Bug report

On a fresh install of 1.8-rc3 with ipv6 enabled, cilium-agents goes into a crashloop. logs provide some warnings and then a fatal message about failing to determine IPv6 for the host interface

level=fatal msg="Failed to initialize NodePort addrs" error="Failed to determine IPv6 of enp0s8 for NodePort" subsys=daemon

Deployed with helm template cilium/cilium --version 1.8.0-rc3 --namespace=kube-system --set global.ipv4.enabled=true --set global.ipv6.enabled=true > cilium-1.8-rc3.yaml

Some additional level=warning messages precede the fatal log are included below

level=warning msg="Device with a default route cannot be found for BPF NodePort device detection" error="Default route not found for family 10" subsys=daemon
level=info msg="Using auto-derived devices for BPF node port" devices="[enp0s8]" directRoutingDevice=enp0s8 subsys=daemon
level=warning msg="Session affinity for host reachable services needs kernel 5.7.0 or newer to work properly when accessed from inside cluster: the same service endpoint will be selected from all network namespaces on the host." subsys=daemon

General Information

• Cilium version (run cilium version)
  1.8-rc3
• Kernel version (run uname -a)
  Linux agent1 5.3.0-51-generic cilium: update helpers to newer kernel #44-Ubuntu SMP Wed Apr 22 21:09:44 UTC 2020 x86_64 x86_64 x86_64 GNU/Linux
• Orchestration system version in use (e.g. kubectl version, Mesos, ...)
  v1.18.4+k3s1
• Link to relevant artifacts (policies, deployments scripts, ...)
• Generate and upload a system zip:

curl -sLO https://git.io/cilium-sysdump-latest.zip && python cilium-sysdump-latest.zip

How to reproduce the issue

1. Deploy 1.8-rc3 with IPv6 enabled