bgp ipv6 - incorrect ipv6 nht

[pornmailbox]

I'm trying to establish ipv6 public bgp connection with my uplink.

For my point-to-point I use eth1. eth0 is my management interface.
Bgp session is ok. but my output of "sh ipv6 nht" reports me -

2001:678:c5c::42(Connected)
resolved via connected
is directly connected, eth1 (vrf default)
Client list: bgp(fd 28)
2a03:500::
resolved via static
unreachable (blackhole) (vrf default), weight 1
Client list: bgp(fd 28)
fe80::ee38:7305:fae8:8830(Connected)
resolved via connected
is directly connected, eth0 (vrf default)
Client list: bgp(fd 28)

arp -an indicates that mac-address of my uplink is on eth1 (5.188.236.120 is it's ipv4).
.......
? (5.188.236.120) at ec:38:73:e8:88:30 [ether] on eth1
.......

but fe80::ee38:7305:fae8:8830(Connected) in vrf is on eth0.

It seems to me that this error in vrf deny me to establish correct ipv6 bgp peer with my uplink.

My running config -

Current configuration:
!
frr version 9.0-dev-gentoo
frr defaults traditional
hostname frr-ipv6-bgp-unicast
log file /var/log/frr/frr.log
service integrated-vtysh-config
!
debug mgmt all
debug zebra nexthop
debug bgp neighbor-events
debug bgp updates in
debug bgp updates out
debug bgp zebra
debug bgp conditional-advertisement
!
debug vrf
!
debug route-map
!
ipv6 route 2a03:500::/32 Null0
!
router bgp 25096
bgp default ipv6-unicast
neighbor 2001:678:c5c::42 remote-as 50509
neighbor 2001:678:c5c::42 description PIN-FV-ipv6
!
address-family ipv6 unicast
network 2a03:500::/32
neighbor 2001:678:c5c::42 soft-reconfiguration inbound
neighbor 2001:678:c5c::42 prefix-list PIN-import-v6 in
neighbor 2001:678:c5c::42 prefix-list PIN-export-v6 out
exit-address-family
exit
!
ipv6 prefix-list PIN-export-v6 seq 9998 permit 2a03:500::/32
ipv6 prefix-list PIN-import-v6 seq 9998 permit any
!
end

output of bgp summary -

IPv6 Unicast Summary (VRF default):
BGP router identifier 172.31.153.225, local AS number 25096 vrf-id 0
BGP table version 2
RIB entries 2, using 384 bytes of memory
Peers 1, using 725 KiB of memory

Neighbor V AS MsgRcvd MsgSent TblVer InQ OutQ Up/Down State/PfxRcd PfxSnt Desc
2001:678:c5c::42 4 50509 34 33 0 0 0 00:14:49 1 1 PIN-FV-ipv6

Total number of neighbors 1

My uplink announce me default only.
I compile latest frr master branch - https://github.com/FRRouting/frr/tree/c53ab57b79f1ddd8424e671bb00a8cf05fa7bcee

[ton31337]

I don't understand the question. From the output, you have established connection and even sending/receiving routes.

[pornmailbox]

I can't sent ping packets - because next-hop is waiting for a reply from eth0, not from my eth1.

[pornmailbox]

the error is that bgp is waiting a nht reply from eth0, not from eth1.

[ton31337]

How show ip route looks like?

[pornmailbox]

Hi.
frr-ipv6-bgp-unicast# sh ipv6 route
Codes: K - kernel route, C - connected, S - static, R - RIPng,
O - OSPFv3, I - IS-IS, B - BGP, T - Table, v - VNC,
V - VNC-Direct, A - Babel, F - PBR, f - OpenFabric,
> - selected route, * - FIB route, q - queued, r - rejected, b - backup
t - trapped, o - offload failure

B>* ::/0 [20/0] via fe80::ee38:7305:fae8:8830, eth1, weight 1, 2d04h45m
C>* 2001:678:c5c::42/127 is directly connected, eth1, 2d04h45m
S>* 2a03:500::/32 [1/0] unreachable (blackhole), weight 1, 2d04h45m
C>* 2a03:500::1000/127 is directly connected, eth3, 2d04h45m
C>* 2a03:500::2000/127 is directly connected, eth4, 2d04h45m
C>* 2a03:500::3000/127 is directly connected, eth5, 2d04h45m
C>* 2a03:500::4000/127 is directly connected, eth6, 2d04h45m
C>* 2a03:500::5000/127 is directly connected, eth7, 2d04h45m
C * fe80::/64 is directly connected, eth8, 2d04h45m
C * fe80::/64 is directly connected, eth7, 2d04h45m
C * fe80::/64 is directly connected, eth6, 2d04h45m
C * fe80::/64 is directly connected, eth5, 2d04h45m
C * fe80::/64 is directly connected, eth4, 2d04h45m
C * fe80::/64 is directly connected, eth3, 2d04h45m
C * fe80::/64 is directly connected, eth2, 2d04h45m
C * fe80::/64 is directly connected, eth1, 2d04h45m
C>* fe80::/64 is directly connected, eth0, 2d04h45m