Re: Basic q regarding STP initial convergence

Hmm ok... I can see where I have been getting confused...
It has been mentioned by various sources including Cisco that " A switch
always enters the blocking state following switch
initialization."<http://www.cisco.com/univercd/cc/td/doc/product/rtrmgmt/sw_ntman/cwsimain/cwsi2/cwsiug2/vlan2/stpapp.htm>

I took this to mean whenever a switch is rebooted...but it is also stated by
various sources that the blocking state is not entered if the port has ever
been running stp b4...so perhaps when the switch was first unboxed...a
reload or shut/no sh will not result in the initial blocking state...
In truth I've never seen the initial blocking state during my own labbing, &
also never checked whenever I installed/configured new switches in the past.

Oh well.
Has noone actually seen an initial blocking state?! :)

Cheers for the replies.

2009/8/9 Swap <ccie19804_at_gmail.com>

> Roy,
>
> Your understanding is not fully correct.
>
> When Switch powers up, it starts from listening state, (there is no 20 sec
> blocked delay). It goes via Listening (15sec) and Learning states (15Sec)
> and hence takes 30 Secs to reach the forwarding state. (Portfast can save
> this 30 secs by directly putting the port to forwarding state).
>
> Blocked state role is only attained if the port knows that it is part of a
> physical loop.
>
> Remember that a port must receive BPDUs in order to stay blocked else it'll
> again start assuming that it is the root and go via listening+learning.
>
> HTH
> Swap
> #19804
>
>
> -----Original Message-----
> From: nobody_at_groupstudy.com [mailto:nobody_at_groupstudy.com] On Behalf Of
> Roy
> Waterman
> Sent: Sunday, August 09, 2009 8:51 AM
> To: Keegan.Holley_at_sungard.com
> Cc: ccielab_at_groupstudy.com; Danshtr; nobody_at_groupstudy.com
> Subject: Re: Basic q regarding STP initial convergence
>
> Hi Keegan
>
> I think you're confusing the blocking state with blocked ports.
> There is an initial blocking state, when a switch is 1st powered on, which
> lasts 20 secs (as far as im aware).
> This time cannot be modified.
> During this time, a switch can only receive bpdus, it cannot transmit bpdus
> or send/rx data.
>
> I was referrring to the initial stp convergence.
> You are referring to stp topology changes causing a blocked port to
> transition to forwarding state.
> The only blocking state occurs at initialisation..ports which are blocked
> after convergence are not in the blocking state.
>
> HTH,
>
> Roy
>
>
>
> There is a difference between the initial blocking state of 20 secs
>
> 2009/8/9 <Keegan.Holley_at_sungard.com>
>
> > I think you're confusing the listening and blocking states. The blocking
> > state is when it has already converged and knows to block a port.
> Listening
> > is when it's trying to see if it's the root. For example if the STP
> > algorithm decided a blocking port was going to transition to forwarding
> why
> > would it block for an additional 20s before listening and not passing
> > traffic for an additional forwarding-delay?
> >
> > HTH,
> >
> > Keegan
> >
> >
> >
> > *Re: Basic q regarding STP initial convergence*
> > *Roy Waterman * to: Danshtr 08/08/09 04:10 PM
> >
> > Sent by: *nobody_at_groupstudy.com* Cc: ccielab
> > *Please respond to Roy Waterman*
> >
> >
> >
> > ------------------------------
> >
> >
> >
> > Never mind, after thinking about it I know the answer ;)
> > Posted prematurely.
> >
> > 2009/8/8 Danshtr <danshtr_at_gmail.com>
> >
> > > What will happen when you boot all your switches at the same time?
> > >
> > > Best regards,
> > > Dan
> > >
> > >
> > >
> > > On Sat, Aug 8, 2009 at 4:25 PM, Roy Waterman <roy.waterman_at_gmail.com
> > >wrote:
> > >
> > >> Hi all
> > >> In several texts I've read that switches upon convergence each think
> > they
> > >> are the root bridge AND send config pdus out all ports.
> > >> However in Cisco's implementation, with the intial blocking state
> being
> > 20
> > >> secs, and having the ability to only receive config pdus from
> > neighboring
> > >> switches,
> > >> while each switch upon initial convergence may THINK they are the root
> > >> bridge is it true that in general they do not have a chance to
> advertise
> > >> that as a result of possibly receiving a better bpdu from a
> neighboring
> > >> switch?
> > >>
> > >> Plus, was the initial 20 sec blocking state proposed by Cisco to
> > actually
> > >> allow switches enough time to learn the identity of the root bridge
> > before
> > >> sending bpdus of their own?
> > >>
> > >> Please advise.
> > >>
> > >> --
> > >> Regards
> > >> Roy
> > >>
> > >>
> > >> Blogs and organic groups at http://www.ccie.net
> > >>
> > >>
> _______________________________________________________________________
> > >> Subscription information may be found at:
> > >> <http://www.ccie.net/>http://www.groupstudy.com/list/CCIELab.html
> > >>
> > >>
> > >>
> > >>
> > >>
> > >>
> > >>
> > >>
> > >
> >
> >
> > --
> > Regards
> > Roy
> >
> >
> > Blogs and organic groups at <http://www.groupstudy.com/list/CCIELab.html
> >
> > http://www.ccie.net
> >
> > _______________________________________________________________________
> > Subscription information may be found at:
> > <http://www.ccie.net/>http://www.groupstudy.com/list/CCIELab.html
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
>
>
> --
> Regards
> Roy
>
>
> Blogs and organic groups at http://www.ccie.net
>
> _______________________________________________________________________
> Subscription information may be found at:
> http://www.groupstudy.com/list/CCIELab.html
>
>
>
>
>
>
>
>
>

-- 
Regards
Roy
Blogs and organic groups at http://www.ccie.net