Minikube: kubectl connection refused - did you specify the right host or port?

Question

I try to run minikube v0.22.1 and kubectl v1.7.5 on MacOS with Virtualbox.

$ minikube start
Starting local Kubernetes v1.7.5 cluster...
Starting VM...
Getting VM IP address...
Moving files into cluster...
Setting up certs...
Connecting to cluster...
Setting up kubeconfig...
Starting cluster components...
Kubectl is now configured to use the cluster.

$ minikube version
minikube version: v0.22.1

$ minikube status
minikube: Running
cluster: Running
kubectl: Correctly Configured: pointing to minikube-vm at 192.168.99.100

However all kubectl commands fail with "connection refused - did you specify the right host or port?"

$ kubectl version
Client Version: version.Info{Major:"1", Minor:"7", GitVersion:"v1.7.5", GitCommit:"17d7182a7ccbb167074be7a87f0a68bd00d58d97", GitTreeState:"clean", BuildDate:"2017-08-31T19:32:26Z", GoVersion:"go1.9", Compiler:"gc", Platform:"darwin/amd64"}
The connection to the server 192.168.99.100:8443 was refused - did you specify the right host or port?

The solution proposed here (sudo ifconfig vboxnet0 up) did not help, the vboxnet0 interface is up.

Any ideas or suggestions are highly appreciated.

Answer 1

When debugging the minikube commands, e.g.

$ minikube dashboard --loglevel 0 --logtostderr

some proxy issues became visible and could be solved.

Answer 2

If you run

kubectl config get-contexts

Do you get the following?

CURRENT   NAME       CLUSTER    AUTHINFO   NAMESPACE
*         minikube   minikube   minikube 

If not that means your kubectl context is not correctly setup. To setup the context correctly run this

kubectl config use-context minikube

Answer 3

I ran into this situation this morning (another Monday!) on MacOS 11.3 with minikube v1.19.0.

I ran minikube status and got the following:

E0503 14:15:43.912005    7308 status.go:412] kubeconfig endpoint: got: 127.0.0.1:64041, want: 127.0.0.1:56537
minikube
type: Control Plane
host: Running
kubelet: Stopped
apiserver: Stopped
kubeconfig: Misconfigured


WARNING: Your kubectl is pointing to stale minikube-vm.
To fix the kubectl context, run `minikube update-context`

Seemed like good advice, so I did run minikube update-context and got this:

  "minikube" context has been updated to point to 127.0.0.1:56537
  Current context is "minikube"

After which everything worked like it did on Friday.

Answer 4

Update: I am using minikube version: v1.25.2

The command mentioned in this thread did NOT work:

minikube start --memory=10000 --cpu 4 #this will FAIL

This, however, DID WORK (use cpus instead. I also changed values to show minimum requirement for Docker):

minikube start --memory=1800 --cpus=2 # this will work
minikube start --memory=1800 --cpus 2 # this will also work

Answer 5

minikube delete && minikube start

sudo minikube start --vm-driver=none (start minikube again)

Answer 6

I will just drop this in here in case anyone find this question.

As of right now I don't know the versions of the OP's setup. So I'm going to assume he has the latest version that was available when he posted, which was: 0.22.1

Description

I had a similar issue. The cluster was timing-out irregularly. One moment I got answers using kubectl cluster-info dump another I didn't. Then it worked again, and then it didn't. I found a github bug report with a solution.

Solution

1. Remove your VirtualBox VM.
2. Remove the ./minikube folder.
3. Remove the minikube executable.
4. Install version 0.19.0.
5. Verify that minikube is working with: kubectl

Versions

• OS: Windows 10 (Home edition)
• Minikube bugged version: 0.22.2
• Minikube working version: 0.19.0
• Kubectl (client): v1.7.0
• Kubectl (server): v1.6.0

EDIT:
I kept having some issue with minikube after I posted this original answer. I found something that fixed the issue completely. It's related to the dynamic memory setting in Hyper-V.

Solution
1. Turn off the hyper-v minikube VM.
2. Go to the VM's settings.
3. Turn off dynamic memory allocation.
4. Assign a decent amount of memory.
5. Save and turn the VM on again.

This should work with any minikube version. See this github issue for progress on an automatated solution

Answer 7

After the Linux Security OS patching and reboot we are unable to start kubernetes service received below error.

Error message: The connection to the server 192.168.1.101:8443 received while starting the kubernetes service.

This issue happened due to systemd package got updated during the security patching.

So We did below action to bring up the application On each master nodes

1. Update the /usr/lib/systemd/system/kubelet.service fie by removing the below two lines:

ExecStartPost=/bin/bash -c 'umask 0022; pgrep -x kubelet > /run/kubelet.pid'
ExecStopPost=/bin/bash -c 'rm -f /run/kubelet.pid'


2. Update the /usr/lib/systemd/system/kube-proxy.service fie by removing the below two lines:

ExecStartPost=/bin/bash -c 'umask 0022; pgrep -x kubelet > /run/kubelet.pid'
ExecStopPost=/bin/bash -c 'rm -f /run/kubelet.pid'

3. Run the kube-restart.sh on the master nodes. 

4. run the kube-restart.sh on the worker nodes.

Answer 8

You may have it stopped or saved for any reason. sometimes, after you enable/disable addons you may need to restart it.

1) Restart minikube VM, stop it

$ minikube stop 

2) Start it again, make sure you assign enough cpu/memory (the following is just an example of how to pass it, you need to adjust it based on available resources in your machine)

$ minikube start --memory=10000 --cpu 4 

If this didn't work out, you can do the following that will help you to know more about the underlying cause of problem:

Check minikube status and make sure the status is Running

$ minikube status 

Or, check minkube logs:

minikube logs

Finally, if you couldn't fix it, you may need to delete and start it from scratch

$ minikube delete && minikube start 

Ref: https://github.com/kubernetes/minikube/issues/1498

Answer 9

This solved my problem

minikube delete
minikube start

just restarted the container