vport0 is a member of veb0, along with em0, em1, em2, and em3,
with rad(8) running on vport0 announcing 2620:ba:6000:3::
vport0 only has a link-local address.

The premise of not having a dedicated route for 2620:ba:6000:3::/64 is
that multiple routers in various locations can advertise a prefix to hosts
such as 2620:ba:6000:3:: without a route to one subnet, thus creating a
single logical subnet across multiple routers. Hosts connected to the router
would then assign themselves a /128 such as 2620:ba:6000:3:58d2:48ff:fee6:270a
and bpgd(8) would then announce these /128 hosts to all other routers.

On Mon, May 6, 2024 at 2:58 PM Claudio Jeker <cje...@diehard.n-r-g.com> wrote:
>
> On Mon, May 06, 2024 at 02:36:07PM -0400, Benjamin Raskin wrote:
> > Hello, Claudio;
> >
> > Sorry about the mistake, I meant to paste the route entry for
> > 2620:ba:6000:3:58d2:48ff:fee6:270a
> > and instead pasted the link local address.
>
> I looked at your route output and it makes little sense.
> How is vport0 setup? It seems you just get ND entries for things like
> 2620:ba:6000:3:58d2:48ff:fee6:270a but there is no network route for
> 2620:ba:6000:3::/64 or similar setup. So very strange...
> This is for sure not how I would route stuff.
>
> > Here is the output of the two commands
> >
> > prod-router-wat-01$ bgpctl show fib 2620:ba:6000:3:58d2:48ff:fee6:270a
> > flags: B = BGP, C = Connected, S = Static
> >        N = BGP Nexthop reachable via this route
> >        r = reject route, b = blackhole route
> >
> > flags prio destination                      gateway
> > B       48 ::/0                             fe80::9ab7:85ff:fe00:3726%mgre0
> > prod-router-wat-01$ bgpctl show rib 2620:ba:6000:3:58d2:48ff:fee6:270a
> > flags: * = Valid, > = Selected, I = via IBGP, A = Announced,
> >        S = Stale, E = Error
> > origin validation state: N = not-found, V = valid, ! = invalid
> > aspa validation state: ? = unknown, V = valid, ! = invalid
> > origin: i = IGP, e = EGP, ? = Incomplete
> >
> > flags  vs destination          gateway          lpref   med aspath origin
> > AI*>  N-? 2620:ba:6000::/48    ::                100     0 i
> > I*    N-? 2620:ba:6000::/48    fe80::9ab7:85ff:fe00:3726%mgre0   100
> >   0 10261 i
> > I*    N-? 2620:ba:6000::/48    fe80::9ab7:85ff:fe00:3727%mgre0   100
> >   0 10261 i
> >
> > I'm not seeing any output with the global address in question, which
> > is pretty weird.
> >
> >
> > On Mon, May 6, 2024 at 2:29 PM Claudio Jeker <cje...@diehard.n-r-g.com> 
> > wrote:
> > >
> > > On Mon, May 06, 2024 at 02:03:52PM -0400, Benjamin Raskin wrote:
> > > > As mentioned in my previous email, I'm looking to advertise global
> > > > addresses such as 2620:ba:6000:3:58d2:48ff:fee6:270a, but then
> > > > I took a look at my routing table and noticed that gateway/nexthop
> > > > for this global address is a MAC address
> > > >
> > > > fe80::58d2:48ff:fee6:270a%vport0            56:af:97:0f:66:6e
> > > >              UHLc       0       75     -     3 vport0
> > >
> > > This is a link local route and therefor not distributed. Also this is a
> > > cloned ND6 entry which is also skipped.
> > >
> > > Your 2620:ba:6000:3:58d2:48ff:fee6:270a routes look like host routes to
> > > me. They don't have C or S and so are neither connected nor static routes.
> > > So you bgpctl show fib command does not show them.
> > >
> > > Please send output of `bgpctl show fib 2620:ba:6000:3:58d2:48ff:fee6:270a`
> > > and `bgpctl show rib 2620:ba:6000:3:58d2:48ff:fee6:270a` maybe that helps
> > > to get closer to the issue.
> > >
> > > > So I'm curious, does bgpd(8) even detect that this address' nexthop
> > > > is a MAC address? I'm taking a look at the source code of bgpd and
> > > > I don't see any mention of MAC or hardware addresses.
> > > >
> > > > I'm not sure where to go from here or who to get in touch with, but I
> > > > want to make sure that this is reproducible first and then go from 
> > > > there.
> > > > This might be a bug or it might be something that I'm doing incorrectly.
> > > >
> > > > On Mon, May 6, 2024 at 11:55 AM Peter Hessler <phess...@theapt.org> 
> > > > wrote:
> > > > >
> > > > > On 2024 May 06 (Mon) at 10:14:21 -0400 (-0400), Benjamin Raskin wrote:
> > > > > :Hello, all;
> > > > > :
> > > > > :I've been having some issues getting bgpd to announce IPv6 routes,
> > > > >
> > > > > ...
> > > > >
> > > > > :
> > > > > :bgpd(8) is configued to advertise all connected and static routes,
> > > > > :however bgpd(8) only advertises routes that are connected to the wg0
> > > > > :interface and none that are connected on the vport0 interface. Below 
> > > > > is
> > > > > :the output of `bgpctl show fib connected inet6` for reference.
> > > > > :
> > > > > :
> > > > > :flags prio destination                      gateway
> > > > > :C        1 ::1/128                          link#8
> > > > > :C        4 fd80::/64                        link#11
> > > > > :C        1 fd80::fce1:baff:fe6e:d685/128    link#11
> > > > > :C        3 fd80::fce1:baff:fea6:bf3a/128    link#11
> > > > > :C        3 fd80::fce1:baff:fed1:1740/128    link#11
> > > > > :C        4 fe80::%vport0/64                 link#10
> > > > > :C        4 fe80::%mgre0/64                  link#12
> > > > > :C        1 fe80::1%lo0/128                  link#8
> > > > > :CN       1 fe80::1efd:8ff:fe7e:6b38%mgre0/128 link#12
> > > > > :C        8 fe80::9ab7:85ff:fe00:3726%mgre0/128 link#12
> > > > > :C        8 fe80::9ab7:85ff:fe00:3727%mgre0/128 link#12
> > > > > :C        4 ff01::%lo0/32                    link#8
> > > > > :C        4 ff01::%vport0/32                 link#10
> > > > > :C        4 ff01::%wg0/32                    link#11
> > > > > :C        4 ff01::%mgre0/32                  link#12
> > > > > :C        4 ff02::%lo0/32                    link#8
> > > > > :C        4 ff02::%vport0/32                 link#10
> > > > > :C        4 ff02::%wg0/32                    link#11
> > > > > :C        4 ff02::%mgre0/32                  link#12
> > > > > :
> > > > > :
> > > > > :As far as I can tell bgpd(8) is configured correctly, and there are 
> > > > > no
> > > > > :anomalies when it comes to routes. Below is a sample of my bgpd(8)
> > > > > :configuration for reference.
> > > > > :
> > > > >
> > > > > fe80:: addresses are "link-local" addressess.  Which means they are 
> > > > > only
> > > > > local to the link, and cannot be announced to other links.
> > > > >
> > > > > You'll need to assign ULA or Global addresses to the links in order 
> > > > > for
> > > > > them to be announced.
> > > > >
> > > > >
> > > > > :
> > > > > :AS 10261
> > > > > :
> > > > > :neighbor fe80::9ab7:85ff:fe00:3726%mgre0 {
> > > > > :        remote-as 10261
> > > > > :}
> > > > > :neighbor fe80::9ab7:85ff:fe00:3727%mgre0 {
> > > > > :        remote-as 10261
> > > > > :}
> > > > > :
> > > > > :network inet6 priority 4
> > > > > :network inet6 connected
> > > > > :network inet6 static
> > > > > :
> > > > > :allow from ibgp
> > > > > :allow to ibgp set { nexthop fe80::1efd:8ff:fe7e:6b38%mgre0 
> > > > > prepend-self 1 }
> > > > > :
> > > > > :
> > > > > :Am I missing something? Am I making some assumption when it comes to 
> > > > > how
> > > > > :bgpd(8) works with IPv6 addresses? Thank you in advance.
> > > > > :
> > > > > :
> > > > > :Ben Raskin
> > > > > :
> > > > >
> > > > > --
> > > > > With a rubber duck, one's never alone.
> > > > >                 -- "The Hitchhiker's Guide to the Galaxy"
> > > >
> > >
> > > --
> > > :wq Claudio
>
> --
> :wq Claudio

Reply via email to