Error: Kubernetes cluster unreachable with helm 3.0

[rubiktubik]

Version:
k3s version v1.0.0 (18bd921)

Describe the bug
I want to use helm version 3 with k3s but i when type helm install stable/postgresql --generate-name for example i get:
Error: Kubernetes cluster unreachable

To Reproduce

1. Installed helm 3 with script](https://helm.sh/docs/intro/install/#from-script).
2. Add repo with helm repo add stable https://kubernetes-charts.storage.googleapis.com/
3. Update repo with helm repo update
4. Install posgresql-chart with helm install stable/postgresql --generate-name

Expected behavior
Installation should work.

Actual behavior
Error: Kubernetes cluster unreachable

Additional context

[r1chjames]

Same issue here on k3s version v1.0.0 (18bd921).

[grawin]

Try setting the KUBECONFIG environment variable.
export KUBECONFIG=/etc/rancher/k3s/k3s.yaml

[r1chjames]

That worked for me, but I only tried on a fresh CentOS/k3s/Helm install. Thanks @grawin

[Genubath]

Same issue. The fix @grawin posted doesn't solve it for me though.

EDIT: I also tried the steps in https://github.com/ibrokethecloud/rancher-helm3 but to no avail.

[rubiktubik]

The fix @grawin posted doesn't worked for me either, i'am using a ubuntu 18.04 system.

[sixcorners]

If you add "-v 20" to your helm command line it will show it's connecting to port 8080.
Running this seems to fix it:
kubectl config view --raw >~/.kube/config

This lets helm use the same config kubectl is using I think.

[galal-hussein]

@rubiktubik looks like helm can't reach the k3s cluster, can you try to use --kubeconfig with helm command or using ~/.kube/config as @sixcorners suggested, please reopen the issue if the problem still persists.

[ghost]

If you add "-v 20" to your helm command line it will show it's connecting to port 8080.
Running this seems to fix it:
kubectl config view --raw >~/.kube/config

This lets helm use the same config kubectl is using I think.

can confirm this solution works for me as well

[pcgeek86]

This resolved the error message for me.

sudo helm install harbor/harbor --version 1.3.0 --generate-name --kubeconfig /etc/rancher/k3s/k3s.yaml

[michelcameroon]

with k3s i had the same problem, system tells me the file
/etc/rancher/k3s/k3s,yaml
is not reachable

the file had rw for root i change to 744 and it works. please tell me if it is correct,

[aureq]

If you are using sudo, be aware that this command doesn't preserve environment variables (such as KUBECONFIG) by default when switching to a different context.

If you wish to preserve specific environment variables when using sudo then:

cat << EOF > /etc/sudoers.d/env
Defaults env_keep += "http_proxy https_proxy no_proxy"
Defaults env_keep += "HTTP_PROXY HTTPS_PROXY NO_PROXY"
Defaults env_keep += "KUBECONFIG"
EOF

[Vesnica]

If you are using sudo, be aware that this command doesn't preserve environment variables (such as KUBECONFIG) by default when switching to a different context.

If you wish to preserve specific environment variables when using sudo then:

cat << EOF > /etc/sudoers.d/env
Defaults env_keep += "http_proxy https_proxy no_proxy"
Defaults env_keep += "HTTP_PROXY HTTPS_PROXY NO_PROXY"
Defaults env_keep += "KUBECONFIG"
EOF

Just use sudo -E, which will preserve the environment variables.

[jordanburke]

@Vesnica thanks that worked for me.

[xbmono]

I'm using helm 3.2.4 on Windows and have the same issue. Setting environment variable KUBECONFIG didn't help either. Without --kubeconfig it doesn't fail but it returns no result on helm ls

[montaro]

what worked for me is to set the KUBECONFIG with absoult path value after changing directory to the chart directory.

[cdperdomo]

If you add "-v 20" to your helm command line it will show it's connecting to port 8080.
Running this seems to fix it:
kubectl config view --raw >~/.kube/config

This lets helm use the same config kubectl is using I think.

This kubectl config view --raw >~/.kube/config works for me. thaks

[poojabolla]

I tried this command,

kubectl config view --raw >~/.kube/config

but after running this config file became empty.

Can anyone suggest how to recover my config file with all values?

[schunduEA]

@poojabolla.. Its gone, you must use >> instead > on appending something in existing file

[sachinmalanki]

@poojabolla.. Its gone, you must use >> instead > on appending something in existing file

I got this issue when using azure kubernetes

az aks get-credentials -n myCluster -g myResourceGroup
The config file is autogenerated and placed in '~/.kube/config' file as per OS

[MrMYHuang]

For microk8s, the k8s config can be generated by this command:
microk8s.kubectl config view --raw > ~/.kube/config

[shing1211]

This resolved the error message for me.

sudo helm install harbor/harbor --version 1.3.0 --generate-name --kubeconfig /etc/rancher/k3s/k3s.yaml

thankyou this works for me

[sufiyanpk7]

For microk8s you can try with

export KUBECONFIG=/var/snap/microk8s/4094/credentials/client.config