灾难恢复后将 VMware VM 故障回复到 AzureFailback of VMware VMs after disaster recovery to Azure

在灾难恢复过程中故障转移到 Azure 以后,可以故障回复到本地站点。After you have failed over to Azure as part of your disaster recovery process, you can fail back to your on-premises site. 使用 Azure Site Recovery 可以执行两种不同类型的故障回复:There are two different types of failback that are possible with Azure Site Recovery:

  • 故障回复到原始位置Fail back to the original location
  • 故障回复到备用位置Fail back to an alternate location

故障转移 VMware 虚拟机后,可以故障回复到同一个本地源虚拟机(如果仍然存在)。If you failed over a VMware virtual machine, you can fail back to the same source on-premises virtual machine if it still exists. 在这种情况下,只会复制更改。In this scenario, only the changes are replicated back. 这种情况称为原始位置恢复This scenario is known as original location recovery. 如果本地虚拟机不存在,则这种情况是一种备用位置恢复If the on-premises virtual machine does not exist, the scenario is an alternate location recovery.

备注

仅可故障回复到原始 vCenter 和配置服务器。You can only fail back to the original vCenter and Configuration server. 无法部署新配置服务器并通过它进行故障回复。You cannot deploy a new Configuration server and fail back using it. 此外,无法向现有配置服务器添加新 vCenter 并故障回复到新 vCenter。Also, you cannot add a new vCenter to the existing Configuration server and failback into the new vCenter.

原始位置恢复 (OLR)Original Location Recovery (OLR)

如果选择故障回复到原始虚拟机,需要满足以下条件:If you choose to fail back to the original virtual machine, the following conditions need to be met:

  • 如果虚拟机由 vCenter 服务器管理,主目标的 ESX 主机应该可以访问虚拟机数据存储。If the virtual machine is managed by a vCenter server, then the master target's ESX host should have access to the virtual machine's datastore.
  • 如果虚拟机位于 ESX 主机上,但不由 vCenter 管理,则虚拟机的硬盘必须位于主目标的主机可访问的数据存储中。If the virtual machine is on an ESX host but isn't managed by vCenter, then the hard disk of the virtual machine must be in a datastore that the master target's host can access.
  • 如果虚拟机位于 ESX 主机上但不使用 vCenter,则应完成针对主目标的 ESX 主机的发现操作,才能进行重新保护。If your virtual machine is on an ESX host and doesn't use vCenter, then you should complete discovery of the ESX host of the master target before you reprotect. 若要对物理服务器进行故障回复,此规则也同样适用。This applies if you're failing back physical servers, too.
  • 可以故障回复到虚拟存储区网络 (vSAN) 或基于原始设备映射 (RDM) 的磁盘(如果磁盘已存在并且已连接到本地虚拟机)。You can fail back to a virtual storage area network (vSAN) or a disk that based on raw device mapping (RDM) if the disks already exist and are connected to the on-premises virtual machine.

重要

请务必启用 disk.enableUUID= TRUE,以便故障回复期间 Azure Site Recovery 服务能够识别将向其写入未完成更改的虚拟机上的原始 VMDK。It is important to enable disk.enableUUID= TRUE so that during failback, the Azure Site Recovery service is able to identify the original VMDK on the virtual machine to which the pending changes will be written. 如果此值未设置为 TRUE,则该服务会尽力尝试识别相应的本地 VMDK。If this value is not set to be TRUE, then the service tries to identify the corresponding on-premises VMDK on a best effort basis. 如果找不到合适的 VMDK,它会创建额外的磁盘,并将数据写入到该磁盘。If the right VMDK is not found, it creates an extra disk and the data gets written on to that.

备用位置恢复 (ALR)Alternate location recovery (ALR)

如果在重新保护本地虚拟机之前该虚拟机不存在,则该方案称为备用位置恢复。If the on-premises virtual machine does not exist before reprotecting the virtual machine, the scenario is called an alternate location recovery. 重新保护工作流将重新创建本地虚拟机。The reprotect workflow creates the on-premises virtual machine again. 这还会导致完整数据下载。This will also cause a full data download.

  • 故障回复到备用位置时,会将虚拟机恢复到主目标服务器所部署到的同一 ESX 主机。When you fail back to an alternate location, the virtual machine is recovered to the same ESX host on which the master target server is deployed. 用于创建磁盘的数据存储将与重新保护虚拟机时选择的数据存储相同。The datastore that's used to create the disk will be the same datastore that was selected when reprotecting the virtual machine.
  • 只能故障回复到虚拟机文件系统 (VMFS) 或 vSAN 数据存储。You can fail back only to a virtual machine file system (VMFS) or vSAN datastore. 如果有 RDM,重新保护和故障回复将不起作用。If you have an RDM, reprotect and failback will not work.
  • 重新保护涉及一次较大的初始数据传输,并会进行更改。Reprotect involves one large initial data transfer that's followed by the changes. 之所以要执行此过程,是因为本地没有虚拟机。This process exists because the virtual machine does not exist on premises. 需要复制回完整的数据。The complete data has to be replicated back. 此重新保护所需的时间还比原始位置恢复长。This reprotect will also take more time than an original location recovery.
  • 无法故障回复到基于 RDM 的磁盘。You cannot fail back to RDM-based disks. 只能在 VMFS/vSAN 数据存储中创建新的虚拟机磁盘 (VMDK)。Only new virtual machine disks (VMDKs) can be created on a VMFS/vSAN datastore.

备注

故障转移到 Azure 时,物理计算机只能故障回复为 VMware 虚拟机。A physical machine, when failed over to Azure, can be failed back only as a VMware virtual machine. 这需要执行与备用位置恢复相同的工作流。This follows the same workflow as the alternate location recovery. 确保除了需要故障回复到的必要 ESX/ESXi 主机之外,还发现至少一台主目标服务器。Ensure that you discover at least one master target server and the necessary ESX/ESXi hosts to which you need to fail back.

后续步骤Next steps

按照步骤执行故障回复操作Follow the steps to perform the failback operation.