Hi all, I have a problem with multicasting on my Brocade/Foundry MLX core (yes yes, I know I'm - so far - on the wrong maillist).
It's a very simple PIM-SM on top of an OSPF topology. The problem is that after a while of running multicast fine, it suddenly "freezes" channels, leaving only the *,G entries in the mcache list - as opposed to both S,G and *,G. I contacted Brocade and after a few efficient escalations, they've concluded that the problem is that the downstream neighbour, doesn't send the right S,G join at some point. In the end the MLX doesn't know the S,G entry and ends up not forwarding the traffic and ends up in rather high CPU utilization on the LP. Since Junipers are the only thing I have beyond the MLX's, they couldn't help me much further, which is fair enough. So, my questions are: Is there problems with the way Junos sends the S,G upstream? Sounds a bit unlikely... Is there any knobs I can turn to change the behaviour in this area? Has anyone had the same issues, possibly including a Juniper/Brocade setup? Cheers, Bjørn _______________________________________________ juniper-nsp mailing list [email protected] https://puck.nether.net/mailman/listinfo/juniper-nsp
