spanning-tree portfast trunk

Peter.D.Brown · ‎11-09-2007

Hi, I'm a bit confused about the spanning-tree portfast trunk command. Portfast should only be used on access ports connected to a single end device (PC/printer/server). What is the point then in having a portfast trunk command? When you configure it it gives the same error message and basically tells you not to do it:

%Warning: portfast should only be enabled on ports connected to a single

host. Connecting hubs, concentrators, switches, bridges, etc... to this

interface when portfast is enabled, can cause temporary bridging loops.

Use with CAUTION

Anyone got any idea?

Thanks

Pete.

bvsnarayana03 · ‎11-09-2007

A switchport before coming to the forwarding state, transitions through states of blocking, listening, learning. This takes about 50 sec for the port to come to forwarding state. But if the port is access port, where a pc or printer is connected, why should it wait so long. Access ports doesnt need any stp calculation & needs to come up as soon as posible for data transmission. portfast on an interface transitions it from directly blocking to forwarding.

However, as u mentioned when a switch is connected to it & the port recv a bpdu, stp would put the port in blocking state if bpduguard is configured.

Edison Ortiz · ‎11-09-2007

With VoIP, there are times when you need to configure a trunk on switchports connected to an IP phone.

The IP phone will bridge the connection to the PC connected right behind it.

The IP Phone would belong to Vlan A while PC would belong to Vlan B but on the switch, the switchport must be trunk with Vlan B being the native Vlan.

Spanning-tree portfast trunk is needed to accelerate the port opening so the IP Phone can obtain its IP address and configuration information from the tftp server.

In short, spanning-tree porfast trunk is needed on switchports configured as trunk which are connected to end devices (not switches) while the regular spanning-tree portfast is needed on access ports.

HTH,

Francois Tallet · ‎11-09-2007

Well, you can have a trunk to what looks like a host for a switch: a router for instance. A router is an L3 device that cannot introduce a bridging loop, that does not run STP but that can trunk with 802.1Q or even ISL. Portfast is clearly indicated in that case.

It used to be that trunks could only be made between switches, that's why trunks where not considered as valid for portfast. That's also the reason why the plain portfast CLI did not have any effect on trunks. That's also why a separate keyword had to be added to the CLI when Cisco evenutally realized that it could be useful to run portfast on trunks;-)

Regards,

Francois

Peter.D.Brown · ‎11-12-2007

Thanks Francois, (and also the other 2 guys for your input).

I had a think over the weekend and came up with another reason which seems quite obvious to me now. If you've only got 1 uplink (which is a trunk) and therefore no chance of bridging loops then I think one should put portfast-trunk on that too.

Cheers, have a great day.

Pete.