From: Pavel Bykov (slidersv@gmail.com)
Date: Tue Mar 24 2009 - 23:01:50 ART
1st: Transition from DOWN state in 802.1d is only 2xFWD timers, therefore
only 30seconds in total, not 50. MAXAGE timer is run only for BPDU timeout.
So if the port is BLK, then it will first wait MAX AGE (to timeout - e.g.
error has occured) before going through 2xFWD.
2nd: Using debugs, the transition to FORWARDING state from link detection
takes about 150ms when PORTFAST is enabled.
3rd: What is primary purpose of spanning tree? To create a network without
any loops - a tree does not loop (unless it's mutated), therefore it's a
tree, and is spanning when devices are connected, but does not loop.
Spanning tree.
Disabling spanning tree and then looping the network is counter productive.
If guards will not work right, and loop does occur, your whole L2 network
(inside that VLAN at least) will flood, overload and go down.
If you see an advantage (a real quantifiable advantage) of disabling
spanning tree towards access switch - sure, that's possible, and it could be
perfectly ok. But not when you have a physical loop there. The loop iss
getting dangerous. We need controllable and predictable network.
On Wed, Mar 25, 2009 at 2:05 AM, ccie2k4@gmail.com <ccie2k4@gmail.com>wrote:
> Hi i have read that it is not recommended to enable portfast on a switch to
> switch interlink however if one switch is a L2 Access Switch and the other
> is a central switch if portfast is enabled on the L2 access switch than
> there don't seem to be any issues even though physically a loop is created
> by connecting to 2 interfaces on the L2 from the central. One just goes
> into
> blocking. The central side is configured for root guard
>
> If portfast is enabled on ports on L2 than wouldn't this provide faster
> connectivity since the port won't have to go through all the states and
> transition to forwarding instantly. Is there a best practice for such
> connectivity even though there are possibilities that a loop might occur.
> Also if there was no physical loop (only 1 port) than would it make
> difference.
>
> if enable bpdu guard on this port than it will disable it because there are
> bpdu's being sent even though both sides are connected as access ports. I
> see this happen but is it the correct way
>
> if portfast is enabled than does the port transition instantly. I see that
> it does but books always say it bypasses listening and learning stage which
> is only 30 sec. spanning-tree is 50 sec hence that still leaves 20 seconds.
> Is this true because 20 it seems like ports go to forwarding a lost faster
> than 20 seconds or does the 20 seconds affect only during convergence.
>
> thnks
>
>
> Blogs and organic groups at http://www.ccie.net
>
> _______________________________________________________________________
> Subscription information may be found at:
> http://www.groupstudy.com/list/CCIELab.html
>
>
>
>
>
>
>
>
-- Pavel Bykov ---------------- Don't forget to help stopping the braindumps, use of which reduces value of your certifications. Sign the petition at http://www.stopbraindumps.com/Blogs and organic groups at http://www.ccie.net
This archive was generated by hypermail 2.1.4 : Mon Apr 06 2009 - 06:44:07 ART