Data recovery is the process of taking a computer that cannot start itself and returning it to a previous operating state. This means a recovery has to start the computer and then recover its operating system, data and applications.
Examples of recovery scenarios include:
- A server has lost its primary hard drive and has no RAID to restore it.
- A server will not start because its operating system is corrupted.
- A server has been lost in a flood and has to be replaced.
- Your building has lost power and you need to recover your server at an alternate site.
The loss of a server means the loss of all the data such as files, emails on an Exchange Server, databases on an SQL Server and the services the server supplied to your network.
It is therefore critical that you implement and test a recovery plan for each of your servers.
Identify your requirements
To recover a server you need:
- A bootable recovery media that can start your server and load a temporary environment.
- A backup of the operating system, data and applications that the bootable recovery media can use.
Before you create your bootable recovery media, you need to consider the following:
- The serverâ€™s firmware: You need to know if the server you backed up and the server you will recover to use BIOS or EFI. Bootable media made on a server that uses one kind of firmware cannot be used to start a server that uses a different kind of firmware.
- The media you use: EFI is designed to be bootable from a removable USB media, so the media you use must be a removable USB drive. You also need to make sure that whatever media you use has the correct drivers and that the server supports those drivers.
When creating your bare-metal backup, you need to consider the following:
- The backup must contain all of the data needed to perform a bare-metal recovery. For example, all the operating system files must be included, as well as the data.
- The backup type must be supported by the bootable recovery media. For example, most bootable media create recovery environments that require image (VHD) backups.
- The disk configuration on the server you backed up must be compatible with the disk configuration on the server you are recovering to. For example, the computer being recovered must have a physical disk the same size, or larger than, the physical disk that the backup was made from.
Create the bootable recovery media and bare-metal backup
Create the bootable media that will start the server and load a temporary environment, and a bare-metal backup of the server that the recovery environment can use.
BackupAssistâ€™s Recovery tab uses RecoverAssist, a built-in feature that creates a customized, bootable recovery media, such as an optical disk or USB drive.
BackupAssistâ€™s System Protection can create an image backup using the bare-metal option on the data selection screen.
Perform a test recovery
A recovery has a lot of dependencies and components such as the hardware, firmware, operating system and media, and they all have to work together for the recovery to be successful. It is therefore important that once you have created your bootable recovery media and backup, that you test them. This means providing a server that you can recover in a test environment. By doing this you can identify and resolve any problems that arise, and document the process.
BackupAssistâ€™s bootable recovery media comes with a suite of useful tools that can diagnose problems and assist with your recovery. You can even add your own files and custom drivers to the recovery media.
Safely store your recovery kit
Both the bootable recovery media and the image backup must be stored in a safe location, and not the same location as the server that you backed up. If your server room is lost in a natural disaster, you donâ€™t want to lose your recovery media and backup as well.
This has been a brief overview of what is a critical function. Losing a server is an event you need to plan for and, due to the hardware and software dependencies, the only way to know if your recovery plan will work, is to test it – so please do.
If you would like to read more about recoveries and the considerations mentioned in this article, please see our Recovery guides.