terraform-provider-helm: Error: stat /Users/xxxxx/.kube/config: no such file or directory when trying to use with EKS

Terraform Version

12.18

Affected Resource(s)

• helm_release

Terraform Configuration Files

data "aws_eks_cluster_auth" "cluster-auth" {
  depends_on = [aws_eks_cluster.default]
  name       = "aws_eks_cluster.default"
}

provider "helm" {
  alias = "eks"
  kubernetes {
    # config_path = "${path.module}/.kube/config"
    host                   = aws_eks_cluster.default.endpoint
    cluster_ca_certificate = aws_eks_cluster.default.certificate_authority[0].data #base64decode(aws_eks_cluster.default.certificate_authority.0.data)
    token                  = data.aws_eks_cluster_auth.cluster-auth.token
    load_config_file       = false
  }
}

resource "helm_release" "controllers" {
  name  = "controllers"
  atomic = true
  chart = "./controllers"
  depends_on = [
    aws_eks_cluster.default,
    local_file.kubeconfig,
    aws_autoscaling_group.default
  ]

  set {
    name  = "clusterName"
    value = aws_eks_cluster.default.id
  }

  set {
    name  = "region"
    value = var.aws_region
  }

  set {
    name  = "zone"
    value = var.route53zone

Expected Behavior

It creates a Helm release with my the YAML in my controllers folder.

Actual Behavior

Receive this errors instead:

_Error: stat /Users/xxxxx/.kube/config: no such file or directory

on main.tf line 610, in resource “helm_release” “controllers”: 610: resource “helm_release” “controllers” {_

Here’s what I know; if I have a ~/.kube/config file present, everything works just fine. Thus, when running terraform apply locally on my laptop, it works. However, I’m trying to replicate running terraform apply in a fresh container in our CD platform, which won’t have that file.

Now, oddly enough, even though the rest of my EKS TF module has commands all over the place using “${path.module}/.kube/config” as the source for the kubeconfig and work just fine, Helm3 provider does not. It INSISTS I have a config file at ~/.kube/config, even though I:

1. Have a perfectly useable one at ${path.module}/.kube/config, path.module being the root of my module repo. This is useable with literally everything else, including various kubectl commands.
2. Tried, as you can see above, not even using a file and just configuring the provider with the cert, token, etc., as well as specifying load_config_file to false. The provider seems to not care about any of this, everything be damned if there is no ~/.kube.config.

Steps to Reproduce

terraform apply against an EKS cluster.