From: riaz shams (riazshams@gmail.com)
Date: Sun Oct 08 2006 - 19:16:25 ART
Thanks Victor for the answer.
but i m asking exact opposite...i m using sparse-mode.
my source ip address which is pinging to the multicast group address is
advertised in ospf and the subnet address of the interface on which
multicast join-group address has been configured, is not advertised. is that
required to be advertised ?
second, if u mean same and u r saying to advertise that subnet...then y
would u redistribute and not advertise the same subnet with network command
in ospf (igp)
many thanks,
riaz.
On 10/8/06, Victor Cappuccio <cvictor@protokolgroup.com> wrote:
>
> Hi Riaz.
>
> Well I would think that I would depend on the IP PIM Mode you are using,
> Let's assume that we are using PIM Dense Mode as the first case.
> Since PIM Dense mode would flood MCast Stream to all interfaces when
> neighbors Router are present (& if the RPF toward the Source is success)
> and
> when that Traffic reaches to neighbors that haves (*,G) Entry then it
> would
> not be necessary to know this information, but if the Neighbor is not
> directly connected (or Local) then you would have problems since the RPF
> would fail, also take in consideration that in the HUB Router you need a
> way
> to forward that multicast feed if it came to the same physical interface
> (Yes use tunnel, because you would have a OIL Problem).
>
> On the other side if are using PIM Sparse, and if you have a Source
> located
> in that unicast not reachable Network Segment by the Routing Protocol
> (assuming that the DR is the same F router in that Network Segment), then
> you would have problems, since the Register Message would arrive at the
> RP,
> but the RP would not know how to reach that Message when sending the
> Register Stop, or to send the S,G Join towards the source. So I would
> think
> that the information would arrive to Groups known by the RP, but The PIM
> SM
> Switchover would fail, and the S,G Joins would not have a Path to the
> Source
> (Yes all kinds of Problems) also take in consideration that All routers
> perform RFC Check to see if the current Mcast Data Stream Camed from the
> Correct Interface (That is how the Routers in the Path knows that they
> have
> joined the correct Source Stree, in that Jumbo Mumbo combination of Shared
> and Source trees created by PIM Sparse Mode (but yet it works)
>
> IMO, I would redistribute the information to the current IGP to have
> global
> reachability
>
> Just my 2 cents
> Victor.-
>
>
> -----Mensaje original-----
> De: nobody@groupstudy.com [mailto:nobody@groupstudy.com] En nombre de riaz
> shams
> Enviado el: Sabado, 07 de Octubre de 2006 05:28 p.m.
> Para: ccielab@groupstudy.com
> Asunto: multicast question.
>
> hi group ,
>
> can anyone pls clear my doubt on this,
>
> three routers, R1 R2 and R3 connected on frame -relay hub and spoke style.
> R2 is hub. OSPF is running as a routing protocol and R3's one interface is
> not part of ospf.
>
> i have given igmp-join group on ethernet's interface of R3 and multicast
> is
> running on all three routers.
>
> now if i ping from R1 to that multicast address , it sometime pings and
> sometimes not.
> i have also checked RPFs and RP mappings.
>
> do i need to bring that interface's subnet in routing domain for multicast
> address to be pingable ?
>
> Tia
> riaz.
>
> _______________________________________________________________________
> Subscription information may be found at:
> http://www.groupstudy.com/list/CCIELab.html
This archive was generated by hypermail 2.1.4 : Wed Nov 01 2006 - 07:29:04 ART