What were you trying to do that didn't work?

After rebooting OpenShift node an interface gets link local address generated/assigned.
This address is used when advertising routes. After some time link local address is regenerated but is not used later by the frr-k8s pod.
 

What is the impact of this issue to you?

Routing table on the external router has broken routes as it points to the address that doesn't exist any more

Please provide the package NVR for which the bug is seen:

frr-8.5.3-4.el9.x86_64

How reproducible is this bug?:

Quite often. There seems to be a race condition so if the link local address is regenerated before frr-k8s pod starts then the issue doesn't reproduce.

Steps to reproduce

1.  Deploy and configure dual stack baremetal cluster
2.  Configure symmetric routing with VRFs and MetalLB
3.  Reboot a node
4. Check routing table on the remote router

Expected results

If link local address changes this change is picked up by FRR

Actual results

Changed link local address is not picked up by FRR