Merge pull request #2056 from CharlesQQ/leader-elect

feat: add leader election options

cmd/agent/app/agent.go

@@ -147,6 +147,9 @@ func run(ctx context.Context, karmadaConfig karmadactl.KarmadaConfig, opts *opti
 		LeaderElectionID:           fmt.Sprintf("karmada-agent-%s", opts.ClusterName),
 		LeaderElectionNamespace:    opts.LeaderElection.ResourceNamespace,
 		LeaderElectionResourceLock: opts.LeaderElection.ResourceLock,
+		LeaseDuration:              &opts.LeaderElection.LeaseDuration.Duration,
+		RenewDeadline:              &opts.LeaderElection.RenewDeadline.Duration,
+		RetryPeriod:                &opts.LeaderElection.RetryPeriod.Duration,
 		HealthProbeBindAddress:     net.JoinHostPort(opts.BindAddress, strconv.Itoa(opts.SecurePort)),
 		LivenessEndpointName:       "/healthz",
 		MetricsBindAddress:         opts.MetricsBindAddress,

cmd/agent/app/options/options.go

@@ -21,6 +21,12 @@ const (
 	defaultPort                    = 10357
 )
 
+var (
+	defaultElectionLeaseDuration = metav1.Duration{Duration: 15 * time.Second}
+	defaultElectionRenewDeadline = metav1.Duration{Duration: 10 * time.Second}
+	defaultElectionRetryPeriod   = metav1.Duration{Duration: 2 * time.Second}
+)
+
 // Options contains everything necessary to create and run controller-manager.
 type Options struct {
 	// Controllers contains all controller names.
@@ -111,6 +117,19 @@ func (o *Options) AddFlags(fs *pflag.FlagSet, allControllers []string) {
 		"The secure port on which to serve HTTPS.")
 	fs.BoolVar(&o.LeaderElection.LeaderElect, "leader-elect", true, "Start a leader election client and gain leadership before executing the main loop. Enable this when running replicated components for high availability.")
 	fs.StringVar(&o.LeaderElection.ResourceNamespace, "leader-elect-resource-namespace", util.NamespaceKarmadaSystem, "The namespace of resource object that is used for locking during leader election.")
+	fs.DurationVar(&o.LeaderElection.LeaseDuration.Duration, "leader-elect-lease-duration", defaultElectionLeaseDuration.Duration, ""+
+		"The duration that non-leader candidates will wait after observing a leadership "+
+		"renewal until attempting to acquire leadership of a led but unrenewed leader "+
+		"slot. This is effectively the maximum duration that a leader can be stopped "+
+		"before it is replaced by another candidate. This is only applicable if leader "+
+		"election is enabled.")
+	fs.DurationVar(&o.LeaderElection.RenewDeadline.Duration, "leader-elect-renew-deadline", defaultElectionRenewDeadline.Duration, ""+
+		"The interval between attempts by the acting master to renew a leadership slot "+
+		"before it stops leading. This must be less than or equal to the lease duration. "+
+		"This is only applicable if leader election is enabled.")
+	fs.DurationVar(&o.LeaderElection.RetryPeriod.Duration, "leader-elect-retry-period", defaultElectionRetryPeriod.Duration, ""+
+		"The duration the clients should wait between attempting acquisition and renewal "+
+		"of a leadership. This is only applicable if leader election is enabled.")
 	fs.StringVar(&o.KarmadaKubeConfig, "karmada-kubeconfig", o.KarmadaKubeConfig, "Path to karmada control plane kubeconfig file.")
 	fs.StringVar(&o.KarmadaContext, "karmada-context", "", "Name of the cluster context in karmada control plane kubeconfig file.")
 	fs.StringVar(&o.ClusterName, "cluster-name", o.ClusterName, "Name of member cluster that the agent serves for.")

cmd/controller-manager/app/controllermanager.go

@@ -109,6 +109,9 @@ func Run(ctx context.Context, opts *options.Options) error {
 		LeaderElection:             opts.LeaderElection.LeaderElect,
 		LeaderElectionID:           opts.LeaderElection.ResourceName,
 		LeaderElectionNamespace:    opts.LeaderElection.ResourceNamespace,
+		LeaseDuration:              &opts.LeaderElection.LeaseDuration.Duration,
+		RenewDeadline:              &opts.LeaderElection.RenewDeadline.Duration,
+		RetryPeriod:                &opts.LeaderElection.RetryPeriod.Duration,
 		LeaderElectionResourceLock: opts.LeaderElection.ResourceLock,
 		HealthProbeBindAddress:     net.JoinHostPort(opts.BindAddress, strconv.Itoa(opts.SecurePort)),
 		LivenessEndpointName:       "/healthz",

cmd/controller-manager/app/options/options.go

@@ -20,6 +20,12 @@ const (
 	defaultPort        = 10357
 )
 
+var (
+	defaultElectionLeaseDuration = metav1.Duration{Duration: 15 * time.Second}
+	defaultElectionRenewDeadline = metav1.Duration{Duration: 10 * time.Second}
+	defaultElectionRetryPeriod   = metav1.Duration{Duration: 2 * time.Second}
+)
+
 // Options contains everything necessary to create and run controller-manager.
 type Options struct {
 	// Controllers is the list of controllers to enable or disable
@@ -134,6 +140,19 @@ func (o *Options) AddFlags(flags *pflag.FlagSet, allControllers, disabledByDefau
 		"Specifies how often karmada-controller-manager posts cluster status to karmada-apiserver.")
 	flags.BoolVar(&o.LeaderElection.LeaderElect, "leader-elect", true, "Start a leader election client and gain leadership before executing the main loop. Enable this when running replicated components for high availability.")
 	flags.StringVar(&o.LeaderElection.ResourceNamespace, "leader-elect-resource-namespace", util.NamespaceKarmadaSystem, "The namespace of resource object that is used for locking during leader election.")
+	flags.DurationVar(&o.LeaderElection.LeaseDuration.Duration, "leader-elect-lease-duration", defaultElectionLeaseDuration.Duration, ""+
+		"The duration that non-leader candidates will wait after observing a leadership "+
+		"renewal until attempting to acquire leadership of a led but unrenewed leader "+
+		"slot. This is effectively the maximum duration that a leader can be stopped "+
+		"before it is replaced by another candidate. This is only applicable if leader "+
+		"election is enabled.")
+	flags.DurationVar(&o.LeaderElection.RenewDeadline.Duration, "leader-elect-renew-deadline", defaultElectionRenewDeadline.Duration, ""+
+		"The interval between attempts by the acting master to renew a leadership slot "+
+		"before it stops leading. This must be less than or equal to the lease duration. "+
+		"This is only applicable if leader election is enabled.")
+	flags.DurationVar(&o.LeaderElection.RetryPeriod.Duration, "leader-elect-retry-period", defaultElectionRetryPeriod.Duration, ""+
+		"The duration the clients should wait between attempting acquisition and renewal "+
+		"of a leadership. This is only applicable if leader election is enabled.")
 	flags.DurationVar(&o.ClusterLeaseDuration.Duration, "cluster-lease-duration", 40*time.Second,
 		"Specifies the expiration period of a cluster lease.")
 	flags.Float64Var(&o.ClusterLeaseRenewIntervalFraction, "cluster-lease-renew-interval-fraction", 0.25,