Failing over a server to a virtual target machine on VMware ESX

The Virtual Recovery Assistant provides the ability to failover to a replica virtual machine in the event of a failure of the source machine. Source servers may be failed over in two different ways:

Automatic Failover—The application detects that a failover condition is met and automatically initiates the failover process.

Manual Failover—The user manually initiates the failover process by clicking the Failover button in the user interface.

The type of failover for the job is selected during job creation. For more information, see Set protection options.

Note:

For optimal virtual machine performance, Double-Take Software recommends that you install VMtools on the replica virtual machine after failover is complete. In most cases, the installation of VMtools will require the virtual machine to be rebooted.

When setting up a protection using a virtual recovery appliance that has an out-of-date VMware tools installation, Virtual Recovery Assistant prompts the user to correct the issue. After the issue is corrected, the console displays an error. You must cancel the protection and restart the console.

During the failover process, the following steps are initiated by the Virtual Recovery Assistant:

The Double-Take connection between the source machine and Virtual Recovery Appliance is disconnected.

The VMware virtual drivers for video, mouse, disk controllers, and network adapters are injected into the Windows registry on the target virtual disk.

The target virtual disk volumes are un-mounted and removed from the Virtual Recovery Appliance.

If you are performing manual failover, the source machine is shut down.

The target virtual disk volumes are connected to the replica virtual machine.

The replica virtual machine is powered on.

Note:

If the source is running Windows 2008 and its SAN policy is not OnlineAll, the replica machine's non-boot volumes will need to be mounted manually after the replica machine is powered on after failover.

If you have a drive C: and D: and a CDROM drive E: and you failover a source with the SAN policy of OfflineShared, the replica will come up with only drive C: online and the CDROM on drive D:. If you online the disk, it will take drive letter E: .

During manual failover from a normally-functioning source machine, in some cases the replica virtual machine may come online before the source has fully shut down, resulting in an IP address conflict. In most cases, this situation will be resolved once the source machine shutdown is complete.

Undoing failover

Note:

If a VirtualCenter server is used during job creation and failover is attempted while VirtualCenter is offline, the failover process will not complete properly and the job will be in an error state. To resolve this issue:

Using the VMware VI client, edit the virtual machine settings of the recovery appliance. Remove the disk drives of the replica virtual machine, but DO NOT delete the drives from disk.
Start the replica virtual machine.
Delete the protection job from the Virtual Recovery Assistant client. If asked whether or not to delete the associated replica virtual machine, DO NOT delete the replica, then confirm the selection.