From: Erick B. (erickbe@xxxxxxxxx)
Date: Sat Dec 08 2001 - 13:42:08 GMT-3
Hi,
Some more details on portfast.
If a port is configured for portfast it will go to
forwarding state right away then spanning tree runs as
normal. So if a BPDU comes in on this port then
spanning tree reconverges as usual and this port is no
longer in portfast mode.
So, in your disucussion about having a port connected
to another switch and one side had portfast on and
other switch didn't. When the port on the other switch
comes up spanning tree runs as usual and sends BPDUs
and the port w/portfast on see's them and reverts back
to normal spanning tree operation.
I haven't seen portfast ran between switches, it is
usually used and recommened for ports connecting to
workstations and servers. Theres also uplinkfast and
backbonefast which may be more suitable for
switch-to-switch links. I don't have much experience
with those options however.
--- Michael Popovich <m.popovich@home.com> wrote:
> It is my understanding of the portfast command that
> if both ports on
> both sides are configured for portfast it will
> indeed immediately start
> passing packets even if there is a loop in the
> network. However, that is
> only temporary. I understand that the spanning tree
> algorithm is still
> running and after the 50 seconds it takes to run
> that loops will be
> detected and spanning tree will run as normal.
> Portfast doesn't disable
> spanning tree.
>
> MP
>
> -----Original Message-----
> From: nobody@groupstudy.com
> [mailto:nobody@groupstudy.com] On Behalf Of
> Phil
> Sent: Saturday, December 08, 2001 8:16 AM
> To: Taz Kazam; Rivron Francois; thegroup
> Subject: RE: Spantree Timer
>
>
> If you have a port with portfast enabled when you
> connect it to another
> switch port it will be automatically placed in
> disable, unless that
> other port os also configured with portfast enabled
> (and in this case
> they will not block and a loop will occur). Phil.
>
> Taz Kazam <taz_kazam@yahoo.co.uk> escreveu: Thanks
> Francois
>
> I managed to find out. I think that the Cat1900 has
> the ports status set to PORTFAST by default
> (command: start-forwarding),
> hence, it removes the foward delay time (15+15) = 30
> sec, only using
> MaxAge = 20 sec.
> If you configure the port with no start-forwarding
> i.e. disable the PORTFAST, it then takes the 50
> seconds to change over.
>
> Thanks you you reply.
>
> Taz.
>
>
> --- Rivron Francois
> wrote: > Hello,
> > my think is as the switch is able to detect a
> fault
> > at a physical level, the
> > port state of the redondant link is move directly
> to listening state.
> > So the redondant path should be on in 30s.
> >
> > Francois
> >
> > > -----Message d'origine-----
> > > De: Taz Kazam [SMTP:taz_kazam@yahoo.co.uk]
> > > Date: vendredi 7 dicembre 2001 10:38
> > > @: thegroup
> > > Objet: Spantree Timer
> > >
> > > Can anyone help.
> > >
> > > I've step up four 1900 Catalyst switch in a
> loop,
> > to
> > > test spanning tree. One port changes its state
> to
> > > blocking, as I expected, however, when I take
> the
> > > cable out of the non-redundant path, which cause
> > the
> > > redundant path to change from blocking to
> > fowarding -
> > > it does, but it only take 20 seconds. From my
> > reading
> > > this should take 50 seconds (15sec fwd delay -
> > 20sec
> > > MaxAge) Blocking -----15-----Listen---
> > > 15-----Learning----20------Fowarding. Im using
> the default factory
> > > settings!
> > >
> > > Anyone can help?
> > >
> > > Taz.
This archive was generated by hypermail 2.1.4 : Thu Jun 13 2002 - 10:32:40 GMT-3