add advanced deployment api (#106)

Signed-off-by: mingzhou.swx <mingzhou.swx@alibaba-inc.com>

Signed-off-by: mingzhou.swx <mingzhou.swx@alibaba-inc.com>
Co-authored-by: mingzhou.swx <mingzhou.swx@alibaba-inc.com>

api/v1alpha1/deployment_types.go

@@ -0,0 +1,81 @@
+package v1alpha1
+
+import (
+	apps "k8s.io/api/apps/v1"
+	"k8s.io/apimachinery/pkg/util/intstr"
+)
+
+const (
+	// DeploymentStrategyAnnotation is annotation for deployment,
+	// which is strategy fields of Advanced Deployment.
+	DeploymentStrategyAnnotation = "rollouts.kruise.io/deployment-strategy"
+
+	// DeploymentExtraStatusAnnotation is annotation for deployment,
+	// which is extra status field of Advanced Deployment.
+	DeploymentExtraStatusAnnotation = "rollouts.kruise.io/deployment-extra-status"
+)
+
+// DeploymentStrategy is strategy field for Advanced Deployment
+type DeploymentStrategy struct {
+	// RollingStyle define the behavior of rolling for deployment.
+	RollingStyle RollingStyleType `json:"rollingStyle,omitempty"`
+	// original deployment strategy rolling update fields
+	RollingUpdate *apps.RollingUpdateDeployment `json:"rollingUpdate,omitempty"`
+	// Paused = true will block the upgrade of Pods
+	Paused bool `json:"paused,omitempty"`
+	// Partition describe how many Pods should be updated during rollout.
+	// We use this field to implement partition-style rolling update.
+	Partition intstr.IntOrString `json:"partition,omitempty"`
+}
+
+type RollingStyleType string
+
+const (
+	// PartitionRollingStyleType means rolling in batches just like CloneSet, and will NOT create any extra Deployment;
+	PartitionRollingStyleType RollingStyleType = "Partition"
+	// CanaryRollingStyleType means rolling in canary way, and will create a canary Deployment.
+	CanaryRollingStyleType RollingStyleType = "Canary"
+)
+
+// DeploymentExtraStatus is extra status field for Advanced Deployment
+type DeploymentExtraStatus struct {
+	// ObservedGeneration record the generation of deployment this status observed.
+	ObservedGeneration int64 `json:"observedGeneration,omitempty"`
+	// UpdatedReadyReplicas the number of pods that has been updated and ready.
+	UpdatedReadyReplicas int32 `json:"updatedReadyReplicas,omitempty"`
+	// ExpectedUpdatedReplicas is an absolute number calculated based on Partition
+	// and Deployment.Spec.Replicas, means how many pods are expected be updated under
+	// current strategy.
+	// This field is designed to avoid users to fall into the details of algorithm
+	// for Partition calculation.
+	ExpectedUpdatedReplicas int32 `json:"expectedUpdatedReplicas,omitempty"`
+}
+
+func SetDefaultDeploymentStrategy(strategy *DeploymentStrategy) {
+	if strategy.RollingStyle == CanaryRollingStyleType {
+		return
+	}
+	if strategy.RollingUpdate == nil {
+		strategy.RollingUpdate = &apps.RollingUpdateDeployment{}
+	}
+	if strategy.RollingUpdate.MaxUnavailable == nil {
+		// Set MaxUnavailable as 25% by default
+		maxUnavailable := intstr.FromString("25%")
+		strategy.RollingUpdate.MaxUnavailable = &maxUnavailable
+	}
+	if strategy.RollingUpdate.MaxSurge == nil {
+		// Set MaxSurge as 25% by default
+		maxSurge := intstr.FromString("25%")
+		strategy.RollingUpdate.MaxUnavailable = &maxSurge
+	}
+
+	// Cannot allow maxSurge==0 && MaxUnavailable==0, otherwise, no pod can be updated when rolling update.
+	maxSurge, _ := intstr.GetScaledValueFromIntOrPercent(strategy.RollingUpdate.MaxSurge, 100, true)
+	maxUnavailable, _ := intstr.GetScaledValueFromIntOrPercent(strategy.RollingUpdate.MaxUnavailable, 100, true)
+	if maxSurge == 0 && maxUnavailable == 0 {
+		strategy.RollingUpdate = &apps.RollingUpdateDeployment{
+			MaxSurge:       &intstr.IntOrString{Type: intstr.Int, IntVal: 0},
+			MaxUnavailable: &intstr.IntOrString{Type: intstr.Int, IntVal: 1},
+		}
+	}
+}

api/v1alpha1/rollout_types.go

@@ -41,6 +41,13 @@ const (
 	// RollbackInBatchAnnotation is set to rollout annotations.
 	// RollbackInBatchAnnotation allow use disable quick rollback, and will roll back in batch style.
 	RollbackInBatchAnnotation = "rollouts.kruise.io/rollback-in-batch"
+
+	// DeploymentRolloutStyleAnnotation define the rolling behavior for Deployment.
+	// must be "partition" or "canary":
+	// * "partition" means rolling Deployment in batches just like CloneSet, and will NOT create any extra Deployment;
+	// * "canary" means rolling in canary way, and will create a canary Deployment.
+	// Defaults to canary
+	DeploymentRolloutStyleAnnotation = "rollouts.kruise.io/deployment-rolling-style"
 )
 
 // RolloutSpec defines the desired state of Rollout

api/v1alpha1/zz_generated.deepcopy.go

@@ -22,6 +22,7 @@ limitations under the License.
 package v1alpha1
 
 import (
+	"k8s.io/api/apps/v1"
 	"k8s.io/apimachinery/pkg/runtime"
 	"k8s.io/apimachinery/pkg/util/intstr"
 	"sigs.k8s.io/gateway-api/apis/v1alpha2"
@@ -265,6 +266,42 @@ func (in *CanaryStrategy) DeepCopy() *CanaryStrategy {
 	return out
 }
 
+// DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.
+func (in *DeploymentExtraStatus) DeepCopyInto(out *DeploymentExtraStatus) {
+	*out = *in
+}
+
+// DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new DeploymentExtraStatus.
+func (in *DeploymentExtraStatus) DeepCopy() *DeploymentExtraStatus {
+	if in == nil {
+		return nil
+	}
+	out := new(DeploymentExtraStatus)
+	in.DeepCopyInto(out)
+	return out
+}
+
+// DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.
+func (in *DeploymentStrategy) DeepCopyInto(out *DeploymentStrategy) {
+	*out = *in
+	if in.RollingUpdate != nil {
+		in, out := &in.RollingUpdate, &out.RollingUpdate
+		*out = new(v1.RollingUpdateDeployment)
+		(*in).DeepCopyInto(*out)
+	}
+	out.Partition = in.Partition
+}
+
+// DeepCopy is an autogenerated deepcopy function, copying the receiver, creating a new DeploymentStrategy.
+func (in *DeploymentStrategy) DeepCopy() *DeploymentStrategy {
+	if in == nil {
+		return nil
+	}
+	out := new(DeploymentStrategy)
+	in.DeepCopyInto(out)
+	return out
+}
+
 // DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.
 func (in *GatewayTrafficRouting) DeepCopyInto(out *GatewayTrafficRouting) {
 	*out = *in

pkg/controller/deployment/controller.go

@@ -19,12 +19,10 @@ package deployment
 
 import (
 	"context"
+	"encoding/json"
 	"flag"
 	"reflect"
 
-	"github.com/openkruise/rollouts/pkg/feature"
-	clientutil "github.com/openkruise/rollouts/pkg/util/client"
-	utilfeature "github.com/openkruise/rollouts/pkg/util/feature"
 	appsv1 "k8s.io/api/apps/v1"
 	v1 "k8s.io/api/core/v1"
 	"k8s.io/apimachinery/pkg/api/errors"
@@ -44,6 +42,12 @@ import (
 	"sigs.k8s.io/controller-runtime/pkg/predicate"
 	"sigs.k8s.io/controller-runtime/pkg/reconcile"
 	"sigs.k8s.io/controller-runtime/pkg/source"
+
+	rolloutsv1alpha1 "github.com/openkruise/rollouts/api/v1alpha1"
+	deploymentutil "github.com/openkruise/rollouts/pkg/controller/deployment/util"
+	"github.com/openkruise/rollouts/pkg/feature"
+	clientutil "github.com/openkruise/rollouts/pkg/util/client"
+	utilfeature "github.com/openkruise/rollouts/pkg/util/feature"
 )
 
 func init() {
@@ -135,6 +139,9 @@ func add(mgr manager.Manager, r reconcile.Reconciler) error {
 	updateHandler := func(e event.UpdateEvent) bool {
 		oldObject := e.ObjectOld.(*appsv1.Deployment)
 		newObject := e.ObjectNew.(*appsv1.Deployment)
+		if !deploymentutil.IsUnderRolloutControl(newObject) {
+			return false
+		}
 		if oldObject.Generation != newObject.Generation || newObject.DeletionTimestamp != nil {
 			klog.V(3).Infof("Observed updated Spec for Deployment: %s/%s", newObject.Namespace, newObject.Name)
 			return true
@@ -153,7 +160,7 @@ func add(mgr manager.Manager, r reconcile.Reconciler) error {
 // Reconcile reads that state of the cluster for a Deployment object and makes changes based on the state read
 // and what is in the Deployment.Spec and Deployment.Annotations
 // Automatically generate RBAC rules to allow the Controller to read and write ReplicaSets
-func (r *ReconcileDeployment) Reconcile(_ context.Context, request reconcile.Request) (res reconcile.Result, retErr error) {
+func (r *ReconcileDeployment) Reconcile(_ context.Context, request reconcile.Request) (reconcile.Result, error) {
 	deployment := new(appsv1.Deployment)
 	err := r.Get(context.TODO(), request.NamespacedName, deployment)
 	if err != nil {
@@ -167,12 +174,12 @@ func (r *ReconcileDeployment) Reconcile(_ context.Context, request reconcile.Req
 	}
 
 	// TODO: create new controller only when deployment is under our control
-	dc, err := r.controllerFactory.NewController(deployment)
-	if err != nil {
+	dc := r.controllerFactory.NewController(deployment)
+	if dc == nil {
 		return reconcile.Result{}, nil
 	}
 
-	err = dc.syncDeployment(context.Background(), request.NamespacedName.String())
+	err = dc.syncDeployment(context.Background(), deployment)
 	return ctrl.Result{}, err
 }
 
@@ -180,7 +187,27 @@ type controllerFactory DeploymentController
 
 // NewController create a new DeploymentController
 // TODO: create new controller only when deployment is under our control
-func (f *controllerFactory) NewController(_ *appsv1.Deployment) (*DeploymentController, error) {
+func (f *controllerFactory) NewController(deployment *appsv1.Deployment) *DeploymentController {
+	if !deploymentutil.IsUnderRolloutControl(deployment) {
+		klog.Warningf("Deployment %v is not under rollout control, ignore", klog.KObj(deployment))
+		return nil
+	}
+
+	strategy := rolloutsv1alpha1.DeploymentStrategy{}
+	strategyAnno := deployment.Annotations[rolloutsv1alpha1.DeploymentStrategyAnnotation]
+	if err := json.Unmarshal([]byte(strategyAnno), &strategy); err != nil {
+		klog.Errorf("Failed to unmarshal strategy for deployment %v: %v", klog.KObj(deployment), strategyAnno)
+		return nil
+	}
+
+	// We do NOT process such deployment with canary rolling style
+	if strategy.RollingStyle == rolloutsv1alpha1.CanaryRollingStyleType {
+		return nil
+	}
+
+	marshaled, _ := json.Marshal(&strategy)
+	klog.V(4).Infof("Processing deployment %v strategy %v", klog.KObj(deployment), string(marshaled))
+
 	return &DeploymentController{
 		client:           f.client,
 		eventBroadcaster: f.eventBroadcaster,
@@ -188,5 +215,6 @@ func (f *controllerFactory) NewController(_ *appsv1.Deployment) (*DeploymentCont
 		dLister:          f.dLister,
 		rsLister:         f.rsLister,
 		podLister:        f.podLister,
-	}, nil
+		strategy:         strategy,
+	}
 }

pkg/controller/deployment/deployment_controller.go

@@ -22,20 +22,24 @@ package deployment
 
 import (
 	"context"
+	"encoding/json"
 	"fmt"
 	"reflect"
+	"strings"
 	"time"
 
 	apps "k8s.io/api/apps/v1"
 	v1 "k8s.io/api/core/v1"
-	"k8s.io/apimachinery/pkg/api/errors"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
+	"k8s.io/apimachinery/pkg/types"
 	clientset "k8s.io/client-go/kubernetes"
 	appslisters "k8s.io/client-go/listers/apps/v1"
 	corelisters "k8s.io/client-go/listers/core/v1"
-	"k8s.io/client-go/tools/cache"
 	"k8s.io/client-go/tools/record"
 	"k8s.io/klog/v2"
+
+	rolloutsv1alpha1 "github.com/openkruise/rollouts/api/v1alpha1"
+	deploymentutil "github.com/openkruise/rollouts/pkg/controller/deployment/util"
 )
 
 const (
@@ -64,6 +68,9 @@ type DeploymentController struct {
 	rsLister appslisters.ReplicaSetLister
 	// podLister can list/get pods from the shared informer's store
 	podLister corelisters.PodLister
+
+	// we will use this strategy to replace spec.strategy of deployment
+	strategy rolloutsv1alpha1.DeploymentStrategy
 }
 
 // getReplicaSetsForDeployment uses ControllerRefManager to reconcile
@@ -81,28 +88,13 @@ func (dc *DeploymentController) getReplicaSetsForDeployment(ctx context.Context,
 
 // syncDeployment will sync the deployment with the given key.
 // This function is not meant to be invoked concurrently with the same key.
-func (dc *DeploymentController) syncDeployment(ctx context.Context, key string) error {
-	namespace, name, err := cache.SplitMetaNamespaceKey(key)
-	if err != nil {
-		klog.ErrorS(err, "Failed to split meta namespace cache key", "cacheKey", key)
-		return err
-	}
-
+func (dc *DeploymentController) syncDeployment(ctx context.Context, deployment *apps.Deployment) (err error) {
 	startTime := time.Now()
-	klog.V(4).InfoS("Started syncing deployment", "deployment", klog.KRef(namespace, name), "startTime", startTime)
+	klog.V(4).InfoS("Started syncing deployment", "deployment", klog.KObj(deployment), "startTime", startTime)
 	defer func() {
-		klog.V(4).InfoS("Finished syncing deployment", "deployment", klog.KRef(namespace, name), "duration", time.Since(startTime))
+		klog.V(4).InfoS("Finished syncing deployment", "deployment", klog.KObj(deployment), "duration", time.Since(startTime))
 	}()
 
-	deployment, err := dc.dLister.Deployments(namespace).Get(name)
-	if errors.IsNotFound(err) {
-		klog.V(2).InfoS("Deployment has been deleted", "deployment", klog.KRef(namespace, name))
-		return nil
-	}
-	if err != nil {
-		return err
-	}
-
 	// Deep-copy otherwise we are mutating our cache.
 	// TODO: Deep-copy only when needed.
 	d := deployment.DeepCopy()
@@ -114,39 +106,81 @@ func (dc *DeploymentController) syncDeployment(ctx context.Context, key string)
 			d.Status.ObservedGeneration = d.Generation
 			dc.client.AppsV1().Deployments(d.Namespace).UpdateStatus(ctx, d, metav1.UpdateOptions{})
 		}
-		return nil
+		return
 	}
 
 	// List ReplicaSets owned by this Deployment, while reconciling ControllerRef
 	// through adoption/orphaning.
 	rsList, err := dc.getReplicaSetsForDeployment(ctx, d)
 	if err != nil {
-		return err
+		return
 	}
 
 	if d.DeletionTimestamp != nil {
 		return dc.syncStatusOnly(ctx, d, rsList)
 	}
 
+	defer func() {
+		err = dc.updateExtraStatus(deployment, rsList)
+	}()
+
 	// Update deployment conditions with an Unknown condition when pausing/resuming
 	// a deployment. In this way, we can be sure that we won't timeout when a user
 	// resumes a Deployment with a set progressDeadlineSeconds.
 	if err = dc.checkPausedConditions(ctx, d); err != nil {
-		return err
+		return
 	}
 
 	if d.Spec.Paused {
-		return dc.sync(ctx, d, rsList)
+		err = dc.sync(ctx, d, rsList)
+		return
 	}
 
 	scalingEvent, err := dc.isScalingEvent(ctx, d, rsList)
 	if err != nil {
-		return err
+		return
 	}
 
 	if scalingEvent {
-		return dc.sync(ctx, d, rsList)
+		err = dc.sync(ctx, d, rsList)
+		return
 	}
 
-	return dc.rolloutRolling(ctx, d, rsList)
+	err = dc.rolloutRolling(ctx, d, rsList)
+	return
+}
+
+// updateExtraStatus will update extra status for advancedStatus
+func (dc *DeploymentController) updateExtraStatus(deployment *apps.Deployment, rsList []*apps.ReplicaSet) error {
+	newRS, _, err := dc.getAllReplicaSetsAndSyncRevision(context.TODO(), deployment, rsList, false)
+	if err != nil {
+		return err
+	}
+
+	updatedReadyReplicas := int32(0)
+	if newRS != nil {
+		updatedReadyReplicas = newRS.Status.ReadyReplicas
+	}
+
+	extraStatus := &rolloutsv1alpha1.DeploymentExtraStatus{
+		ObservedGeneration:      deployment.Generation,
+		UpdatedReadyReplicas:    updatedReadyReplicas,
+		ExpectedUpdatedReplicas: deploymentutil.NewRSReplicasLimit(dc.strategy.Partition, deployment),
+	}
+
+	extraStatusByte, err := json.Marshal(extraStatus)
+	if err != nil {
+		klog.Errorf("Failed to marshal extra status for Deployment %v, err: %v", klog.KObj(deployment), err)
+		return nil // no need to retry
+	}
+
+	extraStatusAnno := string(extraStatusByte)
+	if deployment.Annotations[rolloutsv1alpha1.DeploymentExtraStatusAnnotation] == extraStatusAnno {
+		return nil // no need to update
+	}
+
+	extraStatusAnno = strings.Replace(extraStatusAnno, `"`, `\"`, -1)
+	body := []byte(fmt.Sprintf(`{"metadata":{"annotations":{"%s":"%s"}}}`, rolloutsv1alpha1.DeploymentExtraStatusAnnotation, extraStatusAnno))
+	_, err = dc.client.AppsV1().Deployments(deployment.Namespace).Patch(context.TODO(), deployment.Name, types.MergePatchType, body, metav1.PatchOptions{})
+	return err
 }

pkg/controller/deployment/util/deployment_util.go

@@ -34,6 +34,8 @@ import (
 	intstrutil "k8s.io/apimachinery/pkg/util/intstr"
 	"k8s.io/klog/v2"
 	"k8s.io/utils/integer"
+
+	"github.com/openkruise/rollouts/pkg/util"
 )
 
 const (
@@ -894,3 +896,25 @@ func (o ReplicaSetsBySizeNewer) Less(i, j int) bool {
 	--------------------------------- END ---------------------------------------
 	**** Copied from "k8s.io/kubernetes/pkg/controller/controller_utils.go" ****
 */
+
+// IsUnderRolloutControl return true if this deployment should be controlled by our controller.
+func IsUnderRolloutControl(deployment *apps.Deployment) bool {
+	if deployment.Annotations[util.BatchReleaseControlAnnotation] == "" {
+		return false
+	}
+	if deployment.Spec.Strategy.Type != apps.RecreateDeploymentStrategyType {
+		return false
+	}
+	return deployment.Spec.Paused
+}
+
+// NewRSReplicasLimit return a limited replicas of new RS calculated via partition.
+func NewRSReplicasLimit(partition intstrutil.IntOrString, deployment *apps.Deployment) int32 {
+	replicas := int(*deployment.Spec.Replicas)
+	replicaLimit, _ := intstrutil.GetScaledValueFromIntOrPercent(&partition, replicas, true)
+	replicaLimit = integer.IntMax(integer.IntMin(replicaLimit, replicas), 0)
+	if replicas > 1 && partition.Type == intstrutil.String && partition.String() != "100%" {
+		replicaLimit = integer.IntMin(replicaLimit, replicas-1)
+	}
+	return int32(replicaLimit)
+}

pkg/controller/deployment/util/deployment_util_test.go

@@ -31,6 +31,8 @@ import (
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 	"k8s.io/apimachinery/pkg/types"
 	"k8s.io/apimachinery/pkg/util/intstr"
+	"k8s.io/utils/integer"
+	"k8s.io/utils/pointer"
 )
 
 func newDControllerRef(d *apps.Deployment) *metav1.OwnerReference {
@@ -1167,3 +1169,30 @@ func TestReplicasAnnotationsNeedUpdate(t *testing.T) {
 		})
 	}
 }
+
+func TestNewRSReplicasLimit(t *testing.T) {
+	for partitionInt := 0; partitionInt < 1000; partitionInt++ {
+		partition := intstr.FromInt(partitionInt)
+		deployment := apps.Deployment{Spec: apps.DeploymentSpec{Replicas: pointer.Int32(100)}}
+		result := NewRSReplicasLimit(partition, &deployment)
+		expected := integer.Int32Min(int32(partitionInt), 100)
+		if result != expected {
+			t.Errorf("case[1]: Expected %v, Got: %v", expected, result)
+		}
+	}
+
+	for replicas := 0; replicas < 1000; replicas++ {
+		for partitionPercent := 0; partitionPercent <= 100; partitionPercent++ {
+			partition := intstr.FromString(fmt.Sprintf("%d%%", partitionPercent))
+			deployment := apps.Deployment{Spec: apps.DeploymentSpec{Replicas: pointer.Int32(int32(replicas))}}
+			result := NewRSReplicasLimit(partition, &deployment)
+			expected, _ := intstr.GetScaledValueFromIntOrPercent(&partition, replicas, true)
+			if partitionPercent != 100 && replicas > 1 {
+				expected = integer.IntMin(expected, replicas-1)
+			}
+			if result != int32(expected) {
+				t.Errorf("case[2]: Expected %v, Got: %v, replicas %d, partition %d%%", expected, result, replicas, partitionPercent)
+			}
+		}
+	}
+}