diff --git a/internal/controllers/scheduler/controller.go b/internal/controllers/scheduler/controller.go
index 8fe108c..2cb4a1a 100644
--- a/internal/controllers/scheduler/controller.go
+++ b/internal/controllers/scheduler/controller.go
@@ -170,8 +170,17 @@ func (r *ClusterScheduler) handleCreateOrUpdate(ctx context.Context, req reconci
 		return rr
 	}
 
+	// if no cluster was found, check if there is an existing cluster that qualifies for the request
 	if cDef.Template.Spec.Tenancy == clustersv1alpha1.TENANCY_SHARED {
 		log.Debug("Cluster template allows sharing, checking for fitting clusters", "purpose", purpose, "tenancyCount", cDef.TenancyCount)
+		// remove all clusters with a non-zero deletion timestamp from the list of candidates
+		clusters = filters.FilterSlice(clusters, func(args ...any) bool {
+			c, ok := args[0].(*clustersv1alpha1.Cluster)
+			if !ok {
+				return false
+			}
+			return c.DeletionTimestamp.IsZero()
+		})
 		// unless the cluster template for the requested purpose allows unlimited sharing, filter out all clusters that are already at their tenancy limit
 		if cDef.TenancyCount > 0 {
 			clusters = filters.FilterSlice(clusters, func(args ...any) bool {
diff --git a/internal/controllers/scheduler/controller_test.go b/internal/controllers/scheduler/controller_test.go
index 6a773e8..cb66866 100644
--- a/internal/controllers/scheduler/controller_test.go
+++ b/internal/controllers/scheduler/controller_test.go
@@ -446,4 +446,42 @@ var _ = Describe("Scheduler", func() {
 		Expect(cluster.DeletionTimestamp).To(BeZero(), "Cluster should not be marked for deletion")
 	})
 
+	It("should not consider clusters that are in deletion for scheduling", func() {
+		// verify that the cluster is usually considered for scheduling
+		_, env := defaultTestSetup("testdata", "test-01")
+
+		c := &clustersv1alpha1.Cluster{}
+		c.SetName("shared-1")
+		c.SetNamespace("shared-twice")
+		Expect(env.Client().Get(env.Ctx, client.ObjectKeyFromObject(c), c)).To(Succeed())
+
+		cr := &clustersv1alpha1.ClusterRequest{}
+		cr.SetName("shared")
+		cr.SetNamespace("foo")
+		Expect(env.Client().Get(env.Ctx, client.ObjectKeyFromObject(cr), cr)).To(Succeed())
+		env.ShouldReconcile(testutils.RequestFromObject(cr))
+		Expect(env.Client().Get(env.Ctx, client.ObjectKeyFromObject(cr), cr)).To(Succeed())
+		Expect(cr.Status.Cluster).ToNot(BeNil())
+		Expect(cr.Status.Cluster.Name).To(Equal(c.Name))
+		Expect(cr.Status.Cluster.Namespace).To(Equal(c.Namespace))
+
+		// repeat, but with the cluster in deletion
+		_, env = defaultTestSetup("testdata", "test-01")
+
+		Expect(env.Client().Get(env.Ctx, client.ObjectKeyFromObject(c), c)).To(Succeed())
+		c.Finalizers = []string{"foo"}
+		Expect(env.Client().Update(env.Ctx, c)).To(Succeed())
+		Expect(env.Client().Delete(env.Ctx, c)).To(Succeed())
+		Expect(env.Client().Get(env.Ctx, client.ObjectKeyFromObject(c), c)).To(Succeed())
+		Expect(c.DeletionTimestamp).ToNot(BeZero(), "Cluster should be marked for deletion")
+
+		Expect(env.Client().Get(env.Ctx, client.ObjectKeyFromObject(cr), cr)).To(Succeed())
+		env.ShouldReconcile(testutils.RequestFromObject(cr))
+		Expect(env.Client().Get(env.Ctx, client.ObjectKeyFromObject(cr), cr)).To(Succeed())
+		Expect(cr.Status.Cluster).ToNot(BeNil())
+		Expect(cr.Status.Cluster.Name).ToNot(Equal(c.Name), "Cluster is in deletion and should not be considered for scheduling")
+		Expect(cr.Status.Cluster.Namespace).To(Equal(c.Namespace))
+
+	})
+
 })