Hope everyone is having a great weekend.
I'm feeling a little puzzled as to the best way to implement a nat gateway in my GCP Project.
I was looking to simply implement this: https://github.com/GoogleCloudPlatform/terraform-google-nat-gateway
I've thrown this in my main.tf:
module "nat" {
source = "GoogleCloudPlatform/nat-gateway/google"
region = "us-central1"
network = "default"
subnetwork = "default"
}
And I've added this tag (as per the readme), to my compute instances:
${module.nat.routing_tag_regional}
However, I'm seeing the following (which kinda makes sense I guess):
Error: resource 'google_compute_instance.ds3-build' config: reference to undefined module "nat"
As I have the module in my root/main.tf, and the instances are being build via another custom written module. Do I need to output the value from root/main.tf for it to be available within another module?
root/
├── app1-deploy/
│ ├── main.tf
│ ├── outputs.tf
│ └── variables.tf
├── app2-deploy
│ ├── main.tf
│ ├── outputs.tf
│ └── variables.tf
├── app3-deploy
│ ├── main.tf
│ ├── outputs.tf
│ └── variables.tf
├── creds
│ └── account.json
├── scripts
│ └── startup.sh
├── main.tf
├── outputs.tf
├── variables.tf
└── terraform.tfvars
Machines are provisioned inside app1-deploy/main.tf
app2-deploy/main.tf
etc... and the module "nat"
is in root/main.tf
1) Are the compute resources separate Terraform projects? If so, they cannot see module.nat
because your root
project is a different project with its own state.
2) If your nat
module is created in the root
project, I'd pass along ${module.nat.routing_tag_regional}
using a variable on your compute modules.