multicast across branch sites

[san.carlos]

Hello there.... looking to see if anyone could shine a light on what could be causing the problem that I am currently having with multicasting not working across two branch sites.

multicast routing enabled

igmp v2 enabled

multicast source = Branch1

multicasting at Branch1 works well for clients located at Branch1. multicast from Branch1 to clients at Branch2 does not work. Branches are connected via a Metro Ethernet PTP link through the provider. Provider confirms that multicast traffic is allowed and routable. Multicasting is enabled at Branch2's core router and PIM enabled in sparse mode on vlan interface where clients reside.

Performed a packet trace on a client located at Branch2. On the trace, I can see that the client does attempt to join the correct group, but never receives a response from the source. I notice the following:

The TTL on this IGMP packet is set to 1, which will be theoretically dropped before reaching the source as it is 4 hops away - is this my problem and is there any way to increase the TTL on an IGMP packet or how does one remedy this?

A ping test from a client located at Branch1 to the multicast source is successful. A ping test from a client located at Branch2 to the multicast source fails. A traceroute from the client located at Branch2 to the multicast source does not go beyond the first hop. Not sure I understand the reasonning for this failure.

If anyone could share a thought or two on what could be the problem, it would be greatly appreciated.

Thanks -

[Nandan Mathure]

Hi!

1] How is your PIM configured? is it PIM sparse mode, PIM sparse dense mode, etc.

2] If its sparse mode, then is the RP information reaching Branch 2 like BSR, Static RP, auto rp etc?

3] If no then check for RPF failures.

Thanks,

Nandan Mathure

[san.carlos]

Nandan,

1) Using pim sparse mode

2) Doesnt' appear that the informaiton is coming across. A "sh ip pim rp" at Branch2 shows no relevant Groups-RP mappings. Same command at Branch1 shows active Groups-RP mappings. So I am assuming that I am not receiving any good information. Is there a better command to verify?

3) Ok I will check on RPF failures and post back.

Question:

Can this be remedy by adding a static multicast route (group/rp) at Branch2 for each multicast source located at Branch1?

Thanks -

[Nandan Mathure]

1] Yes you can add multiple mroutes or  you can try using default static multicast if all sources are in Branch 1,

ip mroute 0.0.0.0 255.255.255.255 10.10.10.1 

Here 10.10.10.1 will be next-hop IP address from Branch 2.

2] show ip rpf

3] You can use default mroute or multicast bgp to remedy this.

4] if that doesnt help try debug to check whats happening like "debug ip mpacket detail" "debug ip pim bsr" etc.

Let me know if this worked. Or we can check further.
Thanks,
Nandan Mathure

[san.carlos]

Nandan,

Adding static mroutes do not seem to help resolve this. Debug output on the Branch2 core router of multicast traffic destine to the source located in Branch1 provides an "RPF lookup failed for source or RP".

I found that between the two Branch's core routers, there is a switch running an IPBase version of IOS with routing enabled, but does not support ip multicast-routing. In other words, Branch2 is connected to Branch1 through this switch.

Will be removing this switch and hope that this helps.

Thoughts?

[san.carlos]

Nandan,

After removing the switch that was not routing multicast traffic, multicast related information started to propogate across to Branch2. I appreciate all of your assistance. Your recommendation for debugging really help me identify the issue here. There was no RP information coming across.

Thanks -