01-09-2016 11:08 AM - edited 03-08-2019 03:21 AM
I have a setup similar to this one:
Where "Host 1" and "Host 2" are two machines that communicate with each other using multicast\broadcast traffic. The machines are connected to each other via two switches (Nexus 5K), and the switches are connected via a port channel (physical, not virtual).
Broadcast communication between to two hosts works just fine, but when I switch to multicast, traffic is not passing.
In order to make the switch aware of the multicast group used by the hosts, the hosts send IGMP requests and responses which the switches should learn from (IGMP snooping is enabled on the switches). when I ran "show ip igmp snooping groups" i noticed that the switches learned that there's a multicast router and a multicast client behind port Eth1/1, but did not add the port channel interface to the groups (even though each switch should have heard there is IGMP requests and queries on both Eth1/1 and the port channel).
Will configuring IGMP querier on the port channel (on both switches) resolve this issue?
Does anybody have any idea if this situation can be solved without presenting new configurations to the switches? (are the any messages that the hosts can send to make the switch learn that the port-channel is also part of the multicast group?)
P.S.
When connection both machines to the same switch, multicast traffic works and the switch learns all the ports belonging to the multicast group.
Thanks in advance, and have a pleasant weekend!
Solved! Go to Solution.
01-11-2016 05:04 PM
Both the devices are in the same lan segment. If in case the hosts were connected to the single switch you wouldnt had seen any issue. As the issue arrise bcz of the host connecting between the 2 switches connected together as a trunk you are facing this issue.
This issue is already been reported on the following link:-
http://www.cisco.com/c/en/us/support/docs/switches/catalyst-6500-series-switches/68131-cat-multicast-prob.html
Please configure one of the option mentioned in the above link to resolve this issue.
HTH
REgards
Inayath
01-09-2016 12:26 PM
create a ACL.
01-09-2016 12:36 PM
Hi,
Thanks for the answer. I guess by ACL you mean Access Control List, but as I mentioned, i am trying to avoid any additional configuration on the switches.
Thanks,
Omer.
01-11-2016 05:04 PM
Both the devices are in the same lan segment. If in case the hosts were connected to the single switch you wouldnt had seen any issue. As the issue arrise bcz of the host connecting between the 2 switches connected together as a trunk you are facing this issue.
This issue is already been reported on the following link:-
http://www.cisco.com/c/en/us/support/docs/switches/catalyst-6500-series-switches/68131-cat-multicast-prob.html
Please configure one of the option mentioned in the above link to resolve this issue.
HTH
REgards
Inayath
01-11-2016 08:04 AM
Hello,
I just answered on your other post.
Yes Configuring ip igmp querier will solve this issue
Madhu
01-12-2016 01:13 AM
Thanks for the answer.
I have a little question, though: what happens in situations were both hosts are multicast server and client, for example in routers running routing protocols such as OSPF, in which multicast messages are sent and received by all hosts?
will we still have to configure the IGMP snooping querier?
Does the fact that such protocols use a well-known multicast IP addresses (224.0.0.5/6 in OSPF, for example) has any affect on the way the switches forward this traffic?
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide