[Xorp-users] New to this multcast - Can this work with xorp?

Daniel Paquet daniel.paquet at gmail.com
Wed Oct 31 10:29:31 PDT 2007


show pim neighbors :

Interface    DRpriority NeighborAddr    V Mode   Holdtime Timeout
tun0                  1 10.254.0.13     2 Sparse      105     100



> I think that my setup it okay now and that the problem lies out of my
> linux
> > box. But this part is not under my control.
> >
> > The setup is for ip phones. The phones are on the same network as the
> eth0
> > network card, and my tunnel is linked to a distant 2811 cisco router.
> >
> > The RP is a single machine (10.254.0.1).
> >
> > My local eth0 ip is 10.132.3.20
> > The ip of the tunnel is : 10.254.0.14
> > and the ip of the remote machine on the other end of the tunnel is :
> > 10.254.0.13
> >
> > The thing now is that my linux box do not join any multicast group...
>
> Just to clarify: the receivers are on the eth0 subnet, but are NOT
> running on the Linux box.


Yes the recievers are on eth0 and they are cisco IP phones 7911.

If you run tcpdump on the eth0 interface do you see any IGMP
> membership messages from the receivers (in respond to the Query
> messages from the XORP box)?
>
> tcpdump -i eth0 -n -vvv -s 0 -x proto \\igmp


When running tcpdump I only see igmp query and report from my own box.

If you can see the membership messages from the receivers, a common
> problem is having firewall rules installed in the kernel which
> filter-out the IGMP messages. As a starter, for testing purpose it
> is best to remove all firewall rules.


No firewall rules are set and all chains are set to ACCEPT

> Here is the output of show igmp group
> >
> > Interface    Group           Source          LastReported Timeout V
> State
> > eth0         224.0.0.2       0.0.0.0         10.132.3.20      206 2
> E
> > eth0         224.0.0.13      0.0.0.0         10.132.3.20      205 2
> E
> > eth0         224.0.0.22      0.0.0.0         10.132.3.20      199 2
> E
> > tun0         224.0.0.2       0.0.0.0         10.254.0.14      201 2
> E
> > tun0         224.0.0.13      0.0.0.0         10.254.0.14      203 2
> E
> > tun0         224.0.0.22      0.0.0.0         10.254.0.14      205 2
> E
> >
> >
> > In fact I see that the only machines that answers the IGMP are my local
> > mathines IF
> > I have igmp enabled on both interface (tunnel and eth0)
> >
> > A show pim join dont output anyting I have to send a show pim join all
> and
> > now I get.
> >
> > Group           Source          RP              Flags
> > 224.0.0.0       10.254.0.1      10.254.0.1      RP
> >     Upstream interface (RP):   tun0
> >     Upstream MRIB next hop (RP): UNKNOWN
> >     Upstream state:            NotJoined
> >     Join timer:                -1
> >     Joins RP:                  ...
> >     Join state:                ...
> >     Prune state:               ...
> >     Prune pending state:       ...
> >     Could assert WC:           ...
> >     I am DR:                   OOO
> >     Immediate olist RP:        ...
> >     Inherited olist SG:        ...
> >     Inherited olist SG_RPT:    ...
>
> Yes, if there are no receivers, then in your setup "show pim join"
> won't show any routing entries.
> However, from the above entry I see another problem: the "Upstream
> MRIB next hop (RP)" is unknown. The interface is known, so most
> likely you don't see any PIM Hello messages from the other side the
> tunnel. If PIM-SM is running on the other side of the tunnel, what
> is the output of "show pim neighbors".


This is the output of show pim neighbors :

Interface    DRpriority NeighborAddr    V Mode   Holdtime Timeout
tun0                  1 10.254.0.13     2 Sparse      105     100


> And I read that PIM-SM works with IGMP so am I right when I say that it
> > can't work with anyting else like RIP and OSPF?
>
> IGMP has different purpose: it is to discover directly connected
> receivers and to tell PIM-SM about them.
> The purpose of RIP and OSPF is to tell PIM-SM about the Reverse Path
> Forwarding information. E.g., where to send the PIM Join messages
> towards, say, the RP or the Sender itself.


So since I have only 1 output route I guess that I dont need RIP or OSPF.

Regards,

--
Daniel Paquet
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.ICSI.Berkeley.EDU/pipermail/xorp-users/attachments/20071031/af30077c/attachment.html 


More information about the Xorp-users mailing list