PortFast?

Hey guys-

Just doing some studying here and I came across PortFast. I understand that it reduces the time it takes for a port to go from BLOCKING to FORWARDING modes, but I see everywhere that is should only be used in a switch to workstation/server/host segment and not between two switches... why is this?

Find more posts tagged with

Free for TechExams community: Cybersecurity salary guide

Compare cert salaries and plan your next career move

Button

Comments

MAC_Addy

Because this may create a network loop.

Zartanasaurus

Portfast bypasses the normal spanning tree process. That's obviously fine when you're connecting to a server or computer since you aren't going to get a loop there (exceptions apply). You should enable bpduguard on those ports though.

For obvious reasons, you generally don't want to bypass the spanning tree process between switches since that's where a loop will occur.

Iristheangel

In a switched environment, the ports take about 15-30 seconds to come up because STP is listening and learning through BPDU's. During those 15-30 seconds that it takes for the ports to start forwarding, the switches are determining what the root switch will be and disabling ports that could be a potential loop. By enabling portfast, you are skipping he whole listening/learning process and just having the port immediately start forwarding. This is good for ports directly connected to end-user devices but it can be disastrous if you enable it on ports between other network devices. Hopefully that makes sense

gadav478

Iristheangel wrote: »

In a switched environment, the ports take about 15-30 seconds to come up because STP is listening and learning through BPDU's. During those 15-30 seconds that it takes for the ports to start forwarding, the switches are determining what the root switch will be and disabling ports that could be a potential loop. By enabling portfast, you are skipping he whole listening/learning process and just having the port immediately start forwarding. This is good for ports directly connected to end-user devices but it can be disastrous if you enable it on ports between other network devices. Hopefully that makes sense

I see, so by enabling PortFast, you aren't allowing STP to implement its algorithm to decide which ports to prune. Got it.

Iristheangel

You got it

sratakhin

Can portfast be used on a network without redundant links? I understand it's not the best practice, but I'm just curious

andy4tech

Yea,portfast is usually used between a server and other devices that is connected into your switch that will not create a switching loop if stp is disabled,you know two switches connected together will always have a redundant link with stp enabled on them,redundant links are created so that if there is any problem with the forwarding port,the redundant link can be put in use.

Forsaken_GA

Iristheangel wrote: »

By enabling portfast, you are skipping he whole listening/learning process and just having the port immediately start forwarding

Partially correct. The real trick to PortFast is that it suppresses the TCN's that would normally come with a port changing status, so the rest of the network doesn't go into a reconvergence. Since edge ports will never participate in STP, it's inefficient to send TCN's, since the spanning tree topology hasn't actually changed.

This is also why you don't run portfast on an inter-switch link. When an inter-switch link comes up it is EXTREMELY important TCN's are sent, because the spanning tree forwarding path may very well have changed.

You certainly can put portfast on an inter-switch link, and it won't kill your network. It will possibly form a temporary loop until the portfast port receives a BPDU. At that point, it will drop the port out of forwarding and into blocking and go through the normal STP process. It's bad form, bad configuration, but it's not going to kill your network.

sratakhin

Thank you Forsaken. That's why I meant - using PortFast between switches without redundant links. Not that I'll ever use it though