chore: Adds initial TPF implementation for advanced_cluster (#2813)

* initial create operation

* initial resource implementation

* support returning nil when advanced config fields are null

* test: Add support for overriding http round tripper to allow mocking http request/responses

* minor error renaming

* remove mocked responses

* test: initial TestAdvancedCluster_configSharded requests

* chore: use same admin version as existing cluster

* mock: allow re-reading GETs

* test: support raising error when mock variables are not set

* fix: don't dump GroupID in cluster payloads

* test: update test case to match existing

* refactor: use full request.id() to keep aligned with generator format (differentiate requests with the distinct payloads)

* fix: add `computed` to adv_cluster attributes to avoid provider errors

* test: minor mock fixes

* feat: support state transitions and pause logic

* chore: test data updates

* test: Support overriding diff files by using `_manual` suffix

* style: fmt

* test: delete old mock files

* test: rename basic to replicast

* fix: Deleted should support IDLE --> DELETED

* fix: cluster name update requires replace

* test: Add tenant test case

* chore: include provider changes

* fix(move_state): Use ValueString() and do a proper conversion

* test: Support mocking test case

* refactor: use new `unit.MockTestCase` to simplify resource_test.go

* test: Introduce "live" acceptance test for tpf

* refactor: re-use existing retry strategy states

* refactor: introduce constant.error_codes

* refactor: minor cleanups

* test: Refactor MockTestCase to also run and support setting TF_ACC dynamically

* chore: increase unit test timeout to enable tpf unit tests

* minor improvements

* unit test timeout increase. Used 75s locally

* ci: support running mocked acceptance tests as part of advanced_cluster_tpf job

* ci: as of now, advancedclustertpf unit tests cannot run as part of normal unit test since the resource is not added in provider

* test: Skip broken move_state_tests

* minor fix

* chore: fix imports to latest SDK

* estetic change to workflow file

* refactor: Use provider_mock instead of customizing existing provider (revert all changes to client.go and provider.go

* address PR comments

* use `diags` instead of returning summary, detail

* only removeResource in read

* address immediate PR comments

* refactor: Use a higher abstraction level in update

* refactor: Return TFModel to make purpose more explicit

* refactor: Replace awaitChanges method with AwaitChanges function for improved clarity and consistency

* apply suggestion

* chore: Add step count assertion

* refactor: move out all the unit tests and mocked unit test logic

* remove: accidental provider changes

* remove last changes for unit tests

Author: EspenAlbert

internal/common/retrystrategy/retry_state.go

@@ -1,19 +1,20 @@
 package retrystrategy
 
 const (
-	RetryStrategyPendingState    = "PENDING"
-	RetryStrategyCompletedState  = "COMPLETED"
-	RetryStrategyErrorState      = "ERROR"
-	RetryStrategyPausedState     = "PAUSED"
-	RetryStrategyUpdatingState   = "UPDATING"
-	RetryStrategyDeletingState   = "DELETING"
-	RetryStrategyInitiatingState = "INITIATING"
-	RetryStrategyIdleState       = "IDLE"
-	RetryStrategyFailedState     = "FAILED"
-	RetryStrategyActiveState     = "ACTIVE"
-	RetryStrategyDeletedState    = "DELETED"
-	RetryStrategyCreatingState   = "CREATING"
-
+	RetryStrategyActiveState            = "ACTIVE"
+	RetryStrategyCompletedState         = "COMPLETED"
+	RetryStrategyCreatingState          = "CREATING"
+	RetryStrategyDeletedState           = "DELETED"
+	RetryStrategyDeletingState          = "DELETING"
+	RetryStrategyErrorState             = "ERROR"
+	RetryStrategyFailedState            = "FAILED"
+	RetryStrategyIdleState              = "IDLE"
+	RetryStrategyInitiatingState        = "INITIATING"
+	RetryStrategyPausedState            = "PAUSED"
 	RetryStrategyPendingAcceptanceState = "PENDING_ACCEPTANCE"
 	RetryStrategyPendingRecreationState = "PENDING_RECREATION"
+	RetryStrategyPendingState           = "PENDING"
+	RetryStrategyRepairingState         = "REPAIRING"
+	RetryStrategyRepeatingState         = "REPEATING"
+	RetryStrategyUpdatingState          = "UPDATING"
 )

internal/service/advancedclustertpf/model_ClusterDescription20240805.go

@@ -7,7 +7,7 @@ import (
 	"github.com/hashicorp/terraform-plugin-framework/diag"
 	"github.com/hashicorp/terraform-plugin-framework/types"
 	"github.com/mongodb/terraform-provider-mongodbatlas/internal/common/conversion"
-	"go.mongodb.org/atlas-sdk/v20241113001/admin"
+	"go.mongodb.org/atlas-sdk/v20240805005/admin"
 )
 
 func NewTFModel(ctx context.Context, input *admin.ClusterDescription20240805, timeout timeouts.Value, diags *diag.Diagnostics) *TFModel {
@@ -120,7 +120,8 @@ func NewReplicationSpecsObjType(ctx context.Context, input *[]admin.ReplicationS
 
 func NewTagsObjType(ctx context.Context, input *[]admin.ResourceTag, diags *diag.Diagnostics) types.Set {
 	if input == nil {
-		return types.SetNull(TagsObjType)
+		// API Response not consistent, even when not set in POST/PATCH `[]` is returned instead of null
+		return types.SetValueMust(TagsObjType, nil)
 	}
 	tfModels := make([]TFTagsModel, len(*input))
 	for i, item := range *input {

internal/service/advancedclustertpf/model_ClusterDescriptionProcessArgs20240805.go

@@ -7,7 +7,7 @@ import (
 	"github.com/hashicorp/terraform-plugin-framework/types"
 	"github.com/mongodb/terraform-provider-mongodbatlas/internal/common/conversion"
 	admin20240530 "go.mongodb.org/atlas-sdk/v20240530005/admin"
-	"go.mongodb.org/atlas-sdk/v20241113001/admin"
+	"go.mongodb.org/atlas-sdk/v20240805005/admin"
 )
 
 func AddAdvancedConfig(ctx context.Context, tfModel *TFModel, input *admin.ClusterDescriptionProcessArgs20240805, inputLegacy *admin20240530.ClusterDescriptionProcessArgs, diags *diag.Diagnostics) {

internal/service/advancedclustertpf/model_to_ClusterDescription20240805.go

@@ -8,7 +8,7 @@ import (
 	"github.com/hashicorp/terraform-plugin-framework/types"
 	"github.com/hashicorp/terraform-plugin-framework/types/basetypes"
 	"github.com/mongodb/terraform-provider-mongodbatlas/internal/common/conversion"
-	"go.mongodb.org/atlas-sdk/v20241113001/admin"
+	"go.mongodb.org/atlas-sdk/v20240805005/admin"
 )
 
 func NewAtlasReq(ctx context.Context, input *TFModel, diags *diag.Diagnostics) *admin.ClusterDescription20240805 {
@@ -24,7 +24,6 @@ func NewAtlasReq(ctx context.Context, input *TFModel, diags *diag.Diagnostics) *
 		ConfigServerManagementMode:       conversion.NilForUnknown(input.ConfigServerManagementMode, input.ConfigServerManagementMode.ValueStringPointer()),
 		EncryptionAtRestProvider:         conversion.NilForUnknown(input.EncryptionAtRestProvider, input.EncryptionAtRestProvider.ValueStringPointer()),
 		GlobalClusterSelfManagedSharding: conversion.NilForUnknown(input.GlobalClusterSelfManagedSharding, input.GlobalClusterSelfManagedSharding.ValueBoolPointer()),
-		GroupId:                          input.ProjectID.ValueStringPointer(),
 		Labels:                           newComponentLabel(ctx, input.Labels, diags),
 		MongoDBMajorVersion:              conversion.NilForUnknown(input.MongoDBMajorVersion, input.MongoDBMajorVersion.ValueStringPointer()),
 		Name:                             input.Name.ValueStringPointer(),

internal/service/advancedclustertpf/model_to_ClusterDescriptionProcessArgs20240805.go

@@ -7,7 +7,7 @@ import (
 	"github.com/hashicorp/terraform-plugin-framework/types"
 	"github.com/hashicorp/terraform-plugin-framework/types/basetypes"
 	"github.com/mongodb/terraform-provider-mongodbatlas/internal/common/conversion"
-	"go.mongodb.org/atlas-sdk/v20241113001/admin"
+	"go.mongodb.org/atlas-sdk/v20240805005/admin"
 )
 
 func NewAtlasReqAdvancedConfiguration(ctx context.Context, objInput *types.Object, diags *diag.Diagnostics) *admin.ClusterDescriptionProcessArgs20240805 {

internal/service/advancedclustertpf/move_state.go

@@ -14,6 +14,8 @@ import (
 	"github.com/hashicorp/terraform-plugin-framework/types"
 	"github.com/hashicorp/terraform-plugin-go/tfprotov6"
 	"github.com/hashicorp/terraform-plugin-go/tftypes"
+	"github.com/mongodb/terraform-provider-mongodbatlas/internal/common/conversion"
+	"go.mongodb.org/atlas-sdk/v20240805005/admin"
 )
 
 const (
@@ -140,7 +142,7 @@ func stateMoverTemporaryPreferred(ctx context.Context, req resource.MoveStateReq
 	if resp.Diagnostics.HasError() {
 		return
 	}
-	setMoveState(ctx, state.ProjectID.String(), state.Username.String(), resp)
+	setMoveState(ctx, state.ProjectID.ValueString(), state.Username.ValueString(), resp)
 }
 
 func stateMoverTemporaryRawState(ctx context.Context, req resource.MoveStateRequest, resp *resource.MoveStateResponse) {
@@ -241,12 +243,16 @@ func setMoveState(ctx context.Context, projectID, clusterName string, resp *reso
 			}),
 	}
 	// TODO: we need to have a good state (all attributes known or null) but not need to be the final ones as Read is called after
-	tfNewModel, shouldReturn := mockedSDK(ctx, &resp.Diagnostics, timeout)
-	if shouldReturn {
+	model := NewTFModel(ctx, &admin.ClusterDescription20240805{
+		GroupId: conversion.StringPtr(projectID),
+		Name:    conversion.StringPtr(clusterName),
+	}, timeout, &resp.Diagnostics)
+	if resp.Diagnostics.HasError() {
+		return
+	}
+	AddAdvancedConfig(ctx, model, nil, nil, &resp.Diagnostics)
+	if resp.Diagnostics.HasError() {
 		return
 	}
-	// TODO: setting attributed needed by Read, confirm if ClusterID is needed
-	tfNewModel.ProjectID = types.StringValue(projectID)
-	tfNewModel.Name = types.StringValue(clusterName)
-	resp.Diagnostics.Append(resp.TargetState.Set(ctx, tfNewModel)...)
+	resp.Diagnostics.Append(resp.TargetState.Set(ctx, model)...)
 }

internal/service/advancedclustertpf/move_state_test.go

@@ -2,6 +2,7 @@ package advancedclustertpf_test
 
 import (
 	"fmt"
+	"os"
 	"regexp"
 	"testing"
 
@@ -15,6 +16,7 @@ func TestAccAdvancedCluster_move_preferred(t *testing.T) {
 		projectID   = acc.ProjectIDExecution(t)
 		clusterName = acc.RandomClusterName()
 	)
+	skipIfMoveTestsNotSet(t)
 	t.Setenv(advancedclustertpf.MoveModeEnvVarName, advancedclustertpf.MoveModeValPreferred)
 	// TODO: temporary no parallel tests so t.Setenv can be used
 	resource.Test(t, resource.TestCase{
@@ -34,11 +36,20 @@ func TestAccAdvancedCluster_move_preferred(t *testing.T) {
 	})
 }
 
+// Skip for now as these needs to either run with mock data or be refactored to use a live cluster
+func skipIfMoveTestsNotSet(t *testing.T) {
+	t.Helper()
+	if os.Getenv("TF_RUN_MOVED_TESTS") == "" {
+		t.Skip("Move tests are skipped unless TF_RUN_MOVED_TESTS is set")
+	}
+}
+
 func TestAccAdvancedCluster_move_rawstate(t *testing.T) {
 	var (
 		projectID   = acc.ProjectIDExecution(t)
 		clusterName = acc.RandomClusterName()
 	)
+	skipIfMoveTestsNotSet(t)
 	t.Setenv(advancedclustertpf.MoveModeEnvVarName, advancedclustertpf.MoveModeValRawState)
 	resource.Test(t, resource.TestCase{
 		ProtoV6ProviderFactories: acc.TestAccProviderV6Factories,
@@ -62,6 +73,7 @@ func TestAccAdvancedCluster_move_json(t *testing.T) {
 		projectID   = acc.ProjectIDExecution(t)
 		clusterName = acc.RandomClusterName()
 	)
+	skipIfMoveTestsNotSet(t)
 	t.Setenv(advancedclustertpf.MoveModeEnvVarName, advancedclustertpf.MoveModeValJSON)
 	resource.Test(t, resource.TestCase{
 		ProtoV6ProviderFactories: acc.TestAccProviderV6Factories,