Failed to create a cluster with command kubeadm init

Question

I am tring to install kubenetes on debian 9.3, I followed the instructions on this document https://kubernetes.io/docs/setup/independent/install-kubeadm/, it failed to create the cluster with timeout error, the commands I used are as follows:

export HTTP_PROXY=http://192.168.56.1:1080  # this is my internet proxy
export HTTPS_PROXY=http://192.168.56.1:1080
export NO_PROXY=127.0.0.1,192.168.56.*,10.244.*,10.96.*
kubeadm init --apiserver-advertise-address=192.168.56.101 --pod-network-cidr=10.244.0.0/16

the last command hangs up for 1hour and failed with timeout, I found that several container had been running by command docker ps, the running containers included kube-controller-manager-amd64,etcd-amd64,kube-apiserver-amd64,kube-scheduler-amd64,4 instances of pause-amd64.

the error messages are as follows

duler-debvm01_kube-system(660259102d57385a8043d025ac189c87)": Get https://192.168.56.101:6443/api/v1/namespaces/kube-system/pods/kube-scheduler-debvm01: net/http: TLS handshake timeout
Apr 06 21:44:49 DebVM01 kubelet[10665]: E0406 21:44:49.923017   10665 reflector.go:205] k8s.io/kubernetes/pkg/kubelet/kubelet.go:474: Failed to list *v1.Node: Get https://192.168.56.101:6443/api/v1/nodes?fieldSelector=metadata.name%3Ddebvm01&limit=500&resourceVersion=0: net/http: TLS handshake timeout
Apr 06 21:44:49 DebVM01 kubelet[10665]: E0406 21:44:49.924966   10665 reflector.go:205] k8s.io/kubernetes/pkg/kubelet/kubelet.go:465: Failed to list *v1.Service: Get https://192.168.56.101:6443/api/v1/services?limit=500&resourceVersion=0: net/http: TLS handshake timeout
Apr 06 21:44:49 DebVM01 kubelet[10665]: E0406 21:44:49.925892   10665 reflector.go:205] k8s.io/kubernetes/pkg/kubelet/config/apiserver.go:47: Failed to list *v1.Pod: Get xxx/api/v1/pods?fieldSelector=spec.nodeName%3Ddebvm01&limit=500&resourceVersion=0: net/http: TLS handshake timeout
Apr 06 21:44:50 DebVM01 kubelet[10665]: E0406 21:44:50.029333   10665 eviction_manager.go:238] eviction manager: unexpected err: failed to get node info: node "debvm01" not found
Apr 06 21:44:50 DebVM01 kubelet[10665]: E0406 21:44:50.379543   10665 kubelet_node_status.go:106] Unable to register node "debvm01" with API server: Post xxx: net/http: TLS handshake timeout
Apr 06 21:44:52 DebVM01 kubelet[10665]: E0406 21:44:52.575452   10665 event.go:209] Unable to write event: 'Post xxxx: net/http: TLS handshake timeout' (may retry after sleeping)
Apr 06 21:44:57 DebVM01 kubelet[10665]: I0406 21:44:57.380498   10665 kubelet_node_status.go:273] Setting node annotation to enable volume controller attach/detach
Apr 06 21:44:57 DebVM01 kubelet[10665]: I0406 21:44:57.430059   10665 kubelet_node_status.go:82] Attempting to register node debvm01
Apr 06 21:45:00 DebVM01 kubelet[10665]: E0406 21:45:00.030635   10665 eviction_manager.go:238] eviction manager: unexpected err: failed to get node info: node "debvm01" not found
Apr 06 21:45:01 DebVM01 kubelet[10665]: I0406 21:45:01.484580   10665 kubelet_node_status.go:85] Successfully registered node debvm01

the above error messages has been processed and eliminated a lot of repeated lines as follows:

Apr 06 22:46:20 DebVM01 kubelet[10665]: E0406 22:46:20.773690   10665 kubelet.go:2104] Container runtime network not ready: NetworkReady=false reason:NetworkPluginNotReady message:docker: network plugin is not ready: cni config uninitialized
Apr 06 22:46:25 DebVM01 kubelet[10665]: W0406 22:46:25.779141   10665 cni.go:171] Unable to update cni config: No networks found in /etc/cni/net.d

Kubernetes v1.9.3

could anyone help me?

VASャ VASャ · Accepted Answer · 2018-04-09T15:15:29

kubeadm init --apiserver-advertise-address=192.168.56.101 --pod-network-cidr=10.244.0.0/16

From kubeadm documentation:

--apiserver-advertise-address ip-address The IP address the API Server will advertise it's listening on. Specify '0.0.0.0' to use the address of the default network interface.

Unless otherwise specified, kubeadm uses the default gateway’s network interface to advertise the master’s IP. If you want to use a different network interface, specify --apiserver-advertise-address=ip-address

From kubernetes api-server documentation:

--advertise-address ip-address The IP address on which to advertise the apiserver to members of the cluster. This address must be reachable by the rest of the cluster. If blank, the --bind-address will be used. If --bind-address is unspecified, the host's default interface will be used.

I've done a couple of experiments which confirm that it is necessary for ip-address to be configured (or added as a secondary IP) to one of the master's instance interfaces. Just double check if that interface is up.

The last error message,

network plugin is not ready: cni config uninitialized

means that kubernetes networking subsystem is absent or broken. Try to install/reinstall it with

kubectl apply -f https://docs.projectcalico.org/v3.0/getting-started/kubernetes/installation/hosted/kubeadm/1.7/calico.yaml

This part described in section "(3/4) Installing a pod network" in the document you've mentioned.

If you are stuck, try to reinstall your cluster following this manual.

Failed to create a cluster with command kubeadm init

1 Answers