Error: Failed to install provider when k8s resource is within a submodule

[hadim]

Using TF 0.13.5 and k8s 0.8.3.

When my k8s resource is within a submodule of the root module:

resource "k8s_manifest" "nvidia_installer" {
  content = data.template_file.nvidia_installer.rendered
}

and my root module declares the k8s provider:

provider "k8s" {
  host                   = "https://${module.cluster.endpoint}"
  token                  = data.google_client_config.default.access_token
  cluster_ca_certificate = base64decode(module.cluster.master_ca_certificate)
}

then upon tf init:

\Initializing modules...

Initializing the backend...

Initializing provider plugins...
- Using previously-installed hashicorp/google-beta v3.49.0
- Using previously-installed hashicorp/template v2.2.0
- Using previously-installed hashicorp/google v3.49.0
- Using previously-installed banzaicloud/k8s v0.8.3
- Using previously-installed hashicorp/kubernetes v1.13.3
- Using previously-installed hashicorp/random v3.0.0
- Finding latest version of hashicorp/k8s...

Error: Failed to install provider

Error while installing hashicorp/k8s: provider registry registry.terraform.io
does not have a provider named registry.terraform.io/hashicorp/k8s

Looks like somehow tf is looking for hashicorp/k8s. When I move resource "k8s_manifest" "nvidia_installer" within the root module, everything is fine.

Is that a known issue or am I doing something wrong?

[hadim]

Looks like it's a TF related issue: hashicorp/terraform#26448

[hadim]

Using

terraform {
  required_version = ">= 0.13"

  required_providers {
    k8s = {
      source  = "banzaicloud/k8s"
      version = ">=0.8.3"
    }
  }
}

within the submodule works. (not ideal but working...)