Provider crash due to apiVersion field in kubeconfig
Dudesons opened this issue · 3 comments
Describe the bug
I'm using a this tool https://github.com/imjasonh/gke-auth for managing kubeconfig with gke, it works great with the terraform provider but when I'm trying to apply something with the k8s provider I have the following error:
╷
│ Error: exec plugin: invalid apiVersion "client.authentication.k8s.io/v1"
│
│ with provider["registry.terraform.io/banzaicloud/k8s"],
│ on common.tf line 9, in provider "k8s":
│ 9: provider "k8s" {
│
╵
ERRO[0009] 1 error occurred:
* exit status 1
in my config I grep api
:
grep api .kube/config
apiVersion: v1
apiVersion: client.authentication.k8s.io/v1
apiVersion: client.authentication.k8s.io/v1
apiVersion: client.authentication.k8s.io/v1
apiVersion: client.authentication.k8s.io/v1
apiVersion: client.authentication.k8s.io/v1
apiVersion: client.authentication.k8s.io/v1
apiVersion: client.authentication.k8s.io/v1
apiVersion: client.authentication.k8s.io/v1
apiVersion: client.authentication.k8s.io/v1
apiVersion: client.authentication.k8s.io/v1
apiVersion: client.authentication.k8s.io/v1
apiVersion: client.authentication.k8s.io/v1
and it seems this tools add this field and without this one it can't work
I tried to backport the config in the provider like that:
provider "k8s" {
exec {
api_version = "client.authentication.k8s.io/vbe1"
args = [
"--get"
]
command = "gke-auth"
}
}
it doesn't change anything
Steps to reproduce the issue:
create a gke cluster
install this tool https://github.com/imjasonh/gke-auth
generate the config:
gke-auth --project=[MY_PROJECT] \
--location=[REGION_OR_ZONE] \
--cluster=[CLUSTER_NAME]
Expected behavior
the provider doesn't crash due to the apiVersion field in the kube config
Screenshots
Additional context
Thank you @sagikazarmark for the review of #91
Just to know when do you think we new release will be available ?
I added a few more Kubernetes versions to the test matrix. Once those tests pass I'll tag a new version.
Hi @sagikazarmark, do you have an idea when the provider will be rebuilt in order to fix this issue ?