superstreamconsumer_controller.go

/*
Copyright 2022.

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 controllers

import (
	"context"
	"fmt"
	topologyv1alpha1 "github.com/rabbitmq/messaging-topology-operator/api/v1alpha1"
	topology "github.com/rabbitmq/messaging-topology-operator/api/v1beta1"
	sacv1alpha1 "github.com/rabbitmq/single-active-consumer-operator/api/v1alpha1"
	"github.com/rabbitmq/single-active-consumer-operator/internal/managedresource"
	corev1 "k8s.io/api/core/v1"
	"k8s.io/apimachinery/pkg/fields"
	"k8s.io/apimachinery/pkg/types"
	"k8s.io/client-go/tools/record"
	"k8s.io/client-go/util/retry"
	"sigs.k8s.io/controller-runtime/pkg/builder"
	"sigs.k8s.io/controller-runtime/pkg/controller/controllerutil"
	"sigs.k8s.io/controller-runtime/pkg/handler"
	"sigs.k8s.io/controller-runtime/pkg/predicate"
	"sigs.k8s.io/controller-runtime/pkg/reconcile"
	"sigs.k8s.io/controller-runtime/pkg/source"
	"time"

	"k8s.io/apimachinery/pkg/runtime"
	ctrl "sigs.k8s.io/controller-runtime"
	"sigs.k8s.io/controller-runtime/pkg/client"
	"sigs.k8s.io/controller-runtime/pkg/log"
)

const failedStatusUpdate = "failed to update object status"

// SuperStreamConsumerReconciler reconciles a SuperStreamConsumer object
type SuperStreamConsumerReconciler struct {
	client.Client
	Scheme   *runtime.Scheme
	recorder record.EventRecorder
}

//+kubebuilder:rbac:groups=rabbitmq.com,resources=superstreams,verbs=get;list;watch;create;update;patch;delete
//+kubebuilder:rbac:groups=rabbitmq.com,resources=superstreamconsumers,verbs=get;list;watch;create;update;patch;delete
//+kubebuilder:rbac:groups=rabbitmq.com,resources=superstreamconsumers/status,verbs=get;update;patch
//+kubebuilder:rbac:groups=rabbitmq.com,resources=superstreamconsumers/finalizers,verbs=update
//+kubebuilder:rbac:groups="",resources=pods,verbs=get;create;list;update;delete;patch;watch
//+kubebuilder:rbac:groups="",resources=services,verbs=get;list;watch
//+kubebuilder:rbac:groups="",resources=events,verbs=create;get;patch

// Reconcile is part of the main kubernetes reconciliation loop which aims to
// move the current state of the cluster closer to the desired state.
func (r *SuperStreamConsumerReconciler) Reconcile(ctx context.Context, req ctrl.Request) (ctrl.Result, error) {
	logger := log.FromContext(ctx)

	superStreamConsumer := &sacv1alpha1.SuperStreamConsumer{}
	if err := r.Get(ctx, req.NamespacedName, superStreamConsumer); err != nil {
		return reconcile.Result{}, client.IgnoreNotFound(err)
	}

	logger.Info("Start reconciling")

	superStreamNamespace := superStreamConsumer.Spec.SuperStreamReference.Namespace
	if superStreamNamespace == "" {
		superStreamNamespace = superStreamConsumer.Namespace
	}

	referencedSuperStream := &topologyv1alpha1.SuperStream{}
	if err := r.Get(ctx, types.NamespacedName{Name: superStreamConsumer.Spec.SuperStreamReference.Name, Namespace: superStreamNamespace}, referencedSuperStream); err != nil {
		return reconcile.Result{}, fmt.Errorf("failed to get SuperStream from reference: %w", err)
	}
	if len(referencedSuperStream.Status.Partitions) != referencedSuperStream.Spec.Partitions {
		// The object is likely being reconciled, wait until all the partitions are created
		return reconcile.Result{RequeueAfter: 1 * time.Second}, nil
	}

	existingPods, err := r.existingMatchingPods(ctx, superStreamConsumer)
	if err != nil {
		return ctrl.Result{}, err
	}
	if len(existingPods) > 0 {
		r.logExistingMatchingPods(ctx, existingPods)
	}

	var podBuilders = make(map[*corev1.Pod]*managedresource.SuperStreamConsumerPodBuilder, len(referencedSuperStream.Status.Partitions))
	for _, partition := range referencedSuperStream.Status.Partitions {
		podSpec := r.getPartitionPodSpec(referencedSuperStream, partition, superStreamConsumer)
		if podSpec == nil {
			// There may be an existing Pod for this partition. We must delete the original Pod
			// before creating the new Pod with the new PodSpec.
			existingPods, err := r.getMatchingPods(ctx, superStreamConsumer.Namespace, map[string]string{
				managedresource.AnnotationSuperStream:          referencedSuperStream.Name,
				managedresource.AnnotationSuperStreamPartition: partition,
			})
			if err != nil {
				return ctrl.Result{}, err
			}
			for _, existingPod := range existingPods {
				err := r.deleteExistingPod(ctx, &existingPod, superStreamConsumer)
				if err != nil {
					return ctrl.Result{}, err
				}
				r.recorder.Event(&existingPod, corev1.EventTypeNormal, "SuccessfulDelete", "Successfully deleted pod due to updated podSpec")
			}
			continue
		}

		builder := managedresource.SuperStreamConsumerPod(
			superStreamConsumer,
			r.Scheme,
			*podSpec,
			referencedSuperStream.Name,
			partition,
		)

		resource, err := builder.Build()
		if err != nil {
			return ctrl.Result{}, err
		}
		pod := resource.(*corev1.Pod)
		podBuilders[pod] = builder
	}

	for pod, builder := range podBuilders {
		existingPod, err := r.existingActiveConsumerPod(ctx, superStreamConsumer.Namespace, map[string]string{
			managedresource.AnnotationSuperStream:          pod.Labels[managedresource.AnnotationSuperStream],
			managedresource.AnnotationSuperStreamPartition: pod.Labels[managedresource.AnnotationSuperStreamPartition],
		})
		if err != nil {
			msg := fmt.Sprintf("FailedReconcile%s", builder.ResourceType())
			if writerErr := r.SetReconcileSuccess(ctx, superStreamConsumer, topology.NotReady(msg, superStreamConsumer.Status.Conditions)); writerErr != nil {
				logger.Error(writerErr, failedStatusUpdate, "status", superStreamConsumer.Status)
			}
			return ctrl.Result{}, err
		}
		if podSpecHasNotChanged(existingPod, pod) {
			continue
		}
		if existingPod != nil {
			// There is an existing Pod for this partition, with a different PodSpec. We must delete the original Pod
			// before creating the new Pod with the new PodSpec.
			err := r.deleteExistingPod(ctx, existingPod, superStreamConsumer)
			if err != nil {
				return ctrl.Result{}, err
			}
			r.recorder.Event(existingPod, corev1.EventTypeNormal, "SuccessfulDelete", "Successfully deleted pod due to updated podSpec")
		}

		// Create the new Pod for this partition
		err = retry.RetryOnConflict(retry.DefaultRetry, func() error {
			var apiError error
			_, apiError = controllerutil.CreateOrUpdate(ctx, r.Client, pod, func() error {
				return builder.Update(pod)
			})
			return apiError
		})
		if err != nil {
			msg := fmt.Sprintf("FailedReconcile%s", builder.ResourceType())
			if writerErr := r.SetReconcileSuccess(ctx, superStreamConsumer, topology.NotReady(msg, superStreamConsumer.Status.Conditions)); writerErr != nil {
				logger.Error(writerErr, failedStatusUpdate, "status", superStreamConsumer.Status)
			}
			return ctrl.Result{}, err
		}
	}

	if err := r.SetReconcileSuccess(ctx, superStreamConsumer, topology.Ready(superStreamConsumer.Status.Conditions)); err != nil {
		logger.Error(err, failedStatusUpdate)
	}

	logger.Info("Finished reconciling")
	return ctrl.Result{}, nil
}

func (r *SuperStreamConsumerReconciler) deleteExistingPod(ctx context.Context, existingPod *corev1.Pod, superStreamConsumer *sacv1alpha1.SuperStreamConsumer) error {
	logger := log.FromContext(ctx)
	if err := r.Delete(ctx, existingPod); err != nil {
		if writerErr := r.SetReconcileSuccess(ctx, superStreamConsumer, topology.NotReady("FailedDeleteSuperStreamConsumerPod", superStreamConsumer.Status.Conditions)); writerErr != nil {
			logger.Error(writerErr, failedStatusUpdate, "status", superStreamConsumer.Status)
		}
		return err
	}
	return nil
}

func podSpecHasNotChanged(existingPod *corev1.Pod, pod *corev1.Pod) bool {
	return existingPod != nil && existingPod.Labels[managedresource.AnnotationConsumerPodSpecHash] == pod.Labels[managedresource.AnnotationConsumerPodSpecHash]
}

func (r *SuperStreamConsumerReconciler) getPartitionPodSpec(referencedSuperStream *topologyv1alpha1.SuperStream, partition string, superStreamConsumer *sacv1alpha1.SuperStreamConsumer) *corev1.PodSpec {
	var podSpec *corev1.PodSpec = nil
	routingKey := managedresource.PartitionNameToRoutingKey(referencedSuperStream.Name, partition)
	if foundPodSpec, ok := superStreamConsumer.Spec.ConsumerPodSpec.PerRoutingKey[routingKey]; ok {
		podSpec = foundPodSpec
	} else {
		podSpec = superStreamConsumer.Spec.ConsumerPodSpec.Default
	}

	return podSpec
}

func (r *SuperStreamConsumerReconciler) SetReconcileSuccess(ctx context.Context, superStreamConsumer *sacv1alpha1.SuperStreamConsumer, condition topology.Condition) error {
	superStreamConsumer.Status.Conditions = []topology.Condition{condition}
	superStreamConsumer.Status.ObservedGeneration = superStreamConsumer.GetGeneration()
	return retry.RetryOnConflict(retry.DefaultRetry, func() error {
		return r.Status().Update(ctx, superStreamConsumer)
	})
}

func (r *SuperStreamConsumerReconciler) findConsumersForSuperStream(superStream client.Object) []reconcile.Request {
	consumerList := &sacv1alpha1.SuperStreamConsumerList{}
	listOps := &client.ListOptions{
		FieldSelector: fields.OneTermEqualSelector(".spec.superStreamReference.name", superStream.GetName()),
		Namespace:     superStream.GetNamespace(),
	}
	err := r.List(context.Background(), consumerList, listOps)
	if err != nil {
		return []reconcile.Request{}
	}

	requests := make([]reconcile.Request, len(consumerList.Items))
	for i, item := range consumerList.Items {
		requests[i] = reconcile.Request{
			NamespacedName: types.NamespacedName{
				Name:      item.GetName(),
				Namespace: item.GetNamespace(),
			},
		}
	}
	return requests
}

func (r *SuperStreamConsumerReconciler) getMatchingPods(ctx context.Context, namespace string, podLabels map[string]string) ([]corev1.Pod, error) {
	existingPodList := &corev1.PodList{}
	if err := r.Client.List(ctx, existingPodList, client.InNamespace(namespace), client.MatchingLabels(podLabels)); err != nil {
		return nil, err
	}

	return existingPodList.Items, nil
}

func (r *SuperStreamConsumerReconciler) existingActiveConsumerPod(ctx context.Context, namespace string, podLabels map[string]string) (*corev1.Pod, error) {
	existingPods, err := r.getMatchingPods(ctx, namespace, podLabels)
	if err != nil {
		return nil, fmt.Errorf("failed to lookup matching existing pods: %w", err)
	}

	if len(existingPods) == 0 {
		return nil, nil
	}

	if len(existingPods) > 1 {
		var podNames []string
		for _, pod := range existingPods {
			podNames = append(podNames, pod.Name)
		}
		return nil, fmt.Errorf("expected to find 1 matching consumer pod, but found %d: %s", len(existingPods), podNames)
	}
	return &existingPods[0], nil
}
func (r *SuperStreamConsumerReconciler) existingMatchingPods(ctx context.Context, superStreamConsumer *sacv1alpha1.SuperStreamConsumer) ([]corev1.Pod, error) {
	existingPodList := &corev1.PodList{}
	err := r.Client.List(ctx, existingPodList, client.InNamespace(superStreamConsumer.Namespace), client.MatchingLabels(map[string]string{
		managedresource.AnnotationSuperStream: superStreamConsumer.Spec.SuperStreamReference.Name,
	}))
	return existingPodList.Items, err
}
func (r *SuperStreamConsumerReconciler) logExistingMatchingPods(ctx context.Context, pods []corev1.Pod) {
	logger := log.FromContext(ctx)
	logString := "Existing pods: "
	for _, pod := range pods {
		logString += fmt.Sprintf("%s, ", pod.Name)
	}
	logger.Info(logString)
}

func (r *SuperStreamConsumerReconciler) SetupWithManager(mgr ctrl.Manager) error {
	r.recorder = mgr.GetEventRecorderFor("SuperStreamConsumer")
	if err := mgr.GetFieldIndexer().IndexField(context.Background(), &sacv1alpha1.SuperStreamConsumer{}, ".spec.superStreamReference.name", func(rawObj client.Object) []string {
		// Extract the SuperStream name from the SuperStreamConsumer Spec, if one is provided
		superStreamConsumer := rawObj.(*sacv1alpha1.SuperStreamConsumer)
		if superStreamConsumer.Spec.SuperStreamReference.Name == "" {
			return nil
		}
		return []string{superStreamConsumer.Spec.SuperStreamReference.Name}
	}); err != nil {
		return err
	}
	return ctrl.NewControllerManagedBy(mgr).
		For(&sacv1alpha1.SuperStreamConsumer{}).
		Owns(&corev1.Pod{}).
		Watches(
			&source.Kind{Type: &topologyv1alpha1.SuperStream{}},
			handler.EnqueueRequestsFromMapFunc(r.findConsumersForSuperStream),
			builder.WithPredicates(predicate.GenerationChangedPredicate{}),
		).
		Complete(r)
}