Hello Experts !
We have analyzed and worked through a customer scenario where the following situation was given and did not work properly:
3 vrfs on various PEs where interconnected by matching imp and exp Rts (so they are ONE core unicast domain)
Mc sources, MC receivers and RPs are distributed over those 3 vrfs
each vrf has its own MDT Default and MDT Data setup (so in MC we have 3 differnt core multicast domains )
so because of various problems we found out that:
by using MVPN Extranet (implicitely) the following problems arose:
the follwing scenarios DID not work:
•More than 2 vrfs
•2 vrfs and local RP in one PE
•Crossing the MVPN core 2 times
+++core+++
PE2 (vrf GREEN link to[RP])
+++core+++
PE3 (vrf GREEN [rec])
so as the unicast vrfs are really ONE domain we tried to also use one multicast domain by trying to set the same MDT Default and Data group adresses on all the 3 vrfs.
BUT when committing we failed as IOS XR told us that we are not allowed to use the same MDT default and MDT data tree on more than one vrf
so for me there are 2 possible reasons for that
1.) the programmer wanted to help the user from accidentially selecting the same MC tree for 2 vrf s
[ in that case the restriction would be easy to remove, or reduces to a warning]
OR
2.) the MDT Default group address is a kind of unique pointer inside the MTI and thus no 2 vrf s can have the same (kind of same design fault which was the RD on IOS)
[ in that case the restrictions would need heavy reprogramming to remove]
Maybe someone more knowledgeable could help me clearing that out
By the way, we know that SSM would solve that problem, but.... there are reasons to not use SSM in this customer scenario
with best regards
Alexander Marhold
