0

I using kubeadm to create cluster and successfully join the work node to master's control-plane.

But there are two calico-node pods in master's kube-system namespace can not run, I check into these two pod and they are all have the same error message output.

calico-node-1
Readiness probe failed: calico/node is not ready: BIRD is not ready: BGP not established with XX.XX.XX.XX (master node ip address)
calico-node-2
Readiness probe failed: calico/node is not ready: BIRD is not ready: BGP not established with XX.XX.XX.XX (work node ip address)

Resolution:

According to this answer, in my case I add this line to the calico.yaml file.

            - name: IP_AUTODETECTION_METHOD
              value: "interface=eth0"
ccd
  • 5,788
  • 10
  • 46
  • 96
  • Where have you deployed the cluster? public cloud aws? Have you tried any other cni? – Arghya Sadhu Feb 03 '20 at 15:34
  • can you share the complete list of flags you used while running `kubeadm init` command like value for `--pod-network-cidr` also ip range of your nodes `--apiserver-advertise-address` and also what output did you get when running `kubectl apply -f https://docs.projectcalico.org/v3.11/manifests/calico.yaml` also if you changed any defaults on calico.yaml !! – DT. Feb 03 '20 at 15:42
  • How much network interfaces do you have on your nodes? – Dawid Kruk Feb 04 '20 at 10:49
  • I have only one eth0 interface, but this solution will not work if I reboot the server. – ccd Feb 05 '20 at 08:35
  • Were you following any particular guide for creating this Kubernetes cluster? Please provide output of command: `$ kubectl get nodes -o wide`. What is operating system that your Kubernetes cluster working on? Do you have firewall enabled? – Dawid Kruk Feb 07 '20 at 12:07

0 Answers0