hbmm uses InternalIP, but node uses ExternalIP

[guettli]

/kind bug

Our bm nodes (wl-cluster) have an ExternalIP only:

❯ k get nodes -o wide
NAME                                            STATUS   ROLES           AGE   VERSION   INTERNAL-IP   EXTERNAL-IP     OS-IMAGE             KERNEL-VERSION      CONTAINER-RUNTIME
bm-thomas-sonobuoy333-2105469                   Ready    <none>          10m   v1.34.3   <none>        95.216.111.98   Ubuntu 24.04.3 LTS   6.14.0-37-generic   containerd://2.2.1

Yaml of node (wl-cluster):

status:
  addresses:
  - address: bm-thomas-sonobuoy333-2105469
    type: Hostname
  - address: 95.216.111.98
    type: ExternalIP

But the corresponding capi machine (mgt-cluster) has only an InternalIP, and a ipv6 address:

status:
  addresses:
  - address: 95.216.111.98/28
    type: InternalIP
  - address: 2a01:4f9:2b:dc8::2/64
    type: InternalIP
  - address: thomas-sonobuoy333-md-bm-l7w8s-n2mhz-5l8sd
    type: Hostname
  - address: thomas-sonobuoy333-md-bm-l7w8s-n2mhz-5l8sd
    type: InternalDNS

hbmm:

status:
  addresses:
  - address: 95.216.111.98/28
    type: InternalIP
  - address: 2a01:4f9:2b:dc8::2/64
    type: InternalIP
  - address: thomas-sonobuoy333-md-bm-l7w8s-n2mhz-5l8sd
    type: Hostname
  - address: thomas-sonobuoy333-md-bm-l7w8s-n2mhz-5l8sd
    type: InternalDNS

Two issue:

• Why is InternalIP used?
• Why does the ipv6 appear?

Related: Capi Docs about infra-machine to capi-machine: InfraMachine Addresses - The Cluster API Book