Interconnect different fabricpath networks via CE

Question

Deer Fabricpath Experts
Hi,
I have a problem that derived in Fabricpath and STP.
Please tell me the way to resolve this issue.
Please see below, FP and CE Diagram.
               +--------------------+                  +---------------------+                +---------------------+
               |   FP Network  |                   |  CE Network   |                 |  FP Network   |
Vlan A --- |    Production   | -- Vlan A --  |  Catalyst        | -- Vlan A -- |  Devlopment   | -- Vlan A
               |   Nexus5K/7K |                   |                      |                 | Nexus5K/7K   |
               +---------------------+                 +---------------------+                +---------------------+
Vlan A is a FP Vlan, so that it is need to transit fabricpath network.
FP Network Production and  Development are separated by policy,  for the purpose of no affect convergence each other.
I want to have the same segment Vlan A over the different fabricpath network.
This Vlan is shared of Production / Devlopment FP Network and CE Network.
But, Fabricpath have some limitations,  can not interconnected via CE Network.
FP Network is looked like a root bridge. and if CE port receive superior BPDU, go down blocking state.
As a result, Nexus5K/7K port facing the Catalyst port,  go down same.. blocking state.
This behavior is the normal fabricpath operation.
Look below.
Layer 2 Gateway Spanning Tree Protocol (L2G-STP) builds a loop-free tree topology. However, it has some limitations. One limitation is that the STP root must always (virtually) be in the FabricPath cloud.
For example, it is not possible to have two FabricPath networks connected through a CE cloud. A bridge ID for STP consists of a MAC address and bridge priority. When running in FabricPath mode, the system automatically assigns the edge devices with the MAC address c84c.75fa.6000 from a pool of reserved MAC addresses. As a result, each device has the same MAC address used for the Bridge ID.
http://www.cisco.com/en/US/docs/switches/datacenter/nexus5000/sw/mkt_ops_guides/513_n1_1/n5k_ops_fabricpath.html
But I am prohibited to combine Procuction and Devlopment FP Network directory, using fabricpath adjacency.
I have some ideas, but I don't know right or not.
a.  Apply L2 ACL like a dropping BPDUs to CE port facing Catalyst. If CE port can't receive superior BPDUs , no blocking.
b.  Disable STP of Vlan A in Nexus5K/7K Leaf. If STP can't work, no BPDUs,  no blocking.
Would you please help me.
Regards,
Mot

Answer 1

Carrier sets weight and passes us Sing-primary, Tokyo-Secondary and London-Teriary
We see the route come in with a community of _65000 from Sing, _65001 from Tokyo and _65002 from London
We then could have a route map as follows to set local-pref or med in
Route-map OVERLAP permit 10
match community Sing
set MED or local-pref
Route-map OVERLAP permit 20
match community Tokyo
set MED or local-pref
Route-map OVERLAP permit 30
match community London
set MED or local-pref
Route-map OVERLAP permit
Local-pref would be prefered rather than MED as to determined the exit point. Those the lose rules.
is this the best way, or any other ideas?