From: Carlos Trujillo Jimenez (nergal888@hotmail.com)
Date: Thu Nov 29 2007 - 21:55:06 ART
Thanks Fernando.
Well I Investigated, and the time to converges varies according to the
situation the switch is experiencing.
For example if using the normal spanning tree timers. when we have a direct
link failure in one of its physical interfaces, the switch excludes the max
age timer, so the total convergence time becomes 30 seconds.
If there is an indirect link failure, the switch note about a cut path
towards the root switch because it stops recibing bpdu packets from its root
port, and depends if there are alternative links, it starts recibig bpdu
root packets from an alternative interface.
In this second scenario, the total convergence time becomes 50 or 52 seconds
plus the hello time.
Thanks, again for the response Fernando.
>From: "Fernando Carvalho" <fernando.cagica@t-online.de>
>X-OriginalArrivalTime: 29 Nov 2007 07:56:07.0269 (UTC)
>FILETIME=[4C405150:01C8325D]
>
>Hi,
>
>If you want to use the command spanning-tree vlan <> root
>primary|secondary, do it without the diameter keyword. The command
>itself it is a macro (as already stated in anoter answer to this
>question) and the diameter macro will find an optimized value for the
>timers according to the size of the network (diameter). If it is stated
>to adjust the timers then you want to do it manually, and i think you
>should do it on both the root primary and secondary. The amount of time
>for a port to start forwarding, i think here there are two situations:
>one is when you have a port in the default status or shutdown and then
>you move it to this vlan and activate it, here you will have
>2xforwarding-time until it starts forwarding (for pvst and not sure but
>believe also for rpvst). The second situation is when due to a topology
>change some port (s) in blocking status must change to forwarding, here
>the documentation says 2xforwarding-time + max-age for pvst (for rpvst
>you can have less than 1s).
>
>So finally as an example if you want to ensure that any port added to
>your vlan (according to the first situation i've described) will start
>forwarding in 24s you should put your forwarding -time to 12s, if you
>want it forwarding in 12s you should put your forwarding-time to 6s, ...
>
>HTH
>
>
>-----Urspr|ngliche Nachricht-----
>Von: nobody@groupstudy.com [mailto:nobody@groupstudy.com] Im Auftrag von
>Carlos Trujillo Jimenez
>Gesendet: Mittwoch, 28. November 2007 23:28
>An: ccielab@groupstudy.com
>Betreff: tunning spanning-tree timers.
>
>
>Hi group.
>
>if it is stated to adjust the timers to the minimal delay for a port to
>start forwarding frames, and if in the topology we have 2 root bridges,
>one
>defined as primary, and the other one as secondary, and the rest art
>only
>clients.
>
>My question states:
>MUST WE configure the following only in the primary root switch? or on
>both
>root switches? the primary and the secondary?
>
>sw2(config)#spanning-tree vlan 1 root primary diameter 2
>
>
>
>as a result we have the following:
>
>sw2#show spanning-tree vlan 1
>
>VLAN0001
> Spanning tree enabled protocol ieee
> Root ID Priority 24577
> Address 000a.b86a.0580
> This bridge is the root
> Hello Time 2 sec Max Age 10 sec Forward Delay 7 sec
>
>Another question from the same task is that the total delay time from a
>port
>to start forwarding frames after using this command converts from the
>traditional 50 seconds... to 24 seconds, so we now have fine tuned the
>network, to "converge" in 24 seconds.
>
>Now, if the task states to tune the network to converge in 12 seconds?
>how
>can we do this one? Does rappid spanning tree?
>
>Now Im trying to tune with pervlan spanning tree, but the less value I
>Can
>get is 24 seconds.. any idea? how can I get the 12 seconds?
>
>_________________________________________________________________
>Las mejores tiendas, los precios mas bajos, entregas en todo el mundo,
>YupiMSN Compras: http://latam.msn.com/compras/
>
>_______________________________________________________________________
>Subscription information may be found at:
>http://www.groupstudy.com/list/CCIELab.html
>
>
This archive was generated by hypermail 2.1.4 : Sat Dec 01 2007 - 06:37:32 ART