Allocate dedicated host when a dedicated host doesn't exist

Author: rvanderp3

api/v1beta1/awscluster_conversion.go

@@ -66,6 +66,9 @@ func (src *AWSCluster) ConvertTo(dstRaw conversion.Hub) error {
 		dst.Status.Bastion.HostAffinity = restored.Status.Bastion.HostAffinity
 		dst.Status.Bastion.HostID = restored.Status.Bastion.HostID
 		dst.Status.Bastion.CapacityReservationPreference = restored.Status.Bastion.CapacityReservationPreference
+		if restored.Status.Bastion.DynamicHostAllocation != nil {
+			dst.Status.Bastion.DynamicHostAllocation = restored.Status.Bastion.DynamicHostAllocation
+		}
 	}
 	dst.Spec.Partition = restored.Spec.Partition
 

api/v1beta1/awsmachine_conversion.go

@@ -48,6 +48,9 @@ func (src *AWSMachine) ConvertTo(dstRaw conversion.Hub) error {
 	dst.Spec.HostAffinity = restored.Spec.HostAffinity
 	dst.Spec.CapacityReservationPreference = restored.Spec.CapacityReservationPreference
 	dst.Spec.NetworkInterfaceType = restored.Spec.NetworkInterfaceType
+	if restored.Spec.DynamicHostAllocation != nil {
+		dst.Spec.DynamicHostAllocation = restored.Spec.DynamicHostAllocation
+	}
 	if restored.Spec.ElasticIPPool != nil {
 		if dst.Spec.ElasticIPPool == nil {
 			dst.Spec.ElasticIPPool = &infrav1.ElasticIPPool{}
@@ -60,6 +63,11 @@ func (src *AWSMachine) ConvertTo(dstRaw conversion.Hub) error {
 		}
 	}
 
+	dst.Status.DedicatedHostID = restored.Status.DedicatedHostID
+	dst.Status.HostReleaseAttempts = restored.Status.HostReleaseAttempts
+	dst.Status.LastHostReleaseAttempt = restored.Status.LastHostReleaseAttempt
+	dst.Status.HostReleaseFailedReason = restored.Status.HostReleaseFailedReason
+
 	return nil
 }
 
@@ -115,6 +123,9 @@ func (r *AWSMachineTemplate) ConvertTo(dstRaw conversion.Hub) error {
 	dst.Spec.Template.Spec.HostAffinity = restored.Spec.Template.Spec.HostAffinity
 	dst.Spec.Template.Spec.CapacityReservationPreference = restored.Spec.Template.Spec.CapacityReservationPreference
 	dst.Spec.Template.Spec.NetworkInterfaceType = restored.Spec.Template.Spec.NetworkInterfaceType
+	if restored.Spec.Template.Spec.DynamicHostAllocation != nil {
+		dst.Spec.Template.Spec.DynamicHostAllocation = restored.Spec.Template.Spec.DynamicHostAllocation
+	}
 	if restored.Spec.Template.Spec.ElasticIPPool != nil {
 		if dst.Spec.Template.Spec.ElasticIPPool == nil {
 			dst.Spec.Template.Spec.ElasticIPPool = &infrav1.ElasticIPPool{}

api/v1beta1/conversion.go

@@ -103,3 +103,8 @@ func Convert_v1beta2_S3Bucket_To_v1beta1_S3Bucket(in *v1beta2.S3Bucket, out *S3B
 func Convert_v1beta2_Ignition_To_v1beta1_Ignition(in *v1beta2.Ignition, out *Ignition, s conversion.Scope) error {
 	return autoConvert_v1beta2_Ignition_To_v1beta1_Ignition(in, out, s)
 }
+
+func Convert_v1beta2_AWSMachineStatus_To_v1beta1_AWSMachineStatus(in *v1beta2.AWSMachineStatus, out *AWSMachineStatus, s conversion.Scope) error {
+	// Note: DedicatedHostID is not present in v1beta1, so it will be dropped during conversion
+	return autoConvert_v1beta2_AWSMachineStatus_To_v1beta1_AWSMachineStatus(in, out, s)
+}

api/v1beta1/zz_generated.conversion.go

@@ -213,6 +213,10 @@ type AWSMachineSpec struct {
 	PlacementGroupPartition int64 `json:"placementGroupPartition,omitempty"`
 
 	// Tenancy indicates if instance should run on shared or single-tenant hardware.
+	// When Tenancy=host, AWS will attempt to find a suitable host from:
+	// - Preexisting allocated hosts that have auto-placement enabled
+	// - A specific host ID, if configured
+	// - Allocating a new dedicated host if DynamicHostAllocation is configured
 	// +optional
 	// +kubebuilder:validation:Enum:=default;dedicated;host
 	Tenancy string `json:"tenancy,omitempty"`
@@ -235,17 +239,27 @@ type AWSMachineSpec struct {
 	MarketType MarketType `json:"marketType,omitempty"`
 
 	// HostID specifies the Dedicated Host on which the instance must be started.
+	// This field is mutually exclusive with DynamicHostAllocation.
+	// +kubebuilder:validation:Pattern=`^h-[0-9a-f]{17}$`
 	// +optional
 	HostID *string `json:"hostID,omitempty"`
 
 	// HostAffinity specifies the dedicated host affinity setting for the instance.
-	// When hostAffinity is set to host, an instance started onto a specific host always restarts on the same host if stopped.
-	// When hostAffinity is set to default, and you stop and restart the instance, it can be restarted on any available host.
+	// When HostAffinity is set to host, an instance started onto a specific host always restarts on the same host if stopped.
+	// When HostAffinity is set to default, and you stop and restart the instance, it can be restarted on any available host.
 	// When HostAffinity is defined, HostID is required.
 	// +optional
 	// +kubebuilder:validation:Enum:=default;host
+	// +kubebuilder:default=default
 	HostAffinity *string `json:"hostAffinity,omitempty"`
 
+	// DynamicHostAllocation enables automatic allocation of a single dedicated host.
+	// This field is mutually exclusive with HostID and always allocates exactly one host.
+	// Cost effectiveness of allocating a single instance on a dedicated host may vary
+	// depending on the instance type and the region.
+	// +optional
+	DynamicHostAllocation *DynamicHostAllocationSpec `json:"dynamicHostAllocation,omitempty"`
+
 	// CapacityReservationPreference specifies the preference for use of Capacity Reservations by the instance. Valid values include:
 	// "Open": The instance may make use of open Capacity Reservations that match its AZ and InstanceType
 	// "None": The instance may not make use of any Capacity Reservations. This is to conserve open reservations for desired workloads
@@ -255,6 +269,14 @@ type AWSMachineSpec struct {
 	CapacityReservationPreference CapacityReservationPreference `json:"capacityReservationPreference,omitempty"`
 }
 
+// DynamicHostAllocationSpec defines the configuration for dynamic dedicated host allocation.
+// This specification always allocates exactly one dedicated host per machine.
+type DynamicHostAllocationSpec struct {
+	// Tags to apply to the allocated dedicated host.
+	// +optional
+	Tags map[string]string `json:"tags,omitempty"`
+}
+
 // CloudInit defines options related to the bootstrapping systems where
 // CloudInit is used.
 type CloudInit struct {
@@ -432,6 +454,23 @@ type AWSMachineStatus struct {
 	// Conditions defines current service state of the AWSMachine.
 	// +optional
 	Conditions clusterv1.Conditions `json:"conditions,omitempty"`
+
+	// DedicatedHostID tracks the dynamically allocated dedicated host ID.
+	// This field is populated when DynamicHostAllocation is used.
+	// +optional
+	DedicatedHostID *string `json:"dedicatedHostID,omitempty"`
+
+	// HostReleaseAttempts tracks the number of attempts to release the dedicated host.
+	// +optional
+	HostReleaseAttempts *int32 `json:"hostReleaseAttempts,omitempty"`
+
+	// LastHostReleaseAttempt tracks the timestamp of the last attempt to release the dedicated host.
+	// +optional
+	LastHostReleaseAttempt *metav1.Time `json:"lastHostReleaseAttempt,omitempty"`
+
+	// HostReleaseFailedReason tracks the reason for the last host release failure.
+	// +optional
+	HostReleaseFailedReason *string `json:"hostReleaseFailedReason,omitempty"`
 }
 
 // +kubebuilder:object:root=true

api/v1beta2/awsmachine_types.go

@@ -75,11 +75,11 @@ func (*awsMachineWebhook) ValidateCreate(_ context.Context, obj runtime.Object)
 	allErrs = append(allErrs, r.validateNonRootVolumes()...)
 	allErrs = append(allErrs, r.validateSSHKeyName()...)
 	allErrs = append(allErrs, r.validateAdditionalSecurityGroups()...)
-	allErrs = append(allErrs, r.validateHostAffinity()...)
 	allErrs = append(allErrs, r.Spec.AdditionalTags.Validate()...)
 	allErrs = append(allErrs, r.validateNetworkElasticIPPool()...)
 	allErrs = append(allErrs, r.validateInstanceMarketType()...)
 	allErrs = append(allErrs, r.validateCapacityReservation()...)
+	allErrs = append(allErrs, r.validateHostAllocation()...)
 
 	return nil, aggregateObjErrors(r.GroupVersionKind().GroupKind(), r.Name, allErrs)
 }
@@ -109,7 +109,7 @@ func (*awsMachineWebhook) ValidateUpdate(ctx context.Context, oldObj, newObj run
 	allErrs = append(allErrs, r.validateCloudInitSecret()...)
 	allErrs = append(allErrs, r.validateAdditionalSecurityGroups()...)
 	allErrs = append(allErrs, r.Spec.AdditionalTags.Validate()...)
-	allErrs = append(allErrs, r.validateHostAffinity()...)
+	allErrs = append(allErrs, r.validateHostAllocation()...)
 
 	newAWSMachineSpec := newAWSMachine["spec"].(map[string]interface{})
 	oldAWSMachineSpec := oldAWSMachine["spec"].(map[string]interface{})
@@ -466,14 +466,17 @@ func (r *AWSMachine) validateAdditionalSecurityGroups() field.ErrorList {
 	return allErrs
 }
 
-func (r *AWSMachine) validateHostAffinity() field.ErrorList {
+func (r *AWSMachine) validateHostAllocation() field.ErrorList {
 	var allErrs field.ErrorList
 
-	if r.Spec.HostAffinity != nil {
-		if r.Spec.HostID == nil || len(*r.Spec.HostID) == 0 {
-			allErrs = append(allErrs, field.Required(field.NewPath("spec.hostID"), "hostID must be set when hostAffinity is configured"))
-		}
+	// Check if both hostID and dynamicHostAllocation are specified
+	hasHostID := r.Spec.HostID != nil && len(*r.Spec.HostID) > 0
+	hasDynamicHostAllocation := r.Spec.DynamicHostAllocation != nil
+
+	if hasHostID && hasDynamicHostAllocation {
+		allErrs = append(allErrs, field.Forbidden(field.NewPath("spec.hostID"), "hostID and dynamicHostAllocation are mutually exclusive"), field.Forbidden(field.NewPath("spec.dynamicHostAllocation"), "hostID and dynamicHostAllocation are mutually exclusive"))
 	}
+
 	return allErrs
 }
 

api/v1beta2/awsmachine_webhook.go

@@ -465,16 +465,6 @@ func TestAWSMachineCreate(t *testing.T) {
 			},
 			wantErr: true,
 		},
-		{
-			name: "configure host affinity without Host ID",
-			machine: &AWSMachine{
-				Spec: AWSMachineSpec{
-					InstanceType: "test",
-					HostAffinity: ptr.To("default"),
-				},
-			},
-			wantErr: true,
-		},
 		{
 			name: "create with valid BYOIPv4",
 			machine: &AWSMachine{
@@ -543,6 +533,45 @@ func TestAWSMachineCreate(t *testing.T) {
 			},
 			wantErr: true,
 		},
+		{
+			name: "hostID and dynamicHostAllocation are mutually exclusive",
+			machine: &AWSMachine{
+				Spec: AWSMachineSpec{
+					InstanceType: "test",
+					HostID:       aws.String("h-1234567890abcdef0"),
+					DynamicHostAllocation: &DynamicHostAllocationSpec{
+						Tags: map[string]string{
+							"Environment": "test",
+						},
+					},
+				},
+			},
+			wantErr: true,
+		},
+		{
+			name: "hostID alone is valid",
+			machine: &AWSMachine{
+				Spec: AWSMachineSpec{
+					InstanceType: "test",
+					HostID:       aws.String("h-1234567890abcdef0"),
+				},
+			},
+			wantErr: false,
+		},
+		{
+			name: "dynamicHostAllocation alone is valid",
+			machine: &AWSMachine{
+				Spec: AWSMachineSpec{
+					InstanceType: "test",
+					DynamicHostAllocation: &DynamicHostAllocationSpec{
+						Tags: map[string]string{
+							"Environment": "test",
+						},
+					},
+				},
+			},
+			wantErr: false,
+		},
 	}
 	for _, tt := range tests {
 		t.Run(tt.name, func(t *testing.T) {

api/v1beta2/awsmachine_webhook_test.go

@@ -172,6 +172,22 @@ func (r *AWSMachineTemplate) validateIgnitionAndCloudInit() field.ErrorList {
 
 	return allErrs
 }
+func (r *AWSMachineTemplate) validateHostAllocation() field.ErrorList {
+	var allErrs field.ErrorList
+
+	spec := r.Spec.Template.Spec
+
+	// Check if both hostID and dynamicHostAllocation are specified
+	hasHostID := spec.HostID != nil && len(*spec.HostID) > 0
+	hasDynamicHostAllocation := spec.DynamicHostAllocation != nil
+
+	if hasHostID && hasDynamicHostAllocation {
+		allErrs = append(allErrs, field.Forbidden(field.NewPath("spec.template.spec.hostID"), "hostID and dynamicHostAllocation are mutually exclusive"), field.Forbidden(field.NewPath("spec.template.spec.dynamicHostAllocation"), "hostID and dynamicHostAllocation are mutually exclusive"))
+	}
+
+	return allErrs
+}
+
 func (r *AWSMachineTemplate) validateSSHKeyName() field.ErrorList {
 	return validateSSHKeyName(r.Spec.Template.Spec.SSHKeyName)
 }
@@ -205,6 +221,7 @@ func (r *AWSMachineTemplateWebhook) ValidateCreate(_ context.Context, raw runtim
 	allErrs = append(allErrs, obj.validateSSHKeyName()...)
 	allErrs = append(allErrs, obj.validateAdditionalSecurityGroups()...)
 	allErrs = append(allErrs, obj.Spec.Template.Spec.AdditionalTags.Validate()...)
+	allErrs = append(allErrs, obj.validateHostAllocation()...)
 
 	return nil, aggregateObjErrors(obj.GroupVersionKind().GroupKind(), obj.Name, allErrs)
 }

api/v1beta2/awsmachinetemplate_webhook.go

@@ -80,6 +80,26 @@ func TestAWSMachineTemplateValidateCreate(t *testing.T) {
 			},
 			wantError: false,
 		},
+		{
+			name: "hostID and dynamicHostAllocation are mutually exclusive",
+			inputTemplate: &AWSMachineTemplate{
+				ObjectMeta: metav1.ObjectMeta{},
+				Spec: AWSMachineTemplateSpec{
+					Template: AWSMachineTemplateResource{
+						Spec: AWSMachineSpec{
+							InstanceType: "test",
+							HostID:       aws.String("h-1234567890abcdef0"),
+							DynamicHostAllocation: &DynamicHostAllocationSpec{
+								Tags: map[string]string{
+									"Environment": "test",
+								},
+							},
+						},
+					},
+				},
+			},
+			wantError: true,
+		},
 	}
 	for _, tt := range tests {
 		t.Run(tt.name, func(t *testing.T) {

api/v1beta2/awsmachinetemplate_webhook_test.go

@@ -146,6 +146,10 @@ const (
 	// InstanceReadyCondition reports on current status of the EC2 instance. Ready indicates the instance is in a Running state.
 	InstanceReadyCondition clusterv1.ConditionType = "InstanceReady"
 
+	// DedicatedHostReleaseCondition reports on the status of dedicated host release operations.
+	// This condition tracks whether the dedicated host has been successfully released or if there are failures.
+	DedicatedHostReleaseCondition clusterv1.ConditionType = "DedicatedHostRelease"
+
 	// InstanceNotFoundReason used when the instance couldn't be retrieved.
 	InstanceNotFoundReason = "InstanceNotFound"
 	// InstanceTerminatedReason instance is in a terminated state.
@@ -191,4 +195,13 @@ const (
 
 	// S3BucketFailedReason is used when any errors occur during reconciliation of an S3 bucket.
 	S3BucketFailedReason = "S3BucketCreationFailed"
+
+	// DedicatedHostReleaseSucceededReason used when the dedicated host is successfully released.
+	DedicatedHostReleaseSucceededReason = "DedicatedHostReleaseSucceeded"
+
+	// DedicatedHostReleaseFailedReason used when the dedicated host release fails.
+	DedicatedHostReleaseFailedReason = "DedicatedHostReleaseFailed"
+
+	// DedicatedHostReleaseRetryingReason used when the dedicated host release is being retried.
+	DedicatedHostReleaseRetryingReason = "DedicatedHostReleaseRetrying"
 )