disk_type is not required under auto_provisioning_defaults for node autoprovisioning

[ttruong-actian]

Community Note

• Please vote on this issue by adding a 👍 reaction to the original issue to help the community and maintainers prioritize this request.
• Please do not leave +1 or me too comments, they generate extra noise for issue followers and do not help prioritize the request.
• If you are interested in working on this issue or have submitted a pull request, please leave a comment.
• If an issue is assigned to a user, that user is claiming responsibility for the issue.
• Customers working with a Google Technical Account Manager or Customer Engineer can ask them to reach out internally to expedite investigation and resolution of this issue.

Terraform Version & Provider Version(s)

Terraform v1.8.3 on darwin_arm64

• provider registry.terraform.io/hashicorp/google v5.27.0

Affected Resource(s)

node autoprovisioning does not required disk_type to be defined with default settings. that way, we NAP can create node pools with N2/N4 where different disk_type are required.

Terraform Configuration

      for_each = var.cluster_autoscaling.enabled ? [1] : []
      content {
        service_account   = google_service_account.main_cluster.email
        oauth_scopes      = local.node_oauth_scopes
        boot_disk_kms_key = var.cluster_autoscaling.manual_updated_nap ? var.kms_encryption_key : null
        management {
          auto_repair  = lookup(var.cluster_autoscaling, "auto_repair", false)
          auto_upgrade = lookup(var.cluster_autoscaling, "auto_upgrade", false)
        }
        upgrade_settings {
          strategy        = "SURGE"
          max_surge       = 1
          max_unavailable = 0
        }
        disk_size = lookup(var.cluster_autoscaling, "disk_size", 100)
      }```

### Debug Output

``` ~ cluster_autoscaling {
            # (2 unchanged attributes hidden)

          ~ auto_provisioning_defaults {
              ~ disk_size         = 0 -> 100
              + disk_type         = "pd-balanced"```

### Expected Behavior

I do not want to define the `disk_type` especially with N4 instance family, there's a new `hyerdisk-balanced` that i want to use and not support in older generations of family

### Actual Behavior

``` ~ cluster_autoscaling {
            # (2 unchanged attributes hidden)

          ~ auto_provisioning_defaults {
              ~ disk_size         = 0 -> 100
              + disk_type         =  "pd-standard"```

Terraform want to set the `disk_type` to the default value.

### Steps to reproduce

1. `terraform apply`

### Important Factoids

_No response_

### References

_No response_

[ggtisc]

Hi @ttruong-actian!

To replicate this issue we need the full name of the terraform resource, I found the service google_service_account in terraform registry, but it don't contain any argument or attribute called disk_type

Also please share the full code of the resource, because we can't replicate the issue with information that only you know. For sensitive information you could change it with examples like:

project = "my-project" org_id = "1234567890" email = "my-user@my-domain.com"

[ttruong-actian]

hi @ggtisc ,

my bad about the missing details. Here's my sample code for the cluster_autoscaling block with auto_provisioning_defaults. I wanted to use this block to define the ServiceAccount and KMSEncryptionKey for all node pools. In this block, the disk_type and disk_size are indicated as optional but they're set with a Default Value of 100Gb and pd-standard.

    enabled = var.cluster_autoscaling.enabled
    dynamic "auto_provisioning_defaults" {
      for_each = var.cluster_autoscaling.enabled ? [1] : []
      content {
        service_account   = google_service_account.main_cluster.email
        oauth_scopes      = local.node_oauth_scopes
        boot_disk_kms_key = var.cluster_autoscaling.manual_updated_nap ? var.kms_encryption_key : null
        management {
          auto_repair  = lookup(var.cluster_autoscaling, "auto_repair", false)
          auto_upgrade = lookup(var.cluster_autoscaling, "auto_upgrade", false)
        }
        upgrade_settings {
          strategy        = "SURGE"
          max_surge       = 1
          max_unavailable = 0
        }
      }
    }

When I tried to enable node-autoprovisioning via gcloud cli with a config file, the disk_type and disk_size are not required. my sample nap-config file below

autoprovisioningLocations:
  - us-central1-a
  - us-central1-c
  bootDiskKmsKey: projects/<name>/locations/us-central1/keyRings/<ring>/cryptoKeys/<key>
  imageType: COS_CONTAINERD
  serviceAccount: <name>@<project>.iam.gserviceaccount.com
  scopes:
    - https://www.googleapis.com/auth/logging.write
    - https://www.googleapis.com/auth/monitoring
  resourceLimits:
    - resourceType: 'cpu'
      maximum: 100
    - resourceType: 'memory'
      maximum: 100

[ggtisc]

Thanks!

As you can see in Google Cloud documentation looking for the diskType field you'll notice that the default disk type is pd-standard if you don't set a value for this optional field.

If you're looking to use a hyperdisk-balanced you need to define it. If you need help on how to use it you could check the documentation here, and the different disk types here.

If after this you continue with issues please share a full code of your google_container_cluster like this.