AzureStack: handle missing avail set sku cache

The Resource SKU API for availability sets may not be available in
an Azure Stack environment. The cache is used to determine the
fault domain count. For Azure Stack, we can default to 2. Future
work could potentially set this programatically or expose the
fault domain count in the API.

Author: patrickdillon

azure/scope/machine.go

@@ -150,7 +150,8 @@ func (m *MachineScope) InitMachineCache(ctx context.Context) error {
 		}
 
 		m.cache.availabilitySetSKU, err = skuCache.Get(ctx, string(armcompute.AvailabilitySetSKUTypesAligned), resourceskus.AvailabilitySets)
-		if err != nil {
+		// Resource SKU API for availability sets may not be available in Azure Stack environments.
+		if err != nil && !strings.EqualFold(m.CloudEnvironment(), "HybridEnvironment") {
 			return errors.Wrapf(err, "failed to get availability set SKU %s in compute api", string(armcompute.AvailabilitySetSKUTypesAligned))
 		}
 	}
@@ -494,12 +495,13 @@ func (m *MachineScope) AvailabilitySetSpec() azure.ResourceSpecGetter {
 	}
 
 	spec := &availabilitysets.AvailabilitySetSpec{
-		Name:           availabilitySetName,
-		ResourceGroup:  m.NodeResourceGroup(),
-		ClusterName:    m.ClusterName(),
-		Location:       m.Location(),
-		SKU:            nil,
-		AdditionalTags: m.AdditionalTags(),
+		Name:             availabilitySetName,
+		ResourceGroup:    m.NodeResourceGroup(),
+		ClusterName:      m.ClusterName(),
+		Location:         m.Location(),
+		CloudEnvironment: m.CloudEnvironment(),
+		SKU:              nil,
+		AdditionalTags:   m.AdditionalTags(),
 	}
 
 	if m.cache != nil {

azure/services/availabilitysets/spec.go

@@ -19,24 +19,27 @@ package availabilitysets
 import (
 	"context"
 	"strconv"
+	"strings"
 
 	"github.com/Azure/azure-sdk-for-go/sdk/resourcemanager/compute/armcompute/v5"
 	"github.com/pkg/errors"
 	"k8s.io/utils/ptr"
 
 	infrav1 "sigs.k8s.io/cluster-api-provider-azure/api/v1beta1"
+	"sigs.k8s.io/cluster-api-provider-azure/azure"
 	"sigs.k8s.io/cluster-api-provider-azure/azure/converters"
 	"sigs.k8s.io/cluster-api-provider-azure/azure/services/resourceskus"
 )
 
 // AvailabilitySetSpec defines the specification for an availability set.
 type AvailabilitySetSpec struct {
-	Name           string
-	ResourceGroup  string
-	ClusterName    string
-	Location       string
-	SKU            *resourceskus.SKU
-	AdditionalTags infrav1.Tags
+	Name             string
+	ResourceGroup    string
+	ClusterName      string
+	Location         string
+	CloudEnvironment string
+	SKU              *resourceskus.SKU
+	AdditionalTags   infrav1.Tags
 }
 
 // ResourceName returns the name of the availability set.
@@ -64,20 +67,10 @@ func (s *AvailabilitySetSpec) Parameters(_ context.Context, existing interface{}
 		return nil, nil
 	}
 
-	if s.SKU == nil {
-		return nil, errors.New("unable to get required availability set SKU from machine cache")
-	}
-
-	var faultDomainCount *int32
-	faultDomainCountStr, ok := s.SKU.GetCapability(resourceskus.MaximumPlatformFaultDomainCount)
-	if !ok {
-		return nil, errors.Errorf("unable to get required availability set SKU capability %s", resourceskus.MaximumPlatformFaultDomainCount)
-	}
-	count, err := strconv.ParseInt(faultDomainCountStr, 10, 32)
+	faultDomainCount, err := getFaultDomainCount(s.SKU, s.CloudEnvironment)
 	if err != nil {
-		return nil, errors.Wrapf(err, "unable to parse availability set fault domain count")
+		return nil, err
 	}
-	faultDomainCount = ptr.To[int32](int32(count))
 
 	asParams := armcompute.AvailabilitySet{
 		SKU: &armcompute.SKU{
@@ -98,3 +91,27 @@ func (s *AvailabilitySetSpec) Parameters(_ context.Context, existing interface{}
 
 	return asParams, nil
 }
+
+func getFaultDomainCount(SKU *resourceskus.SKU, cloudEnvironment string) (*int32, error) {
+	// Azure Stack environments may not implement the resource SKU API
+	// for availability sets. Use a default value instead.
+	if strings.EqualFold(cloudEnvironment, azure.StackCloudName) {
+		return ptr.To(int32(2)), nil
+	}
+
+	if SKU == nil {
+		return nil, errors.New("unable to get required availability set SKU from machine cache")
+	}
+
+	var faultDomainCount *int32
+	faultDomainCountStr, ok := SKU.GetCapability(resourceskus.MaximumPlatformFaultDomainCount)
+	if !ok {
+		return nil, errors.Errorf("unable to get required availability set SKU capability %s", resourceskus.MaximumPlatformFaultDomainCount)
+	}
+	count, err := strconv.ParseInt(faultDomainCountStr, 10, 32)
+	if err != nil {
+		return nil, errors.Wrapf(err, "unable to parse availability set fault domain count")
+	}
+	faultDomainCount = ptr.To[int32](int32(count))
+	return faultDomainCount, nil
+}