From: Erick B. (erickbe@xxxxxxxxx)
Date: Thu Mar 07 2002 - 02:56:25 GMT-3
Just to expand on this and get back to the original
issue... I'm not sure if the original poster tried one
of the workarounds. I guess I'm trying to make sure
they understand the actual issue here (classful vs
classless).
Anyway, on R1 (the RIP router) are you seeing a RIP
route for the 172.16.1.x network?
The reason you're able to ping from the added
loopbacks are because the loopbacks are /24 and the
subnet mask matches the subnet on the serial interface
and they get propagated to the RIP router. The
172.16.1.x/30 didn't match the serial interface with
/24 subnet so doesn't get sent. Now, if you had a
171.16.1.x/30 network for example that would get
auto-summarized to 172.16.x.x/16 and passed along
because it's a different classful subnet.
Erick
--- "Chua, Parry" <Parry.Chua@compaq.com> wrote:
> Using normal ping, the source address is the
> interface resolved to next hop in your local router
> routing table.
>
> For example, let said at R3, to reach 176.16.11.0/24
> network, next hop is 176.16.2.1 and it is the direct
> connected interface at R3, it use this as source
> address and destination
> 172.16.11.1. When is packet reach R1, for R1 to
> reply this packet, it use 172.16.11.1 as
> source address and 172.16.2.1 as destination.
> Looking at R1 routing table, how to reach
> network 172.16.2.0 and if it cannot find one,
> drop....
>
> If it can find one but go to wrong place, drop....
>
> Hope this help you..
>
> Parry Chua
>
> -----Original Message-----
> From: Jeongwoo Park [mailto:jpark@wams.com]
> Sent: Thursday, March 07, 2002 11:42 AM
> To: Chua, Parry; Mohamed I. Lubbad;
> ccielab@groupstudy.com
> Subject: RE: It is in the routing table, but Can't
> ping it
>
>
> Wow
> Parry
> It is weird.
> Normal ping was not getting through, but extended
> ping went through. I got
> response back.
> Now, why? why? why?
> I didn't set ip ospf network type point to point.
> I didn't put another loopback interface on any of
> routers.
> Only thing that i did was purely extended ping from
> R3 to R1 (sourced from
> loopback address on R3)
> Previously I had a problem with normal pinging from
> R3 to R1, because it
> didn't going through even though the routes were in
> the routing table.
> If you could explain this, it would be great.
>
> Now, why did you want me to set ip ospf network type
> point to point? what
> difference could it have made?
>
> Regarding redistribution, I was able to redistribute
> 172.16.1.0/30 network
> by simple summary on redistribution router (R2).
> I did not use trick by adding extra area nor extra
> ospf process.
> I was wondering in what case your tricky way would
> help redistributing.
>
> Thanks very much.
>
> JP
>
> -----Original Message-----
> From: Chua, Parry [mailto:Parry.Chua@compaq.com]
> Sent: Wednesday, March 06, 2002 5:35 PM
> To: Jeongwoo Park; Mohamed I. Lubbad;
> ccielab@groupstudy.com
> Subject: RE: It is in the routing table, but Can't
> ping it
>
>
> Hi JP,
>
> Have you receive my respond ?
>
> I suggest you to do some test :
>
> 1. create a loopback int at R3 said 172.16.3.3
> 255.255.255.0 and adverise it
> in OSPF.
> set the ip ospf network type point-to-point
> 2. Create a loopback interface at R1 said
> 172.16.11.1 255.255.255.0
>
> 3. Routing table, R1 should have 172.16.3.0 and at
> R3 should have
> 172.16.11.0
>
> 4. Do a extended ping from R1 to 172.16.3.3 using
> 172.16.11.1 as R1 source
> address
> 5. Do a extended ping from R3 to 172.16.11.1 using
> 172.16.3.3 as R3 source
> address.
>
> Below is my last reply to you, any comment ?
>
> "I guess Jeonqwoo has some problem with this setup,
> let me try to explain
> and see it help.
>
> OSPF
> ====
> Shold run on R2 and R3, this has not problem except
> it has different mask as
> compare to RIP
>
> RIP
> ====
> should run on R1 and R2 but RIP is a classful
> protocol so it will enable on
> all interface
> in R2. In clude the serial link between R2 amd R3,
> this should be in OSPF
> only, but no way to exclude fro RIP. But R1 will not
> see this network due to
> the mask different.
>
> If you want R1 to see the network 172.16.1.0/30
> without redistribution,
> create a secondary address
> between R1 and R2 with /30 mask. Or create a tunnel
> between R1 and R2 with
> /30 mask will do the same.
>
>
> Mutual redistribution at R2
> ===========================
> When you redistbute RIP to OSPF,at least you should
> block network
> 172.16.1.0/30, it belong to OSPF.
>
> If you want 172.16.1.0/30 to redistribute from OSPF
> into RIP as /24, then I
> know two ways :
> 1. At R2, create a dummay area, R2 become ABR, do
> area 0 range 172.16.1.0
> 255.255.255.0
> 2. Create another OSPF process and redistrbute
> 172.17.1.0/30 to it and
> summary address at this
> new OSPF process., At R2 RIP, redistribute this
> OSPF process to RIP"
> > Parry Chua
> >"
>
>
> Parry Chua
>
This archive was generated by hypermail 2.1.4 : Thu Jun 13 2002 - 10:56:55 GMT-3