How to fix Error: Error creating Cloud SQL Instance: OperationError in Terraform

TerraformINTERMEDIATEHIGHUpdated Dec 31, 2025· 115 views

This error occurs when Terraform fails to create a GCP Cloud SQL instance due to state conflicts, permission issues, or timeout problems. The operation failed because another operation was in progress, the instance name was recently deleted, or required APIs were not enabled.

Quick fix

Fixes Error: Error creating Cloud SQL Instance: OperationError

resource "google_sql_database_instance" "example" {
  name             = "my-new-instance-name-${formatdate("YYYYMMDDhhmmss", timestamp())}"
  database_version = "POSTGRES_15"
  region           = "us-central1"
}
  1. Verify instance name availability
  2. Enable required GCP APIs
  3. Verify service account permissions
  4. Serialize Cloud SQL operations
terraform — error
resource "google_sql_database_instance" "example" {
  name             = "my-new-instance-name-${formatdate("YYYYMMDDhhmmss", timestamp())}"
  database_version = "POSTGRES_15"
  region           = "us-central1"
}
Error: Error creating Cloud SQL Instance: OperationError

What this error means

The OperationError in Cloud SQL indicates that the instance creation operation could not proceed due to one of several state or configuration issues. GCP Cloud SQL prevents concurrent operations on the same instance, meaning if another operation (creation, deletion, or modification) is already in progress, new operations will fail. Additionally, instance names cannot be reused for approximately one week after deletion, and certain required APIs must be enabled in your GCP project before Terraform can interact with Cloud SQL.

Typical symptoms

  • Terraform apply fails with "OperationError" when creating google_sql_database_instance
  • Error message indicates operation is not in an appropriate state
  • Instance creation was successful in GCP Console but Terraform still reports failure
  • Creating multiple Cloud SQL instances simultaneously results in operation conflicts
  • Attempt to recreate a recently deleted instance fails with 409 error

Common causes

  1. Another Cloud SQL operation is already in progress on the same instance (concurrent operations are not allowed)
  2. Attempting to reuse an instance name that was deleted less than a week ago
  3. Required GCP APIs (Cloud SQL Admin API, Compute Engine API, Service Networking API) are not enabled
  4. Insufficient service account permissions (missing roles/cloudsql.admin or related roles)
  5. Terraform timeout settings too short for instance initialization to complete
  6. Network connectivity issues preventing the Cloud SQL service from reaching provisioning infrastructure

How to fix "Error: Error creating Cloud SQL Instance: OperationError"

1Verify instance name availability

Check if an instance with the same name was recently deleted. Instance names cannot be reused for approximately one week after deletion. If you recently deleted an instance, either wait a week or use a different name:

hcl
resource "google_sql_database_instance" "example" {
  name             = "my-new-instance-name-${formatdate("YYYYMMDDhhmmss", timestamp())}"
  database_version = "POSTGRES_15"
  region           = "us-central1"
}
2Enable required GCP APIs

Ensure the following APIs are enabled in your GCP project:
- Cloud SQL Admin API
- Compute Engine API
- Cloud Resource Manager API
- Service Networking API (if using private IP)

Enable via gcloud:

bash
gcloud services enable sqladmin.googleapis.com
gcloud services enable compute.googleapis.com
gcloud services enable cloudresourcemanager.googleapis.com
gcloud services enable servicenetworking.googleapis.com
3Verify service account permissions

Ensure your service account has the required roles to create and manage Cloud SQL instances. The minimum required role is roles/cloudsql.admin. If using private IP or VPC networking, additional roles may be needed:

bash
gcloud projects add-iam-policy-binding PROJECT_ID \
  --member="serviceAccount:SERVICE_ACCOUNT_EMAIL" \
  --role="roles/cloudsql.admin"

gcloud projects add-iam-policy-binding PROJECT_ID \
  --member="serviceAccount:SERVICE_ACCOUNT_EMAIL" \
  --role="roles/compute.networkAdmin"
4Serialize Cloud SQL operations

When creating multiple Cloud SQL instances or related resources, force serialization to prevent concurrent operation conflicts. Use depends_on to ensure resources are created sequentially:

hcl
resource "google_sql_database_instance" "instance1" {
  name = "instance-1"
  # ... configuration
}

resource "google_sql_database_instance" "instance2" {
  name = "instance-2"
  # ... configuration
  depends_on = [google_sql_database_instance.instance1]
}

resource "google_sql_database" "example" {
  name     = "example_db"
  instance = google_sql_database_instance.instance1.name
  depends_on = [google_sql_database_instance.instance1]
}

Alternatively, use parallelism flag:

bash
terraform apply -parallelism=1
5Increase Terraform timeout

Cloud SQL instance creation can take 10+ minutes. Increase the timeout to prevent premature failure:

hcl
resource "google_sql_database_instance" "example" {
  name             = "my-instance"
  database_version = "POSTGRES_15"
  region           = "us-central1"
  settings {
    tier = "db-f1-micro"
  }

  timeouts {
    create = "20m"
    delete = "20m"
    update = "20m"
  }
}
6Enable Terraform debug logging

If the issue persists, enable detailed logging to identify the exact failure point:

bash
export TF_LOG=DEBUG
terraform apply 2>&1 | tee terraform.log

# Check logs for specific error details
grep -i "operation\|error\|409\|conflict" terraform.log

Review the full error message in the logs to identify whether the issue is permissions-related, state-related, or a service outage.

Advanced notes

When using Cloud SQL with private IP addresses, you must first establish a Service Networking connection before Terraform can create instances. This requires enabling the Service Networking API and creating a private service connection. The error "Failed to create subnetwork" when using private IPs indicates this connection is missing. Also note that instance names follow global uniqueness rules within a GCP project, and deletion is not immediate—deleted instance names are reserved for approximately one week. Some users have reported needing the roles/compute.networkAdmin role in addition to roles/cloudsql.admin when using VPC networking. Finally, 409 Conflict errors typically indicate state misalignment; if this occurs, check the GCP Console to see the actual state of instances and clean up any orphaned resources before retrying Terraform operations.

Frequently asked questions

What does "Error: Error creating Cloud SQL Instance: OperationError" mean?
The OperationError in Cloud SQL indicates that the instance creation operation could not proceed due to one of several state or configuration issues. GCP Cloud SQL prevents concurrent operations on the same instance, meaning if another operation (creation, deletion, or modification) is already in progress, new operations will fail. Additionally, instance names cannot be reused for approximately one week after deletion, and certain required APIs must be enabled in your GCP project before Terraform can interact with Cloud SQL.
How do I fix "Error: Error creating Cloud SQL Instance: OperationError"?
This error occurs when Terraform fails to create a GCP Cloud SQL instance due to state conflicts, permission issues, or timeout problems. The operation failed because another operation was in progress, the instance name was recently deleted, or required APIs were not enabled.
What causes "Error: Error creating Cloud SQL Instance: OperationError"?
Another Cloud SQL operation is already in progress on the same instance (concurrent operations are not allowed) Attempting to reuse an instance name that was deleted less than a week ago Required GCP APIs (Cloud SQL Admin API, Compute Engine API, Service Networking API) are not enabled
Is the "Error: Error creating Cloud SQL Instance: OperationError" error serious?
Yes. This is a high-severity error that usually blocks work until it is fixed, but the steps above resolve it.

Official resources & further reading