排查附加 VHD 的 VM 意外重启Troubleshoot unexpected reboots of VMs with attached VHDs

如果 Azure 虚拟机 (VM) 具有大量位于同一存储帐户的附加 VHD,则可能会超过单个存储帐户的可伸缩性目标,从而导致 VM 意外重启。If an Azure Virtual Machine (VM) has a large number of attached VHDs that are in the same storage account, you may exceed the scalability targets for an individual storage account, causing the VM to reboot unexpectedly. 检查存储帐户的分钟指标 (TotalRequests/TotalIngress/TotalEgress),了解超过存储帐户的可伸缩性目标的峰值。Check the minute metrics for the storage account (TotalRequests/TotalIngress/TotalEgress) for spikes that exceed the scalability targets for a storage account. 若要帮助确定是否已对存储帐户进行限制,请参阅指标显示 PercentThrottlingError 增加See Metrics show an increase in PercentThrottlingError for assistance in determining whether throttling has occurred on your storage account.

通常,虚拟机对 VHD 进行的每个单独的输入或输出操作都会转换为对基础页 Blob 进行的“Get 页” 或“Put 页” 操作。In general, each individual input or output operation on a VHD from a Virtual Machine translates to Get Page or Put Page operations on the underlying page blob. 因此,可以根据应用程序的特定行为,对环境使用估计的 IOPS 以优化可以在单个存储帐户中设置的 VHD 数。Therefore, you can use the estimated IOPS for your environment to tune how many VHDs you can have in a single storage account based on the specific behavior of your application. Azure 建议在单个存储帐户中设置 40 个或更少的磁盘。Azure recommends having 40 or fewer disks in a single storage account. 有关标准存储帐户的可伸缩性目标的详细信息,请参阅标准存储帐户的可伸缩性目标For more information about scalability targets for standard storage accounts, see Scalability targets for standard storage accounts. 若要详细了解高级页 Blob 存储帐户的可伸缩性目标,请参阅高级页 Blob 存储帐户的可伸缩性目标For more information about scalability targets for premium page blob storage accounts, see Scalability targets for premium page blob storage accounts.

如果即将超过存储帐户的可伸缩性目标,请将 VHD 放入多个存储帐户中,以减少每个帐户中的活动。If you are exceeding the scalability targets for your storage account, place your VHDs in multiple storage accounts to reduce the activity in each individual account.