Restore VMware virtual machine data
Updated over a week ago

Enterprise Workloads Editions: βœ… Business | βœ… Enterprise | βœ… Elite

If you have taken a standard backup (non-application aware), you can restore:

  • Full virtual machine: Restores the entire virtual machine. For more information, see Restore full virtual machine.

  • Data Restore: Restores VMDKs, files and folders.

To know more about the restore workflow, see About VMware restores.

Prerequisites for restoring VMs

Before you proceed with the restore, ensure you have reviewed the following articles:

Click to review the articles


Full virtual machine restore

A Full VM Restore restores the entire VM with all its VMDKs from a selected recovery point.

Click to review considerations for a full virtual machine restore

During original restore, the virtual machine will be restored at the same data store and ESXi as it was present before the restore. The original virtual machine will be overwritten. If the target virtual machine has new VM disks attached which were not present at the time of backup then they will be detached.


πŸ“ Note


​Virtual machine restore to original location is a Guest Repair operation. This requires the original virtual machine to be available during the restore operation.


You can restore virtual machines to an alternate location in the following scenarios:

  • Restore virtual machines present on one VMware Cloud on AWS (VMC) to another (existing or new) VMware Cloud on AWS (VMC) provided they are configured in the same Organization.

  • Restore virtual machines present on one On-premise vCenter to another (existing or new) On-premise vCenter provided they are configured in the same Organization.

  • Restore virtual machines present on VMware Cloud on AWS (VMC) to On-premise vCenter (existing or new) and vice-versa provided they are configured in the same Organization.

For backup and restore of VMs with NVMe controllers or disks, the following scenarios and limitations are applicable:

  • If you have upgraded to version 6.0.0 -147290 of the VMware backup proxy, you can back up and restore VMs with NVMe controllers or disks. Restores from the new recovery points created after the agent upgrade are subject to some Limitations.

  • For restores of virtual machines with NVMe controllers or disks from older recovery points, you need to add ENABLE_NVME_CONVERSION = True parameter in the Phoenix.cfg file on the VMware backup proxy. The restores are subject to some Limitations.

Limitations

  • MS SQL database restores from VMs with NVMe controllers or disk are not supported. Full VM restores having app-aware policy configured will restore such databases as part of VM restores.

  • The backups will fail if you have set HotAdd as a transport mode in the Phoenix.cfg file for the VMs having NVMe disks. However, if no changes are made to transport mode in the Phoenix.cfg file, though HotAdd is prioritized, the transport mode will fall back to NBDSSL. All sorts of restores will work from these recovery points.

Procedure to restore a full virtual machine

  1. Log in to the Management Console.

  2. From the top menu bar, select your organization if organizations are enabled.

  3. Click Protect > VMware.
    The vCenter/ESXi host page appears that lists all the registered vCenter/hypervisors.

  4. You can either select the registered vCenter/ESXi host from the card view or list view or select it from the vCenter/ESXi host list in the left navigation pane.

  5. In the left navigation pane, click Configured VMs.

  6. Select the virtual machine you want to restore.

  7. Click Restore > VM Restore.

  8. Click Full VMRestore and then select a recovery point.
    To know more about restoring cold recovery points, see Restoration of cold-tier data.
    ​

    Full VM Restore.png

  9. Click Proceed to Restore.

  10. Select the location to restore.

    • Original Location: Restores the entire VM to the original hypervisor with original settings.

    • Alternate Location: Creates a new VM on any selected hypervisor with the customized settings.
      Enter the following details for an alternate location.


      πŸ“ Note


      ​ Druva does not power on a full virtual machine restored to an alternate location.


      • Destination VMware Setup:From the dropdown list, select the VMware setup where you want to restore the virtual machine.

      • Backup Proxy Pool: Lists the backup proxy pools available in the vCenter server.

        It is recommended that you select a backup proxy pool with backup proxies that are locally connected to the destination hypervisor for restore.

      • Destination Hypervisor:From the dropdown list, select the ESXi hypervisor where you want to restore this virtual machine.

      • Destination Datastore:From the dropdown list, select the data store where you want to restore this virtual machine.

      • Available Space: In the case of a full restore, the destination must at least have 1 GB space excess than the files being restored.

        In the case of disk restore, the destination must have at least have space equal to the files being restored.

      • Compute Resource: Select host, cluster, or a resource pool where you want to restore the virtual machine. You cannot select a datacenter or a folder.

        This option is not available if you are restoring a virtual machine to a standalone ESXi host.

      • Folder: Select a folder under the data center hierarchy where you want to restore the virtual machine.

        This option is not available if you are restoring a virtual machine to a standalone ESXi host.

      • Network: Select a network setting available at the destination hypervisor.

        For on-premises, If a network adapter was not attached to the virtual machine at the time of backup, Druva will attach the VMXNET3 or VMXNET2 type adapter and restore the virtual machine with the selected network setting.

        If a network adapter was attached to the virtual machine at the time of backup, Druva will restore the virtual machine with the backed up network adapter and selected network setting.

        If multiple network adapters were attached to the virtual machine at the time of backup, Druva will restore the virtual machine with all the adapters and selected network settings.

        For troubleshooting network issues, see Troubleshooting.
        ​

      • Recovered VM Name (Optional):

        • Enter the name to be given to the recovered VM. If this field is left blank, or a folder with the recovered VM name already exists in the datastore, the recovered VM will be named <Name of the original virtual machine>_<counter>.

        • - |?*:/\"<>' are not allowed in the recovered VM name.

        • The VM name must not exceed 80 characters

  11. Review the restore details, and click Finish. After clicking Finish, Druva initiates restore pre-checks. These checks look for issues that can cause your restore job to fail before the restore has even started. Druva does not initiate a restore job if the restore pre-checks identify any issues. You must fix any identified issues before restoring your data. For more information, see VMware restore pre-checks.
    Druva creates a new virtual machine with the same configuration as the original virtual machine using the following naming syntax:
    ​<Name of the original virtual machine>_<counter>
    The <counter> increments by 1 for subsequent restores of the virtual machine.
    ​Example:
    Suppose you want to perform a full VM restore of a virtual machine called Test Virtual Machine to an alternate location and the Recovered VM name is left blank or a folder called Test Virtual Machine already exists in the datastore, Druva will name the recovered virtual machine Test Virtual Machine_1.
    If you perform a second full VM restore of the same virtual machine to an alternate location and leave the Recovered VM name field blank or a folder with the name Test Virtual Machine already exists in the datastore, Druva will name the recovered virtual machine Test Virtual Machine_2.

How the restore request is processed

The triggered restored request is processed in one of the following manners:

  • If the target virtual machine has no backup or restores request running: The triggered restore request will be processed.

  • If the target virtual machine is processing a backup request: The following message is displayed on the screen.
    ​A backup process is currently running on target VM. Do you wish to cancel backup and trigger restore?

    • If you select Yes, the backup in progress will be canceled and the restore request will be triggered. On the Jobs page this backup job will be marked as canceled and an error PHOENIX247 will be displayed. For more information on the error, see Enterprise Workloads Errors.

    • If you select No, the triggered restore request will be canceled and you will be directed to the restore screen.

  • If the target virtual machine is already processing a restore request: As two restore requests cannot run in parallel on the same virtual machine, the following message will be displayed on the screen.
    ​
    ​Failed to schedule restore. Error : Cannot proceed as another restore operation is in progress for the same VM.

Impact of post backup configuration changes

If you change the virtual machine configuration such as name, memory, and vCPUs after backing up the virtual machine:

  • Druva overwrites the updated configuration at the time of restore

  • The changed configuration is rolled back to what it was when the virtual machine was backed up.

  • In addition, the VMware hypervisor changes the thin disk to a thick disk in the following scenarios:

If Changed Block Tracking (CBT) was disabled at the time of backup.

  • The virtual disk was attached to a virtual machine that was created on top of an NFS datastore at the time of backup. For more information, see the VMware KB article.

  • All data blocks of the thin disk were allocated at the time of backup.


Data restore

Restore files and folders

Druva provides you with the capability to restore files to CIFS/SMB share and any virtual machine in your vCenter/ESXi.

Click to review considerations for a data restore

  • If the Hybrid Workloads agent restarts during the file restore operation, the job is not retried as Druva does not store any guest OS credential. This job will fail with PHOENIX250. For more information see, Druva errors.

  • The target virtual machine should have enough space to store the restored files. If the target virtual machine does not have enough space, the triggered restore job will be marked as successful with error.

  • File-level restore for deduplicated volume is not supported. Although it is possible to browse the disk successfully, it does not guarantee that files/folders can be restored successfully. As a workaround, you must restore the full VMDK.

Procedure to restore files and folders

  1. Log in to the Management Console.

  2. From the top menu bar, select your organization if organizations are enabled.

  3. Click Protect > VMware.
    The All vCenter/ESXi host page appears that lists all the registered vCenter/hypervisors.

  4. You can either select the registered vCenter/ESXi host from the card view or list view or select it from the vCenter/ESXi host list in the left navigation pane.

  5. In the left navigation pane, click Configured VMs.

  6. Select the virtual machine you want to restore.

  7. Click Restore > VM Restore.

  8. Click DataRestore and then select a recovery point.
    To know more about restoring cold recovery points, see Restoration of cold-tier data.

  9. Click Files & Folders and select the file or folder you want to restore.

    vmware_load_more.png


    πŸ“ Note


    ​A single click of expanding a particular folder will load a total of 1024 files/folders. If there are more than 1024 files/folders, then the Load More option will be visible to display the next set of results in that folder.



    ​Suppose all the backup proxies for the source backup set vCenter or standalone ESXi host are in a disconnected state. In that case, you can still use a different vCenter or standalone ESXi host proxies that are in a connected state for file-level browse and restore. Select the alternate vCenter or standalone ESXi host for file-level and click Save. The file-level browse is triggered on the alternate vCenter or standalone ESXi host Backup proxy.

    alternate_environment.png


    If the disk configured for the selected VM is a Dynamic GUID (Global Unique Identifier) Partition Table (GPT) disk, an error is displayed with a warning, for example,

    Failed to mount 2 out of 4 volumes.

    mount error.png


    File-level restore eliminates the need to perform time-consuming disk restores when you want to restore a single file. However, during file-level restore, the volumes listing is excluded for Dynamic GPT disks in the backup set. Therefore you can use a Windows Server staging VM to perform a file-level browse restore.
    A staging VM can be another Windows Server VM present in your environment. For more information, see prerequisites for a staging VM.


    πŸ“ Note


    ​During FLR browse for restoring data from Linux VM, Druva considers and displays every logical partition as a separate volume. Also, the volume/mount point under the root is visible as an empty directory.



    Perform the following:

    1. Click Select Staging VM. The Select Staging VM dialog is displayed with various topology details.
      ​

      select staging vm_full screen.png
      • Destination vCenter:

        Select the required destination vCenter where you want to restore the files or, in the case of the CIFS share to which you intend to restore the files. Ensure that the Backup proxy installed on this vCenter can access the CIFS shares.
        ​

      • Destination Hypervisor:

        Select the required hypervisor.
        ​

      • Backup Proxy Pool:

        From the drop-down list, select a Backup proxy pool. It is recommended that you select a Backup proxy pool with backup proxies that are locally connected to the destination hypervisor for restore.


        πŸ“ Note


        ​ If the virtual machine has a Backup proxy pool mapped to it, the mapped Backup proxy pool is selected by default.


      • Staging VM:

        Select a staging VM with credentials assigned or credentials not assigned.
        If you select a staging VM without credentials, then the Credentials field is displayed.


        πŸ“ Note


        ​ This field lists the staging VMs only if it follows all the pre-requisites.
        ​ The staging VM is used to mount volumes for file-level restores to the target VM.


      • ENTER CREDENTIALS FOR THE SELECTED STAGING VM

      • Credentials:Select or enter credentials for the selected staging VM.
        If you click the Custom value, then you can create credentials and the Label, Username, and Password fields are displayed.

      • Label:Enter a label for the staging VM.

      • Username:Enter a valid username for the staging VM.

      • Password:Enter a valid password for the staging VM.

    2. Click Save.


      πŸ“ Note


      ​After you click Save, the Dynamic GPT disks browse workflow is started. A browse job is triggered on the Backup proxy that mounts the VMDKs and attach those disks on the selected staging VM. The directories of the attached staging VM are then displayed. Select the files to be restored and click Restore.


      select files.png
  10. On the File Restore page, you can choose one of the following restore targets.
    ​

    File_restore.png
    • Restore to CIFS share:Restores the files and folders to a specific network share.

      The restored file will be available on UNC share in the Phoenix-VMWare-FLR-[jobid] folder.

      • UNC Path:

        Valid UNC share (network share) path that is reachable from the selected vCenter/ESXi.

        Valid sharename format:

        • \\<hostname> or<Ip address>\<Sharename>

        • //Host/Folder
          ​


        πŸ“ Note


        ​ UNC sharename must not have any Unicode characters.


      • Username:

        Valid username of UNC share.


        πŸ“ Note


        ​ You can specify domain name in the username of the UNC share in the "d omain@username" format.


      • Password:

        Valid password for UNC share.
        ​

      • Destination VMware Setup:

        From the drop-down list, select avCenter/ESXi thathas at least one backup proxy in the connected state.
        ​

      • Backup Proxy Pool:From the drop-down list, select a backup proxy pool. It is recommended that you select a backup proxy pool with backup proxies that are locally connected to the destination hypervisor for restore.
        ​

      • Staging VM:

        Select a staging VM with credentials assigned or credentials not assigned.
        If you select a staging VM without credentials, then the Credentials field is displayed.


        πŸ“ Notes
        ​ This field lists the staging VMs only if it follows all the prerequisites.


        ​ The staging VM is used to mount volumes for file-level restores to the target VM.
        ​
        ​An administrator can assign credentials for the staging VM from All Virtual Machines page on the Management Console. For more information, see Manage credentials for VMware servers.


      • ENTER CREDENTIALS FOR THE SELECTED STAGING VM

      • Credentials:Select or enter credentials for the selected staging VM.
        If you click the Custom value, then you can create credentials and the Label, Username, and Password fields are displayed.

      • Label:Enter a label for the staging VM.

      • Username:Enter a valid username for the staging VM.

      • Password:Enter a valid password for the staging VM.
        ​

    • Restore to Virtual Machine: Restores the files and folders to the original or specified VM.

      • Restore to original VM: Select this option if you want to restore the selected files to the same virtual machine.
        ​

        • GUEST OS CREDENTIALS

        • Username:Valid username.

        • Password:Valid password.

        • Location:Browse to the location where you want to restore files.

        • Backup Proxy Pool:From the drop-down list, select a Backup proxy pool. It is recommended that you select a Backup proxy pool with backup proxies that are locally connected to the destination hypervisor for restore.

        • Staging VM:

          Select a staging VM with credentials assigned or credentials not assigned.
          If you select a staging VM without credentials, then the Credentials field is displayed.


          πŸ“ Notes


          ​ This field lists the staging VMs only if it follows all the prerequisites.

          The staging VM is used to mount volumes for file-level restores to the target VM.

          This field is pre-populated by the same staging VM used in the previous flow to list the available files and folders. However, you can modify this.


        • ENTER CREDENTIALS FOR THE SELECTED STAGING VM

        • Credentials:Select or enter credentials for the selected staging VM.
          If you click the Custom value, then you can create credentials and the Label, Username, and Password fields are displayed.

        • Label:Enter a label for the staging VM.

        • Username:Enter a valid username for the staging VM.

        • Password:Enter a valid password for the staging VM.

        • IF ORIGINAL FILE (S) EXIST

        • Always overwrite:The files with the same name will always be overwritten at the destination regardless of the file version or date. By default, this option is always selected.

        • Always skip:The files with the same name will not be restored to the destination. The number of files skipped will be displayed in the progress logs.

        • Overwrite if original is older:

          The files with the same name will be overwritten only if the files are older than the version being restored.

        • Restore and move original to .orig:If the file being restored already exists at the destination, the version being restored will be saved with the same filename. While the existing file will be renamed to filename.ext.orig. On subsequent restores, the existing file versions at the destination will be appended with an incremental number such as filename.ext.orig1, filename.ext.orig2, and so on.

        • Preserve original directory structure:You can preserve the entire source location directory structure, that is tree path while restoring a file to an alternate location.
          An example of original directory structure: If you are restoring folder β€œB” from the source location Volume1/A/B/C/D to F:/.
          The restored file structure will be F:/A/B/C/D.
          ​

      • Restore to an Alternate VM:Select this option if you want to restore the selected files to a different virtual machine.

        • Destination VMware Setup:Select the required vCenter/ESXi.

        • Destination Hypervisor: Select the required Hypervisor.

        • Destination VM: Select the required virtual machine where you want to restore the files.


          πŸ“ Note
          ​The timestamp under the Destination VM drop-down lists the last time the virtual machine listing page was refreshed.


        • Backup Proxy Pool:

          From the drop-down list, select a backup proxy pool. It is recommended that you select a backup proxy pool with backup proxies that are locally connected to the destination hypervisor for restore.


          πŸ“ Note


          ​If the virtual machine has a backup proxy pool mapped to it, the mapped backup proxy pool is selected by default.


        • GUEST OS CREDENTIALS

        • Username:Valid username.

        • Password:Valid password.

        • Location:Browse to the location where you want to restore files.

        • Staging VM:

          You can select a staging VM with credentials assigned or credentials not assigned.
          If you select a staging VM without credentials, then the Credentials field is displayed.


          πŸ“ Notes
          ​

          This field lists the staging VMs only if it follows all the prerequisites.
          ​ The staging VM is used to mount volumes for file-level restores to the target VM.
          ​This field is pre-populated by the same staging VM used in the previous flow to list the available files and folders. However, you can modify this.


        • ENTER CREDENTIALS FOR THE SELECTED STAGINGVM

        • Credentials:Select or enter credentials for the selected staging VM.
          If you click the Custom value, then you can create credentials and the Label, Username, and Password fields are displayed.

        • Label:Enter a label for the staging VM.

        • Username:Enter a valid username for the staging VM.

        • Password:Enter a valid password for the staging VM.

        • IF ORIGINAL FILE(S) EXIST

        • Always overwrite:The files with the same name will always be overwritten at the destination regardless of the file version or date. By default, this option is always selected.

        • Always skip:The files with the same name will not be restored to the destination. The number of files skipped will be displayed in the progress logs.

        • Overwrite if original is older:

          The files with the same name will be overwritten only if the files are older than the version being restored.

        • Restore and move original to .orig:If the file being restored already exists at the destination, the version being restored will be saved with the same filename. While the existing file will be renamed to filename.ext.orig. On subsequent restores, the existing file versions at the destination will be appended with an incremental number such as filename.ext.orig1, filename.ext.orig2, and so on.

        • Preserve original directory structure:You can preserve the entire source location directory structure, which is tree path while restoring a file to an alternate location.
          ​
          An example of original directory structure: If you are restoring folder β€œB” from the source location Volume1/A/B/C/D to F:/.
          The restored file structure will be F:/A/B/C/D.

Click Finish. After clicking Finish, Druva initiates restore pre-checks. These checks look for issues that can cause your restore job to fail before the restore has even started. Druva does not initiate a restore job if the restore pre-checks identify any issues. You must fix any identified issues before restoring your data. For more information, see VMware restore pre-checks.


πŸ“ Note


​During FLR, zero-byte files are skipped, and the count is not displayed in the Job logs on the Management Console. To see the skipped files, download the restore job logs and unzip all the files, including the FLR job logs. Next,
1. Go to the directory where the log files are unzipped.
2. Open the file with a name starting with ' Phoenix-Flr' using a text editor and then use the search string ' Skipping empty file ' to search the skipped file names. Alternatively, you can use the following Linux command to search the skipped file names:
​grep -i "Skipping empty file" *Phoenix-Flr*


Restoration of ACL Parameters with files

The following ACL parameters are restored along with the files:

File-level Restore to CIFS share:

For Linux source virtual machine

The ACL parameters restored are:

  • File modification time

  • Last access time

For Windows source virtual machine

The ACL parameters restored are:

  • File modification time

  • Last access time

File-level Restore to a virtual machine:

For Linux source virtual machine

  • For the original virtual machine restore type, the ACL parameters restored are:

    • Modification time

    • Last access time

    • User id

    • Group id

    • File permission

  • For the alternate Windows virtual machine restore type, the ACL parameters restored are:

    • File modification time

    • Last access time

  • For the alternate Linux virtual machine restore type, the ACL parameters restored are:

    • Modification time

    • Last access time

    • User id

    • Group id

    • File permission

For Windows source virtual machine

  • For the original virtual machine restore type, the ACL parameters restored are:

    • File modification time

    • Last access time

  • For the alternate Windows virtual machine restore type, the ACL parameters restored are:

    • File modification time

    • Last access time

  • For the alternate Linux virtual machine restore type, the ACL parameters restored are:

    • File modification time

    • Last access time

Restore a VMware virtual disk

Click to review considerations for a virtual disk restore

After a restore completes, you should expect the following:

If you choose to restore a virtual disk/VMDK, Druva creates a new virtual machine with minimum configuration (100 MB and 1 CPU) and associates the VMDK files that you selected to it.​ Configuration does not matter if a virtual disk is not bootable. The new VM is named as per the following convention: <Name of the original virtual machine>_<counter >. The counter increments by 1 for subsequent restores of the virtual machine if a VM folder with the same name preexists in the datastore. For example, for the first restore of a virtual machine named Test Virtual Machine, Druva creates a virtual machine called Test Virtual Machine_1. For the second restore, Druva creates a virtual machine called Test Virtual Machine_2. This process checks the VMware datastore before creating the virtual machine. Multiple restores require a consistent datastore selection. The presence of the error The name β€˜vm_name_1’ already exists in the vCenter/ESXi console may cause restores to fail if the datastore selection is inconsistent.

In addition, the VMware hypervisor changes the thin disk to a thick disk in the following scenarios:

  • If Changed Block Tracking (CBT) was disabled at the time of backup.

  • The virtual disk was attached to a virtual machine that was created on top of an NFS datastore at the time of backup. For more information, see the VMware KB article.

  • All data blocks of the thin disk were allocated at the time of backup.

Procedure to restore virtual disk

  1. Log in to the Management Console.

  2. From the top menu bar, select your organization if organizations are enabled.

  3. Click Protect > VMware.
    The vCenter/ESXi host page appears that lists all the registered vCenter/hypervisors.

  4. You can either select the registered vCenter/ESXi host from the card view or list view or select it from the vCenter/ESXi host list in the left navigation pane.

  5. In the left navigation pane, click Configured VMs.

  6. Select the virtual machine you want to restore.

  7. Click Restore > VM Restore.

  8. Click DataRestore and then select a recovery point.
    To know more about restoring cold recovery points, see Restoration of cold-tier data.

  9. Click VMDK and select the virtual disk you want to restore.

    Restore_vmdk.png

  10. Click Restore.

  11. In the Virtual Disk Restore dialog box, enter the following details.

    • Destination VMware Setup:

      From the dropdown list, select the VMware setup where you want to restore the virtual disk.
      ​

    • Backup Proxy Pool:

      Lists the backup proxy pools available in the vCenter server.

      It is recommended that you select a backup proxy pool with backup proxies that are locally connected to the destination hypervisor for restore.
      ​

    • Destination Hypervisor:

      From the dropdown list, select the ESXi hypervisor where you want to restore this virtual disk.
      ​

    • Destination Datastore:

      From the dropdown list, select the data store where you want to restore this virtual disk.
      ​

    • Compute Resource:

      Select host, cluster, or a resource pool where you want to restore the virtual disk. You cannot select a datacenter or a folder.
      This option is not available if you are restoring a virtual disk to a standalone ESXi host.
      ​

    • Folder:

      Select a folder under the data center hierarchy where you want to restore the virtual disk.
      This option is not available if you are restoring a virtual disk to a standalone ESXi host.

  12. Review the restore details, and click Finish. After clicking Finish, Druva initiates restore pre-checks. These checks look for issues that can cause your restore job to fail before the restore has even started. Druva does not initiate a restore job if the restore pre-checks identify any issues. You must fix any identified issues before restoring your data. For more information, see VMware restore pre-checks.


πŸ“ Note


​To access the restored VMWARE virtual disks, please refer to the article Adding existing Virtual Disks to a Virtual Machine.


Did this answer your question?