Does anyone have this combination working? I am using ROS 7.7.

It appears to be unrelated to my other issue* because it affects every router and it has never worked for me. Furthermore, all the necessary PTP, etc, routes are showing in OSPF-v3 at this time.

I have a route reflector setup for BGP and IPv4. It works as expected. I have a pretty much identical setup for IPv6 and it doesn't work at all. The BGP peers are established but no IPv6 routes are present at the RR. The route filters are a duplicate of the IPv4 setup with just the address ranges changed.

I have Googled this a few times and have only been able to find substantial stuff relating to ROS6.

Direct router to router IBGP is working as expected.

*ROS7 and OSPF-v3 viewtopic.php?t=195048

It's working for me. My connections to the RR's are dual-AFI on a single BGP connection, with a route filter to change the gateway of IPv6 routes to the IPv6 loopback of the router. Otherwise it tries to use the IPv4 literal (::ffff:x.x.x.x), which could also work if you add it as a loopback on the IPv6 side.

It's working for me. My connections to the RR's are dual-AFI on a single BGP connection, with a route filter to change the gateway of IPv6 routes to the IPv6 loopback of the router. Otherwise it tries to use the IPv4 literal (::ffff:x.x.x.x), which could also work if you add it as a loopback on the IPv6 side.

Thanks for the info. In my case its not trying to use anything - the routes simply don't show at the RR.

I have peered the RR WAN IP to the loopback of the other BGP routers in the network, except the one its directly connected to. With this I simply peered across the PTP link. In all cases it shows established sessions so its not a connectivity problem.

OSPF-v3 routes, BGP route filters and IBGP-RR / IBGP-RR-Client types appear to be all set correctly. A corresponding setup works as expected for IPv4.

You do have an explicit accept rule in the filters, correct? Does it show any routes are being filtered (/routing/route/print, look for F in front of the routes)? Do the clients show that they are advertising any routes (/routing/bgp/advertisements print where dst in 2000::/3)?

An example of the RR clients (a border router)

100 ;;; [redacted]
chain=bgp-rr-v6-in rule="if (dst in [redacted]/45 and dst-len in 45-48) { accept; }"
101 chain=bgp-rr-v6-in rule="reject;"
Explanation: Borders should accept local origin route information from RR

102 X ;;; Filter out Team Cymru blackhole routes
chain=bgp-rr-v6-out rule="if (bgp-communities includes 65332:888) { reject; }"
103 chain=bgp-rr-v6-out rule="accept;"
104 chain=bgp-rr-v6-out rule="reject;"
Explanation: Borders should send all routes to RR

At the RR

10 I ;;; BorderRouter1 and BorderRouter2 - IPv6
11 chain=bgp-br-v6-in rule="accept;"
12 chain=bgp-br-v6-in rule="reject;"
Explanation: RR should accept all incoming routes from borders

13 ;;; [redacted]
chain=bgp-br-v6-out rule="if (dst in [redacted - same as above range]/45 && dst-len in 45-48) { accept; }"
14 chain=bgp-br-v6-out rule="reject;"
Explanation: RR should send local origin route information to borders

Does it show any routes are being filtered (/routing/route/print, look for F in front of the routes)?

No BGP origin routes are showing at the RR with command: /routing/route/print where afi=ip6 and bgp

Do the clients show that they are advertising any routes (/routing/bgp/advertisements print where dst in 2000::/3)?

No entries showing for: /routing/bgp/advertisements print where dst in 2000::/3 and peer=[RR peer].
It should also be noted the client peer is set to redistribute. The client peer I am using as an example has full global BGP tables from upline suppliers so plenty of content there.

Its working now and as sometimes happens it was a small oversight.

I had not set the RR peers address families = IPv6. I know this setting well but didn't notice it unset even on multiple look overs. Once set suddenly it works as expected.

Thanks for the help

D'oh!