Routing

Hola amigo's,

I have a customer who is eBGP dual-homed to me over IPv6 for transit. He has 2 routers and I have 1 that supports both sessions for the moment. We are directly connected over a VLAN that also supports his IPv4 eBGP connectivity in the same manner.

He can ping me from both of his redundant routers. I can only ping one of his addresses (all in the same subnet I THINK) unless I specify the outgoing interface for one when I ping.  Recall, It's DC over a VLAN.

I see him in "sh ipv6 neigh" as reachable.

The eBGP peers are DC but I might use other IP addressing (loopbacks etc.) on each end so that we have multi-hop with TTL10 or something.

Not sure what to think, sessions stuck in ACTIVE on both sides.

Any ideas?

Hi,

Normally, having to specify the outgoing interface suggests that the destination may be a link local address (FE80::).  That might not be usable for a BGP session because there is no way to specify which interface the BGP messages should use.

If you're using globally unique addresses, then that's obviously not the case.

Can you provide the configuration excerpts so we can see what's happening?  Have you tried using the same BGP session to exchange both IPv4 and IPv6 NLRI?

Rgds,

Guy

Please excuse me if it takes a while to answer any responses.  Might be away from my desk for a while. 

Hi,

It was a control plane firewall filter that permitted only specific IPv6 end-points for tcp/179. The IPv4 filter permitted any.

Still getting used to this Juniper gear but I love the logic even if the configuration is a bit long-winded.

Thanks to Guy for the reply.