x670v stacking with RP multicast join over backbone

[Erwin_van_Hoof]

At a client we are using 3 stacks with 3x x670v with XOS 15.4.3 configured with native stacking in V160. Between the 3 stacks we use a 40Gb connection (QSP+ with LACP L2).

Each stack makes use of a rendez-vous point.

rendez-vous point configuration lines:
configure pim crp vlan "Local_1" "rp-list1" 30
configure pim cbsr vlan "Local_1"
enable pimrp-list of the first stack:
entry extreme_rp1 { if match any { } then { nlri 239.1.13.0/24 ; nlri 239.1.14.0/24 ; nlri 239.1.15.0/24 ; nlri 239.1.16.0/24 ; nlri 239.249.0.0/24 ; nlri 239.255.255.0/24 ; nlri 224.2.127.254/32 ; } }
I have the following questions:

1. Is it correct that a new multicast stream will (for a short period) flow to the master slot (switch) in a stack? If yes, what is the dispatch time period?

2. If a destination join a multicast stream this will (for a short period) flow to the master slot (switch)? If yes, what is the dispatch time period?

3. If a destination switches a multicast stream this will (for a short period) flow to the master slot (switch)? If yes, what is the dispatch time period?

4. There is much joining/dropping of streams in this setup, with a high number (40+) streams of 3,6Gb. Will this exceed the bandwidth of the backbone within a stack?

[Patrick_Voss]

Hello Erwin,

A stack should function like one so I wouldn't say the stream is flowing to the master. It will come in on the port and VLAN. All of these question should be pretty much instant. The bandwidth utilization is not one we will be able to answer. You can check the port utilization to see the percentage. This percentage is an average over 5 seconds though.