K
Q

Kubernetes failed to discover supported resources: getsockopt: connection refused

6/27/2018

I am going through the kubernetes tutorial at Udacity. When i run the the nginx image using the following command

kubectl run nginx --image=nginx:1.10.0

It given me the error

error: failed to discover supported resources: Get http://localhost:8080/apis/extensions/v1beta1: dial tcp 127.0.0.1:8080: getsockopt: connection refused

If i try to get pods using the following command

kubectl get pods

it says

The connection to the server localhost:8080 was refused - did you specify the right host or port?

The nginx server is running, i can tell because i can get the appropriate output by running curl http://127.0.0.1

I am not able to figure out what the issue is, and there are not a lot of resources on the internet for this problem. Can anyone please tell me how do i resolve it?

-- sid0972
docker
google-kubernetes-engine
kubernetes

Similar Questions

Is it possible to network two kubernetes clusters such that resources not publicly exposed in one can be accessed by the other?
Dynamically assign port numbers in Kubernetes?
Kubernetes Python Client - Find pending jobs and schedule all pods at a time or how to schedule a pending job
What does Kubernetes cronjobs `startingDeadlineSeconds` exactly mean?
Where do pod labels get injected in prometheus-operator
Error parsing command line: unrecognised option '--wiredTigerCacheSizeGB 2'
Kubernetes ClusterIP versus headless service scenario

6 Answers

4/25/2019

In some cases, it is simply because you need the kubectl run command as root (e.g. sudo it).

-- Farshid
Source: StackOverflow
8/25/2019

You need to set up the zone first:

gcloud config set compute/zone us-central1-b

then add a cluster there :

gcloud container clusters create io

now you can run the commands . Let me know if found a problem there :)

-- vansh Kapoor
Source: StackOverflow
4/4/2019

This issue often occurs when kubectl can not find the configuration credential for the intended cluster.

Check $HOME/.kube/config for cluster configuration. If configuration is empty or configuration is set for a wrong cluster, regenerate configuration by running,

gcloud container clusters get-credentials <CLUSTER_NAME> --zone <ZONE>

This will update the configuration credential in $HOME/.kube/config.

Now, everything should work as expected.

Reference: https://github.com/googlecodelabs/feedback/issues/537

-- Shakil
Source: StackOverflow
6/27/2018

Check your kubectl config file (~/.kube/config)

For testing purposes, you can use the admin one:

kubectl --kubeconfig /etc/kubernetes/admin.conf get po

Or (again, for testing)

sudo cp /etc/kubernetes/admin.conf $HOME/
sudo chown $(id -u):$(id -g) $HOME/admin.conf
export KUBECONFIG=$HOME/admin.conf

You can see more suggestions in kubernetes/kubernetes issue 23726

As commented below, that requires kubernetes to be installed, for the node to be able to join a cluster:

sudo kubeadm join --token TOKEN MASTER_IP:6443
-- VonC
Source: StackOverflow
4/29/2019

failed to discover supported resources ......

kubectl command line tools connects with kube-apiserver at port 8443 for its operations.

To proble whether the apiserver is up, try curl https://192.168.99.100:8443

If it fails, it means kube-apiserver is not running. Most probably minikube would not be running.

So try:

  1. minikube status

    minikube start

OR

  1. restart the VM
-- Muralidharan.rade
Source: StackOverflow
6/28/2018

The solution was simple, as @VonC suggested, i did not have kubernetes installed, i followed this tutorial, and now i can proceed with my work.

-- sid0972
Source: StackOverflow