Don't publish duplicate DNS records

Before attempting to publish a domain to a zone, check if that domain is
already being published to the same zone.

Author: rfredette

pkg/operator/controller/dns/controller.go

@@ -62,6 +62,8 @@ const (
 	kubeCloudConfigName = "kube-cloud-config"
 	// cloudCABundleKey is the key in the kube cloud config ConfigMap where the custom CA bundle is located
 	cloudCABundleKey = "ca-bundle.pem"
+	// dnsRecordIndexFieldName is the key for the DNSRecord index, used to identify conflicting domain names.
+	dnsRecordIndexFieldName = "Spec.DNSName"
 )
 
 var log = logf.Logger.WithName(controllerName)
@@ -81,6 +83,17 @@ func New(mgr manager.Manager, config Config) (runtimecontroller.Controller, erro
 	if err := c.Watch(source.Kind[client.Object](operatorCache, &iov1.DNSRecord{}, &handler.EnqueueRequestForObject{}, predicate.GenerationChangedPredicate{})); err != nil {
 		return nil, err
 	}
+	// When a DNS record is deleted, there may be a conflicting record that should be published. Add a second watch
+	// exclusively for deletes so that in addition to the normal on-delete cleanup, queue a reconcile for the next
+	// record with the same domain name (if one exists) so that it can be published.
+	if err := c.Watch(source.Kind[client.Object](operatorCache, &iov1.DNSRecord{}, handler.EnqueueRequestsFromMapFunc(reconciler.mapOnRecordDelete), predicate.Funcs{
+		CreateFunc:  func(e event.CreateEvent) bool { return false },
+		DeleteFunc:  func(e event.DeleteEvent) bool { return true },
+		UpdateFunc:  func(e event.UpdateEvent) bool { return false },
+		GenericFunc: func(e event.GenericEvent) bool { return false },
+	})); err != nil {
+		return nil, err
+	}
 	if err := c.Watch(source.Kind[client.Object](operatorCache, &configv1.DNS{}, handler.EnqueueRequestsFromMapFunc(reconciler.ToDNSRecords))); err != nil {
 		return nil, err
 	}
@@ -102,6 +115,12 @@ func New(mgr manager.Manager, config Config) (runtimecontroller.Controller, erro
 	})); err != nil {
 		return nil, err
 	}
+	if err := operatorCache.IndexField(context.Background(), &iov1.DNSRecord{}, dnsRecordIndexFieldName, func(o client.Object) []string {
+		dnsRecord := o.(*iov1.DNSRecord)
+		return []string{dnsRecord.Spec.DNSName}
+	}); err != nil {
+		return nil, err
+	}
 	return c, nil
 }
 
@@ -337,17 +356,33 @@ func (r *reconciler) publishRecordToZones(zones []configv1.DNSZone, record *iov1
 
 		var err error
 		var condition iov1.DNSZoneCondition
+		var isDomainPublished bool
 		if dnsPolicy == iov1.UnmanagedDNS {
-			log.Info("DNS record not published", "record", record.Spec)
+			log.Info("DNS record not published: DNS management policy is unmanaged", "record", record.Spec)
 			condition = iov1.DNSZoneCondition{
-				Message:            "DNS record is currently not being managed by the operator",
-				Reason:             "UnmanagedDNS",
-				Status:             string(operatorv1.ConditionUnknown),
-				Type:               iov1.DNSRecordPublishedConditionType,
-				LastTransitionTime: metav1.Now(),
+				Message: "DNS record is currently not being managed by the operator",
+				Reason:  "UnmanagedDNS",
+				Status:  string(operatorv1.ConditionUnknown),
+				Type:    iov1.DNSRecordPublishedConditionType,
 			}
 		} else if isRecordPublished {
 			condition, err = r.replacePublishedRecord(zones[i], record)
+		} else if isDomainPublished, err = domainIsAlreadyPublishedToZone(context.Background(), r.cache, record, &zones[i]); err != nil {
+			log.Error(err, "failed to validate DNS record", "record", record.Spec)
+			condition = iov1.DNSZoneCondition{
+				Message: "Pre-publish validation failed",
+				Reason:  "InternalError",
+				Status:  string(operatorv1.ConditionFalse),
+				Type:    iov1.DNSRecordPublishedConditionType,
+			}
+		} else if isDomainPublished {
+			log.Info("DNS record not published: domain name already used by another DNS record", "record", record.Spec)
+			condition = iov1.DNSZoneCondition{
+				Message: "Domain name is already in use",
+				Reason:  "DomainAlreadyInUse",
+				Status:  string(operatorv1.ConditionFalse),
+				Type:    iov1.DNSRecordPublishedConditionType,
+			}
 		} else {
 			condition, err = r.publishRecord(zones[i], record)
 		}
@@ -386,6 +421,34 @@ func recordIsAlreadyPublishedToZone(record *iov1.DNSRecord, zoneToPublish *confi
 	return false
 }
 
+// domainIsAlreadyPublishedToZone returns true if the domain name in the provided DNSRecord is already published by
+// another existing dnsRecord.
+func domainIsAlreadyPublishedToZone(ctx context.Context, cache cache.Cache, record *iov1.DNSRecord, zone *configv1.DNSZone) (bool, error) {
+	records := iov1.DNSRecordList{}
+	if err := cache.List(ctx, &records, client.MatchingFields{dnsRecordIndexFieldName: record.Spec.DNSName}); err != nil {
+		return false, err
+	}
+
+	if len(records.Items) == 0 {
+		return false, nil
+	}
+
+	for _, existingRecord := range records.Items {
+		// we only care if the domain name is published by a different record, so ignore the matching record if it
+		// already exists.
+		if record.UID == existingRecord.UID {
+			continue
+		}
+		if record.Spec.DNSName != existingRecord.Spec.DNSName {
+			continue
+		}
+		if recordIsAlreadyPublishedToZone(&existingRecord, zone) {
+			return true, nil
+		}
+	}
+	return false, nil
+}
+
 func (r *reconciler) delete(record *iov1.DNSRecord) error {
 	var errs []error
 	for i := range record.Status.Zones {
@@ -576,6 +639,44 @@ func (r *reconciler) ToDNSRecords(ctx context.Context, o client.Object) []reconc
 	return requests
 }
 
+// mapOnRecordDelete finds another DNSRecord (if any) that uses the same domain name as the deleted record, and queues a
+// reconcile for that record, so that it can be published. If multiple matching records are found, a request for the
+// oldest record is returned.
+func (r *reconciler) mapOnRecordDelete(ctx context.Context, o client.Object) []reconcile.Request {
+	deletedRecord, ok := o.(*iov1.DNSRecord)
+	if !ok {
+		log.Error(nil, "Got unexpected type of object", "expected", "DNSRecord", "actual", fmt.Sprintf("%T", o))
+		return []reconcile.Request{}
+	}
+	otherRecords := iov1.DNSRecordList{}
+	if err := r.cache.List(ctx, &otherRecords, client.MatchingFields{dnsRecordIndexFieldName: deletedRecord.Spec.DNSName}); err != nil {
+		log.Error(err, "failed to list DNS records")
+		return []reconcile.Request{}
+	}
+	if len(otherRecords.Items) == 0 {
+		// Nothing to do.
+		return []reconcile.Request{}
+	}
+	oldestExistingRecord := iov1.DNSRecord{}
+	for _, existingRecord := range otherRecords.Items {
+		// Exclude records that are marked for deletion.
+		if !existingRecord.DeletionTimestamp.IsZero() {
+			continue
+		}
+		if oldestExistingRecord.CreationTimestamp.IsZero() || existingRecord.CreationTimestamp.Before(&oldestExistingRecord.CreationTimestamp) {
+			oldestExistingRecord = existingRecord
+		}
+	}
+	return []reconcile.Request{
+		reconcile.Request{
+			NamespacedName: types.NamespacedName{
+				Name:      oldestExistingRecord.Name,
+				Namespace: oldestExistingRecord.Namespace,
+			},
+		},
+	}
+}
+
 // createDNSProvider creates a DNS manager compatible with the given cluster
 // configuration.
 func (r *reconciler) createDNSProvider(dnsConfig *configv1.DNS, platformStatus *configv1.PlatformStatus, infraStatus *configv1.InfrastructureStatus, creds *corev1.Secret, AzureWorkloadIdentityEnabled bool) (dns.Provider, error) {

pkg/operator/controller/dns/controller_test.go

@@ -9,9 +9,12 @@ import (
 	operatorv1 "github.com/openshift/api/operator/v1"
 	iov1 "github.com/openshift/api/operatoringress/v1"
 	"github.com/openshift/cluster-ingress-operator/pkg/dns"
+	testutil "github.com/openshift/cluster-ingress-operator/pkg/operator/controller/test/util"
 	corev1 "k8s.io/api/core/v1"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 	"k8s.io/apimachinery/pkg/runtime"
+	"sigs.k8s.io/controller-runtime/pkg/cache/informertest"
+	"sigs.k8s.io/controller-runtime/pkg/client"
 	"sigs.k8s.io/controller-runtime/pkg/client/fake"
 )
 
@@ -117,6 +120,7 @@ func Test_publishRecordToZones(t *testing.T) {
 			r := &reconciler{
 				// TODO To write a fake provider that can return errors and add more test cases.
 				dnsProvider: &dns.FakeProvider{},
+				cache:       buildFakeCache(t),
 			}
 
 			_, actual := r.publishRecordToZones(test.zones, record)
@@ -128,9 +132,9 @@ func Test_publishRecordToZones(t *testing.T) {
 	}
 }
 
-// TestPublishRecordToZonesMergesStatus verifies that publishRecordToZones
+// Test_publishRecordToZonesMergesStatus verifies that publishRecordToZones
 // correctly merges status updates.
-func TestPublishRecordToZonesMergesStatus(t *testing.T) {
+func Test_publishRecordToZonesMergesStatus(t *testing.T) {
 	testCases := []struct {
 		description     string
 		oldZoneStatuses []iov1.DNSZoneStatus
@@ -215,7 +219,10 @@ func TestPublishRecordToZonesMergesStatus(t *testing.T) {
 				},
 				Status: iov1.DNSRecordStatus{Zones: tc.oldZoneStatuses},
 			}
-			r := &reconciler{dnsProvider: &dns.FakeProvider{}}
+			r := &reconciler{
+				dnsProvider: &dns.FakeProvider{},
+				cache:       buildFakeCache(t),
+			}
 			zone := []configv1.DNSZone{{ID: "zone2"}}
 			oldStatuses := record.Status.DeepCopy().Zones
 			_, newStatuses := r.publishRecordToZones(zone, record)
@@ -840,3 +847,28 @@ func Test_customCABundle(t *testing.T) {
 		})
 	}
 }
+
+// buildFakeCache returns a fake cache, with the necessary schema and index function(s) to mimic the parts of the cache
+// that the dns controller interacts with.
+func buildFakeCache(t *testing.T) *testutil.FakeCache {
+	t.Helper()
+	scheme := runtime.NewScheme()
+	iov1.AddToScheme(scheme)
+	fakeClient := fake.NewClientBuilder().
+		WithScheme(scheme).
+		WithIndex(&iov1.DNSRecord{}, dnsRecordIndexFieldName, func(o client.Object) []string {
+			dnsRecord := o.(*iov1.DNSRecord)
+			return []string{dnsRecord.Spec.DNSName}
+		}).
+		Build()
+	cl := &testutil.FakeClientRecorder{
+		Client:  fakeClient,
+		T:       t,
+		Added:   []client.Object{},
+		Updated: []client.Object{},
+		Deleted: []client.Object{},
+	}
+	informer := informertest.FakeInformers{Scheme: scheme}
+	cache := testutil.FakeCache{Informers: &informer, Reader: cl}
+	return &cache
+}