Merge pull request #34019 from deads2k/rbac-12-default-bindings

Automatic merge from submit-queue

Add default cluster role bindings

Add default cluster roles bindings to rbac bootstrapping.  Also adds a case for allowing escalation when you have no authenticator.

@liggitt I expect you may need to make peace with this.

hack/make-rules/test-cmd.sh

@@ -419,6 +419,10 @@ runTests() {
 
   kubectl get "${kube_flags[@]}" --raw /version
 
+  # make sure the server was properly bootstrapped with clusterroles and bindings
+  kube::test::get_object_assert clusterroles/cluster-admin "{{.metadata.name}}" 'cluster-admin'
+  kube::test::get_object_assert clusterrolebindings/cluster-admin "{{.metadata.name}}" 'cluster-admin'
+
   ###########################
   # POD creation / deletion #
   ###########################

pkg/apis/rbac/helpers.go

@@ -20,6 +20,7 @@ import (
 	"fmt"
 	"strings"
 
+	"k8s.io/kubernetes/pkg/api"
 	"k8s.io/kubernetes/pkg/api/unversioned"
 )
 
@@ -162,3 +163,61 @@ func (r *PolicyRuleBuilder) Rule() (PolicyRule, error) {
 
 	return r.PolicyRule, nil
 }
+
+// +k8s:deepcopy-gen=false
+// ClusterRoleBindingBuilder let's us attach methods.  A no-no for API types.
+// We use it to construct bindings in code.  It's more compact than trying to write them
+// out in a literal.
+type ClusterRoleBindingBuilder struct {
+	ClusterRoleBinding ClusterRoleBinding
+}
+
+func NewClusterBinding(clusterRoleName string) *ClusterRoleBindingBuilder {
+	return &ClusterRoleBindingBuilder{
+		ClusterRoleBinding: ClusterRoleBinding{
+			ObjectMeta: api.ObjectMeta{Name: clusterRoleName},
+			RoleRef: RoleRef{
+				APIGroup: GroupName,
+				Kind:     "ClusterRole",
+				Name:     clusterRoleName,
+			},
+		},
+	}
+}
+
+func (r *ClusterRoleBindingBuilder) Groups(groups ...string) *ClusterRoleBindingBuilder {
+	for _, group := range groups {
+		r.ClusterRoleBinding.Subjects = append(r.ClusterRoleBinding.Subjects, Subject{Kind: GroupKind, Name: group})
+	}
+	return r
+}
+
+func (r *ClusterRoleBindingBuilder) Users(users ...string) *ClusterRoleBindingBuilder {
+	for _, user := range users {
+		r.ClusterRoleBinding.Subjects = append(r.ClusterRoleBinding.Subjects, Subject{Kind: UserKind, Name: user})
+	}
+	return r
+}
+
+func (r *ClusterRoleBindingBuilder) SAs(namespace string, serviceAccountNames ...string) *ClusterRoleBindingBuilder {
+	for _, saName := range serviceAccountNames {
+		r.ClusterRoleBinding.Subjects = append(r.ClusterRoleBinding.Subjects, Subject{Kind: ServiceAccountKind, Namespace: namespace, Name: saName})
+	}
+	return r
+}
+
+func (r *ClusterRoleBindingBuilder) BindingOrDie() ClusterRoleBinding {
+	ret, err := r.Binding()
+	if err != nil {
+		panic(err)
+	}
+	return ret
+}
+
+func (r *ClusterRoleBindingBuilder) Binding() (ClusterRoleBinding, error) {
+	if len(r.ClusterRoleBinding.Subjects) == 0 {
+		return ClusterRoleBinding{}, fmt.Errorf("subjects are required: %#v", r.ClusterRoleBinding)
+	}
+
+	return r.ClusterRoleBinding, nil
+}

pkg/generated/openapi/zz_generated.openapi.go

@@ -5004,6 +5004,23 @@ var OpenAPIDefinitions *common.OpenAPIDefinitions = &common.OpenAPIDefinitions{
 		Dependencies: []string{
 			"api.ObjectMeta", "rbac.RoleRef", "rbac.Subject", "unversioned.TypeMeta"},
 	},
+	"rbac.ClusterRoleBindingBuilder": {
+		Schema: spec.Schema{
+			SchemaProps: spec.SchemaProps{
+				Description: "ClusterRoleBindingBuilder let's us attach methods.  A no-no for API types. We use it to construct bindings in code.  It's more compact than trying to write them out in a literal.",
+				Properties: map[string]spec.Schema{
+					"ClusterRoleBinding": {
+						SchemaProps: spec.SchemaProps{
+							Ref: spec.MustCreateRef("#/definitions/rbac.ClusterRoleBinding"),
+						},
+					},
+				},
+				Required: []string{"ClusterRoleBinding"},
+			},
+		},
+		Dependencies: []string{
+			"rbac.ClusterRoleBinding"},
+	},
 	"rbac.ClusterRoleBindingList": {
 		Schema: spec.Schema{
 			SchemaProps: spec.SchemaProps{

pkg/registry/rbac/clusterrole/policybased/storage.go

@@ -23,7 +23,7 @@ import (
 	"k8s.io/kubernetes/pkg/api/rest"
 	"k8s.io/kubernetes/pkg/apis/rbac"
 	"k8s.io/kubernetes/pkg/apis/rbac/validation"
-	"k8s.io/kubernetes/pkg/auth/user"
+	rbacregistry "k8s.io/kubernetes/pkg/registry/rbac"
 	"k8s.io/kubernetes/pkg/runtime"
 )
 
@@ -43,18 +43,8 @@ func NewStorage(s rest.StandardStorage, ruleResolver validation.AuthorizationRul
 }
 
 func (s *Storage) Create(ctx api.Context, obj runtime.Object) (runtime.Object, error) {
-	if u, ok := api.UserFrom(ctx); ok {
-		if s.superUser != "" && u.GetName() == s.superUser {
-			return s.StandardStorage.Create(ctx, obj)
-		}
-
-		// system:masters is special because the API server uses it for privileged loopback connections
-		// therefore we know that a member of system:masters can always do anything
-		for _, group := range u.GetGroups() {
-			if group == user.SystemPrivilegedGroup {
-				return s.StandardStorage.Create(ctx, obj)
-			}
-		}
+	if rbacregistry.EscalationAllowed(ctx, s.superUser) {
+		return s.StandardStorage.Create(ctx, obj)
 	}
 
 	clusterRole := obj.(*rbac.ClusterRole)
@@ -66,10 +56,8 @@ func (s *Storage) Create(ctx api.Context, obj runtime.Object) (runtime.Object, e
 }
 
 func (s *Storage) Update(ctx api.Context, name string, obj rest.UpdatedObjectInfo) (runtime.Object, bool, error) {
-	if user, ok := api.UserFrom(ctx); ok {
-		if s.superUser != "" && user.GetName() == s.superUser {
-			return s.StandardStorage.Update(ctx, name, obj)
-		}
+	if rbacregistry.EscalationAllowed(ctx, s.superUser) {
+		return s.StandardStorage.Update(ctx, name, obj)
 	}
 
 	nonEscalatingInfo := wrapUpdatedObjectInfo(obj, func(ctx api.Context, obj runtime.Object, oldObj runtime.Object) (runtime.Object, error) {

pkg/registry/rbac/clusterrolebinding/policybased/storage.go

@@ -23,7 +23,7 @@ import (
 	"k8s.io/kubernetes/pkg/api/rest"
 	"k8s.io/kubernetes/pkg/apis/rbac"
 	"k8s.io/kubernetes/pkg/apis/rbac/validation"
-	"k8s.io/kubernetes/pkg/auth/user"
+	rbacregistry "k8s.io/kubernetes/pkg/registry/rbac"
 	"k8s.io/kubernetes/pkg/runtime"
 )
 
@@ -43,18 +43,8 @@ func NewStorage(s rest.StandardStorage, ruleResolver validation.AuthorizationRul
 }
 
 func (s *Storage) Create(ctx api.Context, obj runtime.Object) (runtime.Object, error) {
-	if u, ok := api.UserFrom(ctx); ok {
-		if s.superUser != "" && u.GetName() == s.superUser {
-			return s.StandardStorage.Create(ctx, obj)
-		}
-
-		// system:masters is special because the API server uses it for privileged loopback connections
-		// therefore we know that a member of system:masters can always do anything
-		for _, group := range u.GetGroups() {
-			if group == user.SystemPrivilegedGroup {
-				return s.StandardStorage.Create(ctx, obj)
-			}
-		}
+	if rbacregistry.EscalationAllowed(ctx, s.superUser) {
+		return s.StandardStorage.Create(ctx, obj)
 	}
 
 	clusterRoleBinding := obj.(*rbac.ClusterRoleBinding)
@@ -69,10 +59,8 @@ func (s *Storage) Create(ctx api.Context, obj runtime.Object) (runtime.Object, e
 }
 
 func (s *Storage) Update(ctx api.Context, name string, obj rest.UpdatedObjectInfo) (runtime.Object, bool, error) {
-	if user, ok := api.UserFrom(ctx); ok {
-		if s.superUser != "" && user.GetName() == s.superUser {
-			return s.StandardStorage.Update(ctx, name, obj)
-		}
+	if rbacregistry.EscalationAllowed(ctx, s.superUser) {
+		return s.StandardStorage.Update(ctx, name, obj)
 	}
 
 	nonEscalatingInfo := wrapUpdatedObjectInfo(obj, func(ctx api.Context, obj runtime.Object, oldObj runtime.Object) (runtime.Object, error) {

pkg/registry/rbac/escalation_check.go

@@ -0,0 +1,45 @@
+/*
+Copyright 2016 The Kubernetes Authors.
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.
+*/
+
+package rbac
+
+import (
+	"k8s.io/kubernetes/pkg/api"
+	"k8s.io/kubernetes/pkg/auth/user"
+)
+
+func EscalationAllowed(ctx api.Context, superUser string) bool {
+	u, ok := api.UserFrom(ctx)
+	if !ok {
+		// the only way to be without a user is to either have no authenticators by explicitly saying that's your preference
+		// or to be connecting via the insecure port, in which case this logically doesn't apply
+		return true
+	}
+
+	// check to see if this subject is allowed to escalate
+	if len(superUser) != 0 && u.GetName() == superUser {
+		return true
+	}
+	// system:masters is special because the API server uses it for privileged loopback connections
+	// therefore we know that a member of system:masters can always do anything
+	for _, group := range u.GetGroups() {
+		if group == user.SystemPrivilegedGroup {
+			return true
+		}
+	}
+
+	return false
+}

pkg/registry/rbac/rest/storage_rbac.go

@@ -131,5 +131,25 @@ func PostStartHook(hookContext genericapiserver.PostStartHookContext) error {
 		glog.Infof("Created clusterrole.%s/%s", rbac.GroupName, clusterRole.Name)
 	}
 
+	existingClusterRoleBindings, err := clientset.ClusterRoleBindings().List(api.ListOptions{})
+	if err != nil {
+		utilruntime.HandleError(fmt.Errorf("unable to initialize clusterrolebindings: %v", err))
+		return nil
+	}
+	// if clusterrolebindings already exist, then assume we don't have work to do because we've already
+	// initialized or another API server has started this task
+	if len(existingClusterRoleBindings.Items) > 0 {
+		return nil
+	}
+
+	for _, clusterRoleBinding := range append(bootstrappolicy.ClusterRoleBindings(), bootstrappolicy.ControllerRoleBindings()...) {
+		if _, err := clientset.ClusterRoleBindings().Create(&clusterRoleBinding); err != nil {
+			// don't fail on failures, try to create as many as you can
+			utilruntime.HandleError(fmt.Errorf("unable to initialize clusterrolebindings: %v", err))
+			continue
+		}
+		glog.Infof("Created clusterrolebinding.%s/%s", rbac.GroupName, clusterRoleBinding.Name)
+	}
+
 	return nil
 }

pkg/registry/rbac/role/policybased/storage.go

@@ -23,7 +23,7 @@ import (
 	"k8s.io/kubernetes/pkg/api/rest"
 	"k8s.io/kubernetes/pkg/apis/rbac"
 	"k8s.io/kubernetes/pkg/apis/rbac/validation"
-	"k8s.io/kubernetes/pkg/auth/user"
+	rbacregistry "k8s.io/kubernetes/pkg/registry/rbac"
 	"k8s.io/kubernetes/pkg/runtime"
 )
 
@@ -43,18 +43,8 @@ func NewStorage(s rest.StandardStorage, ruleResolver validation.AuthorizationRul
 }
 
 func (s *Storage) Create(ctx api.Context, obj runtime.Object) (runtime.Object, error) {
-	if u, ok := api.UserFrom(ctx); ok {
-		if s.superUser != "" && u.GetName() == s.superUser {
-			return s.StandardStorage.Create(ctx, obj)
-		}
-
-		// system:masters is special because the API server uses it for privileged loopback connections
-		// therefore we know that a member of system:masters can always do anything
-		for _, group := range u.GetGroups() {
-			if group == user.SystemPrivilegedGroup {
-				return s.StandardStorage.Create(ctx, obj)
-			}
-		}
+	if rbacregistry.EscalationAllowed(ctx, s.superUser) {
+		return s.StandardStorage.Create(ctx, obj)
 	}
 
 	role := obj.(*rbac.Role)
@@ -66,10 +56,8 @@ func (s *Storage) Create(ctx api.Context, obj runtime.Object) (runtime.Object, e
 }
 
 func (s *Storage) Update(ctx api.Context, name string, obj rest.UpdatedObjectInfo) (runtime.Object, bool, error) {
-	if user, ok := api.UserFrom(ctx); ok {
-		if s.superUser != "" && user.GetName() == s.superUser {
-			return s.StandardStorage.Update(ctx, name, obj)
-		}
+	if rbacregistry.EscalationAllowed(ctx, s.superUser) {
+		return s.StandardStorage.Update(ctx, name, obj)
 	}
 
 	nonEscalatingInfo := wrapUpdatedObjectInfo(obj, func(ctx api.Context, obj runtime.Object, oldObj runtime.Object) (runtime.Object, error) {

pkg/registry/rbac/rolebinding/policybased/storage.go

@@ -23,7 +23,7 @@ import (
 	"k8s.io/kubernetes/pkg/api/rest"
 	"k8s.io/kubernetes/pkg/apis/rbac"
 	"k8s.io/kubernetes/pkg/apis/rbac/validation"
-	"k8s.io/kubernetes/pkg/auth/user"
+	rbacregistry "k8s.io/kubernetes/pkg/registry/rbac"
 	"k8s.io/kubernetes/pkg/runtime"
 )
 
@@ -43,18 +43,8 @@ func NewStorage(s rest.StandardStorage, ruleResolver validation.AuthorizationRul
 }
 
 func (s *Storage) Create(ctx api.Context, obj runtime.Object) (runtime.Object, error) {
-	if u, ok := api.UserFrom(ctx); ok {
-		if s.superUser != "" && u.GetName() == s.superUser {
-			return s.StandardStorage.Create(ctx, obj)
-		}
-
-		// system:masters is special because the API server uses it for privileged loopback connections
-		// therefore we know that a member of system:masters can always do anything
-		for _, group := range u.GetGroups() {
-			if group == user.SystemPrivilegedGroup {
-				return s.StandardStorage.Create(ctx, obj)
-			}
-		}
+	if rbacregistry.EscalationAllowed(ctx, s.superUser) {
+		return s.StandardStorage.Create(ctx, obj)
 	}
 
 	roleBinding := obj.(*rbac.RoleBinding)
@@ -69,10 +59,8 @@ func (s *Storage) Create(ctx api.Context, obj runtime.Object) (runtime.Object, e
 }
 
 func (s *Storage) Update(ctx api.Context, name string, obj rest.UpdatedObjectInfo) (runtime.Object, bool, error) {
-	if user, ok := api.UserFrom(ctx); ok {
-		if s.superUser != "" && user.GetName() == s.superUser {
-			return s.StandardStorage.Update(ctx, name, obj)
-		}
+	if rbacregistry.EscalationAllowed(ctx, s.superUser) {
+		return s.StandardStorage.Update(ctx, name, obj)
 	}
 
 	nonEscalatingInfo := wrapUpdatedObjectInfo(obj, func(ctx api.Context, obj runtime.Object, oldObj runtime.Object) (runtime.Object, error) {

plugin/pkg/auth/authorizer/rbac/bootstrappolicy/controller_policy.go

@@ -25,14 +25,18 @@ import (
 	rbac "k8s.io/kubernetes/pkg/apis/rbac"
 )
 
+const saRolePrefix = "system:controller:"
+
 var (
 	// controllerRoles is a slice of roles used for controllers
 	controllerRoles = []rbac.ClusterRole{}
+	// controllerRoleBindings is a slice of roles used for controllers
+	controllerRoleBindings = []rbac.ClusterRoleBinding{}
 )
 
 func addControllerRole(role rbac.ClusterRole) {
-	if !strings.HasPrefix(role.Name, "system:controller:") {
-		glog.Fatalf(`role %q must start with "system:controller:"`, role.Name)
+	if !strings.HasPrefix(role.Name, saRolePrefix) {
+		glog.Fatalf(`role %q must start with %q`, role.Name, saRolePrefix)
 	}
 
 	for _, existingRole := range controllerRoles {
@@ -42,6 +46,8 @@ func addControllerRole(role rbac.ClusterRole) {
 	}
 
 	controllerRoles = append(controllerRoles, role)
+	controllerRoleBindings = append(controllerRoleBindings,
+		rbac.NewClusterBinding(role.Name).SAs("kube-system", role.Name[len(saRolePrefix):]).BindingOrDie())
 }
 
 func eventsRule() rbac.PolicyRule {
@@ -50,7 +56,7 @@ func eventsRule() rbac.PolicyRule {
 
 func init() {
 	addControllerRole(rbac.ClusterRole{
-		ObjectMeta: api.ObjectMeta{Name: "system:controller:replication-controller"},
+		ObjectMeta: api.ObjectMeta{Name: saRolePrefix + "replication-controller"},
 		Rules: []rbac.PolicyRule{
 			rbac.NewRule("get", "list", "watch", "update").Groups(legacyGroup).Resources("replicationcontrollers").RuleOrDie(),
 			rbac.NewRule("update").Groups(legacyGroup).Resources("replicationcontrollers/status").RuleOrDie(),
@@ -64,3 +70,8 @@ func init() {
 func ControllerRoles() []rbac.ClusterRole {
 	return controllerRoles
 }
+
+// ControllerRoleBindings returns the role bindings used by controllers
+func ControllerRoleBindings() []rbac.ClusterRoleBinding {
+	return controllerRoleBindings
+}

plugin/pkg/auth/authorizer/rbac/bootstrappolicy/policy.go

@@ -19,6 +19,7 @@ package bootstrappolicy
 import (
 	"k8s.io/kubernetes/pkg/api"
 	rbac "k8s.io/kubernetes/pkg/apis/rbac"
+	"k8s.io/kubernetes/pkg/auth/user"
 )
 
 var (
@@ -48,3 +49,11 @@ func ClusterRoles() []rbac.ClusterRole {
 		},
 	}
 }
+
+// ClusterRoleBindings return default rolebindings to the default roles
+func ClusterRoleBindings() []rbac.ClusterRoleBinding {
+	return []rbac.ClusterRoleBinding{
+		rbac.NewClusterBinding("cluster-admin").Groups(user.SystemPrivilegedGroup).BindingOrDie(),
+		rbac.NewClusterBinding("system:discovery").Groups(user.AllAuthenticated, user.AllUnauthenticated).BindingOrDie(),
+	}
+}