Kubernetes源码分析之kube-apiserver

本节所有的代码基于最新的1.13.4版本。

启动分析

同Kubernetes所有的组件启动代码一致,apiserver启动使用的是cobra的命令行方式

如图,启动主要完成三个步骤:
1、完成参数的配置;
2、判断配置是否合法;
3、执行最终的 Run方法。
Run方法比较简单 如图,主要执行两个步骤:
1、创建server端;
2、启动server。
因为apiserver本质上就是一个server服务器,所有代码核心就是如何配置server,包括路由、访问权限以及同数据库(etcd)的交互等。先看一下server端是如何创建起来的

Server端创建

Server端的创建集中在CreateServerChain方法。方法代码如下:

// CreateServerChain creates the apiservers connected via delegation.
// CreateServerChain创建通过委托连接的apiservers,创建一系列的server
func CreateServerChain(completedOptions completedServerRunOptions, stopCh <-chan struct{}) (*genericapiserver.GenericAPIServer, error) {
	nodeTunneler, proxyTransport, err := CreateNodeDialer(completedOptions)
	if err != nil {
		return nil, err
	}

	// 1.创建kubeAPIServerConfig配置
	kubeAPIServerConfig, insecureServingInfo, serviceResolver, pluginInitializer, admissionPostStartHook, err := CreateKubeAPIServerConfig(completedOptions, nodeTunneler, proxyTransport)
	if err != nil {
		return nil, err
	}

	// If additional API servers are added, they should be gated.
	// 2.判断是否配置了扩展API server,创建apiExtensionsConfig配置
	apiExtensionsConfig, err := createAPIExtensionsConfig(*kubeAPIServerConfig.GenericConfig, kubeAPIServerConfig.ExtraConfig.VersionedInformers, pluginInitializer, completedOptions.ServerRunOptions, completedOptions.MasterCount,
		serviceResolver, webhook.NewDefaultAuthenticationInfoResolverWrapper(proxyTransport, kubeAPIServerConfig.GenericConfig.LoopbackClientConfig))
	if err != nil {
		return nil, err
	}

	// apiExtensionsServer,可扩展的API server
	// 3.启动扩展的API server
	apiExtensionsServer, err := createAPIExtensionsServer(apiExtensionsConfig, genericapiserver.NewEmptyDelegate())
	if err != nil {
		return nil, err
	}

	// 4.启动最核心的kubeAPIServer
	kubeAPIServer, err := CreateKubeAPIServer(kubeAPIServerConfig, apiExtensionsServer.GenericAPIServer, admissionPostStartHook)
	if err != nil {
		return nil, err
	}

	// otherwise go down the normal path of standing the aggregator up in front of the API server
	// this wires up openapi
	kubeAPIServer.GenericAPIServer.PrepareRun()

	// This will wire up openapi for extension api server
	apiExtensionsServer.GenericAPIServer.PrepareRun()

	// aggregator comes last in the chain
	// 5.聚合层的配置aggregatorConfig
	aggregatorConfig, err := createAggregatorConfig(*kubeAPIServerConfig.GenericConfig, completedOptions.ServerRunOptions, kubeAPIServerConfig.ExtraConfig.VersionedInformers, serviceResolver, proxyTransport, pluginInitializer)
	if err != nil {
		return nil, err
	}
	// 6.aggregatorServer,聚合服务器,对所有的服务器访问的整合
	aggregatorServer, err := createAggregatorServer(aggregatorConfig, kubeAPIServer.GenericAPIServer, apiExtensionsServer.Informers)
	if err != nil {
		// we don't need special handling for innerStopCh because the aggregator server doesn't create any go routines
		return nil, err
	}

	// 7.启动非安全端口的server
	if insecureServingInfo != nil {
		insecureHandlerChain := kubeserver.BuildInsecureHandlerChain(aggregatorServer.GenericAPIServer.UnprotectedHandler(), kubeAPIServerConfig.GenericConfig)
		if err := insecureServingInfo.Serve(insecureHandlerChain, kubeAPIServerConfig.GenericConfig.RequestTimeout, stopCh); err != nil {
			return nil, err
		}
	}

	// 8.返回GenericAPIServer,后续启动安全端口的server
	return aggregatorServer.GenericAPIServer, nil
}
复制代码

创建过程主要有以下步骤:
1、根据配置构造apiserver的配置,调用方法CreateKubeAPIServerConfig
2、根据配置构造扩展的apiserver的配置,调用方法为createAPIExtensionsConfig
3、创建server,包括扩展的apiserver和原生的apiserver,调用方法为createAPIExtensionsServerCreateKubeAPIServer。主要就是将各个handler的路由方法注册到Container中去,完全遵循go-restful的设计模式,即将处理方法注册到Route中去,同一个根路径下的Route注册到WebService中去,WebService注册到Container中,Container负责分发。访问的过程为Container-->WebService-->Route。更加详细的go-restful使用可以参考其代码;
4、聚合server的配置和和创建。主要就是将原生的apiserver和扩展的apiserver的访问进行整合,添加后续的一些处理接口。调用方法为createAggregatorConfigcreateAggregatorServer
5、创建完成,返回配置的server信息。
以上几个步骤,最核心的就是apiserver如何创建,即如何按照go-restful的模式,添加路由和相应的处理方法,以CreateKubeAPIServer方法为例,createAPIExtensionsServer类似。

创建

CreateKubeAPIServer方法如下

func CreateKubeAPIServer(kubeAPIServerConfig *master.Config, delegateAPIServer genericapiserver.DelegationTarget, admissionPostStartHook genericapiserver.PostStartHookFunc) (*master.Master, error) {
	kubeAPIServer, err := kubeAPIServerConfig.Complete().New(delegateAPIServer)
	if err != nil {
		return nil, err
	}

	kubeAPIServer.GenericAPIServer.AddPostStartHookOrDie("start-kube-apiserver-admission-initializer", admissionPostStartHook)

	return kubeAPIServer, nil
}
复制代码

通过Complete方法完成配置的最终合法化,New方法生成kubeAPIServer的配置,进入New方法,

// New returns a new instance of Master from the given config.
// Certain config fields will be set to a default value if unset.
// Certain config fields must be specified, including:
//   KubeletClientConfig
// 通过给定的配置,返回一个新的Master实例。对于部分未配置的选项,可以使用默认配置;但是对于KubeletClientConfig这样的配置,必须手动指定
func (c completedConfig) New(delegationTarget genericapiserver.DelegationTarget) (*Master, error) {
	if reflect.DeepEqual(c.ExtraConfig.KubeletClientConfig, kubeletclient.KubeletClientConfig{}) {
		return nil, fmt.Errorf("Master.New() called with empty config.KubeletClientConfig")
	}

	// 1.初始化,创建go-restful的Container,初始化apiServerHandler
	s, err := c.GenericConfig.New("kube-apiserver", delegationTarget)
	if err != nil {
		return nil, err
	}

	if c.ExtraConfig.EnableLogsSupport {
		routes.Logs{}.Install(s.Handler.GoRestfulContainer)
	}

	m := &Master{
		GenericAPIServer: s,
	}

	// install legacy rest storage
	// /api开头的版本api注册到Container中去,如Pod、Namespace等资源
	if c.ExtraConfig.APIResourceConfigSource.VersionEnabled(apiv1.SchemeGroupVersion) {
		legacyRESTStorageProvider := corerest.LegacyRESTStorageProvider{
			StorageFactory:              c.ExtraConfig.StorageFactory,
			ProxyTransport:              c.ExtraConfig.ProxyTransport,
			KubeletClientConfig:         c.ExtraConfig.KubeletClientConfig,
			EventTTL:                    c.ExtraConfig.EventTTL,
			ServiceIPRange:              c.ExtraConfig.ServiceIPRange,
			ServiceNodePortRange:        c.ExtraConfig.ServiceNodePortRange,
			LoopbackClientConfig:        c.GenericConfig.LoopbackClientConfig,
			ServiceAccountIssuer:        c.ExtraConfig.ServiceAccountIssuer,
			ServiceAccountMaxExpiration: c.ExtraConfig.ServiceAccountMaxExpiration,
			APIAudiences:                c.GenericConfig.Authentication.APIAudiences,
		}
		m.InstallLegacyAPI(&c, c.GenericConfig.RESTOptionsGetter, legacyRESTStorageProvider)
	}

	// The order here is preserved in discovery.
	// If resources with identical names exist in more than one of these groups (e.g. "deployments.apps"" and "deployments.extensions"),
	// the order of this list determines which group an unqualified resource name (e.g. "deployments") should prefer.
	// This priority order is used for local discovery, but it ends up aggregated in `k8s.io/kubernetes/cmd/kube-apiserver/app/aggregator.go
	// with specific priorities.
	// TODO: describe the priority all the way down in the RESTStorageProviders and plumb it back through the various discovery
	// handlers that we have.
	// /apis开头版本的api注册到Container中
	restStorageProviders := []RESTStorageProvider{
		auditregistrationrest.RESTStorageProvider{},
		authenticationrest.RESTStorageProvider{Authenticator: c.GenericConfig.Authentication.Authenticator, APIAudiences: c.GenericConfig.Authentication.APIAudiences},
		authorizationrest.RESTStorageProvider{Authorizer: c.GenericConfig.Authorization.Authorizer, RuleResolver: c.GenericConfig.RuleResolver},
		autoscalingrest.RESTStorageProvider{},
		batchrest.RESTStorageProvider{},
		certificatesrest.RESTStorageProvider{},
		coordinationrest.RESTStorageProvider{},
		extensionsrest.RESTStorageProvider{},
		networkingrest.RESTStorageProvider{},
		policyrest.RESTStorageProvider{},
		rbacrest.RESTStorageProvider{Authorizer: c.GenericConfig.Authorization.Authorizer},
		schedulingrest.RESTStorageProvider{},
		settingsrest.RESTStorageProvider{},
		storagerest.RESTStorageProvider{},
		// keep apps after extensions so legacy clients resolve the extensions versions of shared resource names.
		// See https://github.com/kubernetes/kubernetes/issues/42392
		appsrest.RESTStorageProvider{},
		admissionregistrationrest.RESTStorageProvider{},
		eventsrest.RESTStorageProvider{TTL: c.ExtraConfig.EventTTL},
	}
	m.InstallAPIs(c.ExtraConfig.APIResourceConfigSource, c.GenericConfig.RESTOptionsGetter, restStorageProviders...)

	if c.ExtraConfig.Tunneler != nil {
		m.installTunneler(c.ExtraConfig.Tunneler, corev1client.NewForConfigOrDie(c.GenericConfig.LoopbackClientConfig).Nodes())
	}

	m.GenericAPIServer.AddPostStartHookOrDie("ca-registration", c.ExtraConfig.ClientCARegistrationHook.PostStartHook)

	return m, nil
}
复制代码

包含以下步骤:
1、按照go-restful的模式,调用c.GenericConfig.New方法初始化化Container,即gorestfulContainer,初始方法为NewAPIServerHandler。初始化之后,添加路由。

func installAPI(s *GenericAPIServer, c *Config) {

	// 添加"/"与"/index.html"路由
	if c.EnableIndex {
		routes.Index{}.Install(s.listedPathProvider, s.Handler.NonGoRestfulMux)
	}
	// 添加"/swagger-ui/"路由
	if c.SwaggerConfig != nil && c.EnableSwaggerUI {
		routes.SwaggerUI{}.Install(s.Handler.NonGoRestfulMux)
	}
	// 添加"/debug"相关路由
	if c.EnableProfiling {
		routes.Profiling{}.Install(s.Handler.NonGoRestfulMux)
		if c.EnableContentionProfiling {
			goruntime.SetBlockProfileRate(1)
		}
		// so far, only logging related endpoints are considered valid to add for these debug flags.
		routes.DebugFlags{}.Install(s.Handler.NonGoRestfulMux, "v", routes.StringFlagPutHandler(logs.GlogSetter))
	}
	// 添加"/metrics"路由
	if c.EnableMetrics {
		if c.EnableProfiling {
			routes.MetricsWithReset{}.Install(s.Handler.NonGoRestfulMux)
		} else {
			routes.DefaultMetrics{}.Install(s.Handler.NonGoRestfulMux)
		}
	}

	// 添加"/version"路由
	routes.Version{Version: c.Version}.Install(s.Handler.GoRestfulContainer)

	if c.EnableDiscovery {
		s.Handler.GoRestfulContainer.Add(s.DiscoveryGroupManager.WebService())
	}
}
复制代码

该方法中添加了包括/、/swagger-ui、/debug/*、/metrics、/version几条路由,通过访问apiserver即可看到相关的信息

2、判断是否支持logs相关的路由,如果支持,则添加/logs路由;
3、添加以 api开头的路由 在集群中对应的路由有 即/api和/api/v1,比较常用的资源像Pods就是该路由对应的资源;
4、添加以 apis开头的路由 在集群中对应的路由有 可以看到apis开头的路由明显较多。应该是由于kubernetes设计之初的版本都是以api/v1开头,后续扩展的版本以apis开头命名。现在更多的是通过CRD与自定义Controller的方法扩展API,不再进行api版本的扩展。基本上代码中的名称都可以在实际集群中找到对应的API。

路由添加(api开头)

api开头的路由通过InstallLegacyAPI方法添加。进入InstallLegacyAPI方法,如下:

func (m *Master) InstallLegacyAPI(c *completedConfig, restOptionsGetter generic.RESTOptionsGetter, legacyRESTStorageProvider corerest.LegacyRESTStorageProvider) {
	legacyRESTStorage, apiGroupInfo, err := legacyRESTStorageProvider.NewLegacyRESTStorage(restOptionsGetter)
	if err != nil {
		klog.Fatalf("Error building core storage: %v", err)
	}

	controllerName := "bootstrap-controller"
	coreClient := corev1client.NewForConfigOrDie(c.GenericConfig.LoopbackClientConfig)
	bootstrapController := c.NewBootstrapController(legacyRESTStorage, coreClient, coreClient, coreClient)
	m.GenericAPIServer.AddPostStartHookOrDie(controllerName, bootstrapController.PostStartHook)
	m.GenericAPIServer.AddPreShutdownHookOrDie(controllerName, bootstrapController.PreShutdownHook)

	if err := m.GenericAPIServer.InstallLegacyAPIGroup(genericapiserver.DefaultLegacyAPIPrefix, &apiGroupInfo); err != nil {
		klog.Fatalf("Error in registering group versions: %v", err)
	}
}
复制代码

通过NewLegacyRESTStorage方法创建各个资源的RESTStorage。RESTStorage是一个结构体,具体的定义在vendor/k8s.io/apiserver/pkg/registry/generic/registry/store.go下,结构体内主要包含NewFunc返回特定资源信息、NewListFunc返回特定资源列表、CreateStrategy特定资源创建时的策略、UpdateStrategy更新时的策略以及DeleteStrategy删除时的策略等重要方法。
NewLegacyRESTStorage内部,可以看到创建了多种资源的RESTStorage

常见的像event、secret、namespace、endpoints等,统一调用 NewREST方法构造相应的资源。待所有资源的store创建完成之后,使用 restStorageMap的Map类型将每个资源的路由和对应的store对应起来,方便后续去做路由的统一规划,代码如下:

restStorageMap := map[string]rest.Storage{
		"pods":             podStorage.Pod,
		"pods/attach":      podStorage.Attach,
		"pods/status":      podStorage.Status,
		"pods/log":         podStorage.Log,
		"pods/exec":        podStorage.Exec,
		"pods/portforward": podStorage.PortForward,
		"pods/proxy":       podStorage.Proxy,
		"pods/binding":     podStorage.Binding,
		"bindings":         podStorage.Binding,

		"podTemplates": podTemplateStorage,

		"replicationControllers":        controllerStorage.Controller,
		"replicationControllers/status": controllerStorage.Status,

		"services":        serviceRest,
		"services/proxy":  serviceRestProxy,
		"services/status": serviceStatusStorage,

		"endpoints": endpointsStorage,

		"nodes":        nodeStorage.Node,
		"nodes/status": nodeStorage.Status,
		"nodes/proxy":  nodeStorage.Proxy,

		"events": eventStorage,

		"limitRanges":                   limitRangeStorage,
		"resourceQuotas":                resourceQuotaStorage,
		"resourceQuotas/status":         resourceQuotaStatusStorage,
		"namespaces":                    namespaceStorage,
		"namespaces/status":             namespaceStatusStorage,
		"namespaces/finalize":           namespaceFinalizeStorage,
		"secrets":                       secretStorage,
		"serviceAccounts":               serviceAccountStorage,
		"persistentVolumes":             persistentVolumeStorage,
		"persistentVolumes/status":      persistentVolumeStatusStorage,
		"persistentVolumeClaims":        persistentVolumeClaimStorage,
		"persistentVolumeClaims/status": persistentVolumeClaimStatusStorage,
		"configMaps":                    configMapStorage,

		"componentStatuses": componentstatus.NewStorage(componentStatusStorage{c.StorageFactory}.serversToValidate),
	}
	if legacyscheme.Scheme.IsVersionRegistered(schema.GroupVersion{Group: "autoscaling", Version: "v1"}) {
		restStorageMap["replicationControllers/scale"] = controllerStorage.Scale
	}
	if legacyscheme.Scheme.IsVersionRegistered(schema.GroupVersion{Group: "policy", Version: "v1beta1"}) {
		restStorageMap["pods/eviction"] = podStorage.Eviction
	}
	if serviceAccountStorage.Token != nil {
		restStorageMap["serviceaccounts/token"] = serviceAccountStorage.Token
	}
	apiGroupInfo.VersionedResourcesStorageMap["v1"] = restStorageMap
复制代码

最终完成以api开头的所有资源的RESTStorage操作。
创建完之后,则开始进行路由的安装,执行InstallLegacyAPIGroup方法,主要调用链为InstallLegacyAPIGroup-->installAPIResources-->InstallREST-->Install-->registerResourceHandlers,最终核心的路由构造在registerResourceHandlers方法内。这是一个非常复杂的方法,整个方法的代码在700行左右。方法的主要功能是通过上一步骤构造的RESTStorage判断该资源可以执行哪些操作(如create、update等),将其对应的操作存入到action,每一个action对应一个标准的rest操作,如create对应的action操作为POST、update对应的action操作为PUT。最终根据actions数组依次遍历,对每一个操作添加一个handler方法,注册到route中去,route注册到webservice中去,完美匹配go-restful的设计模式。

路由添加(apis开头)

api开头的路由主要是对基础资源的路由实现,而对于其他附加的资源,如认证相关、网络相关等各种扩展的api资源,统一以apis开头命名,实现入口为InstallAPIs
InstallAPIsInstallLegacyAPIGroup主要的区别是获取RESTStorage的方式。对于api开头的路由来说,都是/api/v1这种统一的格式;而对于apis开头路由则不一样,它包含了多种不同的格式(Kubernetes代码内叫groupName),如/apis/apps、/apis/certificates.k8s.io等各种无规律的groupName。为此,kubernetes提供了一种RESTStorageProvider的工厂模式的接口

// RESTStorageProvider is a factory type for REST storage.
type RESTStorageProvider interface {
	GroupName() string
	NewRESTStorage(apiResourceConfigSource serverstorage.APIResourceConfigSource, restOptionsGetter generic.RESTOptionsGetter) (genericapiserver.APIGroupInfo, bool)
}
复制代码

所有以apis开头的路由的资源都需要实现该接口。GroupName()方法获取到的就是类似于/apis/apps、/apis/certificates.k8s.io这样的groupName,NewRESTStorage方法获取到的是相对应的RESTStorage封装后的信息。各种资源的NewRESTStorage接口实现如图:

后续的操作和之前的步骤类似,通过构造go-restful格式的路由信息,完成创建,此处不做赘述。

Server端启动

通过CreateServerChain创建完server后,继续调用GenericAPIServer的Run方法完成最终的启动工作。首先通过PrepareRun方法完成启动前的路由收尾工作,该方法主要完成了SwaggerOpenAPI路由的注册工作(SwaggerOpenAPI主要包含了Kubernetes API的所有细节与规范),并完成/healthz路由的注册工作。完成后,开始最终的server启动工作。
Run方法里通过NonBlockingRun方法启动安全的http server(非安全方式的启动在CreateServerChain方法已经完成)

// Run spawns the secure http server. It only returns if stopCh is closed
// or the secure port cannot be listened on initially.
// Run方法会创建一个安全的http server。只有在stopCh关闭或最初无法监听安全端口时返回
func (s preparedGenericAPIServer) Run(stopCh <-chan struct{}) error {
	// NonBlockingRun创建一个安全的http server
	err := s.NonBlockingRun(stopCh)
	if err != nil {
		return err
	}

	<-stopCh

	// 接收到stopCh之后的处理动作
	err = s.RunPreShutdownHooks()
	if err != nil {
		return err
	}

	// Wait for all requests to finish, which are bounded by the RequestTimeout variable.
	s.HandlerChainWaitGroup.Wait()

	return nil
}
复制代码

启动主要工作包括配置各种证书认证、时间参数、报文大小参数之类,之后通过调用net/http库的启动方式启动,代码比较简洁,不一一列出了。

权限相关

ApiServer中与权限相关的主要有三种机制,即常用的认证鉴权准入控制。对apiserver来说,主要提供的就是rest风格的接口,所以各种权限最终还是集中到对接口的权限判断上。
以最核心的kubeAPIServerConfig举例,在CreateServerChain方法中,调用了CreateKubeAPIServerConfig的方法,该方法主要的作用是创建kubeAPIServer的配置。进入该方法,调用了buildGenericConfig创建一些通用的配置,在NewConfig下,返回了DefaultBuildHandlerChain,该方法主要就是用来对apiserver rest接口的链式判断,即俗称的filter操作,先记录下,后续分析。

func DefaultBuildHandlerChain(apiHandler http.Handler, c *Config) http.Handler {
	handler := genericapifilters.WithAuthorization(apiHandler, c.Authorization.Authorizer, c.Serializer)
	handler = genericfilters.WithMaxInFlightLimit(handler, c.MaxRequestsInFlight, c.MaxMutatingRequestsInFlight, c.LongRunningFunc)
	handler = genericapifilters.WithImpersonation(handler, c.Authorization.Authorizer, c.Serializer)
	handler = genericapifilters.WithAudit(handler, c.AuditBackend, c.AuditPolicyChecker, c.LongRunningFunc)
	failedHandler := genericapifilters.Unauthorized(c.Serializer, c.Authentication.SupportsBasicAuth)
	failedHandler = genericapifilters.WithFailedAuthenticationAudit(failedHandler, c.AuditBackend, c.AuditPolicyChecker)
	handler = genericapifilters.WithAuthentication(handler, c.Authentication.Authenticator, failedHandler, c.Authentication.APIAudiences)
	handler = genericfilters.WithCORS(handler, c.CorsAllowedOriginList, nil, nil, nil, "true")
	handler = genericfilters.WithTimeoutForNonLongRunningRequests(handler, c.LongRunningFunc, c.RequestTimeout)
	handler = genericfilters.WithWaitGroup(handler, c.LongRunningFunc, c.HandlerChainWaitGroup)
	handler = genericapifilters.WithRequestInfo(handler, c.RequestInfoResolver)
	handler = genericfilters.WithPanicRecovery(handler)
	return handler
}
复制代码

配置文件创建完成后,再进行创建工作,进入到CreateKubeAPIServer方法,在初始化go-restful的Container的方法内,可以看到

handlerChainBuilder方法就是对返回的 DefaultBuildHandlerChain方法的一种封装,并作为参数传入到 NewAPIServerHandler方法内。进入 NewAPIServerHandler方法,如下:

func NewAPIServerHandler(name string, s runtime.NegotiatedSerializer, handlerChainBuilder HandlerChainBuilderFn, notFoundHandler http.Handler) *APIServerHandler {
	nonGoRestfulMux := mux.NewPathRecorderMux(name)
	if notFoundHandler != nil {
		nonGoRestfulMux.NotFoundHandler(notFoundHandler)
	}

	gorestfulContainer := restful.NewContainer()
	gorestfulContainer.ServeMux = http.NewServeMux()
	gorestfulContainer.Router(restful.CurlyRouter{}) // e.g. for proxy/{kind}/{name}/{*}
	gorestfulContainer.RecoverHandler(func(panicReason interface{}, httpWriter http.ResponseWriter) {
		logStackOnRecover(s, panicReason, httpWriter)
	})
	gorestfulContainer.ServiceErrorHandler(func(serviceErr restful.ServiceError, request *restful.Request, response *restful.Response) {
		serviceErrorHandler(s, serviceErr, request, response)
	})

	director := director{
		name:               name,
		goRestfulContainer: gorestfulContainer,
		nonGoRestfulMux:    nonGoRestfulMux,
	}

	return &APIServerHandler{
		FullHandlerChain:   handlerChainBuilder(director),
		GoRestfulContainer: gorestfulContainer,
		NonGoRestfulMux:    nonGoRestfulMux,
		Director:           director,
	}
}
复制代码

配置中通过将director作为参数传到handlerChainBuilder的回调方法内,完成对gorestfulContainer的handler的注册工作。其实director就是一个实现了http.Handler的变量。所以,整个的处理逻辑就是将类型为http.Handler的director作为参数,传递到链式filterDefaultBuildHandlerChain方法内。通过DefaultBuildHandlerChain对每一个步骤的filter操作,完成权限控制等之类的操作。如何通过net/http包实现filter的功能,可以参考这篇文章。完成类似于filter的功能之后,后续就是做启动工作,包括证书验证、TLS认证之类的工作,不做过多赘述。主要看下filterDefaultBuildHandlerChain方法是如何处理接口的鉴权操作。

RBAC启动

Kubernetes中比较重要的用的比较多的可能就是RBAC了。在DefaultBuildHandlerChain方法内,通过调用genericapifilters.WithAuthorization方法,实现对每个接口的权限的filter操作。WithAuthorization方法如下

func WithAuthorization(handler http.Handler, a authorizer.Authorizer, s runtime.NegotiatedSerializer) http.Handler {
	if a == nil {
		klog.Warningf("Authorization is disabled")
		return handler
	}
	return http.HandlerFunc(func(w http.ResponseWriter, req *http.Request) {
		ctx := req.Context()
		ae := request.AuditEventFrom(ctx)

		attributes, err := GetAuthorizerAttributes(ctx)
		if err != nil {
			responsewriters.InternalError(w, req, err)
			return
		}
		authorized, reason, err := a.Authorize(attributes)
		// an authorizer like RBAC could encounter evaluation errors and still allow the request, so authorizer decision is checked before error here.
		if authorized == authorizer.DecisionAllow {
			audit.LogAnnotation(ae, decisionAnnotationKey, decisionAllow)
			audit.LogAnnotation(ae, reasonAnnotationKey, reason)
			handler.ServeHTTP(w, req)
			return
		}
		if err != nil {
			audit.LogAnnotation(ae, reasonAnnotationKey, reasonError)
			responsewriters.InternalError(w, req, err)
			return
		}

		klog.V(4).Infof("Forbidden: %#v, Reason: %q", req.RequestURI, reason)
		audit.LogAnnotation(ae, decisionAnnotationKey, decisionForbid)
		audit.LogAnnotation(ae, reasonAnnotationKey, reason)
		responsewriters.Forbidden(ctx, attributes, w, req, reason, s)
	})
}
复制代码

1、调用GetAuthorizerAttributes方法获取配置的各种属性值;
2、调用Authorize方法判断权限是否通过,不同的权限实现其接口,完成鉴权任务;

可以看到,就包含有RBAC单独的处理;
3、如果鉴权成功通过,则调用 handler.ServeHTTP方法继续下一步的 filter操作;否则,直接返回错误信息。
以RBAC为例, Authorize方法最终调用 VisitRulesFor方法实现权限的判断,方法在 kubernetes/pkg/registry/rbac/validation/rule.go文件内。 VisitRulesFor主要代码如下

func (r *DefaultRuleResolver) VisitRulesFor(user user.Info, namespace string, visitor func(source fmt.Stringer, rule *rbacv1.PolicyRule, err error) bool) {
	if clusterRoleBindings, err := r.clusterRoleBindingLister.ListClusterRoleBindings(); err != nil {
		if !visitor(nil, nil, err) {
			return
		}
	} else {
		sourceDescriber := &clusterRoleBindingDescriber{}
		for _, clusterRoleBinding := range clusterRoleBindings {
			subjectIndex, applies := appliesTo(user, clusterRoleBinding.Subjects, "")
			if !applies {
				continue
			}
			rules, err := r.GetRoleReferenceRules(clusterRoleBinding.RoleRef, "")
			if err != nil {
				if !visitor(nil, nil, err) {
					return
				}
				continue
			}
			sourceDescriber.binding = clusterRoleBinding
			sourceDescriber.subject = &clusterRoleBinding.Subjects[subjectIndex]
			for i := range rules {
				if !visitor(sourceDescriber, &rules[i], nil) {
					return
				}
			}
		}
	}

	if len(namespace) > 0 {
		if roleBindings, err := r.roleBindingLister.ListRoleBindings(namespace); err != nil {
			if !visitor(nil, nil, err) {
				return
			}
		} else {
			sourceDescriber := &roleBindingDescriber{}
			for _, roleBinding := range roleBindings {
				subjectIndex, applies := appliesTo(user, roleBinding.Subjects, namespace)
				if !applies {
					continue
				}
				rules, err := r.GetRoleReferenceRules(roleBinding.RoleRef, namespace)
				if err != nil {
					if !visitor(nil, nil, err) {
						return
					}
					continue
				}
				sourceDescriber.binding = roleBinding
				sourceDescriber.subject = &roleBinding.Subjects[subjectIndex]
				for i := range rules {
					if !visitor(sourceDescriber, &rules[i], nil) {
						return
					}
				}
			}
		}
	}
}
复制代码

主要工作就是对clusterRoleBinding以及roleBinding与配置的资源进行判断,比较清晰明了,这与我们使用RBAC的思路基本一致。

数据库操作

ApiServer与数据库的交互主要指的是与etcd的交互。Kubernetes所有的组件不直接与etcd交互,都是通过请求apiserver,apiserver与etcd进行交互完成数据的最终落盘。
在之前的路由实现已经说过,apiserver最终实现的handler对应的后端数据是以Store的结构保存的。这里以api开头的路由举例,在NewLegacyRESTStorage方法中,通过NewREST或者NewStorage会生成各种资源对应的Storage,以endpoints为例,生成的方法如下

func NewREST(optsGetter generic.RESTOptionsGetter) *REST {
	store := &genericregistry.Store{
		NewFunc:                  func() runtime.Object { return &api.Endpoints{} },
		NewListFunc:              func() runtime.Object { return &api.EndpointsList{} },
		DefaultQualifiedResource: api.Resource("endpoints"),

		CreateStrategy: endpoint.Strategy,
		UpdateStrategy: endpoint.Strategy,
		DeleteStrategy: endpoint.Strategy,

		TableConvertor: printerstorage.TableConvertor{TablePrinter: printers.NewTablePrinter().With(printersinternal.AddHandlers)},
	}
	options := &generic.StoreOptions{RESTOptions: optsGetter}
	if err := store.CompleteWithOptions(options); err != nil {
		panic(err) // TODO: Propagate error up
	}
	return &REST{store}
}
复制代码

主要看CompleteWithOptions方法,在CompleteWithOptions方法内,调用了RESTOptions的GetRESTOptions方法,依次调用StorageWithCacher-->NewRawStorage-->Create方法创建最终依赖的后端存储:

// Create creates a storage backend based on given config.
func Create(c storagebackend.Config) (storage.Interface, DestroyFunc, error) {
	switch c.Type {
	case "etcd2":
		return nil, nil, fmt.Errorf("%v is no longer a supported storage backend", c.Type)
	case storagebackend.StorageTypeUnset, storagebackend.StorageTypeETCD3:
		return newETCD3Storage(c)
	default:
		return nil, nil, fmt.Errorf("unknown storage type: %s", c.Type)
	}
}
复制代码

可以看到,通过Create方法判断是创建etcd2或是etcd3的后端etcd版本,目前版本默认的是etcd3。
创建完成对应的存储之后,接下来要做的工作就是将对应的handler方法和最终的后台存储实现绑定起来(handler方法处理最终的数据需要落盘)。
还记着之前说的有个比较长的方法registerResourceHandlers,用来处理具体的handler路由。再次回到该方法,

如图,每个case语句代码一种请求类型和相对应的handler方法,以 POST方法为例,对应的是Create操作。 handler参数的调用最终都会走到 createHandler方法处,位于 kubernetes/vendor/k8s.io/apiserver/pkg/endpoints/handlers/crete.go下。最核心的步骤即调用了Create方法 一步步走下去,最终调用的是位于 kubernetes/staging/src/k8s.io/apiserver/pkg/registry/generic/registry/store.go下的 Create方法。该方法主要包含 BeforeCreateStorage.CreateAfterCreate以及 Decorator等主要方法。对应于 POST操作,则最主要的方法为 Storage.Create。由于目前使用基本都是etcd3,所以实现的方法如下

func (s *store) Create(ctx context.Context, key string, obj, out runtime.Object, ttl uint64) error {
	if version, err := s.versioner.ObjectResourceVersion(obj); err == nil && version != 0 {
		return errors.New("resourceVersion should not be set on objects to be created")
	}
	if err := s.versioner.PrepareObjectForStorage(obj); err != nil {
		return fmt.Errorf("PrepareObjectForStorage failed: %v", err)
	}
	data, err := runtime.Encode(s.codec, obj)
	if err != nil {
		return err
	}
	key = path.Join(s.pathPrefix, key)

	opts, err := s.ttlOpts(ctx, int64(ttl))
	if err != nil {
		return err
	}

	newData, err := s.transformer.TransformToStorage(data, authenticatedDataString(key))
	if err != nil {
		return storage.NewInternalError(err.Error())
	}

	txnResp, err := s.client.KV.Txn(ctx).If(
		notFound(key),
	).Then(
		clientv3.OpPut(key, string(newData), opts...),
	).Commit()
	if err != nil {
		return err
	}
	if !txnResp.Succeeded {
		return storage.NewKeyExistsError(key, 0)
	}

	if out != nil {
		putResp := txnResp.Responses[0].GetResponsePut()
		return decode(s.codec, s.versioner, data, out, putResp.Header.Revision)
	}
	return nil
}
复制代码

主要操作为:
1、调用Encode方法序列化;
2、调用path.Join解析Key;
3、调用TransformToStorage将数据类型进行转换;
4、调用客户端方法进行etcd的写入操作。
至此,完成handler处理与对应的etcd数据库操作的绑定,即完成整个路由后端的操作步骤。
对etcd操作更具体的可以参考这篇文章。

以上均为个人学习总结,如果错误欢迎指正!

转载于:https://juejin.im/post/5c934e5a5188252d7c216981

你可能感兴趣的:(Kubernetes源码分析之kube-apiserver)