检查升级就绪情况

您可以使用 sfupgradecheck 命令确认集群是否已做好升级准备。此命令可验证待定节点、磁盘空间和集群故障等信息。如果您的管理节点位于非公开站点,则升级就绪检查需要元数据 JSON 文件才能成功运行。

关于本任务

此过程介绍如何通过以下任一结果判断升级检查是否成功:

步骤

  1. 运行 sfupgradecheck 命令。
    sfupgradecheck -u <cluster-user-name> -p <cluster-password> MVIP
    请参见以下输入命令示例:
    sfupgradecheck -u admin -p admin 10.117.78.244
    示例输出如下。在此结果中,未显示任何错误,您可以进行升级。
    check_pending_nodes:
    Test Description:     Verify no pending nodes in cluster
    More information:     https://kb.netapp.com/support/s/article/ka11A0000008ltOQAQ/pending-nodes
    check_cluster_faults:
    Test Description:     Report any cluster faults
    check_root_disk_space:
    Test Description:     Verify node root directory has at least 12 GBs of available disk space
    Passed node IDs:      1, 2, 3
    More information:     https://kb.netapp.com/support/s/article/ka11A0000008ltTQAQ/SolidFire-Disk-space-error
    check_mnode_connectivity:
    Test Description:     Verify storage nodes can communicate with management node
    Passed node IDs:      1, 2, 3
    More information:     https://kb.netapp.com/support/s/article/ka11A0000008ltYQAQ/mNode-connectivity
    check_files:
    Test Description:     Verify options file exists
    Passed node IDs:      1, 2, 3
    check_cores:
    Test Description:     Verify no core or dump files exists
    Passed node IDs:      1, 2, 3
    check_upload_speed:
    Test Description:     Measure the upload speed between the storage node and the management node
    Node ID: 1 Upload speed: 90063.90 KBs/sec
    Node ID: 3 Upload speed: 106511.44 KBs/sec
    Node ID: 2 Upload speed: 85038.75 KBs/sec
    如果出现错误,则需要执行其他操作:
    • 集群未做好升级准备。查看 sfupgradecheck 错误消息。
      示例输出如下:
      The following tests failed:
      check_root_disk_space:
      Test Description: Verify node root directory has at least 12 GBs of available disk space
      Severity: ERROR
      Failed node IDs: 2
      Remedy: Remove unneeded files from root drive
      More information: https://kb.netapp.com/support/s/article/ka11A0000008ltTQAQ/SolidFire-Disk-space-error
      check_pending_nodes:
      Test Description: Verify no pending nodes in cluster
      More information: https://kb.netapp.com/support/s/article/ka11A0000008ltOQAQ/pending-nodes
      check_cluster_faults:
      Test Description: Report any cluster faults
      check_root_disk_space:
      Test Description: Verify node root directory has at least 12 GBs of available disk space
      Passed node IDs: 1, 3
      More information: https://kb.netapp.com/support/s/article/ka11A0000008ltTQAQ/SolidFire-Disk-space-error
      check_mnode_connectivity:
      Test Description: Verify storage nodes can communicate with management node
      Passed node IDs: 1, 2, 3
      More information: https://kb.netapp.com/support/s/article/ka11A0000008ltYQAQ/mNode-connectivity
      check_files:
      Test Description: Verify options file exists
      Passed node IDs: 1, 2, 3
      check_cores:
      Test Description: Verify no core or dump files exists
      Passed node IDs: 1, 2, 3
      check_upload_speed:
      Test Description: Measure the upload speed between the storage node and the management node
      Node ID: 1 Upload speed: 86518.82 KBs/sec
      Node ID: 3 Upload speed: 84112.79 KBs/sec
      Node ID: 2 Upload speed: 93498.94 KBs/sec
      在此示例中,节点 1 的磁盘空间不足。您可以从此错误消息列出的知识库文章中查看更多信息。
    • HealthTools 已过时。请查看此错误消息,并注意升级检查失败。
      示例输出如下:
      sfupgradecheck failed: HealthTools is out of date: installed version: 2018.02.01.200 latest version: 2018.03.05.901.
      The latest version of the HealthTools can be downloaded from: https://mysupport.netapp.com/NOW/cgi-bin/software/ 
      Or rerun with the -n option
      如果出现这种情况,必须按照输出中所述的步骤进行操作。
    • 管理节点位于非公开站点上。请查看此消息,并注意升级检查失败。
      示例输出如下:
      sfupgradecheck failed: Unable to verify latest available version of healthtools.
      1. 如果出现这种情况,请获取 metadata.json 文件并将其保存在非管理节点计算机上,然后运行以下命令:sfupgradecheck -l --metadata=<path-to-metadata-json>

        有关详细信息,请参见适用于非公开站点的其他升级信息。

        在非公开站点上升级 Element 软件

      2. 运行以下命令以验证 HealthTools 套件是否为最新:sfupgradecheck -u <cluster-user-name> -p <cluster-password> MVIP