Prerequisites and limitations of VMware agentless restores
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, root, or sudo 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.
To use non-root credentials on a Linux target VM it must have sudo installed and the
/etc/sudoersfile configured so that the user has the following permissions:username ALL=(ALL) NOPASSWD: /bin/tar, SETENV: /usr/openv/tmp/rt/netbackup/bin/nbtar_rt
or
username ALL=(ALL) NOPASSWD: ALL
The default staging location on the target VM is
%TEMP%or%TMP%for Windows and the tmp directory (/tmp) 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.
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 . The restore fails because 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 withNB_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 and 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 and you do not select , the restore succeeds. The first file that is restored is present when the restore completes. To prevent this issue, do not select when restoring multiple files with the same name.
The and 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.
Cohesity 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: