[zh] sync v1.24 kubeadm-24

2022-05-23 21:04:48 +08:00 · 2022-05-23 21:04:48 +08:00 · 4c52949427
parent c0cf43b1fe
commit 4c52949427
3 changed files with 229 additions and 91 deletions
--- a/content/zh/docs/reference/setup-tools/kubeadm/implementation-details.md
+++ b/content/zh/docs/reference/setup-tools/kubeadm/implementation-details.md
@ -166,13 +166,8 @@ Kubeadm 在启动 init 之前执行一组预检，目的是验证先决条件并
 - Kubernetes system requirements:
  - if running on linux:
    - [error] if Kernel is older than the minimum required version
-    - [error] if required cgroups subsystem aren't in set up
-  - if using docker:
-    - [warning/error] if Docker service does not exist, if it is disabled, if it is not active.
-    - [error] if Docker endpoint does not exist or does not work
-    - [warning] if docker version is not in the list of validated docker versions
-  - If using other cri engine:
-    - [error] if crictl socket does not answer
+    - [error] if required cgroups subsystem aren't set up
+- [error] if the CRI endpoint does not answer
 -->
 - [警告] 如果要使用的 Kubernetes 版本（由 `--kubernetes-version` 标志指定）比 kubeadm CLI
  版本至少高一个小版本。
@ -180,12 +175,7 @@ Kubeadm 在启动 init 之前执行一组预检，目的是验证先决条件并
  - 如果在 linux上运行：
    - [错误] 如果内核早于最低要求的版本
    - [错误] 如果未设置所需的 cgroups 子系统
-  - 如果使用 docker：
-    - [警告/错误] 如果 Docker 服务不存在、被禁用或未激活。
-    - [错误] 如果 Docker 端点不存在或不起作用
-    - [警告] 如果 docker 版本不在经过验证的 docker 版本列表中
-  - 如果使用其他 cri 引擎：
-    - [错误] 如果 crictl 套接字未应答
+- [错误] 如果 CRI 端点未应答
 <!--  
 - [error] if user is not root
 - [error] if the machine hostname is not a valid DNS subdomain
@ -745,18 +735,20 @@ As soon as the control plane is available, kubeadm executes following actions:
 一旦控制平面可用，kubeadm 将执行以下操作：

 <!-- 
- Labels the node as control-plane with `node-role.kubernetes.io/master=""`
- Taints the node with `node-role.kubernetes.io/master:NoSchedule`
+- Labels the node as control-plane with `node-role.kubernetes.io/control-plane=""`
+- Taints the node with `node-role.kubernetes.io/master:NoSchedule` and `node-role.kubernetes.io/control-plane:NoSchedule`
 -->
- 给节点打上 `node-role.kubernetes.io/master=""` 标签，标记其为控制平面
- 给节点打上 `node-role.kubernetes.io/master:NoSchedule` 污点
+- 给节点打上 `node-role.kubernetes.io/control-plane=""` 标签，标记其为控制平面
+- 给节点打上 `node-role.kubernetes.io/master:NoSchedule` 和 `node-role.kubernetes.io/control-plane:NoSchedule` 污点

 <!-- Please note that: -->
 请注意：

 <!-- 
+1. The `node-role.kubernetes.io/master` taint is deprecated and will be removed in kubeadm version 1.25
 1. Mark control-plane phase can be invoked individually with the [`kubeadm init phase mark-control-plane`](/docs/reference/setup-tools/kubeadm/kubeadm-init-phase/#cmd-phase-mark-control-plane) command
 -->
+1. `node-role.kubernetes.io/master` 污点是已废弃的，将会在 kubeadm 1.25 版本中移除
 1. 可以使用 [`kubeadm init phase mark-control-plane`](/zh/docs/reference/setup-tools/kubeadm/kubeadm-init-phase/#cmd-phase-mark-control-plane) 
  命令单独触发控制平面标记

@ -1026,14 +1018,12 @@ cluster startup problems.

 <!--  
 1. `kubeadm join` preflight checks are basically a subset `kubeadm init` preflight checks
-1. Starting from 1.9, kubeadm provides better support for CRI-generic functionality; in that case, docker specific controls
-   are skipped or replaced by similar controls for crictl.
+1. Starting from 1.24, kubeadm uses crictl to communicate to all known CRI endpoints.
 1. Starting from 1.9, kubeadm provides support for joining nodes running on Windows; in that case, linux specific controls are skipped.
 1. In any case the user can skip specific preflight checks (or eventually all preflight checks) with the `--ignore-preflight-errors` option.
 -->
 1. `kubeadm join` 预检基本上是 `kubeadm init` 预检的一个子集
-2. 从 1.9 开始，kubeadm 为 CRI 通用的功能提供了更好的支持；在这种情况下，
-   Docker 特定的控制参数将跳过或替换为 crictl 中与之相似的控制参数。
+2. 从 1.24 开始，kubeadm 使用 crictl 与所有已知的 CRI 端点进行通信。
 3. 从 1.9 开始，kubeadm 支持加入在 Windows 上运行的节点；在这种情况下，
   将跳过 Linux 特定的控制参数。
 4. 在任何情况下，用户都可以通过 `--ignore-preflight-errors` 选项跳过
--- a/content/zh/docs/reference/setup-tools/kubeadm/kubeadm-config.md
+++ b/content/zh/docs/reference/setup-tools/kubeadm/kubeadm-config.md
@ -8,53 +8,59 @@ weight: 50
 <!--
 During `kubeadm init`, kubeadm uploads the `ClusterConfiguration` object to your cluster
 in a ConfigMap called `kubeadm-config` in the `kube-system` namespace. This configuration is then read during
-`kubeadm join`, `kubeadm reset` and `kubeadm upgrade`. To view this ConfigMap call `kubeadm config view`.
+`kubeadm join`, `kubeadm reset` and `kubeadm upgrade`.
 -->
 在 `kubeadm init` 执行期间，kubeadm 将 `ClusterConfiguration` 对象上传
 到你的集群的 `kube-system` 名字空间下名为 `kubeadm-config` 的 ConfigMap 对象中。
 然后在 `kubeadm join`、`kubeadm reset` 和 `kubeadm upgrade` 执行期间读取此配置。 
-要查看此 ConfigMap，请调用 `kubeadm config view`。

 <!--
-You can use `kubeadm config print` to print the default configuration and `kubeadm config migrate` to
-convert your old configuration files to a newer version. `kubeadm config images list` and
-`kubeadm config images pull` can be used to list and pull the images that kubeadm requires.
+You can use `kubeadm config print` to print the default static configuration that kubeadm
+uses for `kubeadm init` and `kubeadm join`.
 -->
-你可以使用 `kubeadm config print` 命令打印默认配置，
-并使用 `kubeadm config migrate` 命令将旧版本的配置转化成新版本。
-`kubeadm config images list` 和 `kubeadm config images pull`
-命令可以用来列出并拉取 kubeadm 所需的镜像。
+你可以使用 `kubeadm config print` 命令打印默认静态配置，
+kubeadm 运行 `kubeadm init` and `kubeadm join` 时将使用此配置。
+
+<!-- 
+The output of the command is meant to serve as an example. You must manually edit the output
+of this command to adapt to your setup. Remove the fields that you are not certain about and kubeadm
+will try to default them on runtime by examining the host. 
+-->
+{{< note >}}
+此命令的输出旨在作为示例。你必须手动编辑此命令的输出来适配你的设置。
+删除你不确定的字段，kubeadm 将通过检查主机来尝试在运行时给它们设默认值。
+{{< /note >}}

 <!--
-For more information navigate to
+For more information on `init` and `join` navigate to
 [Using kubeadm init with a configuration file](/docs/reference/setup-tools/kubeadm/kubeadm-init/#config-file)
 or [Using kubeadm join with a configuration file](/docs/reference/setup-tools/kubeadm/kubeadm-join/#config-file).
 -->
-更多信息请浏览[使用带配置文件的 kubeadm init](/zh/docs/reference/setup-tools/kubeadm/kubeadm-init/#config-file)
-或[使用带配置文件的 kubeadm join](/zh/docs/reference/setup-tools/kubeadm/kubeadm-join/#config-file).
+更多有关 `init` 和 `join` 的信息请浏览[使用带配置文件的 kubeadm init](/zh/docs/reference/setup-tools/kubeadm/kubeadm-init/#config-file)
+或[使用带配置文件的 kubeadm join](/zh/docs/reference/setup-tools/kubeadm/kubeadm-join/#config-file)。

 <!--
-You can also configure several kubelet-configuration options with `kubeadm init`. These options will be the same on any node in your cluster. 
-See [Configuring each kubelet in your cluster using kubeadm](/docs/setup/production-environment/tools/kubeadm/kubelet-integration/) for details.
+For more information on using the kubeadm configuration API navigate to
+[Customizing components with the kubeadm API](/docs/setup/production-environment/tools/kubeadm/control-plane-flags).
 -->
-你也可以在使用 `kubeadm init` 命令时配置若干 kubelet 配置选项。
-这些选项对于集群中所有节点而言都是相同的。
-参阅[使用 kubeadm 来配置集群中的各个 kubelet](/zh/docs/setup/production-environment/tools/kubeadm/kubelet-integration/)
-了解详细信息。
- 
-<!--
-In Kubernetes v1.13.0 and later to list/pull kube-dns images instead of the CoreDNS image
-the `--config` method described [here](/docs/reference/setup-tools/kubeadm/kubeadm-init-phase/#cmd-phase-addon)
-has to be used.
+有关使用 kubeadm 的配置 API 的更多信息，
+请浏览[使用 kubeadm API 来自定义组件](/zh/docs/setup/production-environment/tools/kubeadm/control-plane-flags)。
+
+<!-- 
+You can use `kubeadm config migrate` to convert your old configuration files that contain a deprecated
+API version to a newer, supported API version. 
 -->
-在 Kubernetes v1.13.0 及更高版本中，要列出/拉取 kube-dns 镜像而不是 CoreDNS 镜像，
-必须使用[这里](/zh/docs/reference/setup-tools/kubeadm/kubeadm-init-phase/#cmd-phase-addon)
-所描述的 `--config` 方法。
+你可以使用 `kubeadm config migrate` 来转换旧配置文件，
+把其中已弃用的 API 版本更新为受支持的 API 版本。
+
+<!-- 
+`kubeadm config images list` and `kubeadm config images pull` can be used to list and pull the images
+that kubeadm requires. 
+-->
+`kubeadm config images list` 和 `kubeadm config images pull` 可以用来列出和拉取 kubeadm 所需的镜像。

 <!-- body -->
-## kubeadm config upload from-file {#cmd-config-from-file}
-
-## kubeadm config print{#cmd-config-view}
+## kubeadm config print {#cmd-config-print}
 {{< include "generated/kubeadm_config_print.md" >}}

 ## kubeadm config print init-defaults {#cmd-config-print-init-defaults}
--- a/content/zh/docs/reference/setup-tools/kubeadm/kubeadm-init.md
+++ b/content/zh/docs/reference/setup-tools/kubeadm/kubeadm-init.md
@ -207,6 +207,13 @@ What this example would do is write the manifest files for the control plane and
 这允许你修改文件，然后使用 `--skip-phases` 跳过这些阶段。
 通过调用最后一个命令，你将使用自定义清单文件创建一个控制平面节点。

+{{< feature-state for_k8s_version="v1.22" state="beta" >}}
+
+<!-- 
+Alternatively, you can use the `skipPhases` field under `InitConfiguration`. 
+-->
+或者，你可以使用 `InitConfiguration` 下的 `skipPhases` 字段。
+
 <!--
 ### Using kubeadm init with a configuration file {#config-file}
 -->
@ -239,8 +246,8 @@ The default configuration can be printed out using the
 If your configuration is not using the latest version it is **recommended** that you migrate using
 the [kubeadm config migrate](/docs/reference/setup-tools/kubeadm/kubeadm-config/) command.

-For more information on the fields and usage of the configuration you can navigate to our API reference
-page and pick a version from [the list](https://pkg.go.dev/k8s.io/kubernetes/cmd/kubeadm/app/apis/kubeadm#section-directories).
+For more information on the fields and usage of the configuration you can navigate to our
+[API reference page](/docs/reference/config-api/kubeadm-config.v1beta3/).
 -->
 可以使用 [kubeadm config print](/zh/docs/reference/setup-tools/kubeadm/kubeadm-config/)
 命令打印出默认配置。
@ -249,10 +256,119 @@ page and pick a version from [the list](https://pkg.go.dev/k8s.io/kubernetes/cmd
 **推荐**使用 [kubeadm config migrate](/zh/docs/reference/setup-tools/kubeadm/kubeadm-config/)
 命令进行迁移。

-有关配置的字段和用法的更多信息，
-你可以访问 API 参考页面并从
-[列表](https://pkg.go.dev/k8s.io/kubernetes/cmd/kubeadm/app/apis/kubeadm#section-directories)
-中选择一个版本。
+关于配置的字段和用法的更多信息，你可以访问 [API 参考页面](/zh/docs/reference/config-api/kubeadm-config.v1beta3/)。
+
+<!-- 
+### Using kubeadm init with feature gates {#feature-gates} 
+-->
+### 使用 kubeadm init 时设置特性门控 {#feature-gates} 
+
+<!-- 
+Kubeadm supports a set of feature gates that are unique to kubeadm and can only be applied
+during cluster creation with `kubeadm init`. These features can control the behavior
+of the cluster. Feature gates are removed after a feature graduates to GA. 
+-->
+Kubeadm 支持一组独有的特性门控，只能在 `kubeadm init` 创建集群期间使用。
+这些特性可以控制集群的行为。特性门控会在毕业到 GA 后被移除。
+
+<!-- 
+To pass a feature gate you can either use the `--feature-gates` flag for
+`kubeadm init`, or you can add items into the `featureGates` field when you pass
+a [configuration file](/docs/reference/config-api/kubeadm-config.v1beta3/#kubeadm-k8s-io-v1beta3-ClusterConfiguration)
+using `--config`. 
+-->
+你可以使用 `--feature-gates` 标志来为 `kubeadm init` 设置特性门控，
+或者你可以在用 `--config` 传递[配置文件](/zh/docs/reference/config-api/kubeadm-config.v1beta3/#kubeadm-k8s-io-v1beta3-ClusterConfiguration)
+时添加条目到 `featureGates` 字段中去。
+
+<!-- 
+Passing [feature gates for core Kubernetes components](/docs/reference/command-line-tools-reference/feature-gates)
+directly to kubeadm is not supported. Instead, it is possible to pass them by
+[Customizing components with the kubeadm API](/docs/setup/production-environment/tools/kubeadm/control-plane-flags/). 
+-->
+直接传递 [Kubernetes 核心组件的特性门控](/zh/docs/reference/command-line-tools-reference/feature-gates)给 kubeadm 是不支持的。
+相反，可以通过[使用 kubeadm API 的自定义组件](/zh/docs/setup/production-environment/tools/kubeadm/control-plane-flags/)来传递。
+
+<!-- 
+List of feature gates: 
+-->
+特性门控的列表：
+
+{{< table caption="kubeadm feature gates" >}}
+特性 | 默认值 | Alpha | Beta
+:-------|:--------|:------|:-----
+`PublicKeysECDSA` | `false` | 1.19 | -
+`RootlessControlPlane` | `false` | 1.22 | -
+`UnversionedKubeletConfigMap` | `true` | 1.22 | 1.23
+{{< /table >}}
+
+<!-- 
+Once a feature gate goes GA it is removed from this list as its value becomes locked to `true` by default. 
+-->
+{{< note >}}
+一旦特性门控变成了 GA，那它将会从这个列表中移除，因为它的值会被默认锁定为 `true` 。
+{{< /note >}}
+
+<!-- 
+Feature gate descriptions: 
+-->
+特性门控的描述：
+
+<!-- 
+`PublicKeysECDSA`
+: Can be used to create a cluster that uses ECDSA certificates instead of the default RSA algorithm.
+Renewal of existing ECDSA certificates is also supported using `kubeadm certs renew`, but you cannot
+switch between the RSA and ECDSA algorithms on the fly or during upgrades. 
+-->
+`PublicKeysECDSA`
+: 可用于创建集群时使用 ECDSA 证书而不是默认 RSA 算法。
+支持用 `kubeadm certs renew` 更新现有 ECDSA 证书，
+但你不能在集群运行期间或升级期间切换 RSA 和 ECDSA 算法。
+
+<!-- 
+`RootlessControlPlane`
+: Setting this flag configures the kubeadm deployed control plane component static Pod containers
+for `kube-apiserver`, `kube-controller-manager`, `kube-scheduler` and `etcd` to run as non-root users.
+If the flag is not set, those components run as root. You can change the value of this feature gate before
+you upgrade to a newer version of Kubernetes. 
+-->
+`RootlessControlPlane`
+: 设置此标志来配置 kubeadm 所部署的控制平面组件中的静态 Pod 容器 
+`kube-apiserver`、`kube-controller-manager`、`kube-scheduler` 和 `etcd` 以非 root 用户身份运行。
+如果未设置该标志，则这些组件以 root 身份运行。
+你可以在升级到更新版本的 Kubernetes 之前更改此特性门控的值。
+
+<!-- 
+`UnversionedKubeletConfigMap`
+: This flag controls the name of the {{< glossary_tooltip text="ConfigMap" term_id="configmap" >}} where kubeadm stores
+kubelet configuration data. With this flag not specified or set to `true`, the ConfigMap is named `kubelet-config`.
+If you set this flag to `false`, the name of the ConfigMap includes the major and minor version for Kubernetes
+(for example: `kubelet-config-{{< skew currentVersion >}}`). Kubeadm ensures that RBAC rules for reading and writing
+that ConfigMap are appropriate for the value you set. When kubeadm writes this ConfigMap (during `kubeadm init`
+or `kubeadm upgrade apply`), kubeadm respects the value of `UnversionedKubeletConfigMap`. When reading that ConfigMap
+(during `kubeadm join`, `kubeadm reset`, `kubeadm upgrade ...`), kubeadm attempts to use unversioned ConfigMap name first;
+if that does not succeed, kubeadm falls back to using the legacy (versioned) name for that ConfigMap. 
+-->
+`UnversionedKubeletConfigMap`
+: 此标志控制 kubeadm 存储 kubelet 配置数据的 {{<glossary_tooltip text="ConfigMap" term_id="configmap" >}} 的名称。
+在未指定此标志或设置为 `true` 的情况下，此 ConfigMap 被命名为 `kubelet-config`。
+如果将此标志设置为 `false`，则此 ConfigMap 的名称会包括 Kubernetes 的主要版本和次要版本（例如：`kubelet-config-{{< skew currentVersion >}}`）。 
+Kubeadm 会确保用于读写 ConfigMap 的 RBAC 规则适合你设置的值。
+当 kubeadm 写入此 ConfigMap 时（在 `kubeadm init` 或 `kubeadm upgrade apply` 期间），
+kubeadm 根据 `UnversionedKubeletConfigMap` 的设置值来执行操作。
+当读取此 ConfigMap 时（在 `kubeadm join`、`kubeadm reset`、`kubeadm upgrade ...` 期间），
+kubeadm 尝试首先使用无版本（后缀）的 ConfigMap 名称；
+如果不成功，kubeadm 将回退到使用该 ConfigMap 的旧（带版本号的）名称。
+
+<!-- 
+Setting `UnversionedKubeletConfigMap` to `false` is supported but **deprecated**. 
+-->
+{{< note >}}
+设置 `UnversionedKubeletConfigMap` 为 `false` 是被支持的特性，但该特性**已被弃用**。
+{{< /note >}}
+
+
+

 <!--
 ### Adding kube-proxy parameters {#kube-proxy}
@ -285,6 +401,37 @@ For information about passing flags to control plane components see:
 有关向控制平面组件传递命令行参数的说明请查看：
 [控制平面命令行参数](/zh/docs/setup/production-environment/tools/kubeadm/control-plane-flags/)

+<!--
+### Running kubeadm without an Internet connection {#without-internet-connection}
+-->
+### 在没有互联网连接的情况下运行 kubeadm {#without-internet-connection}
+
+<!--
+For running kubeadm without an internet connection you have to pre-pull the required control-plane images.
+-->
+要在没有互联网连接的情况下运行 kubeadm，你必须提前拉取所需的控制平面镜像。
+
+<!--
+You can list and pull the images using the `kubeadm config images` sub-command:
+-->
+你可以使用 `kubeadm config images` 子命令列出并拉取镜像：
+
+```shell
+kubeadm config images list
+kubeadm config images pull
+```
+
+<!-- 
+You can pass `--config` to the above commands with a [kubeadm configuration file](#config-file)
+to control the `kubernetesVersion` and `imageRepository` fields. 
+-->
+你可以通过 `--config` 把 [kubeadm 配置文件](#config-file) 传递给上述命令来控制 `kubernetesVersion` 和 `imageRepository` 字段。
+
+<!-- 
+All default `k8s.gcr.io` images that kubeadm requires support multiple architectures. 
+-->
+kubeadm 需要的所有默认 `k8s.gcr.io` 镜像都支持多种硬件体系结构。
+
 <!--
 ### Using custom images {#custom-images}
 -->
@ -306,23 +453,44 @@ You can override this behavior by using [kubeadm with a configuration file](#con
 <!--
 Allowed customization are:

+* To provide `kubernetesVersion` which affects the version of the images.
 * To provide an alternative `imageRepository` to be used instead of
  `k8s.gcr.io`.
-* To set `useHyperKubeImage` to `true` to use the HyperKube image.
-* To provide a specific `imageRepository` and `imageTag` for etcd or DNS add-on.
+* To provide a specific `imageRepository` and `imageTag` for etcd or CoreDNS.
 -->
 允许的自定义功能有：

+* 提供影响镜像版本的 `kubernetesVersion`。 
 * 使用其他的 `imageRepository` 来代替 `k8s.gcr.io`。
-* 将 `useHyperKubeImage` 设置为 `true`，使用 HyperKube 镜像。
-* 为 etcd 或 DNS 附件提供特定的 `imageRepository` 和 `imageTag`。
+* 为 etcd 或 CoreDNS 提供特定的 `imageRepository` 和 `imageTag`。

-<!--
-Please note that the configuration field `kubernetesVersion` or the command line flag
-`-kubernetes-version` affect the version of the images.
+<!-- 
+`imageRepository` may differ for backwards compatibility reasons. For example,
+one image might have a subpath at `k8s.gcr.io/subpath/image`, but be defaulted
+to `my.customrepository.io/image` when using a custom repository. 
 -->
-请注意配置文件中的配置项 `kubernetesVersion` 或者命令行参数 `--kubernetes-version`
-会影响到镜像的版本。
+由于向后兼容的原因，`imageRepository` 可能会有所不同。
+例如，某镜像的子路径可能是 `k8s.gcr.io/subpath/image`，
+但使用自定义仓库时默认为 `my.customrepository.io/image`。
+
+<!-- 
+To ensure you push the images to your custom repository in paths that kubeadm
+can consume, you must:
+-->
+确保将镜像推送到 kubeadm 可以使用的自定义仓库的路径中，你必须：
+
+<!-- 
+* Pull images from the defaults paths at `k8s.gcr.io` using `kubeadm config images {list|pull}`.
+* Push images to the paths from `kubeadm config images list --config=config.yaml`,
+where `config.yaml` contains the custom `imageRepository`, and/or `imageTag`
+for etcd and CoreDNS.
+* Pass the same `config.yaml` to `kubeadm init`. 
+-->
+* 使用 `kubeadm config images {list|pull}` 从 `k8s.gcr.io` 的默认路径中拉取镜像。
+* 将镜像推送到 `kubeadm config images list --config=config.yaml` 的路径，
+其中 `config.yaml` 包含自定义的 `imageRepository` 和/或用于 etcd 和 CoreDNS 的 `imageTag`。 
+* 将相同的 `config.yaml` 传递给 `kubeadm init`。
+

 <!--
 ### Uploading control-plane certificates to the cluster
@ -425,32 +593,6 @@ value to the kubelet.
 [`--hostname-override`](/zh/docs/reference/command-line-tools-reference/kubelet/#options)
 值传递给 kubelet。

-<!--
-### Running kubeadm without an internet connection
-->
-### 在没有互联网连接的情况下运行 kubeadm
-
-<!--
-For running kubeadm without an internet connection you have to pre-pull the required control-plane images.
-->
-要在没有互联网连接的情况下运行 kubeadm，你必须提前拉取所需的控制平面镜像。
-
-<!--
-You can list and pull the images using the `kubeadm config images` sub-command:
-->
-你可以使用 `kubeadm config images` 子命令列出并拉取镜像：
-
-```shell
-kubeadm config images list
-kubeadm config images pull
-```
-
-<!--
-All images that kubeadm requires such as `k8s.gcr.io/kube-*`, `k8s.gcr.io/etcd` and `k8s.gcr.io/pause` support multiple architectures.
-->
-kubeadm 需要的所有镜像，例如 `k8s.gcr.io/kube-*`、`k8s.gcr.io/etcd` 和 `k8s.gcr.io/pause`
-都支持多种架构。
-
 <!--
 ### Automating kubeadm
 -->