Delay in switching the pods from one node to another

[prathyushmr]

I have followed the https://github.com/Mirantis/k8s-externalipcontroller/blob/master/doc/fail-over-optimization.md . But its still taking 4-5 minutes to detect the node failure and restart the pods in another node . On the master node i edited /etc/kubernetes/manifests/kube-controller-manager.yaml to set

• --node-monitor-period=2s
  • --node-monitor-grace-period=16s
  • --pod-eviction-timeout=30s and on worker node /var/lib/kubelet/kubeadm-flags.env to set -node-status-update-frequency=4s