RSS

熟悉又陌生的 k8s 字段:finalizers

前言

经常操作 Kubernetes 集群的同学肯定对 finalizers 字段不陌生,每当删除 namespace 或 pod 等一些 Kubernetes 资源时,有时资源状态会卡在 Terminating,很长时间无法删除,甚至有时增加 --force flag 之后还是无法正常删除。这时就需要 edit 该资源,将 finalizers 字段设置为 [],之后 Kubernetes 资源就正常删除了。

这是一个比较常见的操作,但是当有人问 finalizers 字段的作用是什么的时候,我是懵逼的,我甚至不知道这个熟悉又陌生的单词怎么读!那么这篇文章就来探索一下 finalizers 这个字段到底是做什么的,在实践中应该怎么应用这个字段。(另外,这个单词读作 [‘faɪnəlaɪzər]

Finalizers

Finalizers 字段属于 Kubernetes GC 垃圾收集器,是一种删除拦截机制,能够让控制器实现异步的删除前(Pre-delete)回调。其存在于任何一个资源对象的 Meta 中,在 k8s 源码中声明为 []string,该 Slice 的内容为需要执行的拦截器名称。

对带有 Finalizer 的对象的第一个删除请求会为其 metadata.deletionTimestamp 设置一个值,但不会真的删除对象。一旦此值被设置,finalizers 列表中的值就只能被移除。

metadata.deletionTimestamp 字段被设置时,负责监测该对象的各个控制器会通过轮询对该对象的更新请求来执行它们所要处理的所有 Finalizer。 当所有 Finalizer 都被执行过,资源被删除。

metadata.deletionGracePeriodSeconds 的取值控制对更新的轮询周期。

每个控制器要负责将其 Finalizer 从列表中去除。

每执行完一个就从 finalizers 中移除一个,直到 finalizers 为空,之后其宿主资源才会被真正的删除。

// DeletionTimestamp is RFC 3339 date and time at which this resource will be deleted. This
// field is set by the server when a graceful deletion is requested by the user, and is not
// directly settable by a client. The resource is expected to be deleted (no longer visible
// from resource lists, and not reachable by name) after the time in this field, once the
// finalizers list is empty. As long as the finalizers list contains items, deletion is blocked.
// Once the deletionTimestamp is set, this value may not be unset or be set further into the
// future, although it may be shortened or the resource may be deleted prior to this time.
// For example, a user may request that a pod is deleted in 30 seconds. The Kubelet will react
// by sending a graceful termination signal to the containers in the pod. After that 30 seconds,
// the Kubelet will send a hard termination signal (SIGKILL) to the container and after cleanup,
// remove the pod from the API. In the presence of network partitions, this object may still
// exist after this timestamp, until an administrator or automated process can determine the
// resource is fully terminated.
// If not set, graceful deletion of the object has not been requested.
//
// Populated by the system when a graceful deletion is requested.
// Read-only.
// More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#metadata
// +optional
DeletionTimestamp *Time `json:"deletionTimestamp,omitempty" protobuf:"bytes,9,opt,name=deletionTimestamp"`

// Number of seconds allowed for this object to gracefully terminate before
// it will be removed from the system. Only set when deletionTimestamp is also set.
// May only be shortened.
// Read-only.
// +optional
DeletionGracePeriodSeconds *int64 `json:"deletionGracePeriodSeconds,omitempty" protobuf:"varint,10,opt,name=deletionGracePeriodSeconds"`

// Must be empty before the object is deleted from the registry. Each entry
// is an identifier for the responsible component that will remove the entry
// from the list. If the deletionTimestamp of the object is non-nil, entries
// in this list can only be removed.
// Finalizers may be processed and removed in any order.  Order is NOT enforced
// because it introduces significant risk of stuck finalizers.
// finalizers is a shared field, any actor with permission can reorder it.
// If the finalizer list is processed in order, then this can lead to a situation
// in which the component responsible for the first finalizer in the list is
// waiting for a signal (field value, external system, or other) produced by a
// component responsible for a finalizer later in the list, resulting in a deadlock.
// Without enforced ordering finalizers are free to order amongst themselves and
// are not vulnerable to ordering changes in the list.
// +optional
// +patchStrategy=merge
Finalizers []string `json:"finalizers,omitempty" patchStrategy:"merge" protobuf:"bytes,14,rep,name=finalizers"`

在 Operator 中的应用

知道了 Finalizers 是什么,那么当然也要知道怎么用 Finalizers 了。在实际开发 Operator 时,删除前(Pre-delete)回调是一个比较常见的功能,用于处理一些在资源删除前需要处理的逻辑,如:关联资源释放、释放资源通知、相关数据清理,甚至是阻止资源删除。一般 Finalizers 的处理也是会 Reconcile 中实现的,下面就使用 chaosblade-operator 中的源码,简单介绍一些 Finalizers 的使用方式。

首先要了解的是 ChaosBlade-Operator 的工作原理:每个实验都会以 CR 的形式部署到 k8s 集群中,之后由 chaosblade-operator 来操作以 DaemonSet 形式部署 chaosblade-tool 对具体资源进行混沌实验。停止实验只需删除对应 CR 即可,在删除 CR 时,首先会执行一遍实验恢复逻辑,之后才会将 CR 删除。但如果恢复实验失败,则会将 CR 的 Phase 设置为 Destroying,而在 Reconcile 中观测到 Phase 状态为 Destroying 或者 metadata.deletionTimestamp 不为空时,就会不会移除 finalizers 中的拦截器名称,阻止该 CR 被删除。

这样设计的目的是为了在实验恢复失败后,让用户去主动查看实验恢复失败原因,如果是一些意外原因导致的实验恢复失败,及时去处理。在确认原因后,可使用 CLI 工具增加 --force-remove 进去强制删除,项目维护者在 Issue#368 中也就这个设计给出了解答。

// pkg/controller/chaosblade/controller.go 部分源码
...
const chaosbladeFinalizer = "finalizer.chaosblade.io"
...
func (r *ReconcileChaosBlade) Reconcile(request reconcile.Request) (reconcile.Result, error) {
    reqLogger := logrus.WithField("Request.Name", request.Name)
    forget := reconcile.Result{}
    // Fetch the RC instance
    cb := &v1alpha1.ChaosBlade{}
    err := r.client.Get(context.TODO(), request.NamespacedName, cb)
    if err != nil {
      return forget, nil
    }
    if len(cb.Spec.Experiments) == 0 {
      return forget, nil
    }

    // Destroyed->delete
    // Remove the Finalizer if the CR object status is destroyed to delete it
    if cb.Status.Phase == v1alpha1.ClusterPhaseDestroyed {
      cb.SetFinalizers(remove(cb.GetFinalizers(), chaosbladeFinalizer))
      err := r.client.Update(context.TODO(), cb)
      if err != nil {
        reqLogger.WithError(err).Errorln("remove chaosblade finalizer failed at destroyed phase")
      }
      return forget, nil
    }
    if cb.Status.Phase == v1alpha1.ClusterPhaseDestroying || cb.GetDeletionTimestamp() != nil {
      err := r.finalizeChaosBlade(reqLogger, cb)
      if err != nil {
        reqLogger.WithError(err).Errorln("finalize chaosblade failed at destroying phase")
      }
      return forget, nil
    }
    ...
    return forget, nil
}

如果 Phase 状态为 Destroyed,则从 Finalizers 中移除 finalizer.chaosblade.io,之后正常删除 CR。

结语

在实际工作中,像 Finalizers 这样的东西太多了,很多平时挂在嘴边的东西,深究起来我们可能对其并不了解,甚至原本的理解就是错误的。在今后的文章中,除了各种实践干货,笔者还会将更多的精力投注到基本原理、底层实现、源码剖析中,更聚焦于技术本身,在不重复造轮子的基础上,学习和了解更多产品背后的代码设计和实现原理。最后在分享一句弗兰西斯·培根的话:

“人生如同道路。最近的捷径通常是最坏的路。”

参考


#kubernetes #operator

「一键投喂 小鱼干/小骨头」

小助手毛线

感谢投喂 毛线 & Sunny

使用微信扫描二维码完成支付


「欢迎扫描二维码关注公众号,了解更多云原生知识」



听说,阿里云给它的 OpenAPI 开发了一套编程语言 图解 K8S 源码 - Deployment Controller 篇