From: elping (elpingu@acedsl.com)
Date: Wed Sep 25 2002 - 23:55:54 GMT-3
good point
Balaji Siva wrote:
> inline
>
> -----Original Message-----
> From: nobody@groupstudy.com [mailto:nobody@groupstudy.com]On Behalf Of
> Chris
> Sent: Wednesday, September 25, 2002 10:03 PM
> To: Voss, David; ccielab@groupstudy.com
> Subject: Re: Consensus
>
> What happens when the switch hosting the root guard command goes down, get's
> turned off etc?
>
> ### obviously the switch in question would be isolated and becomes a root.
> but then point of STP root in a isolated switch is meaningless. It becomes a
> useless feature...
>
> The ONLY way to ensure a switch or vlan will NEVER become root is to disable
> spanning-tree on that switch or vlan, or turn off the switch.
>
> ### this is just taking the word too literally..and impractical solution..i
> guess may be that is what is needed for the lab.
>
> ALL spanning-tree decision are based on the concept of a root. If you are
> running spanning-tree there is a root somewhere. You cannot guaruntee that
> any particular switch will never become root when this is the case. You have
> to turn it off.
>
> ### understood..but again this issue to taken too literally...
>
> ----- Original Message -----
> From: "Voss, David" <dvoss@heidrick.com>
> To: <ccielab@groupstudy.com>
> Sent: Wednesday, September 25, 2002 9:03 PM
> Subject: RE: Consensus
>
> > So the consensus for the following (I'm adapting it a bit):
> >
> > Make sure that your switch "never" becomes root for vlan 20.
> >
> >
> >
> > root guard on the port hosting VLAN20 if you have 2 or more switches
> sharing
> > VLAN20
> >
> > and
> >
> > disable spanning tree for VLAN20 if you have 1 switch
> >
> >
> >
> >
> >
> > -----Original Message-----
> > From: Balaji Siva [mailto:bsivasub@cisco.com]
> > Sent: Wednesday, September 25, 2002 7:59 PM
> > To: Nick Shah; Voss, David; ccielab@groupstudy.com
> > Subject: RE: Consensus
> >
> >
> > This makes more sense..though the config is actually made on the
> > neighbouring switches !!!
> >
> > It is probably used on distribution/core to prevent small access switches
> > from becoming root..
> >
> > B
> >
> >
> >
> >
> >
> > -----Original Message-----
> > From: nobody@groupstudy.com [mailto:nobody@groupstudy.com]On Behalf Of
> > Nick Shah
> > Sent: Wednesday, September 25, 2002 8:42 PM
> > To: Voss, David; ccielab@groupstudy.com
> > Subject: Re: Consensus
> >
> >
> > Set spantree guard is the ONLY sure shot way.
> >
> > I am willing to bet my $$$ on this one.
> >
> > To raise the odds on the bet :)
> >
> >
> http://www.cisco.com/univercd/cc/td/doc/product/lan/cat5000/rel_6_1/config/s
> > pantree.htm#xtocid2856623
> >
> > rgds
> > Nick
> > ----- Original Message -----
> > From: "Voss, David" <dvoss@heidrick.com>
> > To: <ccielab@groupstudy.com>
> > Sent: Thursday, September 26, 2002 9:51 AM
> > Subject: Consensus
> >
> >
> > > I've been going through the threads on spanning tree and one thread that
> > > never had a consensus was how to ensure a switch would "never" become
> > root.
> > > There is not a text I have found that addresses this. From what I can
> > tell,
> > > one option is to turn off spanning tree completely, the other to set
> > > priority to 65535. Neither sounds appropriate to me.
This archive was generated by hypermail 2.1.4 : Mon Oct 07 2002 - 07:44:04 GMT-3