On Mon, Jan 09, 2023 at 11:59:22AM -0500, Matt wrote:
> Hello list,
>
>
>
> I've run across an interesting issue which I think might be something I did
> wrong but here goes. Below is my configuration file for bgpd.conf. I will
> also give you the interface configurations for the two tunnels that I am
> running. When I show the RIB using bgpctl show rib, I notice that the set
> localpref parameter is not being applied properly to IPv6.
>
>
>
> #/etc/hostname.wg0
>
> wgkey <my-private-key>
>
> wgpeer <peer-public-key> wgendpoint 47.87.173.98 21764 wgaip
> 192.168.220.190/32 wgaip 172.20.53.98/32 wgaip 172.20.0.0/14 wgaip
> fe80::ade1 wgaip fe80::ade0 wgaip fd00::/8 wgpka 20
>
> inet 192.168.220.190/32
>
> inet6 fe80::ade1%wg0
>
> descr "TO-KIOUBIT"
>
> up
>
> !route add -host 172.20.53.98 192.168.220.190
>
> !route add -inet6 fe80::ade0 fe80::ade1%wg0
>
> !route add -inet6 fd00::/8 fe80::ade1%wg0
>
>
>
> #/etc/hostname.gre0
>
> 172.21.83.84 172.21.83.85
>
> tunnel 173.49.42.100 81.2.241.46
>
> descr "TO-NOP.HU"
>
> up
>
> !ifconfig gre0 inet6 fd40:cc1e:c0de::252 fd40:cc1e:c0de::251
>
>
>
> #/etc/bgpd.conf
>
> ASN="4242421764"
>
>
>
> AS $ASN
>
> router-id 192.168.220.190
>
>
>
> prefix-set mynetworks {
>
> 172.20.165.192/27
>
> fd0b:7449:62d2::/48
>
> }
>
>
>
> prefix-set nothankyou {
>
> 10.0.0.0/8
>
> }
>
>
>
> network prefix-set mynetworks set large-community $ASN:1:1
>
>
>
>
>
> group "kioubit" {
>
> set localpref 20
>
> neighbor 172.20.53.98 {
>
> remote-as 4242423914
>
> descr "TO-KIOUBIT-IPV4-US2"
>
> }
>
>
>
> neighbor fe80::ade0 {
>
> remote-as 4242423914
>
> descr "TO-KIOUBIT-IPV6-US2"
>
> }
>
> }
>
>
>
> group "mc36" {
>
> set localpref 10
>
> neighbor 172.21.83.85 {
>
> remote-as 4242421955
>
> descr "TO-NOP.HU-IPV4"
>
> }
>
>
>
> neighbor fd40:cc1e:c0de::251 {
>
> remote-as 4242421955
>
> descr "TO-NOP.HU-IPV6"
>
> set localpref 10
>
> }
>
> }
>
>
>
> deny quick from ebgp prefix-set mynetworks or-longer
>
> deny quick from ebgp prefix-set nothankyou or-longer
>
> deny quick from any max-as-len 8
>
>
>
> allow to ebgp prefix-set mynetworks large-community $ASN:1:1
>
> allow from ebgp ovs valid
>
>
>
> match from ebgp set { large-community delete $ASN:*:* }
>
> match from any community GRACEFUL_SHUTDOWN set { localpref 0 }
>
>
>
> include "/etc/roa-set.conf"
>
>
>
> When I type bgpctl show rib, I see that the route selected for IPv6 traffic
> is going through the neighbor fd40:cc1e:c0de::251 and not fe80::ade0.
> Ideally, I'd rather have IPv6 go through the neighbor fe80::ade0 as that one
> is on my continent. Below is an example from the show rib statement. I don't
> even know why the fe80::ade0 address is not even showing up in the output.
>
>
>
> *> V fd00:bb:5bf3::/48 fd40:cc1e:c0de::251 10 0 4242421955
> 4242423088 4242420549 i
>
> V fd00:bb:5bf3::/48 :: 20 0 4242423914
> 4242420549 i
>
>
>
> I have verified that the neighbor fe80::ade0 is actually getting a
> connection and sending me route updates. Here is an example:
>
>
>
> V fdff:feed:c0de::/48 :: 20 0 4242423914 4242420585
> 4242422980 210074 64719 65043 4242420138 i
>
>
>
> Any ideas?
Hard to judge from the little information you share but the :: nexthop is
for sure not good. Because of this the route to fd00:bb:5bf3::/48 via AS
4242423914 is not valid and can't be selected as route.
Not sure what exactly goes on there but you need to fix that.
Also check out 'bgpctl show rib nei fe80::ade0 in' to see the unfiltered
routes.
--
:wq Claudio
