BGP: IPv6 peering not working (IPv4 works)

[andsens]

BGP: IPv6 peering not working (IPv4 works)

[lmb]

Hi @andsens, unfortunately we can't provide support on the bugtracker. Have you tried getting help on our Slack?

[andsens]

@lmb that's fair. tbh it didn't even occur to me that you had other help channels, I guess my brain was too burned out from all the troubleshooting to think of that. I'll try the slack channel and post an update here if I find out what the problem is.

[andsens]

@harsimran-pabla helped me figure this out on Slack.
By splitting up the BGP peering into IPv4 and IPv6 and then using either address, the Next Hop is fixed and the routing works:

{
  "protocols": {
    "bgp": {
      "64512": {
        "neighbor": {
          "10.192.205.141": {
            "remote-as": "65000"
          },
          "fd24:28b4:307f:0:2a0:98ff:fe52:9307": {
            "remote-as": "65000",
            "address-family": {
              "ipv6-unicast": "''"
            }
          }
        }
      }
    }
  }

$ show ipv6 bgp
BGP table version is 0, local router ID is 10.192.205.1
Status codes: s suppressed, d damped, h history, * valid, > best, i - internal,
              r RIB-failure, S Stale, R Removed
Origin codes: i - IGP, e - EGP, ? - incomplete
   Network          Next Hop            Metric LocPrf Weight Path
*>ifd24:28b4:307f:cafe:0:32::/112
                    fd24:28b4:307f:0:2a0:98ff:fe52:9307
                                                 100      0 i
*>ifd24:28b4:307f:cafe:0:34:0:1/128
                    fd24:28b4:307f:0:2a0:98ff:fe52:9307
                                                 100      0 i
*>ifd24:28b4:307f:cafe:0:34:0:2/128
                    fd24:28b4:307f:0:2a0:98ff:fe52:9307
                                                 100      0 i
*>ifd24:28b4:307f:cafe:0:34:0:3/128
                    fd24:28b4:307f:0:2a0:98ff:fe52:9307
                                                 100      0 i
Total number of prefixes 4

I've changed the IPs for privacy reasons. I'm actually using my ISP prefix delegated /48 space, so that works as well!

EDIT: I updated the AS number for the cilium node. Using 64512 for both sides results in other hosts being able to route just fine but the node itself is then unable to route to the LB endpoints on it. Again, I'm a total noob at BGP so I haven't the slightest how all this fits together. I'm just a monkey hammering on a typewriter that surprisingly enough ended up typing out a passable haiku.