From: Alexander Arsenyev (GU/ETL) (alexander.arsenyev@ericsson.com)
Date: Mon Sep 26 2005 - 12:51:34 GMT-3
I agree with below statement. In fact, RFC2328 states in section 16(3):
<quote>
The inter-area routes are calculated, through examination of summary-LSAs. If the router is attached to multiple areas
(i.e., it is an area border router), only backbone summary-LSAs are examined.
</quote>
So when there are 2 paths to 1.1.6.6/32 from R3, R3 will use path via area 0/via R1.
If link between R1 and R3 is down, R3 ceases to become an ABR and uses path via R2/area 1.
HTH,
Cheers
Alex
#13405
-----Original Message-----
From: nobody@groupstudy.com [mailto:nobody@groupstudy.com]On Behalf Of
Niche
Sent: 26 September 2005 08:32
To: kevin gannon
Cc: Jaycee Cockburn - BCX SS; Stefan Grey; ccielab@groupstudy.com
Subject: Re: VERY INTERESTING OSPF behaviour??
Hi Stefan,
AFAIK, OSPF will always pick the route which use area 0 as a transit area
instead of other non-area-0 areas.
One simple test you can try, configure a Virtual Link between R3 and R2
(Area 1 - Ethernet) and keep the ospf cost you have modified on the
interface remain unchange. And tell us what you get.
Cheers~
Jacky
On 9/25/05, kevin gannon <kevin@gannons.net> wrote:
>
> Jaycee but from the sh ip route both routes are IA routes no ?
>
> Regards
> Kevin
>
> On 9/25/05, Jaycee Cockburn - BCX SS <Jaycee.Cockburn@bcx.co.za> wrote:
> > Hi Stefan,
> > I think this has to do with OSPF prefering the intra area routes before
> > inter area routes. Area 0 routes being the backbone area will, I
> > suppose, be prefered over other area's routes. What happens, just for
> > kicks, if you add on R2 "area 1 virtual-link <R3 RID>" and on R3 "area 1
> > virtual-link <R2 RID>"?
> >
> > Cheers
> > JC
> >
> > -----Original Message-----
> > From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf Of
> > Stefan Grey
> > Sent: 25 September 2005 04:25 PM
> > To: ccielab@groupstudy.com
> > Subject: VERY INTERESTING OSPF behaviour??
> >
> > Hi, group,
> > I see such behaviou the first time:
> >
> > R4
> > |
> > |
> > R1------R2
> > | __ |
> > | _|
> > | |
> > R3----
> >
> >
> >
> >
> > Between R1 and R4 is area 3, between router R3 and R1 area 0. Between R1
> > and
> > R2 area 0. connection between R3 - R2 area 1. As to physical
> > connections:
> > between R1 and R4 ethernet, between R3 and R2 ethernet. Between R1,R2
> > nbma frame-relay, between R1 -R3 ISDN. Everything is configured
> > correctly. R3 gets two routes about lo0 of R4 from R1 and R2.
> > on R1 is configured the ip ospf 9999 command to make this route less
> > preffereable than from R2. But what strange thing wee see:
> >
> >
> > O IA 1.1.6.6/32 <http://1.1.6.6/32> [110/10001] via
1.1.35.5<http://1.1.35.5>,
> 00:17:32, BRI0/0
> > I wonder why the less prefereble route is used?? Ok I turn of the bri0/0
> > interface and look which other route I get from router R2.
> >
> > O IA 1.1.6.6/32 <http://1.1.6.6/32> [110/76] via
1.1.23.2<http://1.1.23.2>,
> 00:00:29, Ethernet0/1
> > This route is the same as previous but with much better metric. Why it
> > is not choosed?? But the worther one is choosed??
> >
> >
> > Thanks,
> >
> > The only one advantage of the route through R1 is that R1 - R3 is in
> > AREA0.
> > But as we know the route is choosed just in two criterias AD/ Metric.
> > I am totally confused with this.
> >
> > _________________________________________________________________
> > The secret is out - quality Irish singles meet at
Match.com<http://Match.com>
> !
> > http://match.msn.ie/match/mt.cfm?pg=channel&tcid=237596
> >
> > _______________________________________________________________________
> > Subscription information may be found at:
> > http://www.groupstudy.com/list/CCIELab.html
> >
> > _______________________________________________________________________
> > Subscription information may be found at:
> > http://www.groupstudy.com/list/CCIELab.html
>
> _______________________________________________________________________
> Subscription information may be found at:
> http://www.groupstudy.com/list/CCIELab.html
This archive was generated by hypermail 2.1.4 : Sun Oct 02 2005 - 14:40:16 GMT-3