Hi All, we are deploying multicasting in my network over wan & LAN. now we are confused about multicast redundancy for same groups. We have a DR site which is replica of main site with same hardware & configuration. we have deployed sparse-dense mode with static RP. Its working fine but if the main site goes down multicast should work properly from DR site where we have same multicast source as main site with the same group IP.
Any help would be gr8.
As long as you have a route to RP, you will converge. I am assuming you are sending from one source only.
RP needs to be redudant also, MSDP is a good option here.
PS: I have stayed away from talking about switch from shared tree to source tree assuming yoy wish to keep shared tree distribution.
thanks for your reply...I am using 3750 switch everywhere. & only one source group will be activated at the same time. Now we have two 3845 running at main site with HSRP. One of them acting as a multicast RP. There are two more routers on DR site with HSRP & there are clients also in the DR site, so when main site is live DR site clients will have stream from main site servers & if Main site goes down then every client will have stream from DR site sources...
how can I achieve this. I think MSDP heavily rely on BGP & MBGP which we dont have in our network.
Any other way we can achieve this. there are two more site across wan even crossing multiple ISPs..we are providing mcast over GRE tunnel...so for those clients also we have to take care of redundancy...
U can get MSDP to work with no BGP. all what u need is a peer session (TCP) between 2 routers acting as RP and using anycast address.
Hi, Thanks for your reply. I have gone thru the document for MSDP but still have feeling that for RPF check it requires bgp next-hope to configured & for bgp next-hope we need to configure BGP. Well I will try MSDP with mesh group & will check whether it works or not..
still want confirmation from you...if u hv implemented somewhere without BGP...
i m planning to put these commands on Routers which will act as RPs. by the way on 3750 switches where clients are connected wht would be the config... just pim sparse-dense mode or i will hv to configure MSDP here as well...
need your expert comment...
ip pim rp-address 10.100.1.4
ip msdp peer 10.100.1.2 connect-source lo0
ip msdp peer description 10.100.1.2 to RTR02
ip msdp peer 10.100.1.3 connect-source lo0
ip msdp peer description 10.100.1.3 to zRTR01
ip msdp peer 10.100.1.1 connect-source lo0
ip msdp peer description 10.100.1.1 to RTR01
ip msdp mesh-group xxX 10.100.1.2
ip msdp mesh-group xxX 10.100.1.3
ip msdp mesh-group xxX 10.100.1.1
I have configured MSDP based on IGP only (IS-IS in this case).
I have not used mesh groups, so kept at as simple as can be.
I recommend that if possible you make ur changes within a service window which would allow you to fine tune and test.
ip msdp peer x.x.x.x connect-source Loopback10 remote-as 65001
ip msdp description YYYYYY
ip msdp cache-sa-state
Hi, I tried msdp peer but it didnt work for me... I did not define any AS number in MSDP peer command as I mentioned in my earlier msg, i didnt use mesh group. I could see that peer relationship was up but I dont know why I was not able to see traffic at DR site while source was on main site & vice versa...one thing I want to tell u is that between DR & MAIN site I 6 switches & 4 routers on which ATM links are terminated so to avoid multicast confgiuration on these intermediate devices I have created a GRE tunnel between MAIN site HSRP routers & DR site routers..I m not sure weather MSDP creates problem with GRE tunnel or not ?
NW diagram is attached for your reference.
Hi, I managed to get the MSDP peering UP & running between Primary Site router & DR site router over GRE. I can see msdp cache on DR site router & (*,G) & (S,G) entry on DR site router but still DR site client are not able to get multicast stream from Sources server which is in primary site. clients which are on primary site they r getting stream perfectly.. but I m not getting y clients in DR sites are not getting stream...when I say ip mroute on DR site client switch I only can see (*,G) entry not (S,G) entry...but on DR router which is directly connected on switch has both entry in it...I am attaching the config for both the routers & switch config where I am not getting
stream...also attaching the diagram of network..please help me to config ANYCAST so that I can achieve redundancy..
RPF check fails for mcast packets received by DR client switch from DR primary router as the PIM neighbor is 10.200.3.11 (real address) and the route back to the source is via 10.200.3.9 (HSRP address). You would need to configure an mroute static route to fix this.