Virtual Kubelet的工作原理

从Kubernetes API服务器的角度来看，Virtual Kubelet看起来像普通的kubelet，但其关键区别在于它们在其他地方调度容器，例如在云无服务器API中，而不是在节点上。对内向k8s apiserver注册node，保持跟kubelet一样的协议通信，对外提供简化的pleg接口供第三方provider实现。常见的供应商有：

• Alibaba Cloud Elastic Container Instance (ECI)
• AWS Fargate
• Azure Batch
• Azure Container Instances (ACI)
• Kubernetes Container Runtime Interface (CRI)
• Huawei Cloud Container Instance (CCI)
• Hyper.sh
• HashiCorp Nomad
• Service Fabric Mesh
• vSphere Integrated Containers (VIC)

整体架构图：

// PodLifecycleHandler defines the interface used by the PodController to react
// to new and changed pods scheduled to the node that is being managed.
//
// Errors produced by these methods should implement an interface from
// github.com/virtual-kubelet/virtual-kubelet/errdefs package in order for the
// core logic to be able to understand the type of failure.
type PodLifecycleHandler interface {
    // CreatePod takes a Kubernetes Pod and deploys it within the provider.
    CreatePod(ctx context.Context, pod *corev1.Pod) error

    // UpdatePod takes a Kubernetes Pod and updates it within the provider.
    UpdatePod(ctx context.Context, pod *corev1.Pod) error

    // DeletePod takes a Kubernetes Pod and deletes it from the provider.
    DeletePod(ctx context.Context, pod *corev1.Pod) error

    // GetPod retrieves a pod by name from the provider (can be cached).
    // The Pod returned is expected to be immutable, and may be accessed
    // concurrently outside of the calling goroutine. Therefore it is recommended
    // to return a version after DeepCopy.
    GetPod(ctx context.Context, namespace, name string) (*corev1.Pod, error)

    // GetPodStatus retrieves the status of a pod by name from the provider.
    // The PodStatus returned is expected to be immutable, and may be accessed
    // concurrently outside of the calling goroutine. Therefore it is recommended
    // to return a version after DeepCopy.
    GetPodStatus(ctx context.Context, namespace, name string) (*corev1.PodStatus, error)

    // GetPods retrieves a list of all pods running on the provider (can be cached).
    // The Pods returned are expected to be immutable, and may be accessed
    // concurrently outside of the calling goroutine. Therefore it is recommended
    // to return a version after DeepCopy.
    GetPods(context.Context) ([]*corev1.Pod, error)
}


通过这几个pleg接口设计可以看出：

1）对于增删改pod而言，是从控制面发起，由具体provider实现的逻辑。至于provider如何实现，也就对应于底层pod的具体不同实现。

2）而对于查pod而言，却是入参pod的相关信息，输出的是控制面pod的协议。查看pod的信息是通过k8s控制面输出的，k8s apiserver
会从etcd中获取pod的元数据，或者client端通过informer cached获取数据。翻看源码发现，provider提供了接口，但接口实现时机留
给业务方具体问题具体分析。常用的场景为provider端的pod的状态发生变化然后通过查询的接口同步变化到k8s控制面，至于控制面pod
变化同步到provider端，这个在增删改pod的时候会实现。

综上所述，简化的pleg接口实现了pod生命周期的维护，按需实现，能满足业务的基本诉求。而Virtual Kubelet的工作原理则是对内向apiserver
维持Kubelet的协议实现伪装，对外设计pleg以及exec、logs等接口供不同业务实现自定义的provider。