Implement placement shim resource provider apis

cmd/shim/main.go

@@ -22,6 +22,7 @@ import (
 	"k8s.io/apimachinery/pkg/runtime"
 	utilruntime "k8s.io/apimachinery/pkg/util/runtime"
 	clientgoscheme "k8s.io/client-go/kubernetes/scheme"
+	"k8s.io/client-go/rest"
 	ctrl "sigs.k8s.io/controller-runtime"
 	"sigs.k8s.io/controller-runtime/pkg/certwatcher"
 	"sigs.k8s.io/controller-runtime/pkg/cluster"
@@ -112,11 +113,41 @@ func main() {
 	ctrl.SetLogger(zap.New(zap.UseFlagOptions(&opts)))
 
 	// Custom entrypoint for placement shim e2e tests.
+	// Spins up a minimal manager with a multicluster client so that e2e
+	// tests can access the controller-runtime cache for hypervisor lookups.
 	if runPlacementShimE2E {
-		if err := placement.RunE2E(ctx); err != nil {
+		mgrCtx, mgrCancel := context.WithCancel(ctx)
+
+		mgr, err := ctrl.NewManager(restConfig, ctrl.Options{
+			Scheme:                 scheme,
+			Metrics:                metricsserver.Options{BindAddress: "0"},
+			HealthProbeBindAddress: "",
+		})
+		if err != nil {
+			setupLog.Error(err, "unable to start e2e manager")
+			os.Exit(1)
+		}
+		multiclusterClient := setupMulticlusterClient(mgrCtx, mgr, restConfig)
+		if err := placement.IndexFields(mgrCtx, multiclusterClient); err != nil {
+			setupLog.Error(err, "unable to set up e2e field indexes")
+			os.Exit(1)
+		}
+		go func() {
+			if err := mgr.Start(mgrCtx); err != nil {
+				setupLog.Error(err, "e2e manager exited with error")
+			}
+		}()
+		if !mgr.GetCache().WaitForCacheSync(ctx) {
+			setupLog.Error(nil, "e2e cache sync failed")
+			mgrCancel()
+			os.Exit(1)
+		}
+		if err := placement.RunE2E(ctx, multiclusterClient); err != nil {
 			setupLog.Error(err, "E2E tests failed")
+			mgrCancel()
 			os.Exit(1)
 		}
+		mgrCancel()
 		os.Exit(0)
 	}
 
@@ -223,26 +254,7 @@ func main() {
 		os.Exit(1)
 	}
 
-	homeCluster, err := cluster.New(restConfig, func(o *cluster.Options) { o.Scheme = scheme })
-	if err != nil {
-		setupLog.Error(err, "unable to create home cluster")
-		os.Exit(1)
-	}
-	if err := mgr.Add(homeCluster); err != nil {
-		setupLog.Error(err, "unable to add home cluster")
-		os.Exit(1)
-	}
-	multiclusterClient := &multicluster.Client{
-		HomeCluster:     homeCluster,
-		HomeRestConfig:  restConfig,
-		HomeScheme:      scheme,
-		ResourceRouters: multicluster.DefaultResourceRouters,
-	}
-	multiclusterClientConfig := conf.GetConfigOrDie[multicluster.ClientConfig]()
-	if err := multiclusterClient.InitFromConf(ctx, mgr, multiclusterClientConfig); err != nil {
-		setupLog.Error(err, "unable to initialize multicluster client")
-		os.Exit(1)
-	}
+	multiclusterClient := setupMulticlusterClient(ctx, mgr, restConfig)
 
 	// Our custom monitoring registry can add prometheus labels to all metrics.
 	// This is useful to distinguish metrics from different deployments.
@@ -307,3 +319,27 @@ func main() {
 		os.Exit(1)
 	}
 }
+
+func setupMulticlusterClient(ctx context.Context, mgr manager.Manager, restConfig *rest.Config) *multicluster.Client {
+	homeCluster, err := cluster.New(restConfig, func(o *cluster.Options) { o.Scheme = scheme })
+	if err != nil {
+		setupLog.Error(err, "unable to create home cluster")
+		os.Exit(1)
+	}
+	if err := mgr.Add(homeCluster); err != nil {
+		setupLog.Error(err, "unable to add home cluster")
+		os.Exit(1)
+	}
+	mcl := &multicluster.Client{
+		HomeCluster:     homeCluster,
+		HomeRestConfig:  restConfig,
+		HomeScheme:      scheme,
+		ResourceRouters: multicluster.DefaultResourceRouters,
+	}
+	mclConfig := conf.GetConfigOrDie[multicluster.ClientConfig]()
+	if err := mcl.InitFromConf(ctx, mgr, mclConfig); err != nil {
+		setupLog.Error(err, "unable to initialize multicluster client")
+		os.Exit(1)
+	}
+	return mcl
+}

helm/bundles/cortex-placement-shim/values.yaml

@@ -41,6 +41,8 @@ cortex-shim:
     osProjectName:
     osUserDomainName:
     osProjectDomainName:
+    features:
+      enableResourceProviders: false
     auth:
       tokenCacheTTL: "5m"
       policies:

internal/shim/placement/field_index.go

@@ -19,9 +19,6 @@ const (
 	// of the hypervisor in OpenStack. This also corresponds to the uuid of the
 	// resource provider in OpenStack Placement.
 	idxHypervisorOpenStackId = "status.hypervisorId"
-	// idxStatusAggUUIDs is the name of the index for looking up hypervisors
-	// by the UUIDs of the aggregates they belong to.
-	idxStatusAggUUIDs = "status.aggregates[*].uuid"
 	// idxHypervisorKubernetesId is the name of the index for looking up
 	// hypervisors by their uid in Kubernetes.
 	idxHypervisorKubernetesId = "metadata.uid"
@@ -31,9 +28,9 @@ const (
 	idxHypervisorName = "metadata.name"
 )
 
-// indexFields indexes all fields that are needed by the shim to quickly
+// IndexFields indexes all fields that are needed by the shim to quickly
 // look up objects from the controller-runtime cache.
-func indexFields(ctx context.Context, mcl *multicluster.Client) error {
+func IndexFields(ctx context.Context, mcl *multicluster.Client) error {
 	log := logf.FromContext(ctx)
 	log.Info("Setting up field indexes for the multicluster client")
 	h, hl := &hv1.Hypervisor{}, &hv1.HypervisorList{}
@@ -80,22 +77,5 @@ func indexFields(ctx context.Context, mcl *multicluster.Client) error {
 	}
 	log.Info("Successfully set up index for hypervisor name")
 
-	if err := mcl.IndexField(ctx, h, hl, idxStatusAggUUIDs, func(obj client.Object) []string {
-		hv, ok := obj.(*hv1.Hypervisor)
-		if !ok {
-			log.Error(errors.New("unexpected type"), "object", obj)
-			return nil
-		}
-		var uuids []string
-		for _, agg := range hv.Status.Aggregates {
-			uuids = append(uuids, agg.UUID)
-		}
-		return uuids
-	}); err != nil {
-		log.Error(err, "failed to set up index for hypervisor aggregates")
-		return err
-	}
-	log.Info("Successfully set up index for hypervisor aggregates")
-
 	return nil
 }

internal/shim/placement/field_index_test.go

@@ -110,15 +110,14 @@ func TestIndexFields_RegistersAllIndexes(t *testing.T) {
 	cc := &captureCache{}
 	mcl := buildClient(t, cc)
 
-	if err := indexFields(context.Background(), mcl); err != nil {
-		t.Fatalf("indexFields: %v", err)
+	if err := IndexFields(context.Background(), mcl); err != nil {
+		t.Fatalf("IndexFields: %v", err)
 	}
 
 	wantFields := []string{
 		idxHypervisorOpenStackId,
 		idxHypervisorKubernetesId,
 		idxHypervisorName,
-		idxStatusAggUUIDs,
 	}
 	if len(cc.calls) != len(wantFields) {
 		t.Fatalf("got %d IndexField calls, want %d", len(cc.calls), len(wantFields))
@@ -139,7 +138,7 @@ func TestIndexFields_PropagatesError(t *testing.T) {
 	cc := &captureCache{err: wantErr}
 	mcl := buildClient(t, cc)
 
-	err := indexFields(context.Background(), mcl)
+	err := IndexFields(context.Background(), mcl)
 	if err == nil {
 		t.Fatal("expected error, got nil")
 	}
@@ -151,8 +150,8 @@ func TestIndexFields_PropagatesError(t *testing.T) {
 func TestExtractor_HypervisorOpenStackId(t *testing.T) {
 	cc := &captureCache{}
 	mcl := buildClient(t, cc)
-	if err := indexFields(context.Background(), mcl); err != nil {
-		t.Fatalf("indexFields: %v", err)
+	if err := IndexFields(context.Background(), mcl); err != nil {
+		t.Fatalf("IndexFields: %v", err)
 	}
 	fn := extractorByField(t, cc.calls, idxHypervisorOpenStackId)
 
@@ -192,8 +191,8 @@ func TestExtractor_HypervisorOpenStackId(t *testing.T) {
 func TestExtractor_HypervisorKubernetesId(t *testing.T) {
 	cc := &captureCache{}
 	mcl := buildClient(t, cc)
-	if err := indexFields(context.Background(), mcl); err != nil {
-		t.Fatalf("indexFields: %v", err)
+	if err := IndexFields(context.Background(), mcl); err != nil {
+		t.Fatalf("IndexFields: %v", err)
 	}
 	fn := extractorByField(t, cc.calls, idxHypervisorKubernetesId)
 
@@ -233,8 +232,8 @@ func TestExtractor_HypervisorKubernetesId(t *testing.T) {
 func TestExtractor_HypervisorName(t *testing.T) {
 	cc := &captureCache{}
 	mcl := buildClient(t, cc)
-	if err := indexFields(context.Background(), mcl); err != nil {
-		t.Fatalf("indexFields: %v", err)
+	if err := IndexFields(context.Background(), mcl); err != nil {
+		t.Fatalf("IndexFields: %v", err)
 	}
 	fn := extractorByField(t, cc.calls, idxHypervisorName)
 
@@ -271,63 +270,6 @@ func TestExtractor_HypervisorName(t *testing.T) {
 	}
 }
 
-func TestExtractor_StatusAggUUIDs(t *testing.T) {
-	cc := &captureCache{}
-	mcl := buildClient(t, cc)
-	if err := indexFields(context.Background(), mcl); err != nil {
-		t.Fatalf("indexFields: %v", err)
-	}
-	fn := extractorByField(t, cc.calls, idxStatusAggUUIDs)
-
-	tests := []struct {
-		name string
-		obj  client.Object
-		want []string
-	}{
-		{
-			name: "multiple aggregates",
-			obj: &hv1.Hypervisor{
-				Status: hv1.HypervisorStatus{
-					Aggregates: []hv1.Aggregate{
-						{Name: "agg-a", UUID: "uuid-a"},
-						{Name: "agg-b", UUID: "uuid-b"},
-					},
-				},
-			},
-			want: []string{"uuid-a", "uuid-b"},
-		},
-		{
-			name: "single aggregate",
-			obj: &hv1.Hypervisor{
-				Status: hv1.HypervisorStatus{
-					Aggregates: []hv1.Aggregate{
-						{Name: "agg-x", UUID: "uuid-x"},
-					},
-				},
-			},
-			want: []string{"uuid-x"},
-		},
-		{
-			name: "no aggregates",
-			obj:  &hv1.Hypervisor{},
-			want: nil,
-		},
-		{
-			name: "wrong type",
-			obj:  &corev1.ConfigMap{},
-			want: nil,
-		},
-	}
-	for _, tt := range tests {
-		t.Run(tt.name, func(t *testing.T) {
-			got := fn(tt.obj)
-			if !strSliceEqual(got, tt.want) {
-				t.Errorf("got %v, want %v", got, tt.want)
-			}
-		})
-	}
-}
-
 func strSliceEqual(a, b []string) bool {
 	if len(a) == 0 && len(b) == 0 {
 		return true

internal/shim/placement/handle_allocation_candidates_e2e.go

@@ -12,6 +12,7 @@ import (
 	"net/http"
 
 	"github.com/cobaltcore-dev/cortex/pkg/conf"
+	"sigs.k8s.io/controller-runtime/pkg/client"
 	logf "sigs.k8s.io/controller-runtime/pkg/log"
 )
 
@@ -25,7 +26,7 @@ import (
 //     - allocation_requests is non-empty (at least one candidate found).
 //     - provider_summaries contains the test RP's UUID.
 //  4. Cleanup: DELETE the test RP and custom resource class.
-func e2eTestAllocationCandidates(ctx context.Context) error {
+func e2eTestAllocationCandidates(ctx context.Context, _ client.Client) error {
 	log := logf.FromContext(ctx)
 	log.Info("Running allocation candidates endpoint e2e test")
 	config, err := conf.GetConfig[e2eRootConfig]()

internal/shim/placement/handle_allocations_e2e.go

@@ -11,6 +11,7 @@ import (
 	"net/http"
 
 	"github.com/cobaltcore-dev/cortex/pkg/conf"
+	"sigs.k8s.io/controller-runtime/pkg/client"
 	logf "sigs.k8s.io/controller-runtime/pkg/log"
 )
 
@@ -30,7 +31,7 @@ import (
 //  7. GET /allocations/{consumer2} — verify the second allocation exists.
 //  8. DELETE /allocations/{consumer} — remove allocations for both consumers.
 //  9. Cleanup: DELETE the test RP and custom resource class.
-func e2eTestAllocations(ctx context.Context) error {
+func e2eTestAllocations(ctx context.Context, _ client.Client) error {
 	log := logf.FromContext(ctx)
 	log.Info("Running allocations endpoint e2e test")
 	config, err := conf.GetConfig[e2eRootConfig]()

internal/shim/placement/handle_reshaper_e2e.go

@@ -11,6 +11,7 @@ import (
 	"net/http"
 
 	"github.com/cobaltcore-dev/cortex/pkg/conf"
+	"sigs.k8s.io/controller-runtime/pkg/client"
 	logf "sigs.k8s.io/controller-runtime/pkg/log"
 )
 
@@ -30,7 +31,7 @@ import (
 //  7. Verify allocation: GET /allocations/{consumer} — expect it to reference
 //     RP-B (not RP-A).
 //  8. Cleanup: DELETE allocation, both RPs, and custom resource class.
-func e2eTestReshaper(ctx context.Context) error {
+func e2eTestReshaper(ctx context.Context, _ client.Client) error {
 	log := logf.FromContext(ctx)
 	log.Info("Running reshaper endpoint e2e test")
 	config, err := conf.GetConfig[e2eRootConfig]()

internal/shim/placement/handle_resource_classes_e2e.go

@@ -10,6 +10,7 @@ import (
 	"net/http"
 
 	"github.com/cobaltcore-dev/cortex/pkg/conf"
+	"sigs.k8s.io/controller-runtime/pkg/client"
 	logf "sigs.k8s.io/controller-runtime/pkg/log"
 )
 
@@ -23,7 +24,7 @@ import (
 //  5. GET /resource_classes/{name} — verify the custom class now exists.
 //  6. DELETE /resource_classes/{name} — remove the custom class.
 //  7. GET /resource_classes/{name} — confirm deletion returns 404.
-func e2eTestResourceClasses(ctx context.Context) error {
+func e2eTestResourceClasses(ctx context.Context, _ client.Client) error {
 	log := logf.FromContext(ctx)
 	log.Info("Running resource classes endpoint e2e test")
 	config, err := conf.GetConfig[e2eRootConfig]()

internal/shim/placement/handle_resource_provider_aggregates_e2e.go

@@ -12,6 +12,7 @@ import (
 	"slices"
 
 	"github.com/cobaltcore-dev/cortex/pkg/conf"
+	"sigs.k8s.io/controller-runtime/pkg/client"
 	logf "sigs.k8s.io/controller-runtime/pkg/log"
 )
 
@@ -26,7 +27,7 @@ import (
 //  6. PUT /{uuid}/aggregates — clear aggregates by sending an empty list.
 //  7. GET /{uuid}/aggregates — verify aggregates are empty after clear.
 //  8. Cleanup: DELETE the test RP (also runs via deferred cleanup on failure).
-func e2eTestResourceProviderAggregates(ctx context.Context) error {
+func e2eTestResourceProviderAggregates(ctx context.Context, _ client.Client) error {
 	log := logf.FromContext(ctx)
 	log.Info("Running resource provider aggregates endpoint e2e test")
 	config, err := conf.GetConfig[e2eRootConfig]()

internal/shim/placement/handle_resource_provider_allocations_e2e.go

@@ -11,6 +11,7 @@ import (
 	"net/http"
 
 	"github.com/cobaltcore-dev/cortex/pkg/conf"
+	"sigs.k8s.io/controller-runtime/pkg/client"
 	logf "sigs.k8s.io/controller-runtime/pkg/log"
 )
 
@@ -24,7 +25,7 @@ import (
 //  4. GET /resource_providers — list real providers, then GET /{uuid}/allocations
 //     on up to 3 of them to verify the endpoint works with production data.
 //  5. Cleanup: DELETE the test RP.
-func e2eTestResourceProviderAllocations(ctx context.Context) error {
+func e2eTestResourceProviderAllocations(ctx context.Context, _ client.Client) error {
 	log := logf.FromContext(ctx)
 	log.Info("Running resource provider allocations endpoint e2e test")
 	config, err := conf.GetConfig[e2eRootConfig]()