Fabric Path STP Interaction
Fabric Path STP Interaction
Fabric Path and STP interaction is based on core ports and edge ports. Let’s first discuss edge ports, as edge ports are connected to classical ethernet domain so, edge ports process the BPDU and participates in STP topology calculation.
However Core ports are those ports which are part of only fabric path domain so these ports don’t participate in STP and neither process any BPDU and nor transmits any BPDU. Even TCN are also not propagated to other domain connected to same fabric path fabric.
Due to this above behaviour, entire Cisco fabric Path domain appears as a single STP bridge to any Classical Ethernet STP domains and work like the all Cisco fabric Path bridges share a common bridge ID C84C.75FA.6000 and this BID is statically defines and cannot be configured by any administrator.
You must configure all Cisco FabricPath Edge Switch as STP root by following commands spanning-tree vlan X root primary to make the edge switch root for all vlan participating in STP. Even though multiple edge switch are connected to fabric Path domain, then all those switch must have same bridge priority value.
You should also configure STP root-guard on all edge ports which are going to connect to classical STP ethernet switch to make sure that all edge switch should be root.
TCN Forwarding in fabric Path:
By default STP BPDU and TCNs are not forwarded through Cisco fabric Path core port. But there are some configuration option available by which we can enable fabric Path edge switch to forward TCN through Fabric path domain to support certain topologies.
Below Figure states the case why TCNs must be carried over fabric path domain to optimize reconvergence on switch or failed interface.
Here in normal scenario, traffic for Host B is forwarded by S1 due to STP blocked port on ethernet switch. Now the link got failed as shown in second figure which caused to generate the TCNs and as soon as TCN is received by ethernet switch, it will flush the Mac address table, likewise S2 will also flush the MAC table, but S2 has core port and connected to S1, S2 will not able to inform about TCN to S1 due to which S1 will not flush its table and continue to forward the traffic, but S1 port is down due to link failure due to which all traffic from S1 to host B will be dropped .
To optimize the reconvergence when we connect the STP domain to fabric Path core, you must configure spanning tree protocol domain id to all fabric path edge switches. Also assign same domain ID to Cisco fabric Path edge switches interconnected through STP topology.
Now if S2 receives a TCN from an edge port, it should forward the TCN to Cisco FabricPath so that S1 can flush its MAC address entries and also forward the TCN back to the Spanning Tree Protocol domain. If S3 receives a TCN from Spanning Tree Protocol domain 1, it should flush its MAC address entries for S1 and S2, but it should not forward the TCN to Spanning Tree Protocol domain 2. Because the TCN flooded to Cisco FabricPath from S2 originated from a different Spanning Tree Protocol domain ID, S3 and S4 will not propagate the TCN to Spanning Tree Protocol domain 2.