Failing test(s): TestAccSqlDatabaseInstance_Edition_Downgrade

[SarahFrench]

Impacted tests

• TestAccSqlDatabaseInstance_Edition_Downgrade

Affected Resource(s)

• google_sql_database_instance

Failure rates

• 100% since 2024-09-04

Message(s)

------- Stdout: -------
=== RUN   TestAccSqlDatabaseInstance_Edition_Downgrade
=== PAUSE TestAccSqlDatabaseInstance_Edition_Downgrade
=== CONT  TestAccSqlDatabaseInstance_Edition_Downgrade
    resource_sql_database_instance_test.go:1818: Step 3/4 error: Error running apply: exit status 1
        Error: Error, failed to update instance settings for : googleapi: Error 400: Invalid request: Only ENTERPRISE PLUS edition supports data cache.., invalid
          with google_sql_database_instance.instance,
          on terraform_plugin_test.tf line 3, in resource "google_sql_database_instance" "instance":
           3: resource "google_sql_database_instance" "instance" {
--- FAIL: TestAccSqlDatabaseInstance_Edition_Downgrade (525.33s)
FAIL

------- Stdout: -------
=== RUN   TestAccSqlDatabaseInstance_Edition_Downgrade
=== PAUSE TestAccSqlDatabaseInstance_Edition_Downgrade
=== CONT  TestAccSqlDatabaseInstance_Edition_Downgrade
    resource_sql_database_instance_test.go:1792: Step 1/4 error: After applying this test step, the plan was not empty.
        stdout:
        Terraform used the selected providers to generate the following execution
        plan. Resource actions are indicated with the following symbols:
          ~ update in-place
        Terraform will perform the following actions:
          # google_sql_database_instance.instance will be updated in-place
          ~ resource "google_sql_database_instance" "instance" {
                id                             = "tf-test-enterprise-downgrade-txcfd1ya7y"
                name                           = "tf-test-enterprise-downgrade-txcfd1ya7y"
                # (18 unchanged attributes hidden)
              ~ settings {
                    # (17 unchanged attributes hidden)
                  - data_cache_config {
                      - data_cache_enabled = true -> null
                    }
                    # (3 unchanged blocks hidden)
                }
            }
        Plan: 0 to add, 1 to change, 0 to destroy.
--- FAIL: TestAccSqlDatabaseInstance_Edition_Downgrade (346.01s)
FAIL

Nightly build test history

https://hashicorp.teamcity.com/test/-2885316322015289714?currentProjectId=TerraformProviders_GoogleCloud_GOOGLE_NIGHTLYTESTS&expandTestHistoryChartSection=true

b/375381759

[melinath]

This is also failing on all PRs.

[SarahFrench]

There were no changes to the test at the start of September when the test started failing, same is true for the resource.

So looks like an API-side change?

The failing test (code) doesn't set a value for data_cache_config. Note: that field is optional + computed, hence why it's affecting the test despite not being in the config.

---

More info here: https://cloud.google.com/sql/docs/postgres/upgrade-cloud-sql-instance-to-enterprise-plus-in-place#switch-to-enterprise

Note: Cloud SQL disables the data cache automatically when you change an instance from Cloud SQL Enterprise Plus edition to Cloud SQL Enterprise edition.

[SarahFrench]

In the test's logs I can see that the PUT request to change the edition to "ENTERPRISE" includes a dataCacheConfig block, which I believe is causing the error. That field is optional + computed, hence why it's present despite not being in the config. It's presumably set to enabled by default when the resource is first created with edition=ENTERPRISE_PLUS, and then that value persists after config is changed to edition=ENTERPRISE

---[ REQUEST ]---------------------------------------
PUT /sql/v1beta4/projects/PROEJCT_ID/instances/tf-test-enterprise-downgrade-afq6ij34d9?alt=json&prettyPrint=false HTTP/1.1
Host: sqladmin.googleapis.com
User-Agent: google-api-go-client/0.5 Terraform/1.8.3 (+https://www.terraform.io) Terraform-Plugin-SDK/2.33.0 terraform-provider-google/dev6
Content-Length: 884
Content-Type: application/json
X-Goog-Api-Client: gl-go/1.21.13 gdcl/0.201.0
Accept-Encoding: gzip

{
 "instanceType": "CLOUD_SQL_INSTANCE",
 "settings": {
  "activationPolicy": "ALWAYS",
  "availabilityType": "ZONAL",
  "backupConfiguration": {
   "backupRetentionSettings": {
    "retainedBackups": 15,
    "retentionUnit": "COUNT"
   },
   "binaryLogEnabled": false,
   "enabled": false,
   "pointInTimeRecoveryEnabled": false,
   "startTime": "11:00",
   "transactionLogRetentionDays": 7
  },
  "connectorEnforcement": "NOT_REQUIRED",
  "dataCacheConfig": {
   "dataCacheEnabled": true
  },
  "dataDiskSizeGb": "10",
  "dataDiskType": "PD_SSD",
  "edition": "ENTERPRISE",
  "enableDataplexIntegration": false,
  "enableGoogleMlIntegration": false,
  "ipConfiguration": {
   "enablePrivatePathForGoogleCloudServices": false,
   "ipv4Enabled": true,
   "requireSsl": null,
   "serverCaMode": "GOOGLE_MANAGED_INTERNAL_CA",
   "sslMode": "ALLOW_UNENCRYPTED_AND_ENCRYPTED"
  },
  "locationPreference": {
   "zone": "us-east1-b"
  },
  "pricingPlan": "PER_USE",
  "settingsVersion": "6",
  "storageAutoResize": true,
  "tier": "db-custom-2-13312"
 }
}

We should include a CustomizeDiff, etc, to set a new computed value for data_cache_config and its child fields if the edition is being changed from "ENTERPRISE_PLUS" to "ENTERPRISE". This then raises the question of what the behaviour should be if a user upgrades an existing "ENTERPRISE" to "ENTERPRISE_PLUS".

[github-actions]

I'm going to lock this issue because it has been closed for 30 days ⏳. This helps our maintainers find and focus on the active issues.
If you have found a problem that seems similar to this, please open a new issue and complete the issue template so we can capture all the details necessary to investigate further.