main.go

/*
Copyright 2017 Google LLC

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

    https://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 main

import (
	"context"
	"flag"
	"fmt"
	"os"
	"os/signal"
	"strings"
	"syscall"
	"time"

	"github.com/golang/glog"
	apiv1 "k8s.io/api/core/v1"
	apiextensionsclient "k8s.io/apiextensions-apiserver/pkg/client/clientset/clientset"
	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
	"k8s.io/client-go/informers"
	clientset "k8s.io/client-go/kubernetes"
	"k8s.io/client-go/kubernetes/scheme"
	_ "k8s.io/client-go/plugin/pkg/client/auth/gcp"
	"k8s.io/client-go/rest"
	"k8s.io/client-go/tools/clientcmd"
	"k8s.io/client-go/tools/leaderelection"
	"k8s.io/client-go/tools/leaderelection/resourcelock"
	"k8s.io/client-go/tools/record"
	"k8s.io/utils/clock"

	"github.com/kubeflow/spark-operator/pkg/batchscheduler"
	crclientset "github.com/kubeflow/spark-operator/pkg/client/clientset/versioned"
	crinformers "github.com/kubeflow/spark-operator/pkg/client/informers/externalversions"
	operatorConfig "github.com/kubeflow/spark-operator/pkg/config"
	"github.com/kubeflow/spark-operator/pkg/controller/scheduledsparkapplication"
	"github.com/kubeflow/spark-operator/pkg/controller/sparkapplication"
	"github.com/kubeflow/spark-operator/pkg/util"
	"github.com/kubeflow/spark-operator/pkg/webhook"
)

var (
	master                         = flag.String("master", "", "The address of the Kubernetes API server. Overrides any value in kubeconfig. Only required if out-of-cluster.")
	kubeConfig                     = flag.String("kubeConfig", "", "Path to a kube config. Only required if out-of-cluster.")
	controllerThreads              = flag.Int("controller-threads", 10, "Number of worker threads used by the SparkApplication controller.")
	resyncInterval                 = flag.Int("resync-interval", 30, "Informer resync interval in seconds.")
	namespace                      = flag.String("namespace", apiv1.NamespaceAll, "The Kubernetes namespace to manage. Will manage custom resource objects of the managed CRD types for the whole cluster if unset.")
	labelSelectorFilter            = flag.String("label-selector-filter", "", "A comma-separated list of key=value, or key labels to filter resources during watch and list based on the specified labels.")
	enableWebhook                  = flag.Bool("enable-webhook", false, "Whether to enable the mutating admission webhook for admitting and patching Spark pods.")
	webhookTimeout                 = flag.Int("webhook-timeout", 30, "Webhook Timeout in seconds before the webhook returns a timeout")
	enableResourceQuotaEnforcement = flag.Bool("enable-resource-quota-enforcement", false, "Whether to enable ResourceQuota enforcement for SparkApplication resources. Requires the webhook to be enabled.")
	ingressURLFormat               = flag.String("ingress-url-format", "", "Ingress URL format.")
	enableUIService                = flag.Bool("enable-ui-service", true, "Enable Spark service UI.")
	enableLeaderElection           = flag.Bool("leader-election", false, "Enable Spark operator leader election.")
	leaderElectionLockNamespace    = flag.String("leader-election-lock-namespace", "spark-operator", "Namespace in which to create the ConfigMap for leader election.")
	leaderElectionLockName         = flag.String("leader-election-lock-name", "spark-operator-lock", "Name of the ConfigMap for leader election.")
	leaderElectionLeaseDuration    = flag.Duration("leader-election-lease-duration", 15*time.Second, "Leader election lease duration.")
	leaderElectionRenewDeadline    = flag.Duration("leader-election-renew-deadline", 14*time.Second, "Leader election renew deadline.")
	leaderElectionRetryPeriod      = flag.Duration("leader-election-retry-period", 4*time.Second, "Leader election retry period.")
	enableBatchScheduler           = flag.Bool("enable-batch-scheduler", false, fmt.Sprintf("Enable batch schedulers for pods' scheduling, the available batch schedulers are: (%s).", strings.Join(batchscheduler.GetRegisteredNames(), ",")))
	enableMetrics                  = flag.Bool("enable-metrics", false, "Whether to enable the metrics endpoint.")
	metricsPort                    = flag.String("metrics-port", "10254", "Port for the metrics endpoint.")
	metricsEndpoint                = flag.String("metrics-endpoint", "/metrics", "Metrics endpoint.")
	metricsPrefix                  = flag.String("metrics-prefix", "", "Prefix for the metrics.")
	ingressClassName               = flag.String("ingress-class-name", "", "Set ingressClassName for ingress resources created.")
	metricsLabels                  util.ArrayFlags
	metricsJobStartLatencyBuckets  util.HistogramBuckets = util.DefaultJobStartLatencyBuckets
)

func main() {
	flag.Var(&metricsLabels, "metrics-labels", "Labels for the metrics")
	flag.Var(&metricsJobStartLatencyBuckets, "metrics-job-start-latency-buckets",
		"Comma-separated boundary values (in seconds) for the job start latency histogram bucket; "+
			"it accepts any numerical values that can be parsed into a 64-bit floating point")
	flag.Parse()

	// Create the client config. Use kubeConfig if given, otherwise assume in-cluster.
	config, err := buildConfig(*master, *kubeConfig)
	if err != nil {
		glog.Fatal(err)
	}
	kubeClient, err := clientset.NewForConfig(config)
	if err != nil {
		glog.Fatal(err)
	}

	signalCh := make(chan os.Signal, 1)
	signal.Notify(signalCh, syscall.SIGINT, syscall.SIGTERM)

	stopCh := make(chan struct{}, 1)
	startCh := make(chan struct{}, 1)

	if *enableLeaderElection {
		podName := os.Getenv("POD_NAME")
		hostname, err := os.Hostname()
		if err != nil {
			glog.Fatal(err)
		}
		broadcaster := record.NewBroadcaster()
		source := apiv1.EventSource{Component: "spark-operator-leader-elector", Host: hostname}
		recorder := broadcaster.NewRecorder(scheme.Scheme, source)
		resourceLock := &resourcelock.LeaseLock{
			LeaseMeta: metav1.ObjectMeta{
				Namespace: *leaderElectionLockNamespace,
				Name:      *leaderElectionLockName,
			},
			Client: kubeClient.CoordinationV1(),
			LockConfig: resourcelock.ResourceLockConfig{
				Identity:      podName,
				EventRecorder: recorder,
			},
		}
		if err != nil {
			glog.Fatal(err)
		}

		electionCfg := leaderelection.LeaderElectionConfig{
			Lock:          resourceLock,
			LeaseDuration: *leaderElectionLeaseDuration,
			RenewDeadline: *leaderElectionRenewDeadline,
			RetryPeriod:   *leaderElectionRetryPeriod,
			Callbacks: leaderelection.LeaderCallbacks{
				OnStartedLeading: func(c context.Context) {
					close(startCh)
				},
				OnStoppedLeading: func() {
					close(stopCh)
				},
			},
		}

		elector, err := leaderelection.NewLeaderElector(electionCfg)
		if err != nil {
			glog.Fatal(err)
		}

		go elector.Run(context.Background())
	}

	glog.Info("Starting the Spark Operator")

	crClient, err := crclientset.NewForConfig(config)
	if err != nil {
		glog.Fatal(err)
	}
	apiExtensionsClient, err := apiextensionsclient.NewForConfig(config)
	if err != nil {
		glog.Fatal(err)
	}

	if err = util.InitializeIngressCapabilities(kubeClient); err != nil {
		glog.Fatalf("Error retrieving Kubernetes cluster capabilities: %s", err.Error())
	}

	var batchSchedulerMgr *batchscheduler.SchedulerManager
	if *enableBatchScheduler {
		if !*enableWebhook {
			glog.Fatal(
				"failed to initialize the batch scheduler manager as it requires the webhook to be enabled")
		}
		batchSchedulerMgr = batchscheduler.NewSchedulerManager(config)
	}

	crInformerFactory := buildCustomResourceInformerFactory(crClient)
	podInformerFactory := buildPodInformerFactory(kubeClient)

	var metricConfig *util.MetricConfig
	if *enableMetrics {
		metricConfig = &util.MetricConfig{
			MetricsEndpoint:               *metricsEndpoint,
			MetricsPort:                   *metricsPort,
			MetricsPrefix:                 *metricsPrefix,
			MetricsLabels:                 metricsLabels,
			MetricsJobStartLatencyBuckets: metricsJobStartLatencyBuckets,
		}

		glog.Info("Enabling metrics collecting and exporting to Prometheus")
		util.InitializeMetrics(metricConfig)
	}

	applicationController := sparkapplication.NewController(
		crClient, kubeClient, crInformerFactory, podInformerFactory, metricConfig, *namespace, *ingressURLFormat, *ingressClassName, batchSchedulerMgr, *enableUIService)
	scheduledApplicationController := scheduledsparkapplication.NewController(
		crClient, kubeClient, apiExtensionsClient, crInformerFactory, clock.RealClock{})

	// Start the informer factory that in turn starts the informer.
	go crInformerFactory.Start(stopCh)
	go podInformerFactory.Start(stopCh)

	var hook *webhook.WebHook
	if *enableWebhook {
		var coreV1InformerFactory informers.SharedInformerFactory
		if *enableResourceQuotaEnforcement {
			coreV1InformerFactory = buildCoreV1InformerFactory(kubeClient)
		}
		var err error
		// Don't deregister webhook on exit if leader election enabled (i.e. multiple webhooks running)
		hook, err = webhook.New(kubeClient, crInformerFactory, *namespace, !*enableLeaderElection, *enableResourceQuotaEnforcement, coreV1InformerFactory, webhookTimeout)
		if err != nil {
			glog.Fatal(err)
		}

		if *enableResourceQuotaEnforcement {
			go coreV1InformerFactory.Start(stopCh)
		}

		if err = hook.Start(stopCh); err != nil {
			glog.Fatal(err)
		}
	} else if *enableResourceQuotaEnforcement {
		glog.Fatal("Webhook must be enabled to use resource quota enforcement.")
	}

	if *enableLeaderElection {
		glog.Info("Waiting to be elected leader before starting application controller goroutines")
		select {
		case <-signalCh:
			os.Exit(0)
		case <-startCh:
		}
	}

	glog.Info("Starting application controller goroutines")

	if err = applicationController.Start(*controllerThreads, stopCh); err != nil {
		glog.Fatal(err)
	}
	if err = scheduledApplicationController.Start(*controllerThreads, stopCh); err != nil {
		glog.Fatal(err)
	}

	select {
	case <-signalCh:
		close(stopCh)
	case <-stopCh:
	}

	glog.Info("Shutting down the Spark Operator")
	applicationController.Stop()
	scheduledApplicationController.Stop()
	if *enableWebhook {
		if err := hook.Stop(); err != nil {
			glog.Fatal(err)
		}
	}
}

func buildConfig(masterURL string, kubeConfig string) (*rest.Config, error) {
	if kubeConfig != "" {
		return clientcmd.BuildConfigFromFlags(masterURL, kubeConfig)
	}
	return rest.InClusterConfig()
}

func buildCustomResourceInformerFactory(crClient crclientset.Interface) crinformers.SharedInformerFactory {
	var factoryOpts []crinformers.SharedInformerOption
	if *namespace != apiv1.NamespaceAll {
		factoryOpts = append(factoryOpts, crinformers.WithNamespace(*namespace))
	}
	if len(*labelSelectorFilter) > 0 {
		tweakListOptionsFunc := func(options *metav1.ListOptions) {
			options.LabelSelector = *labelSelectorFilter
		}
		factoryOpts = append(factoryOpts, crinformers.WithTweakListOptions(tweakListOptionsFunc))
	}
	return crinformers.NewSharedInformerFactoryWithOptions(
		crClient,
		// resyncPeriod. Every resyncPeriod, all resources in the cache will re-trigger events.
		time.Duration(*resyncInterval)*time.Second,
		factoryOpts...)
}

func buildPodInformerFactory(kubeClient clientset.Interface) informers.SharedInformerFactory {
	var podFactoryOpts []informers.SharedInformerOption
	if *namespace != apiv1.NamespaceAll {
		podFactoryOpts = append(podFactoryOpts, informers.WithNamespace(*namespace))
	}
	tweakListOptionsFunc := func(options *metav1.ListOptions) {
		options.LabelSelector = fmt.Sprintf("%s,%s", operatorConfig.SparkRoleLabel, operatorConfig.LaunchedBySparkOperatorLabel)
		if len(*labelSelectorFilter) > 0 {
			options.LabelSelector = options.LabelSelector + "," + *labelSelectorFilter
		}
	}
	podFactoryOpts = append(podFactoryOpts, informers.WithTweakListOptions(tweakListOptionsFunc))
	return informers.NewSharedInformerFactoryWithOptions(kubeClient, time.Duration(*resyncInterval)*time.Second, podFactoryOpts...)
}

func buildCoreV1InformerFactory(kubeClient clientset.Interface) informers.SharedInformerFactory {
	var coreV1FactoryOpts []informers.SharedInformerOption
	if *namespace != apiv1.NamespaceAll {
		coreV1FactoryOpts = append(coreV1FactoryOpts, informers.WithNamespace(*namespace))
	}
	if len(*labelSelectorFilter) > 0 {
		tweakListOptionsFunc := func(options *metav1.ListOptions) {
			options.LabelSelector = *labelSelectorFilter
		}
		coreV1FactoryOpts = append(coreV1FactoryOpts, informers.WithTweakListOptions(tweakListOptionsFunc))
	}
	return informers.NewSharedInformerFactoryWithOptions(kubeClient, time.Duration(*resyncInterval)*time.Second, coreV1FactoryOpts...)
}