[c-nsp] bgp route-reflector

[Mark Kent]

I'm posting this for someone who isn't familiar with this list.
He just turned off "soft-reconfig" on his bgp sessions and now his
route-reflectors appear broken.  Before telling him to reload,
I wanted to make sure there isn't a bit of magic I've forgotten:

The setting:

{Internet}<-->A<-->B<-->C<-->{Internet}
 AS1             AS34567        AS2

A,B,C are 7206vxr.   The configs are simple:

Router A:

router bgp 34567
 bgp cluster-id 300
 nei x.y.z.77 remote-as 34567

Router B:
router bgp 34567
 bgp cluster-id 300
 nei x.y.z.49 remote-as 34567
 nei x.y.z.49 route-reflector-client
 nei x.y.z.78 remote-as 34567
 nei x.y.z.78 route-reflector-client

Router C:

router bgp 34567
 bgp cluster-id 300
 nei x.y.z.50 remote-as 34567

The router in the middle does have a bunch-o'routes from both AS1 and
AS2.  If I do "sh ip bg n x.y.z.78 ad" then it shows plenty of routes
being advertized, but on the edge routers themselves, we see:

Neighbor V    AS MsgRcvd MsgSent   TblVer  InQ OutQ Up/Down  State/PfxRcd
x.y.z.77 4 34567 8241501 13054604 34176418    0    0 00:34:35        0

Zero prefixes...  If we clear the bgp session then we see the InQ go
non-zero with each "sh ip bg su" we do but the end result is always 0
prefix received.  It's the same situation on both router A and C.

next-hops are known on each router (via ospf).

Versions in use are 12.3(15), 12.2(8)T10 and 12.0(25)S1 (a,b,c).

So, reload time, or a simple oversight in rr config?
[And, Yes, I am suggesting upgrading the IOS before any reload]

Thanks,
-mark