I have created a Kubernetes cluster with 2 nodes, on Master node and on Worker node (2 different VMs).
The worker node has joined succesfully the cluster, so when i run the commana
kubectl get nodes in my master node it appears the 2 nodes exists in the cluster!
However, when i run the command kubectl apply -f https://k8s.io/examples/controllers/nginx-deployment.yaml from my worker node terminal, in order to create a deployment in the worker node, i have the following error:
The connection to the server localhost:8080 was refused. - did you specify the right host or port?
Any help what is going on here?
CodePudding user response：
The easy way to do it is to copy the config from master node usually found here : /etc/kubernetes/admin.conf , to whetever node you want to configure kubectl ( even on master node) . The location to be copied is : $HOME/.kube/config
Also, you can this command from master node by specify nodeselector or label.
CodePudding user response：
It looks like you have issue with your kubeconfig file, as usually localhost:8080 is the default server to connect in absence of this file . Generally, Kubernetes uses this file to store cluster authentication information and a list of contexts to which kubectl refers when running commands - that's why kubectl can't work properly without this file.
To check the presence of kubeconfig file, enter this command: kubectl config view.
Or just check the presence of the file named config in the $HOME/.kube directory, which is the default location for kubeconfig file.
If it is absent, you would need to copy the config file to your node, e.g.:
sudo cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
sudo chown $(id -u):$(id -g) $HOME/.kube/config
sudo service kubelet restart
It is also possible to generate config file in a more difficult way - instead of copying - as described here.