and M1 is connected to an unmanged switch, say U1, which is connected to U2, which is connected to U3... to U5. Then U5 is connected back to M1 to close the ring (daisy chain loop).

I'm pretty sure that if either the M1-U1 link is broken, or the M1-U5 link is broken, then the system will recover by activating the backup link.

What I'm not sure about is if a break occurs, say between U3 and U4. Will the M1 switch be smart enough to notice traffic is not getting through, and enable partial forwarding on the backup link?

Thanks,
Randy

Re: Spanning Tree question

Sorry, I didn't mean it would enable 'partial forwarding,' it should actually fully reactivate the second connection, so both ports are in forwarding mode. As long as the loop is broken in the middle, then it's ok for both ports to be forwarding.

The problem is I tried this out, and as soon as I close the loop, all traffic stops, and both ports show they are 'discarding' rather than forwarding.

Any ideas why this isn't working?

Thanks,
Randy

Re: Spanning Tree question

I found the problem - it was mainly an issue of waiting long enough. I thought "Rapid" spanning tree protocol would recover quickly, but it took almost 60 seconds to recover from any topology change. I tweaked the timeouts and got it down to ~12 seconds. That seems to be the best we can do.

Randy