Being a bit of a multicast newbie, this issue has been stumping me for a while now.
We have a client that has an IPTV setup over an internal network that we manage, multicast from one box to many TVs.
Every 4 and a half minutes, it will drop connection to the stream for about a second.
We have a distribution switch (CRS326-24S+2Q+), which connects to 2 switches at each area (4 areas, each with a CRS354 and CRS328) which then connects to the TVs, among other things.
The IPTV multicast sender is connected to a 4011 off to the side (Have had it straight into the CRS326), which connects straight to the CRS326
All connected on a vlan between the switches on a single subnet, with the CRS326 as the multicast querier.
Everything is running 6.48, except the CRS326 at 6.48.1.
When the dropoffs happen, nothing changes on the MDB. But it’s always the same time from when you start watching a channel/stream. With the time between (Seems to be between 4 minutes 20 and 4 minutes 30), it made me look at the membership interval on the multicast querier, but I have increased that without any changes to the issue.
Any ideas where the issue are coming from will be greatly appreciated.
I’m seeing the same thing now on a CRS520 (7.22.1) and CRS518 (7.20.8). Backstory - I have a multicast source feeding into a CRS520, its bridge is set to “permanent” router but it is NOT the querier. When I use a laptop to view the multicast stream while connected to the CRS520, it will stream for hours without a problem.
I have another CRS520 connected to the multicast source CRS520 (not MLAG yet, poor man’s switch stacking with LAGs) and a CRS518 also connected to the multicast source CRS520 that IS the querier.
I am also running PIM on the multicast source CRS520 as that is my SVI for about 25 VLANs, and PIM is routing from the VLAN where the encoder lives to the SVI VLANs.
When I have a laptop connected to the CRS518, it will pull a stream for exactly 4 minutes 20 seconds - the membership interval - and the stream drops.
I had this working but I had “allow unknown multicast” enabled just about everywhere, as a result the CPU would be spiked for hours when anyone would pull a multicast stream. So I’m eliminating as much “allow unknown multicast” as I can and disabling as many multicast routers on bridge ports as I can to get the CPU usage under control.
Is this interconnect VLAN tagged and are the IGMP queries sent by the CRS326 supposed to be sent tagged? According to the docs the ROS querier cannot send tagged queries:
Only untagged IGMP/MLD general membership queries are generated, IGMP queries are sent with IPv4 0.0.0.0 source address, MLD queries are sent with IPv6 link-local address of the bridge interface. The bridge will not send queries if an external IGMP/MLD querier is detected (see the monitoring values igmp-querier and mld-querier).