From: Champika G (693455@gmail.com)
Date: Wed Sep 12 2007 - 09:30:24 ART
Hi Eric
Lets say HSRP is run between R2-R4 & R3-R5 and we make sure that R2 is
the Active. THen R3 will always send traffic to R2. Also we need to
stop the default propagated by the ABR at R3 & R5 to ensure that R3 &
R5 doesnt see two routes going out.
But when R2-R1 link is down the traffic will come from R3 to R2 then
go to R4 to R1 and coming back from R1 to R4. Now at R4 he will not
pass it on to R2 since the Vlans of R3 are over a directly connected
interface. THerefore R4 will just forward to R5 instead when it should
be sending to R2.
Here the symmetry is lost.
This is where i dont know how to go about it?
On 9/12/07, Champika G <693455@gmail.com> wrote:
> Hi Eric
>
> It can go through R3->L2SW1->L2SW2->R4->R1 provided it comes back the
> same route. But this is not what is needed for this scenario.
>
> The link costs are all the same except for R4-L1SW2 link. THat cost is
> 3. Without that cost R1 will have equal cost routes to the vlan
> segments and will loadbalance between R1-R4 & R1-R2. This is not the
> required behavior.
> When all interface and devices are in an UP state the prefered return
> path is from R1-R2.
> This is actually a live scenario in Customer place and because the
> routers are running IOS FW services routing must be symetric.
> All links are 100mb ethernet.
> Therefore If R1-R2 link is not down traffic should come back that path
> irrespective of other link failures.
> All suggestions are welcome.
> CHeers.
>
>
> On 9/12/07, Champika G <693455@gmail.com> wrote:
> > The return path for (2) should be R1->R2->R4->L2SW2->L2SW1->R3 &
> > The return path for (3) should be R1->R4->R2>L2SW1->R3.
> >
> > Under normal circumstances R3 should route via R2-R1 link. So
> > essentially R2 is the primary router.
> >
> > On 9/12/07, Champika G <693455@gmail.com> wrote:
> > > THe drawing was not very clear.
> > > THe connections are;
> > >
> > > R1-R2 : p2p link
> > > R1-R4 : p2p link
> > > R2-R4 : p2p link
> > > R4-L2SW2 : L2 on SW
> > > R2-L2SW1 : L2 on SW
> > > L2SW2-R5 : L2 on SW
> > > L2SW1-L2SW2 : Trunk
> > > L2SW1-R3 : L2 on SW
> > > R3-R5 : p2p link
> > >
> > >
> > >
> > > On 9/12/07, Champika G <693455@gmail.com> wrote:
> > > > Hi Guys
> > > > Yes it works that way. I was trying a distribute list out in R2. Silly
> > > > me. Should have known better!!!!
> > > > Actually my original scenario is as below. I was trying to achieve
> > > > fault tolerent routing in R3 & R5.
> > > > Two new routers R4 & R5 are there seperated by a L2 switch. Switch to
> > > > switch is a trunk.
> > > > Area 1 is still a Totally stub area.
> > > > Conditions are;
> > > > (1) R3 is the HSRP active for ethernet segment vlans
> > > > (2) Incase of a link failure bet. R2-L2SW the traffic should be routed
> > > > R3->L2SW1->L2SW2->R4->R2->R1. The return path should be the reverse.
> > > > No asymetrical routing is allowed.
> > > > (3) If R1-R2 link is down traffic should be rerouted via
> > > > R3->L2SW1->R2->R4->R1. Return path should be the reverse. Again no
> > > > asymetric routing.
> > > > (4) R1 is the ABR.
> > > >
> > > > The question is if (2) or (3) happens the return path does not follow
> > > > the same outgoing path. How do i achieve that? Any thoughts
> > > >
> > > > ========(area 1)R4(area 1)==L2SW2==(area 1)R5-Vlans
> > > > || || ||
> > > > ||
> > > > ==(Area 0)R1(area 1)==(area 1)R2(area 1)==L2SW1==(area 1)R3-vlans
> > > >
> > > >
> > > > On 9/12/07, Champika G <693455@gmail.com> wrote:
> > > > > Dear All
> > > > >
> > > > > I have a situation where i want to stop the default route from
> > > > > propagating down to R3.
> > > > > R1 is the ABR and area 1 is a totally stub. R1 (ABR) injects a default
> > > > > to Area 1 and this default i can see in R3 as well.
> > > > > I want to stop R3 recieveing this default. ANy ideas how this can be
> > > > > done. I tried several methods but they did not work.
> > > > >
> > > > > OSPF
> > > > >
> > > > > ==(Area 0)R1(area 1)==(area 1)R2(area 1)==(area 1)R3
> > > > >
> > > > > Regards
> > > > > Champ
This archive was generated by hypermail 2.1.4 : Sat Oct 06 2007 - 12:01:11 ART