[f-nsp] Cisco equivalent of 'portfast' on FCX?

Wilbur Smith wsmith at brocade.com
Mon Jun 16 15:39:50 EDT 2014


Hello Folks,

You can do this with "Fast Span" or "Fast Uplink Span". Although it won't display in the running config, this is already enabled by default

Here's the details from our manual:
----------------------------------------------------------------------------------------------------------------------
Fast Port Span

When STP is running on a device, message forwarding is delayed during the spanning tree
recalculation period following a topology change. The STP forward delay parameter specifies the
period of time a bridge waits before forwarding data packets. The forward delay controls the
listening and learning periods of STP reconvergence. You can configure the forward delay to a
value from 4-30 seconds. The default is 15 seconds. Thus, using the standard forward delay,
convergence requires 30 seconds (15 seconds for listening and an additional 15 seconds for
learning) when the default value is used.

This slow convergence is undesirable and unnecessary in some circumstances. The Fast Port
Span feature allows certain ports to enter the forwarding state in four seconds. Specifically, Fast
Port Span allows faster convergence on ports that are attached to end stations and thus do not
present the potential to cause Layer 2 forwarding loops. Because the end stations cannot cause 
forwarding loops, they can safely go through the STP state changes (blocking to listening to learning
to forwarding) more quickly than is allowed by the standard STP convergence time. Fast Port Span
performs the convergence on these ports in four seconds (two seconds for listening and two
seconds for learning).

In addition, Fast Port Span enhances overall network performance in the following ways:

. Fast Port Span reduces the number of STP topology change notifications on the network.

When an end station attached to a Fast Span port comes up or down, the Brocade device does
not generate a topology change notification for the port. In this situation, the notification is
unnecessary since a change in the state of the host does not affect the network topology.

. Fast Port Span eliminates unnecessary MAC cache aging that can be caused by topology
change notifications. 

Bridging devices age out the learned MAC addresses in their MAC caches
if the addresses are unrefreshed for a given period of time, sometimes called the MAC aging
interval. When STP sends a topology change notification, devices that receive the notification
use the value of the STP forward delay to quickly age out their MAC caches. For example, if a
device normal MAC aging interval is 5 minutes, the aging interval changes temporarily to the
value of the forward delay (for example, 15 seconds) in response to an STP topology change.
In normal STP, the accelerated cache aging occurs even when a single host goes up or down.
Because Fast Port Span does not send a topology change notification when a host on a Fast
Port Span port goes up or down, the unnecessary cache aging that can occur in these
circumstances under normal STP is eliminated.

Fast Port Span is a system-wide parameter and is enabled by default. Thus, when you boot a
device, all the ports that are attached only to end stations run Fast Port Span. For ports that are not
eligible for Fast Port Span, such as ports connected to other networking devices, the device
automatically uses the normal STP settings. If a port matches any of the following criteria, the port
is ineligible for Fast Port Span and uses normal STP instead:

. The port is 802.1Q tagged
. The port is a member of a trunk group
. The port has learned more than one active MAC address
. An STP Configuration BPDU has been received on the port, thus indicating the presence of
another bridge on the port.

You also can explicitly exclude individual ports from Fast Port Span if needed. For example, if the
only uplink ports for a wiring closet switch are Gbps ports, you can exclude the ports from Fast Port
Span.

Disabling and re-enabling Fast Port Span

Fast Port Span is a system-wide parameter and is enabled by default. Therefore, all ports that are
eligible for Fast Port Span use it.
To disable or re-enable Fast Port Span, enter the following commands.
Brocade(config)# no fast port-span
Brocade(config)# write memory

To re-enable Fast Port Span, enter the following commands.

Brocade(config)# fast port-span
Brocade(config)# write memory

----------------------------------------------------------------------------------------------

Fast Uplink Span

The Fast Port Span feature described in the previous section enhances STP performance for end
stations. The Fast Uplink Span feature enhances STP performance for wiring closet switches with
redundant uplinks. Using the default value for the standard STP forward delay, convergence
following a transition from an active link to a redundant link can take 30 seconds (15 seconds for
listening and an additional 15 seconds for learning).

You can use the Fast Uplink Span feature on a Brocade device deployed as a wiring closet switch to
decrease the convergence time for the uplink ports to another device to just one second. The new
Uplink port directly goes to forward mode (bypassing listening and learning modes). The wiring
closet switch must be a Brocade device but the device at the other end of the link can be a Brocade
device or another vendor's switch.

Configuration of the Fast Uplink Span feature takes place entirely on the Brocade device. To
configure the Fast Uplink Span feature, specify a group of ports that have redundant uplinks on the
wiring closet switch (Brocade device). If the active link becomes unavailable, the Fast Uplink Span
feature transitions the forwarding to one of the other redundant uplink ports in just one second. All
Fast Uplink Span-enabled ports are members of a single Fast Uplink Span group.

NOTE
To avoid the potential for temporary bridging loops, Brocade recommends that you use the Fast
Uplink feature only for wiring closet switches (switches at the edge of the network cloud). In
addition, enable the feature only on a group of ports intended for redundancy, so that at any given
time only one of the ports is expected to be in the forwarding state.

NOTE
When the wiring closet switch (Brocade device) first comes up or when STP is first enabled, the
uplink ports still must go through the standard STP state transition without any acceleration. This
behavior guards against temporary routing loops as the switch tries to determine the states for all
the ports. Fast Uplink Span acceleration applies only when a working uplink becomes unavailable.

Active uplink port failure

The active uplink port is the port elected as the root port using the standard STP rules. All other
ports in the group are redundant uplink ports. If an active uplink port becomes unavailable, Fast
Uplink Span transitions the forwarding of traffic to one of the redundant ports in the Fast Uplink
Span group in one second bypassing listening and learning port states.
Switchover to the active uplink port

When a failed active uplink port becomes available again, switchover from the redundant port to
the active uplink port is delayed by 30 seconds. The delay allows the remote port to transition to
forwarding mode using the standard STP rules. After 30 seconds, the blocked active uplink port
begins forwarding in just one second and the redundant port is blocked.

NOTE
Use caution when changing the spanning tree priority. If the switch becomes the root bridge, Fast
Uplink Span will be disabled automatically.

Configuring a Fast Uplink port group

To configure a group of ports for Fast Uplink Span, enter the following commands:
Brocade(config)# fast uplink-span ethernet 1/2/1 to 1/2/4
Brocade(config)# write memory


----------------------------------------------------------------------------------------------------------------







-----Original Message-----
From: foundry-nsp [mailto:foundry-nsp-bounces at puck.nether.net] On Behalf Of "Rolf Hanßen"
Sent: Monday, June 16, 2014 12:20 AM
To: rsm at fast-serv.com
Cc: foundry-nsp at puck.nether.net
Subject: Re: [f-nsp] Cisco equivalent of 'portfast' on FCX?

Hi,

btw, there is a "no spanning-tree", but you need to configure it per vlan.

SSH at xxx(config)# vlan 3000
SSH at xxx(config-vlan-3000)#no sp
  spanning-tree              Set spanning tree for this VLAN
SSH at xxx(config-vlan-3000)#no spanning-tree

kind regards
Rolf

> Nothing jumping out at me:
>
> #spanning-tree ?
>    802-1w               Disable or enable IEEE 802.1W parameters
>    designated-protect   Declare the port to be on Designated guard for
> all
>                         spanning trees
>    root-protect         Declare the port to be on Root guard for all
> spanning
>                         trees
>    <cr>
>
> I'm guessing something like:
>
> #no spanning-tree
> #stp-bdpu-guard
>
> Any clues?
>
> --
> Randy M.
>
> _______________________________________________
> foundry-nsp mailing list
> foundry-nsp at puck.nether.net
> http://puck.nether.net/mailman/listinfo/foundry-nsp
>


_______________________________________________
foundry-nsp mailing list
foundry-nsp at puck.nether.net
http://puck.nether.net/mailman/listinfo/foundry-nsp




More information about the foundry-nsp mailing list