使用 Azure 备份即时还原功能获得更高的备份和还原性能Get improved backup and restore performance with Azure Backup Instant Restore capability


根据用户的反馈,我们正在将“VM 备份堆栈 V2” 重命名为“即时还原” ,以减少与 Azure Stack 功能的混淆。Based on feedback from users we are renaming VM backup stack V2 to Instant Restore to reduce confusion with Azure Stack functionality. 所有 Azure 备份用户现在已升级到即时还原All the Azure backup users have now been upgraded to Instant Restore.

即时还原的新模型提供以下功能增强:The new model for Instant Restore provides the following feature enhancements:

  • 可以使用执行备份作业期间创建的用于恢复的快照,而无需等待将数据传输到保管库的操作完成。Ability to use snapshots taken as part of a backup job that is available for recovery without waiting for data transfer to the vault to finish. 它缩短了快照在触发还原之前复制到保管库的等待时间。It reduces the wait time for snapshots to copy to the vault before triggering restore.
  • 默认情况下,可在本地保留快照两天,这缩短了备份和还原时间。Reduces backup and restore times by retaining snapshots locally, for two days by default. 此默认快照保留期值可配置为 1 到 5 天的任何值。This default snapshot retention value is configurable to any value between 1 to 5 days.
  • 最大支持 4 TB 的磁盘。Supports disk sizes up to 4 TB. Azure 备份不建议重设磁盘大小。Resizing of disks is not recommended by Azure Backup. 若要注册 Azure 备份大磁盘的受限公共预览版以支持大于 4 TB 且最大 30 TB 的磁盘,请参阅备份磁盘大小高达 30 TB 的 VMTo sign up for a limited public preview of Azure Backup large disk support for disks greater than 4 TB and up to 30 TB in size, see Backup of VM with disk sizes up to 30 TB.
  • 支持标准 SSD 磁盘、标准 HDD 磁盘和高级 SSD 磁盘。Supports Standard SSD disks along with Standard HDD disks and Premium SSD disks.
  • 还原时可以使用非托管 VM 的原始存储帐户(按磁盘)。Ability to use an unmanaged VMs original storage accounts (per disk), when restoring. 即使 VM 的磁盘跨存储帐户进行分布,也具备此能力。This ability exists even when the VM has disks that are distributed across storage accounts. 这可以加快各种 VM 配置的还原操作。It speeds up restore operations for a wide variety of VM configurations.
  • 若要通过 Instant Restore 备份使用高级存储的 VM,建议从总的已分配存储空间中分配 50% 的可用空间,这在首次备份时是必需的。For backup of VMs that are using premium storage, with Instant Restore, we recommend allocating 50% free space of the total allocated storage space, which is required only for the first backup. 首次备份完成后,50% 的可用空间不再是备份的要求。The 50% free space is not a requirement for backups after the first backup is complete.

功能亮点What's new in this feature

目前,备份作业包括两个阶段:Currently, the backup job consists of two phases:

  1. 获取 VM 快照。Taking a VM snapshot.
  2. 将 VM 快照传输到 Azure 恢复服务保管库。Transferring a VM snapshot to the Azure Recovery Services vault.

只有在完成阶段 1 和 2 之后,才认为已创建恢复点。A recovery point is considered created only after phases 1 and 2 are completed. 在执行此项升级的过程中,完成快照后会立即创建恢复点,可以在相同的还原流中,使用此快照恢复点类型执行还原。As a part of this upgrade, a recovery point is created as soon as the snapshot is finished and this recovery point of snapshot type can be used to perform a restore using the same restore flow. 可以在 Azure 门户中使用“快照”作为恢复点类型来识别此恢复点,快照传输到保管库后,恢复点类型将更改为“快照和保管库”。You can identify this recovery point in the Azure portal by using “snapshot” as the recovery point type, and after the snapshot is transferred to the vault, the recovery point type changes to “snapshot and vault.”

VM 备份堆栈资源管理器部署模型中的备份作业 - 存储和保管库

默认情况下,快照将保留两天。By default, snapshots are retained for two days. 此功能允许从保管库中的这些快照执行还原操作,并可缩短还原时间。This feature allows restores operation from these snapshots there by cutting down the restore times. 对于非托管磁盘方案,此功能减少了转换数据并将数据从保管库复制回到用户存储帐户所需的时间;对于托管磁盘用户,它可以基于恢复服务数据创建托管磁盘。It reduces the time that is required to transform and copy data back from the vault to the user's storage account for unmanaged disk scenarios while for managed disk users, it creates managed disks out of the Recovery Services data.

功能注意事项Feature considerations

  • 快照将连同磁盘一起存储,以提高恢复点的创建速度并加快还原操作。Snapshots are stored along with the disks to boost recovery point creation and to speed up restore operations. 因此,可以查看 7 天内创建的快照的相应存储成本。As a result, you'll see storage costs that correspond to snapshots taken during this period.
  • 增量快照作为页 blob 存储。Incremental snapshots are stored as page blobs. 使用非托管磁盘的所有用户需要为其本地存储帐户中存储的快照付费。All the users using unmanaged disks are charged for the snapshots stored in their local storage account. 由于托管 VM 备份使用的还原点集合在基础存储级别使用 Blob 快照,因此对于托管磁盘,你将看到与 Blob 快照定价对应的成本,并且成本是递增的。Since the restore point collections used by Managed VM backups use blob snapshots at the underlying storage level, for managed disks you will see costs corresponding to blob snapshot pricing and they are incremental.
  • 对于高级存储帐户,为即时恢复点创建的快照计入分配空间的 10 TB 限制。For premium storage accounts, the snapshots taken for instant recovery points count towards the 10 TB limit of allocated space.
  • 可以根据还原需求配置快照保留期。You get an ability to configure the snapshot retention based on the restore needs. 根据要求,可以在备份策略边栏选项卡中将快照保留期设置为最少一天,如下所述。Depending on the requirement, you can set the snapshot retention to a minimum of one day in the backup policy blade as explained below. 如果不经常执行还原,这可帮助节省保留快照的成本。This can help you save cost for snapshot retention if you don’t perform restores frequently.
  • 这是单向升级,一旦升级到即时还原,就不能回退。This is a one directional upgrade, once upgraded to Instant restore, you cannot go back.


使用此即时还原升级,所有客户(包含新客户和现有客户)的快照保留期都将设置为两天的默认值。With this instant restore upgrade, the snapshot retention duration of all the customers (new and existing both included) will be set to a default value of two days. 但是,可以根据需要将持续时间设置为 1 到 5 天之间的任意值。However, you can set the duration as per your requirement to any value between 1 to 5 days.

成本影响Cost impact

增量快照存储在 VM 的存储帐户中,用于即时恢复。The incremental snapshots are stored in VM’s storage account, which are used for instant recovery. 使用增量快照意味着快照占用的空间等于创建该快照后写入的页面所占用的空间。Incremental snapshot means the space occupied by a snapshot is equal to the space occupied by pages that are written after the snapshot was created. 费用仍按快照占用的空间按 GB 计算,每 GB 价格与定价页中所述的价格相同。Billing is still for the per GB used space occupied by the snapshot and the price per GB is same as mentioned in the pricing page.


就每周策略来说,快照保留期固定为 5 天。Snapshot retention is fixed to 5 days for weekly policies.

配置快照保留期Configure snapshot retention

使用 Azure 门户Using Azure portal

在 Azure 门户中可以看到,“即时还原”部分下的“VM 备份策略”边栏选项卡中添加了一个字段。In the Azure portal, you can see a field added in the VM Backup Policy blade under the Instant Restore section. 对于与特定备份策略关联的所有 VM,可以在“VM 备份策略”边栏选项卡中更改快照保留持续时间。You can change the snapshot retention duration from the VM Backup Policy blade for all the VMs associated with the specific backup policy.


使用 PowerShellUsing PowerShell


从 Az PowerShell 版本 1.6.0 开始,可以使用 PowerShell 在策略中更新即时还原快照保留期From Az PowerShell version 1.6.0 onwards, you can update the instant restore snapshot retention period in policy using PowerShell

$bkpPol = Get-AzureRmRecoveryServicesBackupProtectionPolicy -WorkloadType "AzureVM"
Set-AzureRmRecoveryServicesBackupProtectionPolicy -policy $bkpPol

每个策略的默认快照保留期设置为 2 天。The default snapshot retention for each policy is set to 2 days. 用户可以将此值更改为 1 到 5 天。User can change the value to a minimum of 1 and a maximum of 5 days. 就每周策略来说,快照保留期固定为 5 天。For weekly policies, the snapshot retention is fixed to 5 days.

常见问题Frequently asked questions

哪些因素影响即时还原的成本?What are the cost implications of Instant restore?

快照将连同磁盘一起存储,以加速恢复点的创建和还原操作。Snapshots are stored along with the disks to speed up recovery point creation and restore operations. 因此,你会看到与 VM 备份策略中选择的快照保留期相对应的存储成本。As a result, you'll see storage costs that correspond to the snapshot retention selected as a part of VM backup policy.

在高级存储帐户中,为即时恢复点创建的快照是否会占用 10 TB 快照限制?In Premium Storage accounts, do the snapshots taken for instant recovery point occupy the 10 TB snapshot limit?

是的,对于高级存储帐户,为即时恢复点创建的快照会占用 10 TB 的已分配快照空间。Yes, for premium storage accounts the snapshots taken for instant recovery point occupy 10 TB of allocated snapshot space.

在 5 天期限内,快照保留的工作方式是怎样的?How does the snapshot retention work during the five-day period?

如果每天创建一个新快照,则有 5 个单独的增量快照。Each day a new snapshot is taken, then there are five individual incremental snapshots. 快照大小取决于数据变动率,在大多数情况下,变动率大约为 2%-7%。The size of the snapshot depends on the data churn, which are in most cases around 2%-7%.

即时还原快照是增量快照还是完整快照?Is an instant restore snapshot an incremental snapshot or full snapshot?

即时还原功能创建的快照是增量快照。Snapshots taken as a part of instant restore capability are incremental snapshots.

如何计算即时还原功能大约增加的成本?How can I calculate the approximate cost increase due to instant restore feature?

这取决于 VM 的变动率。It depends on the churn of the VM. 在稳定状态下,可以假设增加的成本 = 快照保留期 * 每个 VM 的每日变动率 * 每 GB 存储成本。In a steady state, you can assume the increase in cost is = Snapshot retention period daily churn per VM storage cost per GB.

如果还原点的恢复类型是“快照和保管库”,而我执行了还原操作,那么,将使用哪种恢复类型?If the recovery type for a restore point is “Snapshot and vault” and I perform a restore operation, which recovery type will be used?

如果恢复类型是“快照和保管库”,则会从本地快照自动执行还原,与从保管库执行还原相比,其速度要快得多。If the recovery type is “snapshot and vault”, restore will be automatically done from the local snapshot, which will be much faster compared to the restore done from the vault.

如果选择的还原点(第 2 层)保留期小于快照(第 1 层)保留期,会发生什么情况?What happens if I select retention period of restore point (Tier 2) less than the snapshot (Tier1) retention period?

除非删除快照(第 1 层),否则新模型不允许删除还原点(第 2 层)。The new model does not allow deleting the restore point (Tier2) unless the snapshot (Tier1) is deleted. 建议将还原点(第 2 层)保留期设置为大于快照保留期。We recommend scheduling restore point (Tier2) retention period greater than the snapshot retention period.

为何我即使在备份策略中设置了保留期,我的快照也仍然存在?Why is my snapshot existing even after the set retention period in backup policy?

如果恢复点包含快照并且存在最新可用的 RP,则该快照会一直保留到下一次成功备份为止。If the recovery point has snapshot and that is the latest RP available, it is retained until the time there is a next successful backup. 这符合当前设计的 GC 策略,该策略强制要求始终至少有一个最新的 RP,以防 VM 中的问题导致所有备份进一步出错。This is as per the designed GC policy today that mandates at least one latest RP to be always present in case all backups further on fail due to an issue in the VM. 正常情况下,在 RP 过期后,将在最多 24 小时内予以清理。In normal scenarios, RPs are cleaned up in maximum of 24 hours after their expiry.