KubernetesAPIServer的原理是什么

今天就跟大家聊聊有关Kubernetes APIServer的原理是什么,可能很多人都不太了解,为了让大家更加了解,小编给大家总结了以下内容,希望大家根据这篇文章可以有所收获。

创新互联建站专业为企业提供石鼓网站建设、石鼓做网站、石鼓网站设计、石鼓网站制作等企业网站建设、网页设计与制作、石鼓企业网站模板建站服务,10多年石鼓做网站经验,不只是建网站,更提供有价值的思路和整体网络服务。

前言

整个Kubernetes技术体系由声明式API以及Controller构成,而kube-apiserver是Kubernetes的声明式api server,并为其它组件交互提供了桥梁。因此加深对kube-apiserver的理解就显得至关重要了。

Kubernetes APIServer的原理是什么

整体组件功能

kube-apiserver作为整个Kubernetes集群操作etcd的唯一入口,负责Kubernetes各资源的认证&鉴权,校验以及CRUD等操作,提供RESTful APIs,供其它组件调用:

Kubernetes APIServer的原理是什么

kube-apiserver包含三种APIServer:

  • aggregatorServer:负责处理 apiregistration.k8s.io 组下的APIService资源请求,同时将来自用户的请求拦截转发给aggregated server(AA)

  • kubeAPIServer:负责对请求的一些通用处理,包括:认证、鉴权以及各个内建资源(pod, deployment,service and etc)的REST服务等

  • apiExtensionsServer:负责CustomResourceDefinition(CRD)apiResources以及apiVersions的注册,同时处理CRD以及相应CustomResource(CR)的REST请求(如果对应CR不能被处理的话则会返回404),也是apiserver Delegation的最后一环

另外还包括bootstrap-controller,主要负责Kubernetes default apiserver service的创建以及管理。

接下来将对上述组件进行概览性总结。

bootstrap-controller

  • apiserver bootstrap-controller创建&运行逻辑在k8s.io/kubernetes/pkg/master目录

  • bootstrap-controller主要用于创建以及维护内部kubernetes default apiserver service

  • kubernetes default apiserver service spec.selector为空,这是default apiserver service与其它正常service的最大区别,表明了这个特殊的service对应的endpoints不由endpoints controller控制,而是直接受kube-apiserver bootstrap-controller管理(maintained by this code, not by the pod selector)

  • bootstrap-controller的几个主要功能如下:

    • 创建 default、kube-system 和 kube-public 以及 kube-node-lease 命名空间

    • 创建&维护kubernetes default apiserver service以及对应的endpoint

    • 提供基于Service ClusterIP的检查及修复功能(--service-cluster-ip-range指定范围)

    • 提供基于Service NodePort的检查及修复功能(--service-node-port-range指定范围)

// k8s.io/kubernetes/pkg/master/controller.go:142
// Start begins the core controller loops that must exist for bootstrapping
// a cluster.
func (c *Controller) Start() {
    if c.runner != nil {
        return
    }
    // Reconcile during first run removing itself until server is ready.
    endpointPorts := createEndpointPortSpec(c.PublicServicePort, "https", c.ExtraEndpointPorts)
    if err := c.EndpointReconciler.RemoveEndpoints(kubernetesServiceName, c.PublicIP, endpointPorts); err != nil {
        klog.Errorf("Unable to remove old endpoints from kubernetes service: %v", err)
    }
    repairClusterIPs := servicecontroller.NewRepair(c.ServiceClusterIPInterval, c.ServiceClient, c.EventClient, &c.ServiceClusterIPRange, c.ServiceClusterIPRegistry, &c.SecondaryServiceClusterIPRange, c.SecondaryServiceClusterIPRegistry)
    repairNodePorts := portallocatorcontroller.NewRepair(c.ServiceNodePortInterval, c.ServiceClient, c.EventClient, c.ServiceNodePortRange, c.ServiceNodePortRegistry)
    // run all of the controllers once prior to returning from Start.
    if err := repairClusterIPs.RunOnce(); err != nil {
        // If we fail to repair cluster IPs apiserver is useless. We should restart and retry.
        klog.Fatalf("Unable to perform initial IP allocation check: %v", err)
    }
    if err := repairNodePorts.RunOnce(); err != nil {
        // If we fail to repair node ports apiserver is useless. We should restart and retry.
        klog.Fatalf("Unable to perform initial service nodePort check: %v", err)
    }
    // 定期执行bootstrap controller主要的四个功能(reconciliation)  
    c.runner = async.NewRunner(c.RunKubernetesNamespaces, c.RunKubernetesService, repairClusterIPs.RunUntil, repairNodePorts.RunUntil)
    c.runner.Start()
}

更多代码原理详情,参考 kubernetes-reading-notes 。

kubeAPIServer

KubeAPIServer主要提供对内建API Resources的操作请求,为Kubernetes中各API Resources注册路由信息,同时暴露RESTful API,使集群中以及集群外的服务都可以通过RESTful API操作Kubernetes中的资源

另外,kubeAPIServer是整个Kubernetes apiserver的核心,下面将要讲述的aggregatorServer以及apiExtensionsServer都是建立在kubeAPIServer基础上进行扩展的(补充了Kubernetes对用户自定义资源的能力支持)

kubeAPIServer最核心的功能是为Kubernetes内置资源添加路由,如下:

  • 调用 m.InstallLegacyAPI 将核心 API Resources添加到路由中,在apiserver中即是以 /api 开头的 resource;

  • 调用 m.InstallAPIs 将扩展的 API Resources添加到路由中,在apiserver中即是以 /apis 开头的 resource;

// k8s.io/kubernetes/pkg/master/master.go:332
// 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
func (c completedConfig) New(delegationTarget genericapiserver.DelegationTarget) (*Master, error) {
    ...
    // 安装 LegacyAPI(core API)
    // install legacy rest storage
    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,
            SecondaryServiceIPRange:     c.ExtraConfig.SecondaryServiceIPRange,
            ServiceNodePortRange:        c.ExtraConfig.ServiceNodePortRange,
            LoopbackClientConfig:        c.GenericConfig.LoopbackClientConfig,
            ServiceAccountIssuer:        c.ExtraConfig.ServiceAccountIssuer,
            ServiceAccountMaxExpiration: c.ExtraConfig.ServiceAccountMaxExpiration,
            APIAudiences:                c.GenericConfig.Authentication.APIAudiences,
        }
        if err := m.InstallLegacyAPI(&c, c.GenericConfig.RESTOptionsGetter, legacyRESTStorageProvider); err != nil {
            return nil, err
        }
    }
    ...
    // 安装 APIs(named groups apis)
    if err := m.InstallAPIs(c.ExtraConfig.APIResourceConfigSource, c.GenericConfig.RESTOptionsGetter, restStorageProviders...); err != nil {
        return nil, err
    }
    ...
    return m, nil
}

整个kubeAPIServer提供了三类API Resource接口:

  • core group:主要在 /api/v1 下;

  • named groups:其 path 为 /apis/$GROUP/$VERSION

  • 系统状态的一些 API:如/metrics/version 等;

而API的URL大致以 /apis/{group}/{version}/namespaces/{namespace}/resource/{name} 组成,结构如下图所示:

Kubernetes APIServer的原理是什么

kubeAPIServer会为每种API资源创建对应的RESTStorage,RESTStorage的目的是将每种资源的访问路径及其后端存储的操作对应起来:通过构造的REST Storage实现的接口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中,每一个操作对应一个标准的REST method,如create对应REST method为POST,而update对应REST method为PUT。最终根据actions数组依次遍历,对每一个操作添加一个handler(handler对应REST Storage实现的相关接口),并注册到route,最终对外提供RESTful API,如下:

// m.GenericAPIServer.InstallLegacyAPIGroup --> s.installAPIResources --> apiGroupVersion.InstallREST --> installer.Install --> a.registerResourceHandlers
// k8s.io/kubernetes/staging/src/k8s.io/apiserver/pkg/endpoints/installer.go:181
func (a *APIInstaller) registerResourceHandlers(path string, storage rest.Storage, ws *restful.WebService) (*metav1.APIResource, error) {
    ...
    // 1、判断该 resource 实现了哪些 REST 操作接口,以此来判断其支持的 verbs 以便为其添加路由
    // what verbs are supported by the storage, used to know what verbs we support per path
    creater, isCreater := storage.(rest.Creater)
    namedCreater, isNamedCreater := storage.(rest.NamedCreater)
    lister, isLister := storage.(rest.Lister)
    getter, isGetter := storage.(rest.Getter)
    ...
    // 2、为 resource 添加对应的 actions(+根据是否支持 namespace)
    // Get the list of actions for the given scope.
    switch {
    case !namespaceScoped:
        // Handle non-namespace scoped resources like nodes.
        resourcePath := resource
        resourceParams := params
        itemPath := resourcePath + "/{name}"
        nameParams := append(params, nameParam)
        proxyParams := append(nameParams, pathParam)
        ...
        // Handler for standard REST verbs (GET, PUT, POST and DELETE).
        // Add actions at the resource path: /api/apiVersion/resource
        actions = appendIf(actions, action{"LIST", resourcePath, resourceParams, namer, false}, isLister)
        actions = appendIf(actions, action{"POST", resourcePath, resourceParams, namer, false}, isCreater)
        ...
    }
    ...
    // 3、从 rest.Storage 到 restful.Route 映射
    // 为每个操作添加对应的 handler
    for _, action := range actions {
        ...
        switch action.Verb {
        ...
        case "POST": // Create a resource.
            var handler restful.RouteFunction
            // 4、初始化 handler
            if isNamedCreater {
                handler = restfulCreateNamedResource(namedCreater, reqScope, admit)
            } else {
                handler = restfulCreateResource(creater, reqScope, admit)
            }
            handler = metrics.InstrumentRouteFunc(action.Verb, group, version, resource, subresource, requestScope, metrics.APIServerComponent, handler)
            ...
            // 5、route 与 handler 进行绑定    
            route := ws.POST(action.Path).To(handler).
                Doc(doc).
                Param(ws.QueryParameter("pretty", "If 'true', then the output is pretty printed.")).
                Operation("create"+namespaced+kind+strings.Title(subresource)+operationSuffix).
                Produces(append(storageMeta.ProducesMIMETypes(action.Verb), mediaTypes...)...).
                Returns(http.StatusOK, "OK", producedObject).
                // TODO: in some cases, the API may return a v1.Status instead of the versioned object
                // but currently go-restful can't handle multiple different objects being returned.
                Returns(http.StatusCreated, "Created", producedObject).
                Returns(http.StatusAccepted, "Accepted", producedObject).
                Reads(defaultVersionedObject).
                Writes(producedObject)
            if err := AddObjectParams(ws, route, versionedCreateOptions); err != nil {
                return nil, err
            }
            addParams(route, action.Params)
            // 6、添加到路由中    
            routes = append(routes, route)
        case "DELETE": // Delete a resource.
        ...
        default:
            return nil, fmt.Errorf("unrecognized action verb: %s", action.Verb)
        }
        for _, route := range routes {
            route.Metadata(ROUTE_META_GVK, metav1.GroupVersionKind{
                Group:   reqScope.Kind.Group,
                Version: reqScope.Kind.Version,
                Kind:    reqScope.Kind.Kind,
            })
            route.Metadata(ROUTE_META_ACTION, strings.ToLower(action.Verb))
            ws.Route(route)
        }
        // Note: update GetAuthorizerAttributes() when adding a custom handler.
    }
    ...
}

kubeAPIServer代码结构整理如下:

1. apiserver整体启动逻辑 k8s.io/kubernetes/cmd/kube-apiserver
2. apiserver bootstrap-controller创建&运行逻辑 k8s.io/kubernetes/pkg/master
3. API Resource对应后端RESTStorage(based on genericregistry.Store)创建k8s.io/kubernetes/pkg/registry
4. aggregated-apiserver创建&处理逻辑 k8s.io/kubernetes/staging/src/k8s.io/kube-aggregator
5. extensions-apiserver创建&处理逻辑 k8s.io/kubernetes/staging/src/k8s.io/apiextensions-apiserver
6. apiserver创建&运行 k8s.io/kubernetes/staging/src/k8s.io/apiserver/pkg/server
7. 注册API Resource资源处理handler(InstallREST&Install®isterResourceHandlers) k8s.io/kubernetes/staging/src/k8s.io/apiserver/pkg/endpoints
8. 创建存储后端(etcdv3) k8s.io/kubernetes/staging/src/k8s.io/apiserver/pkg/storage
9. genericregistry.Store.CompleteWithOptions初始化 k8s.io/kubernetes/staging/src/k8s.io/apiserver/pkg/registry

调用链整理如下:

Kubernetes APIServer的原理是什么

更多代码原理详情,参考 kubernetes-reading-notes 。

aggregatorServer

aggregatorServer主要用于处理扩展Kubernetes API Resources的第二种方式Aggregated APIServer(AA),将CR请求代理给AA:

Kubernetes APIServer的原理是什么

这里结合Kubernetes官方给出的aggregated apiserver例子sample-apiserver,总结原理如下:

  • aggregatorServer通过APIServices对象关联到某个Service来进行请求的转发,其关联的Service类型进一步决定了请求转发的形式。aggregatorServer包括一个GenericAPIServer和维护自身状态的Controller。其中GenericAPIServer主要处理apiregistration.k8s.io组下的APIService资源请求,而Controller包括:

    • apiserviceRegistrationController:负责根据APIService定义的aggregated server service构建代理,将CR的请求转发给后端的aggregated server

    • availableConditionController:维护 APIServices 的可用状态,包括其引用 Service 是否可用等;

    • autoRegistrationController:用于保持 API 中存在的一组特定的 APIServices;

    • crdRegistrationController:负责将 CRD GroupVersions 自动注册到 APIServices 中;

    • openAPIAggregationController:将 APIServices 资源的变化同步至提供的 OpenAPI 文档;

  • apiserviceRegistrationController负责根据APIService定义的aggregated server service构建代理,将CR的请求转发给后端的aggregated server。apiService有两种类型:Local(Service为空)以及Service(Service非空)。apiserviceRegistrationController负责对这两种类型apiService设置代理:Local类型会直接路由给kube-apiserver进行处理;而Service类型则会设置代理并将请求转化为对aggregated Service的请求(proxyPath := "/apis/" + apiService.Spec.Group + "/" + apiService.Spec.Version),而请求的负载均衡策略则是优先本地访问kube-apiserver(如果service为kubernetes default apiserver service:443)=>通过service ClusterIP:Port访问(默认) 或者 通过随机选择service endpoint backend进行访问:

    func (s *APIAggregator) AddAPIService(apiService *v1.APIService) error {
      ...
        proxyPath := "/apis/" + apiService.Spec.Group + "/" + apiService.Spec.Version
        // v1. is a special case for the legacy API.  It proxies to a wider set of endpoints.
        if apiService.Name == legacyAPIServiceName {
            proxyPath = "/api"
        }
        // register the proxy handler
        proxyHandler := &proxyHandler{
            localDelegate:   s.delegateHandler,
            proxyClientCert: s.proxyClientCert,
            proxyClientKey:  s.proxyClientKey,
            proxyTransport:  s.proxyTransport,
            serviceResolver: s.serviceResolver,
            egressSelector:  s.egressSelector,
        }
      ...
        s.proxyHandlers[apiService.Name] = proxyHandler
        s.GenericAPIServer.Handler.NonGoRestfulMux.Handle(proxyPath, proxyHandler)
        s.GenericAPIServer.Handler.NonGoRestfulMux.UnlistedHandlePrefix(proxyPath+"/", proxyHandler)
      ...
        // it's time to register the group aggregation endpoint
        groupPath := "/apis/" + apiService.Spec.Group
        groupDiscoveryHandler := &apiGroupHandler{
            codecs:    aggregatorscheme.Codecs,
            groupName: apiService.Spec.Group,
            lister:    s.lister,
            delegate:  s.delegateHandler,
        }
        // aggregation is protected
        s.GenericAPIServer.Handler.NonGoRestfulMux.Handle(groupPath, groupDiscoveryHandler)
        s.GenericAPIServer.Handler.NonGoRestfulMux.UnlistedHandle(groupPath+"/", groupDiscoveryHandler)
        s.handledGroups.Insert(apiService.Spec.Group)
        return nil
    }
    // k8s.io/kubernetes/staging/src/k8s.io/kube-aggregator/pkg/apiserver/handler_proxy.go:109
    func (r *proxyHandler) ServeHTTP(w http.ResponseWriter, req *http.Request) {
        // 加载roxyHandlingInfo处理请求  
        value := r.handlingInfo.Load()
        if value == nil {
            r.localDelegate.ServeHTTP(w, req)
            return
        }
        handlingInfo := value.(proxyHandlingInfo)
      ...
        // 判断APIService服务是否正常
        if !handlingInfo.serviceAvailable {
            proxyError(w, req, "service unavailable", http.StatusServiceUnavailable)
            return
        }
        // 将原始请求转化为对APIService的请求
        // write a new location based on the existing request pointed at the target service
        location := &url.URL{}
        location.Scheme = "https"
        rloc, err := r.serviceResolver.ResolveEndpoint(handlingInfo.serviceNamespace, handlingInfo.serviceName, handlingInfo.servicePort)
        if err != nil {
            klog.Errorf("error resolving %s/%s: %v", handlingInfo.serviceNamespace, handlingInfo.serviceName, err)
            proxyError(w, req, "service unavailable", http.StatusServiceUnavailable)
            return
        }
        location.Host = rloc.Host
        location.Path = req.URL.Path
        location.RawQuery = req.URL.Query().Encode()
        newReq, cancelFn := newRequestForProxy(location, req)
        defer cancelFn()
       ...
        proxyRoundTripper = transport.NewAuthProxyRoundTripper(user.GetName(), user.GetGroups(), user.GetExtra(), proxyRoundTripper)
        handler := proxy.NewUpgradeAwareHandler(location, proxyRoundTripper, true, upgrade, &responder{w: w})
        handler.ServeHTTP(w, newReq)
    }

    $ kubectl get APIService           
    NAME                                   SERVICE                      AVAILABLE   AGE
    ...
    v1.apps                                Local                        True        50d
    ...
    v1beta1.metrics.k8s.io                 kube-system/metrics-server   True        50d
    ...

    # default APIServices
    $ kubectl get -o yaml APIService/v1.apps
    apiVersion: apiregistration.k8s.io/v1
    kind: APIService
    metadata:
      labels:
        kube-aggregator.kubernetes.io/automanaged: onstart
      name: v1.apps
      selfLink: /apis/apiregistration.k8s.io/v1/apiservices/v1.apps
    spec:
      group: apps
      groupPriorityMinimum: 17800
      version: v1
      versionPriority: 15
    status:
      conditions:
      - lastTransitionTime: "2020-10-20T10:39:48Z"
        message: Local APIServices are always available
        reason: Local
        status: "True"
        type: Available
    
    # aggregated server    
    $ kubectl get -o yaml APIService/v1beta1.metrics.k8s.io
    apiVersion: apiregistration.k8s.io/v1
    kind: APIService
    metadata:
      labels:
        addonmanager.kubernetes.io/mode: Reconcile
        kubernetes.io/cluster-service: "true"
      name: v1beta1.metrics.k8s.io
      selfLink: /apis/apiregistration.k8s.io/v1/apiservices/v1beta1.metrics.k8s.io
    spec:
      group: metrics.k8s.io
      groupPriorityMinimum: 100
      insecureSkipTLSVerify: true
      service:
        name: metrics-server
        namespace: kube-system
        port: 443
      version: v1beta1
      versionPriority: 100
    status:
      conditions:
      - lastTransitionTime: "2020-12-05T00:50:48Z"
        message: all checks passed
        reason: Passed
        status: "True"
        type: Available
    
    # CRD
    $ kubectl get -o yaml APIService/v1.duyanghao.example.com
    apiVersion: apiregistration.k8s.io/v1
    kind: APIService
    metadata:
      labels:
        kube-aggregator.kubernetes.io/automanaged: "true"
      name: v1.duyanghao.example.com
      selfLink: /apis/apiregistration.k8s.io/v1/apiservices/v1.duyanghao.example.com
    spec:
      group: duyanghao.example.com
      groupPriorityMinimum: 1000
      version: v1
      versionPriority: 100
    status:
      conditions:
      - lastTransitionTime: "2020-12-11T08:45:37Z"
        message: Local APIServices are always available
        reason: Local
        status: "True"
        type: Available

  • aggregatorServer创建过程中会根据所有kube-apiserver定义的API资源创建默认的APIService列表,名称即是$VERSION.$GROUP,这些APIService都会有标签kube-aggregator.kubernetes.io/automanaged: onstart,例如:v1.apps apiService。autoRegistrationController创建并维护这些列表中的APIService,也即我们看到的Local apiService;对于自定义的APIService(aggregated server),则不会对其进行处理

  • aggregated server实现CR(自定义API资源) 的CRUD API接口,并可以灵活选择后端存储,可以与core kube-apiserver一起公用etcd,也可自己独立部署etcd数据库或者其它数据库。aggregated server实现的CR API路径为:/apis/$GROUP/$VERSION,具体到sample apiserver为:/apis/wardle.example.com/v1alpha1,下面的资源类型有:flunders以及fischers

  • aggregated server通过部署APIService类型资源,service fields指向对应的aggregated server service实现与core kube-apiserver的集成与交互

  • sample-apiserver目录结构如下,可参考编写自己的aggregated server:

    staging/src/k8s.io/sample-apiserver
    ├── artifacts
    │   ├── example
    │   │   ├── apiservice.yaml
          ...
    ├── hack
    ├── main.go
    └── pkg
    ├── admission
    ├── apis
    ├── apiserver
    ├── cmd
    ├── generated
    │   ├── clientset
    │   │   └── versioned
                  ...
    │   │       └── typed
    │   │           └── wardle
    │   │               ├── v1alpha1
    │   │               └── v1beta1
    │   ├── informers
    │   │   └── externalversions
    │   │       └── wardle
    │   │           ├── v1alpha1
    │   │           └── v1beta1
    │   ├── listers
    │   │   └── wardle
    │   │       ├── v1alpha1
    │   │       └── v1beta1
    └── registry

     

    • 其中,artifacts用于部署yaml示例

    • hack目录存放自动脚本(eg: update-codegen)

    • main.go是aggregated server启动入口;pkg/cmd负责启动aggregated server具体逻辑;pkg/apiserver用于aggregated server初始化以及路由注册

    • pkg/apis负责相关CR的结构体定义,自动生成(update-codegen)

    • pkg/admission负责准入的相关代码

    • pkg/generated负责生成访问CR的clientset,informers,以及listers

    • pkg/registry目录负责CR相关的RESTStorage实现

更多代码原理详情,参考 kubernetes-reading-notes 。

apiExtensionsServer

apiExtensionsServer主要负责CustomResourceDefinition(CRD)apiResources以及apiVersions的注册,同时处理CRD以及相应CustomResource(CR)的REST请求(如果对应CR不能被处理的话则会返回404),也是apiserver Delegation的最后一环

原理总结如下:

  • Custom Resource,简称CR,是Kubernetes自定义资源类型,与之相对应的就是Kubernetes内置的各种资源类型,例如Pod、Service等。利用CR我们可以定义任何想要的资源类型

  • CRD通过yaml文件的形式向Kubernetes注册CR实现自定义api-resources,属于第二种扩展Kubernetes API资源的方式,也是普遍使用的一种

  • APIExtensionServer负责CustomResourceDefinition(CRD)apiResources以及apiVersions的注册,同时处理CRD以及相应CustomResource(CR)的REST请求(如果对应CR不能被处理的话则会返回404),也是apiserver Delegation的最后一环

  • crdRegistrationController负责将CRD GroupVersions自动注册到APIServices中。具体逻辑为:枚举所有CRDs,然后根据CRD定义的crd.Spec.Group以及crd.Spec.Versions字段构建APIService,并添加到autoRegisterController.apiServicesToSync中,由autoRegisterController进行创建以及维护操作。这也是为什么创建完CRD后会产生对应的APIService对象

  • APIExtensionServer包含的controller以及功能如下所示:

     - `namingController`:检查 crd obj 中是否有命名冲突,可在 crd `.status.conditions` 中查看;
    
     - `establishingController`:检查 crd 是否处于正常状态,可在 crd `.status.conditions` 中查看;
    
     - `nonStructuralSchemaController`:检查 crd obj 结构是否正常,可在 crd `.status.conditions` 中查看;
    
     - `apiApprovalController`:检查 crd 是否遵循 Kubernetes API 声明策略,可在 crd `.status.conditions` 中查看;
    
     - `finalizingController`:类似于 finalizes 的功能,与 CRs 的删除有关;

    • openapiController:将 crd 资源的变化同步至提供的 OpenAPI 文档,可通过访问 /openapi/v2 进行查看;

    • crdController:负责将 crd 信息注册到 apiVersions 和 apiResources 中,两者的信息可通过 kubectl api-versionskubectl api-resources 查看;

    • kubectl api-versions命令返回所有Kubernetes集群资源的版本信息(实际发出了两个请求,分别是https://127.0.0.1:6443/api以及https://127.0.0.1:6443/apis,并在最后将两个请求的返回结果进行了合并)

      $ kubectl -v=8 api-versions 
      I1211 11:44:50.276446   22493 loader.go:375] Config loaded from file:  /root/.kube/config
      I1211 11:44:50.277005   22493 round_trippers.go:420] GET https://127.0.0.1:6443/api?timeout=32s
      ...
      I1211 11:44:50.290265   22493 request.go:1068] Response Body: {"kind":"APIVersions","versions":["v1"],"serverAddressByClientCIDRs":[{"clientCIDR":"0.0.0.0/0","serverAddress":"x.x.x.x:6443"}]}
      I1211 11:44:50.293673   22493 round_trippers.go:420] GET https://127.0.0.1:6443/apis?timeout=32s
      ...
      I1211 11:44:50.298360   22493 request.go:1068] Response Body: {"kind":"APIGroupList","apiVersion":"v1","groups":[{"name":"apiregistration.k8s.io","versions":[{"groupVersion":"apiregistration.k8s.io/v1","version":"v1"},{"groupVersion":"apiregistration.k8s.io/v1beta1","version":"v1beta1"}],"preferredVersion":{"groupVersion":"apiregistration.k8s.io/v1","version":"v1"}},{"name":"extensions","versions":[{"groupVersion":"extensions/v1beta1","version":"v1beta1"}],"preferredVersion":{"groupVersion":"extensions/v1beta1","version":"v1beta1"}},{"name":"apps","versions":[{"groupVersion":"apps/v1","version":"v1"}],"preferredVersion":{"groupVersion":"apps/v1","version":"v1"}},{"name":"events.k8s.io","versions":[{"groupVersion":"events.k8s.io/v1beta1","version":"v1beta1"}],"preferredVersion":{"groupVersion":"events.k8s.io/v1beta1","version":"v1beta1"}},{"name":"authentication.k8s.io","versions":[{"groupVersion":"authentication.k8s.io/v1","version":"v1"},{"groupVersion":"authentication.k8s.io/v1beta1","version":"v1beta1"}],"preferredVersion":{"groupVersion":"authentication.k8s.io/v1"," [truncated 4985 chars]
      apiextensions.k8s.io/v1
      apiextensions.k8s.io/v1beta1
      apiregistration.k8s.io/v1
      apiregistration.k8s.io/v1beta1
      apps/v1
      authentication.k8s.io/v1beta1
      ...
      storage.k8s.io/v1
      storage.k8s.io/v1beta1
      v1

    • kubectl api-resources命令就是先获取所有API版本信息,然后对每一个API版本调用接口获取该版本下的所有API资源类型

      $ kubectl -v=8 api-resources
       5077 loader.go:375] Config loaded from file:  /root/.kube/config
       I1211 15:19:47.593450   15077 round_trippers.go:420] GET https://127.0.0.1:6443/api?timeout=32s
       I1211 15:19:47.602273   15077 request.go:1068] Response Body: {"kind":"APIVersions","versions":["v1"],"serverAddressByClientCIDRs":[{"clientCIDR":"0.0.0.0/0","serverAddress":"x.x.x.x:6443"}]}
       I1211 15:19:47.606279   15077 round_trippers.go:420] GET https://127.0.0.1:6443/apis?timeout=32s
       I1211 15:19:47.610333   15077 request.go:1068] Response Body: {"kind":"APIGroupList","apiVersion":"v1","groups":[{"name":"apiregistration.k8s.io","versions":[{"groupVersion":"apiregistration.k8s.io/v1","version":"v1"},{"groupVersion":"apiregistration.k8s.io/v1beta1","version":"v1beta1"}],"preferredVersion":{"groupVersion":"apiregistration.k8s.io/v1","version":"v1"}},{"name":"extensions","versions":[{"groupVersion":"extensions/v1beta1","version":"v1beta1"}],"preferredVersion":{"groupVersion":"extensions/v1beta1","version":"v1beta1"}},{"name":"apps","versions":[{"groupVersion":"apps/v1","version":"v1"}],"preferredVersion":{"groupVersion":"apps/v1","version":"v1"}},{"name":"events.k8s.io","versions":[{"groupVersion":"events.k8s.io/v1beta1","version":"v1beta1"}],"preferredVersion":{"groupVersion":"events.k8s.io/v1beta1","version":"v1beta1"}},{"name":"authentication.k8s.io","versions":[{"groupVersion":"authentication.k8s.io/v1","version":"v1"},{"groupVersion":"authentication.k8s.io/v1beta1","version":"v1beta1"}],"preferredVersion":{"groupVersion":"authentication.k8s.io/v1"," [truncated 4985 chars]
       I1211 15:19:47.614700   15077 round_trippers.go:420] GET https://127.0.0.1:6443/apis/batch/v1?timeout=32s
       I1211 15:19:47.614804   15077 round_trippers.go:420] GET https://127.0.0.1:6443/apis/authentication.k8s.io/v1?timeout=32s
       I1211 15:19:47.615687   15077 round_trippers.go:420] GET https://127.0.0.1:6443/apis/auth.tkestack.io/v1?timeout=32s
       https://127.0.0.1:6443/apis/authentication.k8s.io/v1beta1?timeout=32s
       I1211 15:19:47.616794   15077 round_trippers.go:420] GET https://127.0.0.1:6443/apis/coordination.k8s.io/v1?timeout=32s
       I1211 15:19:47.616863   15077 round_trippers.go:420] GET https://127.0.0.1:6443/apis/apps/v1?timeout=32s
       ...
       NAME                              SHORTNAMES   APIGROUP                       NAMESPACED   KIND
       bindings                                                                      true         Binding
       endpoints                         ep                                          true         Endpoints
       events                            ev                                          true         Event
       limitranges                       limits                                      true         LimitRange
       namespaces                        ns                                          false        Namespace
       nodes                             no                                          false        Node
       ...

  • 总结CR CRUD APIServer处理逻辑如下:

    • 解析req(GET /apis/duyanghao.example.com/v1/namespaces/default/students),根据请求路径中的group(duyanghao.example.com),version(v1),以及resource字段(students)获取对应CRD内容(crd, err := r.crdLister.Get(crdName))

    • 通过crd.UID以及crd.Name获取crdInfo,若不存在则创建对应的crdInfo(crdInfo, err := r.getOrCreateServingInfoFor(crd.UID, crd.Name))。crdInfo中包含了CRD定义以及该CRD对应Custom Resource的customresource.REST storage

    • customresource.REST storage由CR对应的Group(duyanghao.example.com),Version(v1),Kind(Student),Resource(students)等创建完成,由于CR在Kubernetes代码中并没有具体结构体定义,所以这里会先初始化一个范型结构体Unstructured(用于保存所有类型的Custom Resource),并对该结构体进行SetGroupVersionKind操作(设置具体Custom Resource Type)

    • 从customresource.REST storage获取Unstructured结构体后会对其进行相应转换然后返回

      // k8s.io/kubernetes/staging/src/k8s.io/apiextensions-apiserver/pkg/apiserver/customresource_handler.go:223
      func (r *crdHandler) ServeHTTP(w http.ResponseWriter, req *http.Request) {
        ctx := req.Context()
        requestInfo, ok := apirequest.RequestInfoFrom(ctx)
        ...
        crdName := requestInfo.Resource + "." + requestInfo.APIGroup
        crd, err := r.crdLister.Get(crdName)
        ...
        crdInfo, err := r.getOrCreateServingInfoFor(crd.UID, crd.Name)
        verb := strings.ToUpper(requestInfo.Verb)
        resource := requestInfo.Resource
        subresource := requestInfo.Subresource
        scope := metrics.CleanScope(requestInfo)
        ...
        switch {
        case subresource == "status" && subresources != nil && subresources.Status != nil:
            handlerFunc = r.serveStatus(w, req, requestInfo, crdInfo, terminating, supportedTypes)
        case subresource == "scale" && subresources != nil && subresources.Scale != nil:
            handlerFunc = r.serveScale(w, req, requestInfo, crdInfo, terminating, supportedTypes)
        case len(subresource) == 0:
            handlerFunc = r.serveResource(w, req, requestInfo, crdInfo, terminating, supportedTypes)
        default:
            responsewriters.ErrorNegotiated(
                apierrors.NewNotFound(schema.GroupResource{Group: requestInfo.APIGroup, Resource: requestInfo.Resource}, requestInfo.Name),
                Codecs, schema.GroupVersion{Group: requestInfo.APIGroup, Version: requestInfo.APIVersion}, w, req,
            )
        }
        if handlerFunc != nil {
            handlerFunc = metrics.InstrumentHandlerFunc(verb, requestInfo.APIGroup, requestInfo.APIVersion, resource, subresource, scope, metrics.APIServerComponent, handlerFunc)
            handler := genericfilters.WithWaitGroup(handlerFunc, longRunningFilter, crdInfo.waitGroup)
            handler.ServeHTTP(w, req)
            return
        }
      }

    • createAPIExtensionsServer=>NewCustomResourceDefinitionHandler=>crdHandler=>注册CR CRUD API接口:

      // New returns a new instance of CustomResourceDefinitions from the given config.
      func (c completedConfig) New(delegationTarget genericapiserver.DelegationTarget) (*CustomResourceDefinitions, error) {
        ...
          crdHandler, err := NewCustomResourceDefinitionHandler(
            versionDiscoveryHandler,
              groupDiscoveryHandler,
            s.Informers.Apiextensions().V1().CustomResourceDefinitions(),
              delegateHandler,
            c.ExtraConfig.CRDRESTOptionsGetter,
              c.GenericConfig.AdmissionControl,
            establishingController,
              c.ExtraConfig.ServiceResolver,
            c.ExtraConfig.AuthResolverWrapper,
              c.ExtraConfig.MasterCount,
              s.GenericAPIServer.Authorizer,
              c.GenericConfig.RequestTimeout,
              time.Duration(c.GenericConfig.MinRequestTimeout)*time.Second,
              apiGroupInfo.StaticOpenAPISpec,
              c.GenericConfig.MaxRequestBodyBytes,
          )
          if err != nil {
              return nil, err
          }
          s.GenericAPIServer.Handler.NonGoRestfulMux.Handle("/apis", crdHandler)
          s.GenericAPIServer.Handler.NonGoRestfulMux.HandlePrefix("/apis/", crdHandler)
          ...
          return s, nil
      }

Conclusion

本文从源码层面对Kubernetes apiserver进行了一个概览性总结,包括:aggregatorServer,kubeAPIServer,apiExtensionsServer以及bootstrap-controller等。

看完上述内容,你们对Kubernetes APIServer的原理是什么有进一步的了解吗?如果还想了解更多知识或者相关内容,请关注创新互联行业资讯频道,感谢大家的支持。


文章名称:KubernetesAPIServer的原理是什么
标题链接:http://azwzsj.com/article/jehppi.html