从源码层面解析kube-scheduler的默认配置是怎么做的

这篇文章将为大家详细讲解有关从源码层面解析kube-scheduler的默认配置是怎么做的，小编觉得挺实用的，因此分享给大家做个参考，希望大家阅读完这篇文章后可以有所收获。

从头来看，在kube-scheduler的main函数中，s := options.NewSchedulerServer()创建SchedulerServer时，是按照默认参数创建的。

--- plugin/cmd/kube-scheduler/scheduler.go:30 ---func main() {        s := options.NewSchedulerServer()        s.AddFlags(pflag.CommandLine)        flag.InitFlags()        logs.InitLogs()        defer logs.FlushLogs()        verflag.PrintAndExitIfRequested()        if err := app.Run(s); err != nil {                glog.Fatalf("scheduler app failed to run: %v", err)        }}--- plugin/cmd/kube-scheduler/app/options/options.go:44 ---// NewSchedulerServer creates a new SchedulerServer with default parametersfunc NewSchedulerServer() *SchedulerServer {        versioned := &v1alpha1.KubeSchedulerConfiguration{}        api.Scheme.Default(versioned)        cfg := componentconfig.KubeSchedulerConfiguration{}        api.Scheme.Convert(versioned, &cfg, nil)        cfg.LeaderElection.LeaderElect = true        s := SchedulerServer{                KubeSchedulerConfiguration: cfg,        }        return &s}

上面NewSchedulerServer中api.Scheme.Default(versioned)就是设置默认参数的操作。看看Default(...)的代码：

--- pkg/runtime/scheme.go:439 ---// Default sets defaults on the provided Object.func (s *Scheme) Default(src Object) {        if fn, ok := s.defaulterFuncs[reflect.TypeOf(src)]; ok {                fn(src)        }}

原来Default(...)做的工作就是从Scheme.defaulterFuncs这个Map中获取&v1alpha1.KubeSchedulerConfiguration{}这个type对应的defaultFunc fn，并执行fn(&v1alpha1.KubeSchedulerConfiguration{})来完成默认参数的配置。

OK，那么问题来了。这些type对应的defaultFunc是怎么register到Scheme.defaulterFuncs这个Map中的呢？

答案就在pkg/apis/componentconfig/v1alpha1/register.go中定义的全局变量SchemeBuilder。SchemeBuilder = runtime.NewSchemeBuilder(addKnownTypes, addDefaultingFuncs)在创建SchemeBuilder时就调用了addDefaultFuncs函数。注册defaultFunc的工作应该就是在addDefaultingFuncs方法中实现的。

看看addDefaultingFuncs的实现，果不其然啊.SetDefaults_KubeSchedulerConfiguration就是&v1alpha1.KubeSchedulerConfiguration{}对应的defaultFuncs。

---- pkg/apis/componentconfig/v1alpha1/register.go ----var (        SchemeBuilder = runtime.NewSchemeBuilder(addKnownTypes, addDefaultingFuncs)        AddToScheme   = SchemeBuilder.AddToScheme)---- pkg/apis/componentconfig/v1alpha1/defaults.go -----func addDefaultingFuncs(scheme *kruntime.Scheme) error {        RegisterDefaults(scheme)        return scheme.AddDefaultingFuncs(                SetDefaults_KubeProxyConfiguration,                SetDefaults_KubeSchedulerConfiguration,                SetDefaults_LeaderElectionConfiguration,                SetDefaults_KubeletConfiguration,        )}func SetDefaults_KubeSchedulerConfiguration(obj *KubeSchedulerConfiguration) {        if obj.Port == 0 {                obj.Port = ports.SchedulerPort        }        if obj.Address == "" {                obj.Address = "0.0.0.0"        }        if obj.AlgorithmProvider == "" {                obj.AlgorithmProvider = "DefaultProvider"        }        if obj.ContentType == "" {                obj.ContentType = "application/vnd.kubernetes.protobuf"        }        if obj.KubeAPIQPS == 0 {                obj.KubeAPIQPS = 50.0        }        if obj.KubeAPIBurst == 0 {                obj.KubeAPIBurst = 100        }        if obj.SchedulerName == "" {                obj.SchedulerName = api.DefaultSchedulerName        }        if obj.HardPodAffinitySymmetricWeight == 0 {                obj.HardPodAffinitySymmetricWeight = api.DefaultHardPodAffinitySymmetricWeight        }        if obj.FailureDomains == "" {                obj.FailureDomains = api.DefaultFailureDomains        }}

再结合plugin/cmd/kube-scheduler/app/options/options.go:57定义的AddFlags，可得kube-scheduler的默认配置如下：

• port = 10251

• address = "0.0.0.0"

• algorithm-provider = "DefaultProvider"

• content-type = "application/vnd.kubernetes.protobuf"

• kube-api-qps = 50

• kube-api-burst = 100

• scheduler-name = "default-scheduler"

• hard-pod-affinity-symmetric-weight = 1

• failure-domains = "kubernetes.io/hostname,failure-domain.beta.kubernetes.io/zone,failure-domain.beta.kubernetes.io/region"

关于"从源码层面解析kube-scheduler的默认配置是怎么做的"这篇文章就分享到这里了，希望以上内容可以对大家有一定的帮助，使各位可以学到更多知识，如果觉得文章不错，请把它分享出去让更多的人看到。