kubernetes
/
website
mirror of https://github.com/kubernetes/website.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity

Merge pull request #32671 from zaunist/docs/concepts 

[zh] Resync device-plugins
pull/32698/head
Kubernetes Prow Robot 2022-03-31 17:40:38 -07:00 committed by GitHub
parent b53955eed4 fd9b3076be
commit e94ce37cc2
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 82 additions and 11 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

93
content/zh/docs/concepts/extend-kubernetes/compute-storage-net/device-plugins.md
View File

@ -40,7 +40,7 @@ The kubelet exports a `Registration` gRPC service:
```gRPC
service Registration {
rpc Register(RegisterRequest) returns (Empty) {}
rpc Register(RegisterRequest) returns (Empty) {}
}
```
@ -63,7 +63,7 @@ and reports two healthy devices on a node, the node status is updated
to advertise that the node has 2 "Foo" devices installed and available.
-->
设备插件可以通过此 gRPC 服务在 kubelet 进行注册。在注册期间，设备插件需要发送下面几样内容：
* 设备插件的 Unix 套接字。
* 设备插件的 API 版本。
* `ResourceName` 是需要公布的。这里 `ResourceName` 需要遵循
@ -92,12 +92,14 @@ specification as they request other types of resources, with the following limit
* 扩展资源仅可作为整数资源使用，并且不能被过量使用
* 设备不能在容器之间共享
### 示例 {#example-pod}
<!--
Suppose a Kubernetes cluster is running a device plugin that advertises resource `hardware-vendor.example/foo`
on certain nodes. Here is an example of a pod requesting this resource to run a demo workload:
-->
假设 Kubernetes 集群正在运行一个设备插件，该插件在一些节点上公布的资源为 `hardware-vendor.example/foo`。
下面就是一个 Pod 示例，请求此资源以运行某演示负载：
下面就是一个 Pod 示例，请求此资源以运行一个工作负载的示例：
```yaml
---
@ -140,8 +142,12 @@ The general workflow of a device plugin includes the following steps:
一个 gRPC 服务，该服务实现以下接口：
<!--
```gRPC
service DevicePlugin {
// GetDevicePluginOptions returns options to be communicated with Device Manager.
rpc GetDevicePluginOptions(Empty) returns (DevicePluginOptions) {}
// ListAndWatch returns a stream of List of Devices
// Whenever a Device state change or a Device disappears, ListAndWatch
// returns the new list
@ -168,6 +174,9 @@ The general workflow of a device plugin includes the following steps:
-->
```gRPC
service DevicePlugin {
// GetDevicePluginOptions 返回与设备管理器沟通的选项。
rpc GetDevicePluginOptions(Empty) returns (DevicePluginOptions) {}
// ListAndWatch 返回 Device 列表构成的数据流。
// 当 Device 状态发生变化或者 Device 消失时，ListAndWatch
// 会返回新的列表。
@ -331,10 +340,12 @@ service PodResourcesLister {
}
```
### `List` gRPC 端点 {#grpc-endpoint-list}
<!--
The `List` endpoint provides information on resources of running pods, with details such as the
id of exclusively allocated CPUs, device id as it was reported by device plugins and id of
the NUMA node where these devices are allocated. Also, for NUMA-based machines, it contains
the NUMA node where these devices are allocated. Also, for NUMA-based machines, it contains
the information about memory and hugepages reserved for a container.
-->
这一 `List` 端点提供运行中 Pods 的资源信息，包括类似独占式分配的
@ -387,6 +398,51 @@ message ContainerDevices {
}
```
<!--
{{< note >}}
cpu_ids in the `ContainerResources` in the `List` endpoint correspond to exclusive CPUs allocated
to a partilar container. If the goal is to evaluate CPUs that belong to the shared pool, the `List`
endpoint needs to be used in conjunction with the `GetAllocatableResources` endpoint as explained
below:
1. Call `GetAllocatableResources` to get a list of all the allocatable CPUs
2. Call `GetCpuIds` on all `ContainerResources` in the system
3. Subtract out all of the CPUs from the `GetCpuIds` calls from the `GetAllocatableResources` call
{{< /note >}}
-->
{{< note >}}
`List` 端点中的 `ContainerResources` 中的 cpu_ids 对应于分配给某个容器的专属 CPU。
如果要统计共享池中的 CPU，`List` 端点需要与 `GetAllocatableResources` 端点一起使用，如下所述:
1. 调用 `GetAllocatableResources` 获取所有可用的 CPUs。
2. 在系统中所有的 `ContainerResources` 上调用 `GetCpuIds`。
3. 用 `GetAllocatableResources` 获取的 CPU 数减去 `GetCpuIds` 获取的 CPU 数。
{{< /note >}}
### `GetAllocatableResources` gRPC 端点 {#grpc-endpoint-getallocatableresources}
{{< feature-state state="beta" for_k8s_version="v1.23" >}}
<!--
{{< note >}}
`GetAllocatableResources` should only be used to evaluate [allocatable](/docs/tasks/administer-cluster/reserve-compute-resources/#node-allocatable)
resources on a node. If the goal is to evaluate free/unallocated resources it should be used in
conjunction with the List() endpoint. The result obtained by `GetAllocatableResources` would remain
the same unless the underlying resources exposed to kubelet change. This happens rarely but when
it does (for example: hotplug/hotunplug, device health changes), client is expected to call
`GetAlloctableResources` endpoint.
However, calling `GetAllocatableResources` endpoint is not sufficient in case of cpu and/or memory
update and Kubelet needs to be restarted to reflect the correct resource capacity and allocatable.
{{< /note >}}
-->
{{< note >}}
`GetAllocatableResources` 应该仅被用于评估一个节点上的[可分配的](/zh/docs/tasks/administer-cluster/reserve-compute-resources/#node-allocatable)
资源。如果目标是评估空闲/未分配的资源，此调用应该与 List() 端点一起使用。
除非暴露给 kubelet 的底层资源发生变化 否则 `GetAllocatableResources` 得到的结果将保持不变。
这种情况很少发生，但当发生时（例如：热插拔，设备健康状况改变），客户端应该调用 `GetAlloctableResources` 端点。
然而，调用 `GetAllocatableResources` 端点在 cpu、内存被更新的情况下是不够的，
Kubelet 需要重新启动以获取正确的资源容量和可分配的资源。
{{< /note >}}
<!--
GetAllocatableResources provides information on resources initially available on the worker node.
It provides more information than kubelet exports to APIServer.
@ -394,7 +450,6 @@ It provides more information than kubelet exports to APIServer.
端点 `GetAllocatableResources` 提供最初在工作节点上可用的资源的信息。
此端点所提供的信息比导出给 API 服务器的信息更丰富。
```gRPC
// AllocatableResourcesResponses 包含 kubelet 所了解到的所有设备的信息
message AllocatableResourcesResponse {
@ -405,6 +460,23 @@ message AllocatableResourcesResponse {
```
<!--
Starting from Kubernetes v1.23, the `GetAllocatableResources` is enabled by default.
You can disable it by turning off the
`KubeletPodResourcesGetAllocatable` [feature gate](/docs/reference/command-line-tools-reference/feature-gates/).
Preceding Kubernetes v1.23, to enable this feature `kubelet` must be started with the following flag:
`--feature-gates=KubeletPodResourcesGetAllocatable=true`
-->
从 Kubernetes v1.23 开始，`GetAllocatableResources` 被默认启用。
你可以通过关闭 `KubeletPodResourcesGetAllocatable`
[特性门控](/zh/docs/reference/command-line-tools-reference/feature-gates/) 来禁用。
在 Kubernetes v1.23 之前，要启用这一功能，`kubelet` 必须用以下标志启动：
`--feature-gates=KubeletPodResourcesGetAllocatable=true`
<!--
`ContainerDevices` do expose the topology information declaring to which NUMA cells the device is affine.
The NUMA cells are identified using a opaque integer ID, which value is consistent to what device
@ -457,7 +529,7 @@ The Topology Manager is a Kubelet component that allows resources to be co-ordin
```gRPC
message TopologyInfo {
repeated NUMANode nodes = 1;
repeated NUMANode nodes = 1;
}
message NUMANode {
@ -507,14 +579,15 @@ Here are some examples of device plugin implementations:
## 设备插件示例 {#examples}
下面是一些设备插件实现的示例：
* [AMD GPU 设备插件](https://github.com/RadeonOpenCompute/k8s-device-plugin)
* [Intel 设备插件](https://github.com/intel/intel-device-plugins-for-kubernetes) 支持 Intel GPU、FPGA 和 QuickAssist 设备
* [KubeVirt 设备插件](https://github.com/kubevirt/kubernetes-device-plugins) 用于硬件辅助的虚拟化
* The [NVIDIA GPU 设备插件](https://github.com/NVIDIA/k8s-device-plugin)
* 需要 [nvidia-docker](https://github.com/NVIDIA/nvidia-docker) 2.0，以允许运行 Docker 容器的时候启用 GPU。
* 需要 [nvidia-docker](https://github.com/NVIDIA/nvidia-docker) 2.0，以允许运行 Docker 容器的时候启用 GPU。
* [为 Container-Optimized OS 所提供的 NVIDIA GPU 设备插件](https://github.com/GoogleCloudPlatform/container-engine-accelerators/tree/master/cmd/nvidia_gpu)
* [RDMA 设备插件](https://github.com/hustcat/k8s-rdma-device-plugin)
* [SocketCAN 设备插件](https://github.com/collabora/k8s-socketcan)
* [Solarflare 设备插件](https://github.com/vikaschoudhary16/sfc-device-plugin)
* [SR-IOV 网络设备插件](https://github.com/intel/sriov-network-device-plugin)
* [Xilinx FPGA 设备插件](https://github.com/Xilinx/FPGA_as_a_Service/tree/master/k8s-fpga-device-plugin)
@ -529,7 +602,5 @@ Here are some examples of device plugin implementations:
-->
* 查看[调度 GPU 资源](/zh/docs/tasks/manage-gpus/scheduling-gpus/) 来学习使用设备插件
* 查看在上如何[公布节点上的扩展资源](/zh/docs/tasks/administer-cluster/extended-resource-node/)
* 阅读如何在 Kubernetes 中使用 [TLS Ingress 的硬件加速](https://kubernetes.io/blog/2019/04/24/hardware-accelerated-ssl/tls-termination-in-ingress-controllers-using-kubernetes-device-plugins-and-runtimeclass/)
* 阅读如何在 Kubernetes 中使用 [TLS Ingress 的硬件加速](https://kubernetes.io/blog/2019/04/24/hardware-accelerated-ssl/tls-termination-in-ingress-controllers-using-kubernetes-device-plugins-and-runtimeclass/)
* 学习[拓扑管理器](/zh/docs/tasks/administer-cluster/topology-manager/)