The following prerequisites exist for VMware agentless restores:

• You must provision VxUpdate packages for all platforms for which you have virtual machines where you want to perform agentless recovery.

• You must have credentials with administrator or root permissions for the target virtual machine.

• The target VM is where the files are recovered. It must be powered on and have the latest version of VMware Tools installed.

• The target VM should have at least one Paravirtual Controller with available LUNs. Or, available space for a Paravirtual SCSI Controller.

• The default staging location on the target VM is %TEMP% or %TMP% for Windows and the root directory (/) for Linux.

• The staging location must exist on the target VM file system.

• If you want to allow the use of instant access for recovery of the files and folders, the recovery point must support instant access.

  See Create an instant access VM.

The following limitations exist for VMware agentless restores:

• Agentless restores to Windows target VMs can fail if you use an account other than the built-in Administrator for Windows Guest OS account as the Target VM Credentials. The restore fails because Run all administrators in Admin Approval Mode is enabled. More information is available:

  https://www.veritas.com/content/support/en_US/article.100046138.html

• VMware agentless restores can only be used for the restore of files and folders.

• In some instances, when you perform an agentless restores, orphaned VMs starting with NB_ are left behind. Using the ESX server credentials to perform the restore on the target VM even though the vCenter manages the ESX server can cause this condition. This condition is a known limitation of VMware. To resolve the problem, register the vCenter in NetBackup and use vCenter credentials for backups and restores. The orphaned VMs starting with NB_ can be removed from inventory manually by logging into the vCenter using VMware vSphere Client.

• Restore job fails if NetBackup is unable to use the staging directory. This directory is specified in the TMP or TEMP environment variable.

• Restore job fails if NetBackup does not have sufficient privileges to the staging directory. Or, if there is insufficient space in the staging directory.

• If you select Flatten existing directory structure and Overwrite existing files options, you risk an incorrect restore if it contains multiple files with the same file name. In this case, the last file that is restored is the one that is present when the restore completes.

  If you select Flatten existing directory structure and you do not select Overwrite existing files, the restore succeeds. The first file that is restored is present when the restore completes. To prevent this issue, do not select Flatten existing directory structure when restoring multiple files with the same name.

• The Flatten existing directory structure and Append string to file names options are only applicable to files. They are not available for directories.

• Multiple restore jobs to the same VM are not supported. The user must start another job as needed for that VM once the first restore job for that VM has completed.

• If a backup and a restore occur simultaneously on the same VM, one or both jobs can have unexpected results. If a backup or a restore exits with a non-zero status code, one possible cause is simultaneous jobs occurring on the same VM.

• Veritas does not recommend VMware agentless restore if a NetBackup client already exists on the target VM. The NetBackup administrator must use the agent-based restore in such cases.

• For the current list of guest operating systems that NetBackup supports for the target VM, see Supported guest operating systems for VMware in the following document:

  Support for NetBackup in virtual environments