From: Taimoor (taimoor@xxxxxxxxxxx)
Date: Tue May 21 2002 - 18:31:07 GMT-3
Jeremy,
It does not matter whether you are sending unicast updates or not, if an
update is learnt from an interface it will not be sent out the same
interface if split horizon is turned on. You already know about the
different scenarios where split horizon is on or off by default, but in
answering your question, no the updates will not be forwarded out the same
interface they were learnt if split horizon is on. (Unless you are running
into some sort of a bug).
Taimoor
> -----Original Message-----
> From: nobody@groupstudy.com [mailto:nobody@groupstudy.com]On Behalf Of
> Wright, Jeremy
> Sent: Tuesday, May 21, 2002 2:11 PM
> To: 'tsabry@slb.com'; Wright, Jeremy; dmadlan@qwest.com
> Cc: 'Mingzhou Nie'; ccielab@groupstudy.com
> Subject: RE: rip & split horizon
>
>
> you are right with what you are saying. my question is, with split horizon
> on, normally the updates would not make it to the other spoke
> because split
> horizon will not allow that. so, with split horizon on it wont
> get there but
> if i specify a neighbor statement will the update be unicast to the the
> other spoke. sorry for the confusion...just getting my weird scenarios
> straight.
>
> -----Original Message-----
> From: Tarek Sabry [mailto:tsabry@houston.sns.slb.com]
> Sent: Tuesday, May 21, 2002 3:59 PM
> To: 'Wright, Jeremy'; dmadlan@qwest.com
> Cc: 'Mingzhou Nie'; ccielab@groupstudy.com
> Subject: RE: rip & split horizon
>
>
> I'm having a hard time following this thread! If I understand correctly,
> Dave is correct; split horizon is on, so disable it in order for RIP to
> work. If you want to use unicase RIP then go ahead and define neighbors.
>
> Tarek
>
> -----Original Message-----
> From: nobody@groupstudy.com [mailto:nobody@groupstudy.com]On Behalf Of
> Wright, Jeremy
> Sent: Tuesday, May 21, 2002 3:33 PM
> To: 'dmadlan@qwest.com'; Wright, Jeremy
> Cc: 'Mingzhou Nie'; 'ccielab@groupstudy.com'
> Subject: RE: rip & split horizon
>
>
> yes a frame network. but cant you go back and manually turn split
> horizon on
> the interface. obviously this is not real world, im trying to
> think of every
> stupid possible scenario for the lab.
>
> -----Original Message-----
> From: MADMAN [mailto:dmadlan@qwest.com]
> Sent: Tuesday, May 21, 2002 3:13 PM
> To: Wright, Jeremy
> Cc: 'Mingzhou Nie'; 'ccielab@groupstudy.com'
> Subject: Re: rip & split horizon
>
>
>
> is this a frame network were talking about?? If so split horizon is
> disable.
>
> Dave
>
> "Wright, Jeremy" wrote:
> >
> > without neighbor statements though wouldnt split horizon stop
> it going out
> > to the other spoke since it is on the same interface
> >
> > -----Original Message-----
> > From: MADMAN [mailto:dmadlan@qwest.com]
> > Sent: Tuesday, May 21, 2002 3:05 PM
> > To: Wright, Jeremy
> > Cc: 'Mingzhou Nie'; 'ccielab@groupstudy.com'
> > Subject: Re: rip & split horizon
> >
> > Yes, without neighbor statements.
> >
> > Dave
> >
> > "Wright, Jeremy" wrote:
> > >
> > > i understand that. im talking straight physical interfaces. so if the
> hub
> > > (r1) receives an update from spoke 1(r2) will it forward the update to
> the
> > > other spoke 2 (r3) if you have neighbor statements at the hub pointing
> to
> > > the spokes.
> > >
> > > -----Original Message-----
> > > From: Mingzhou Nie [mailto:mnie@yahoo.com]
> > > Sent: Tuesday, May 21, 2002 2:07 PM
> > > To: Wright, Jeremy; 'ccielab@groupstudy.com'
> > > Subject: Re: rip & split horizon
> > >
> > > whether split horizon is on or not depends on interface type. main
> > > interface has it on and subinterface has it off. It has nothing to do
> > > with using neighbor or not.
> > >
> > > --- "Wright, Jeremy" <JA_WRIGHT@admworld.com> wrote:
> > > > i understand the concept of rip and split horizon. if you have a hub
> > > > and
> > > > spoke setup and if you do neighbor statements at the hub pointing to
> > > > the 2
> > > > spokes will the update get there with split horizon turned
> on? i dont
> > > > have
> > > > the equipment in front of me to simulate this right now. thanks.
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > > ************************
> > > > Jeremy Wright
> > > > Network Analyst
> > > > Archer Daniels Midland
> > > > ja_wright@admworld.com
> > > > (217)451-4063
> > > >
> > > > ************************
This archive was generated by hypermail 2.1.4 : Thu Jun 13 2002 - 10:59:04 GMT-3