My topology is working perfectly fine and is as follows:

CE1 <-> PE1 <-> P1 <-> P2 <-> PE2 <-> CE2
  • CE1 <-> PE1 and CE2 <-> PE2 are having OSPFv3 with address-family ipv4 and ipv6 both.

  • PE1 <-> PE2 is running MP-iBGP with dual-stack support

Redistribution of OSPF in MP-iBGP is as follows:

Config snippet in PE1 & PE2:

address-family ipv6 vrf custA
redistribute connected
redistribute ospf 6
no sync

address-family ipv4 vrf custA
redistribute connected
redistribute ospfv3 6
no sync

CE1 & CE2 are able to reach the IPv4 and IPv6 subnets of each other.


What is the difference in these two highlighted commands:

In ipv6 address-family it should be redistribute ospf 6 command while in ipv4 address-family it should always be redistribute ospfv3 6. If I interchange these commands within the address-families, both IPv4 & IPv6 networks of CE1 & CE2 become unreachable.

Seems to be a mystery to me. Pl let me know if any other details are required. Any clarification is appreciated.

  • OSPFv3 does not in any way support OSPFv2 since they whant to completely replace IPV4 altogether. Though people have suggested implementing support for IPV4 the most they got was dual stacking or running both IPv4 and IPv6 when needed. This is not an educated guess, read one of the switching and routing CCENT or CCNA books. Just so no one gets confused in the future.
    – user23771
    Mar 16, 2016 at 18:11
  • Did any answer help you? if so, you should accept the answer so that the question doesn't keep popping up forever, looking for an answer. Alternatively, you could provide and accept your own answer.
    – Ron Maupin
    Aug 10, 2017 at 2:50

1 Answer 1


Only OSPFv3 supports IPv6 (not OSPFv2), so in IPv6 address-family I believe they assume that 'redistribute ospf 6' means you would by default only want OSPFv3 protocol-garnered routes, because that's the only source of them and you're in the family that deals with those routes already.

Contrastly, in IPv4 address family, 'redistribute ospfv3 6' specifically says that the protocol ospfv3 should have it's ipv6 routes redistributed. It's like a reminder that you can only get them from OSPFv3, and not OSPFv2 (though you can get ipv4 routes from OSPFv3, as it was intended to not only support ipv6, but also ipv4 so it could replace OSPFv2).

That's my 'educated guess' anyway :)

  • This seems to be the correct logical explanation. However, the surprising thing is that I did not come across any such documentation after extensive search on the internet. As you know that "customer" always wants the proof of behavior and approach, could you pls share if there is any documentation available as reference?
    – Mahesh M
    Jan 19, 2014 at 6:39
  • That's a fair assumption, as there isn't the option to explicitly define ospfv3 under the ipv6 command subset.
    – Ryan Foley
    Jan 19, 2014 at 9:23

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.