[j-nsp] SRX240 Multicast

[Olivier Benghozi]

Hi Keith,

some suggestions (from the elements you gave):

1) remove your igmp config (but not the igmp-snooping config) for vlans where you configure PIM. It's useless here. Igmp querying is automatically assumed where PIM is configured.

2) remove interface ge-0/0/8.0 from the PIM config. It's not a L3 interface.

3) Nothing here suggests neither the use of PIM SSM nor the presence of a rendez-vous point.
So, you problem is that you are trying to route multicast in PIM Sparse Mode without a Rendez-Vous Point. The protocol expects it even on a single router doing everything and I don't remember that Juniper did anything special here.

Conclusion: you must configure the SRX as a rendez-vous point. For that, the best way is usually (to me at least) to have a loopback (in the proper zone, it's an SRX) and to use the BSR (Boot Strap Router) way to distribute the RP information across the network.

But to make it work just now, I suggest you should:
- create a /32 loopback in the same zone as vlan 100, by example
- add the following config:

[edit protocols pim]
set rp local family inet address a.b.c.d  <----- the IP of the loopback

However, for a quick and dirty test, I guess you might use the IP from the layer 3 interface vlan.100, without loopback, and it might do the trick.

You didn't precise this but I also assume that the receiver is sending IGMP Joins :)


regards,
Olivier Benghozi
Wifirst


Le 28 août 2014 à 02:07, Keith <kwoody at citywest.ca> a écrit :

> Hi.
> 
> We have been taking our first foray into multicast the last few days.
> 
> In our test bed we have an SRX240 with all the senders and one receiver
> connected.
> 
> Sender is pushing 600megs of traffic on ge-0/0/3
> 
> Receiver is on ge-0/0/8, both interfaces on vlan 100.
> 
> Receiver used to be on vlan 51, but it was not receiving *any* multicast traffic
> at all. So as a test we moved ge-0/0/8 to the same vlan as the sender to see if
> the test set could see the groups and it can.
> 
> But ge-0/0/8 is drowning in traffic from the sender. I have igmp-snooping, interfaces
> on igmp, pim enabled, but I see 640 megs of traffic input on ge-0/0/3 and 640 megs
> output on ge-0/0/8.
> 
> I have multicast routes showing up in the routing table.
> 
> Couple of show commands says that igmp is not enabled on ge-0/0/8.
> 
>>> run show igmp statistics interface ge-0/0/8
> 
> IGMP not enabled on interface ge-0/0/8.0
> 
> IGMP Global Statistics
> Bad Length                    0
> Bad Checksum                  0
> Bad Receive If                0
> Rx non-local                  0
> Timed out                     3
> Rejected Report               0
> Total Interfaces              4
> 
> I have obviously missed something in the protocols section but I have
> no idea what. Anyone lend a clue?
> 
>>> show protocols
> igmp {
>    accounting;
>    interface ge-0/0/8.0;
>    interface ge-0/0/3.0;
>    interface vlan.100;
>    interface vlan.51;
>    interface vlan.49;
> }
> inactive: ospf {
>    area 0.0.0.0 {
>        interface all;
>    }
> }
> pim {
>    inactive: traceoptions {
>        file pim-trace.log;
>        flag all;
>        flag join detail;
>        flag prune detail;
>    }
>    interface all {
>        mode sparse;
>        version 2;
>    }
>    interface vlan.51 {
>        mode sparse;
>        version 2;
>    }
>    interface vlan.49 {
>        mode sparse;
>        version 2;
>    }
>    interface vlan.100 {
>        mode sparse;
>        version 2;
>    }
>    interface ge-0/0/8.0 {
>        mode sparse;
>        version 2;
>    }
> }
> igmp-snooping {
>    vlan vlan.51;
>    vlan vlan.49;
>    vlan vlan.100;
>   }
> 
> Thanks,
> Keith