Cisco Support Community
cancel
Showing results for 
Search instead for 
Did you mean: 
Announcements

Welcome to Cisco Support Community. We would love to have your feedback.

For an introduction to the new site, click here. If you'd prefer to explore, try our test area to get started. And see here for current known issues.

New Member

Multicast with source redundancy - Stream disrupted

I'm currently building a multicast infrastructure for IPTV and need to provide redundancy with my source.

In my current prototype system I'm using a Cisco 3560-IPBASE-M (SMI) v12.2(25).

Here is the layout:

Source 1 (VLAN X)

\

Cisco 3560 - Set Top Box (VLAN X)

/

Source 2 (VLAN Y)

So Source 1 is connected to the 3560 on VLAN X, Source 2 is connected to the 3560 on VLAN Y and the Set Top Box is connected on VLAN X.

Source 1 and Source 2 do not have the same IP address.

Both sources are sending a multicast stream to 225.1.49.50. The Set Top Box is subscribed to this multicast group and is receiving the stream from Source 1.

When there is a problem with Source 1, I change its VLAN to VLAN Z, and change Source 2's VLAN to VLAN X.

So the Set Top Box is supposed to pick up the stream from source 2.

Going from Source 1 to Source 2 is OK. The Set Top Box is still able to pickup the stream but going from Source 2 to Source 1 (and changing the VLAN accordingly) leads to a 30 seconds (precisely and always 30 seconds) period during which the stream is not received by the Set Top Box.

I've made no other modification to the Cisco configuration (except configuration and modifying VLANs).

What could go wrong?

I've tried to statically configure the 3 interfaces to be member of the multicast groups:

ip igmp snooping vlan 124 static 225.1.49.50 interface Fa0/44 (45 and 46)

but this does not change anything.

I've also tried to give both sources the same IP address without success either. I'm using IGMP v2, not v3, so having different IP addresses should not be a problem.

Thanks for your help!

Herve

  • LAN Switching and Routing
3 REPLIES
Silver

Re: Multicast with source redundancy - Stream disrupted

May be the reson is RPF failure . Multicast routing won't forward a packet unless the source of the packet passes a Reverse Path Forwarding (RPF) check.

So use the following commands verify that :

show ip pim neighbor

show ip mroute

New Member

Re: Multicast with source redundancy - Stream disrupted

Hi aghaznavi,

That's one of the thing I'm thinking about because the source IP address changes.

Currently, I'm running an SMI image on the Cisco. I'm currently in the process of upgrading to EMI (still waiting for this CD....).

Without EMI, I do not have access to show ip pim neighbour or show ip mroute.

Thanks

New Member

Re: Multicast with source redundancy - Stream disrupted

Just to let you know that I found the problem.

When you connect a cable on the switch, there is a 30 seconds period (in fact 15+15) during which the device that you connected is not reachable through the network. That is due to the spanning-tree protocol that is first in listening mode (15s) then in learning mode (15s) before going to a forwarding state in order not to create forwarding loops on the network.

The same thing happens when you change the VLAN of a port.

In order to eliminate this problem, you just need to declare the port as portfast (spanning-tree portfast). The STP will not go through the listening and learning state but directly to the forwarding state when you either connect a new device on the port or you change the VLAN of an already connected device.

Hope it can help other people who would have had the same problem.

Herve Dubreuil

166
Views
4
Helpful
3
Replies
This widget could not be displayed.