The ovnkube-master pods are CrashLoopBackoff status with the "Pod <podname> is invalid: metadata.annotations: Too long: must have at most 262144 byte" message.

After deleting the pod with the error message, the ovnkube-master pod was returned to Running.

When checking the logs of ovnkube-master pod, it seems that routes configuration is set repeatedly.

2024-04-15T05:08:20.146710725Z I0415 05:08:20.032158       1 kube.go:71] Setting annotations map[k8s.ovn.org/pod-networks:{"default":{"ip_addresses":["10.130.0.132/23","fd01:0:0:3::84/64"],"mac_address":"0a:58:0a:82:00:84","gateway_ips":["fd01:0:0:3::1"],"routes":[{"dest":"10.128.0.0/14","nextHop":"10.130.0.1"},{"dest":"172.30.0.0/16","nextHop":"10.130.0.1"},{"dest":"10.128.0.0/14","nextHop":"10.130.0.1"},{"dest":"172.30.0.0/16","nextHop":"10.130.0.1"},{"dest":"10.128.0.0/14","nextHop":"10.130.0.1"},<...snip...>{"dest":"172.30.0.0/16","nextHop":"10.130.0.1"},{"dest":"10.128.0.0/14","nextHop":"10.130.0.1"},{"dest":"172.30.0.0/16","nextHop":"10.130.0.1"},{"dest":"10.128.0.0/14","nextHop":"10.130.0.1"},{"dest":"172.30.0.0/16","nextHop":"10.130.0.1"}]}}] on pod <namespace>/<podname>
2024-04-15T05:08:20.207768306Z E0415 05:08:20.207745       1 kube.go:80] Error in setting annotation on pod <namespace>/<podname>: Pod "<podname>" is invalid: metadata.annotations: Too long: must have at most 262144 bytes

Red Hat OpenShift Container Platform 4.12.39    

Sometimes