IPv6 Dual-stack mesh pod routing works only when bgp graceful restart is enabled

[rkojedzinszky]

What happened?

I've enabled ipv6 in my k8s cluster, and have set up kube-router to do so.

What did you expect to happen?

I expected that pod prefixes will appear in nodes' routing tables, as for ipv4, howewer it did not happen.

How can we reproduce the behavior you experienced?

I think a basic kube-router setup, without external peers. Two nodes in same /64. Following official documentation setting up kube-router. Howewer, when bgp-graceful-restart is disabled, then the ipv6 pod networks dont appear in routing tables.

System Information (please complete the following information)

• Kube-Router Version (kube-router --version): 2.6.3
• Kube-Router Parameters:

        - --enable-ipv4=true
        - --enable-ipv6=true
        - --run-router=true
        - --run-firewall=true
        - --run-service-proxy=true
        - --advertise-cluster-ip=true
        - --advertise-pod-cidr=true
        - --advertise-external-ip
        - --kubeconfig=/var/lib/kube-router/kubeconfig
        - --iptables-sync-period=24h
        - --ipvs-sync-period=24h
        - --routes-sync-period=24h
        - --auto-mtu=false
        - --service-external-ip-range=ipv4/27
        - --service-external-ip-range=ipv4/24
        - --service-cluster-ip-range=a:b:c:d::/64
        - --service-external-ip-range=a:b:c:e::/64
        - --runtime-endpoint=unix:///var/run/crio/crio.sock

• Kubernetes Version (kubectl version) :

$ kubectl version
Client Version: v1.32.3
Kustomize Version: v5.5.0
Server Version: v1.34.0
WARNING: version difference between client (1.32) and server (1.34) exceeds the supported minor version skew of +/-1

• Cloud Type: on premise
• Kubernetes Deployment Type: kubeadm
• Kube-Router Deployment Type: DaemonSet
• Cluster Size: 10 nodes

[aauren]

@rkojedzinszky can you try: cloudnativelabs/kube-router-git:PR-1994 to ensure that it fixes this issue the same as #1993?