bgpd: fix ipv6 nexthop-local unchanged #17037
Conversation
| @@ -0,0 +1,23 @@ | |||
| router bgp 65000 | |||
There was a problem hiding this comment.
All the peers here are iBGP. We are seeing the issue when the prefix is advertised towards an eBGP peer, still on the same network. Because they still share same network, we expect the RR to leave the NH unchanged, which it does, but it also inserts a link-local.
There was a problem hiding this comment.
In terms of debugging, if we enable BGP update debugging we can see the incoming update without a link-local and then us (RR) sending it with a link-local.
There was a problem hiding this comment.
For your use-case, you will need "neighbor XXXX nexthop-local unchanged". If unset, it will keep inserting the link-local from the route-reflector. This is what "neighbor XXXX nexthop-local unchanged" is designed for.
There was a problem hiding this comment.
That also doesn't (or didn't) work to prevent the (F)RR adding in it's link-local. Although we could live with using it this way, it is different behaviour to IPv4 case where nexthop is unchanged when all iBGP and eBGP neighbours share the same segment without any additional flags.
There was a problem hiding this comment.
ah I see the proposed change. Thank you for the solution!
This reverts commit 5f6a61f. Signed-off-by: Louis Scalbert <[email protected]>
louis-6wind
force-pushed
the
fix-ipv6-ll-nexthop
branch
from
1bd3d0f
to
ba7b1ac
Compare
louis-6wind
changed the title
Do not add an IPv6 link-local nexthop if the originating peer does not provide one and the nexthop-local unchanged setting is enabled. Signed-off-by: Louis Scalbert <[email protected]>
If the "nexthop-local unchanged" setting is enabled, it preserves the IPv6 link-local nexthop from the originating peer. However, if the originating and destination peers are not on the same network segment, the originating peer's IPv6 link-local address will be unreachable from the destination peer. In such cases, reset the IPv6 link-local nexthop, even if "nexthop-local unchanged" is set on the destination peer. Signed-off-by: Louis Scalbert <[email protected]>
Add bgp_nexthop_ipv6 to check the ipv6 link-local nexthop conformity in several situations. Signed-off-by: Louis Scalbert <[email protected]>
louis-6wind
force-pushed
the
fix-ipv6-ll-nexthop
branch
from
ba7b1ac
to
3e56ea7
Compare
#16219 attempted to solve #16198 issue but removed ipv6 link-local nexthop in situation it was legitimate. For example, when an ipv4-mapped IPv6 address is present in global IPv6 nexthop, the link-local is also useful if the peer does not understand the ipv4-mapped IPv6 address format.
Revert #16219 and apply fixes to the "ipv6 nexthop-local unchanged" behavior. When the parameter is present:
Add a topotest to check the ipv6 nexthops in various situations including the use case from #16219 (comment) and #17037 (comment)