Instance group creation timeout

[un1x01d]

Hi,
Trying to create a simple nat setup using this simple config:

module "nat" {
  source     = "GoogleCloudPlatform/nat-gateway/google"
  region     = "us-east1"
  network    = "tfstate-vpc-nat"
  subnetwork = "tfstate-vpc-nat-subnet"
}

module "mig" {
  source      = "github.com/GoogleCloudPlatform/terraform-google-managed-instance-group"
  region      = "us-east1"
  zone        = "us-east1-b"
  name        = "nat-group"
  target_tags = ["${module.nat.routing_tag_regional}"]
  network    = "tfstate-vpc-nat"
  subnetwork = "tfstate-vpc-nat-subnet"
  service_port      = 80
  service_port_name = "http"
}

When creating the instance group, i can see it in the GCP console but it shows that it's still creating until it times-out after 20mins.

module.nat.module.nat-gateway.google_compute_instance_group_manager.default: Creating...
  auto_healing_policies.#:                   "" => "1"
  auto_healing_policies.0.health_check:      "" => "https://www.googleapis.com/compute/v1/projects/tfstate/global/healthChecks/nat-gateway-us-east1-b"
  auto_healing_policies.0.initial_delay_sec: "" => "300"
  base_instance_name:                        "" => "nat-gateway-us-east1-b"
  description:                               "" => "compute VM Instance Group"
  fingerprint:                               "" => "<computed>"
  instance_group:                            "" => "<computed>"
  instance_template:                         "" => "https://www.googleapis.com/compute/beta/projects/tfstate/global/instanceTemplates/default-20180725205407178200000001"
  name:                                      "" => "nat-gateway-us-east1-b"
  named_port.#:                              "" => "1"
  named_port.0.name:                         "" => "http"
  named_port.0.port:                         "" => "80"
  project:                                   "" => "<computed>"
  self_link:                                 "" => "<computed>"
  target_size:                               "" => "1"
  update_strategy:                           "" => "NONE"
  version.#:                                 "" => "<computed>"
  wait_for_instances:                        "" => "true"
  zone:                                      "" => "us-east1-b"
module.nat.nat-gateway.google_compute_instance_group_manager.default: Still creating... (10s elapsed)
module.nat.nat-gateway.google_compute_instance_group_manager.default: Still creating... (20s elapsed)
module.nat.nat-gateway.google_compute_instance_group_manager.default: Still creating... (30s elapsed)
module.nat.nat-gateway.google_compute_instance_group_manager.default: Still creating... (40s elapsed)
module.nat.nat-gateway.google_compute_instance_group_manager.default: Still creating... (50s elapsed)
module.nat.nat-gateway.google_compute_instance_group_manager.default: Still creating... (1m0s elapsed)
module.nat.nat-gateway.google_compute_instance_group_manager.default: Still creating... (1m10s elapsed)
module.nat.nat-gateway.google_compute_instance_group_manager.default: Still creating... (1m20s elapsed)
module.nat.nat-gateway.google_compute_instance_group_manager.default: Still creating... (1m30s elapsed)
module.nat.nat-gateway.google_compute_instance_group_manager.default: Still creating... (1m40s elapsed)
module.nat.nat-gateway.google_compute_instance_group_manager.default: Still creating... (1m50s elapsed)
module.nat.nat-gateway.google_compute_instance_group_manager.default: Still creating... (2m0s elapsed)
module.nat.nat-gateway.google_compute_instance_group_manager.default: Still creating... (2m10s elapsed)
module.nat.nat-gateway.google_compute_instance_group_manager.default: Still creating... (2m20s elapsed)

..... .... ...

[JessicaGreben]

I was having the same problem yesterday where module.nat.nat-gateway.google_compute_instance_group_manager.default was timing out. In the GCP console, the Instance Group had an infinite spinner saying it was still trying to "Verify". Magically it seems to have resolved itself today. Wish I had more details why, maybe something on GCP's side. /shrug.

[danisla]

Creating the health check for the self-healing of instances in the managed instance group can take several minutes to configure on the GCP network side. This should work most if not all of the time and I expect it to resolve itself over time.

[danisla]

Fixed with #79

[Logan-Kang]

Hello

As above, I also have timeout issue is also occurring when using managed instance group module.
If the problem is not generated at once, check instance_group_manager for more than 5 minutes during regeneration attempts, and this also does not run, so destroy is also impossible. Please tell me how to solve this problem.