1.3. 订阅报告

软件包级别的受管集群状态位于受管集群上的 <namespace:<your-appsub-namespace> 中，其中包含应用程序部署的所有资源的详细状态。对于部署到受管集群的每个 appsub，在受管集群的 appsub 命名空间中创建了 SubscriptionStatus CR。如果存在错误，则会报告每个资源，并带有详细错误。

请参阅以下 SubscriptionStatus 示例 YAML 文件：

apiVersion: apps.open-cluster-management.io/v1alpha1
kind: SubscriptionStatus
metadata:
  labels:
    apps.open-cluster-management.io/cluster: <your-managed-cluster>
    apps.open-cluster-management.io/hosting-subscription: <your-appsub-namespace>.<your-appsub-name>
  name: <your-appsub-name>
  namespace: <your-appsub-namespace>
statuses:
  packages:
  - apiVersion: v1
    kind: Service
    lastUpdateTime: "2021-09-13T20:12:34Z"
    Message: <detailed error. visible only if the package fails>
    name: frontend
    namespace: test-ns-2
    phase: Deployed
  - apiVersion: apps/v1
    kind: Deployment
    lastUpdateTime: "2021-09-13T20:12:34Z"
    name: frontend
    namespace: test-ns-2
    phase: Deployed
  - apiVersion: v1
    kind: Service
    lastUpdateTime: "2021-09-13T20:12:34Z"
    name: redis-master
    namespace: test-ns-2
    phase: Deployed
  - apiVersion: apps/v1
    kind: Deployment
    lastUpdateTime: "2021-09-13T20:12:34Z"
    name: redis-master
    namespace: test-ns-2
    phase: Deployed
  - apiVersion: v1
    kind: Service
    lastUpdateTime: "2021-09-13T20:12:34Z"
    name: redis-slave
    namespace: test-ns-2
    phase: Deployed
  - apiVersion: apps/v1
    kind: Deployment
    lastUpdateTime: "2021-09-13T20:12:34Z"
    name: redis-slave
    namespace: test-ns-2
    phase: Deployed

集群级状态位于 hub 集群中的 <namespace:<your-managed-cluster-1>，只包含该受管集群上每个应用程序的整体状态。hub 集群上的每个集群命名空间中的 subscriptionReport 报告以下状态之一：

请参阅以下 SubscriptionStatus 示例 YAML 文件：

apiVersion: apps.open-cluster-management.io/v1alpha1
kind: subscriptionReport
metadata:
  labels:
    apps.open-cluster-management.io/cluster: "true"
  name: <your-managed-cluster-1>
  namespace: <your-managed-cluster-1>
reportType: Cluster
results:
- result: deployed
  source: appsub-1-ns/appsub-1                     // appsub 1 to <your-managed-cluster-1>
  timestamp:
    nanos: 0
    seconds: 1634137362
- result: failed
  source: appsub-2-ns/appsub-2                     // appsub 2 to <your-managed-cluster-1>
  timestamp:
    nanos: 0
    seconds: 1634137362
- result: propagationFailed
  source: appsub-3-ns/appsub-3                     // appsub 3 to <your-managed-cluster-1>
  timestamp:
    nanos: 0
    seconds: 1634137362

每个应用程序有一个应用程序级 subscriptionReport，位于 hub 集群的 appsub 命名空间中的 <namespace:<your-appsub-namespace>，并包含以下信息：

备注： inProcess 状态是总数减 deployed, 减 failed `, 并减 `propagationFailed。

请参阅以下 SubscriptionStatus 示例 YAML 文件：

apiVersion: apps.open-cluster-management.io/v1alpha1
kind: subscriptionReport
metadata:
  labels:
    apps.open-cluster-management.io/hosting-subscription: <your-appsub-namespace>.<your-appsub-name>
  name: <your-appsub-name>
  namespace: <your-appsub-namespace>
reportType: Application
resources:
- apiVersion: v1
  kind: Service
  name: redis-master2
  namespace: playback-ns-2
- apiVersion: apps/v1
  kind: Deployment
  name: redis-master2
  namespace: playback-ns-2
- apiVersion: v1
  kind: Service
  name: redis-slave2
  namespace: playback-ns-2
- apiVersion: apps/v1
  kind: Deployment
  name: redis-slave2
  namespace: playback-ns-2
- apiVersion: v1
  kind: Service
  name: frontend2
  namespace: playback-ns-2
- apiVersion: apps/v1
  kind: Deployment
  name: frontend2
  namespace: playback-ns-2
results:
- result: deployed
  source: cluster-1                            //cluster 1 status
  timestamp:
    nanos: 0
    seconds: 0
- result: failed
  source: cluster-3                            //cluster 2 status
  timestamp:
    nanos: 0
    seconds: 0
- result: propagationFailed
  source: cluster-4                            //cluster 3 status
  timestamp:
    nanos: 0
    seconds: 0
summary:
  deployed: 8
  failed: 1
  inProgress: 0
  propagationFailed: 1
  clusters: 10

在第一个失败的集群上报告 ManagedClusterView CR。如果应用程序在带有资源部署失败的多个集群中部署，则只会为 hub 集群上第一个失败的集群命名空间创建一个 managedClusterView CR。managedClusterView CR 从失败集群中检索详细的订阅状态，因此应用程序所有者不需要访问失败的远程集群。

运行以下命令，您可以运行以下命令获取状态：

% oc get managedclusterview -n <failing-clusternamespace> "<app-name>-<app name>"

如果无法访问受管集群来获取订阅状态，您可以使用 CLI。集群级别或应用程序级别订阅报告提供了整个状态，而不是应用程序的详细错误消息。

% getAppSubStatus.sh -c <your-managed-cluster> -s <your-appsub-namespace> -n <your-appsub-name>

当给定受管集群上无法登录到每个受管集群以检索这些信息时，您还可以获得给定受管集群上 AppSub 的 Last Update Time。因此，创建了一个实用程序脚本，以简化受管集群中 AppSub 的 Last Update Time 的检索。这个脚本设计为在 Hub 集群上运行。它会创建一个 managedClusterView 资源，以从受管集群获取 AppSub，并解析数据以获取 Last Update Time。

% getLastUpdateTime.sh -c <your-managed-cluster> -s <your-appsub-namespace> -n <your-appsub-name>