Cisco Support Community
Community Member

Switches Flooding Multicast Streams but IGMP Enabled ?

Hi ,

          I have a 2x Cisco 3750G switches (Stacked) that are part of an IP Video Surveillance System .

All network is set to 1 VLAN (VLAN 1 default).

The Cisco SWs ( Core )  are configured with IGMP Querier + Snooping .

Connected to the Cisco SWs are 4 Fiber Rings(Loops)  for redundancy .

RSTP and IGMP Snooping are enabled on all the network SWs .

Network Architecture Figure :


When all Fiber Rings (Loops ) are connected , the Cisco and Ring SWs spanning-tree tables show the blocked and forwarding ports properly .

I am able to connect all my IP Cams + Encoders to the Ring SWs and connect my Servers + WorkStations to the Cisco SWs to record and view the multicast ip streams .

All streams from Rings 1 ,3 and 4 are getting normally to the Servers + WorkStations through multicast ( IGMP Snooping are filled correctly on the Cisco and non-Cisco SWs ) . All multicast streams are joined normally except for Ring 2 .

The Multicast streams passing through Ring 2 will operate normally for around 2 hours then after that the Fiber SWs begin to flood their multicast traffic causing the cams on the same switch to drop off the network .

I checked the Configuration of all the SWs on Ring 2 but all are are the same .

What is weird is that I have to set all the streams to unicast to stop the flooding just on Ring 2 (All other rings are not affected ) then after some time if I enabled a video multicast stream on one of the SWs of the Ring it will be streamed properly to the Servers + WorkStations ( All IGMP tables along the way will be filled properly ) then again after 2 hours or so , the flooding will start again suddenly and all IGMP  table entries for the SW on Ring2 will be empty . No problem occurs on the other Rings which have more multicast streams .

What might be wrong ?

Any ideas ?

Everyone's tags (4)
CreatePlease to create content