Hello redditors, I have this situation and seeking wisdom from you all.
I've got an ASR1004 that serves as an edge for IPv6 then I have routers from other brands as endpoints, they are connected via an MPLS backbone, those other routers don't support 6PE, 6VPE or anything like this, but they do support L2VPNs.
This is the configuration in the ASR:
pseudowire-class VPLS_IPV6_OVERLAY encapsulation mpls control-word ! l2 vfi VPLS_IPV6_OVERLAY manual vpn id 2000 bridge-domain 2000 neighbor 10.255.0.9 pw-class VPLS_IPV6_OVERLAY neighbor 10.255.0.6 pw-class VPLS_IPV6_OVERLAY neighbor 10.255.0.7 pw-class VPLS_IPV6_OVERLAY ! interface BDI2000 description IPV6 OVERLAY no ip address ipv6 address FD23:AFBC:CFFF::1/64 ipv6 enable end
Here's the situation:
- If there's a single neighbor active (say 10.255.0.9) I can ping that neighbor via IPv6 using the tunnel IPs just fine, no problem whatsoever (say fd23:afbc:cfff::100 can ping fd23:afbc:cfff::1 and viceversa)
- As soon I enable the other neighbors IPv6 only works towards one of the neighbors
- There's no issue with the l2vpn vc, all the routers report the tunnels up/active with the right label ids
- What I am seeing is that ND messages seems to fail on all the routers but one
- It really looks like the ASR is only replicating the ND message on a single tunnel and this is randomly selected, sometimes 10.255.0.6 works fine while 10.255.0.7 and 10.255.0.9 stop working. Other times 10.255.0.7 work fine while the other two don't.
- Say 10.255.0.6 sends an ND message asking for the ASR's IPv6 address, the ASR replies to those messages only to 10.255.0.9 and doesn't flood it to all the pseudo wires (at least this is all I can think by judging the behaviour).
So I am guessing there's something I am missing here, perhaps anykind of multicast/flooding support I have not enabled?
Thank you all for the help in advance
No comments:
Post a Comment