Perform a bare-metal restore

This section of documentation provides you with complete instructions on how to use the Bare-Metal Restore functionality in Server Backup Manager.

Bare metal restore is the process of reformatting a computer/server from scratch after a catastrophic failure. In conventional methods, the process typically involves reinstalling the operating system and software applications and then, if possible, restoring data and settings.

With R1Soft backup technology, you can restore servers directly from disk-based backup. There is no need to first partition your drive and re-install the operating system.

With Server Backup Manager, you can use Bare-Metal Restore with the help of the following tools:

• Live CD Agent
• Live CD Server
• PXE Network Boot Agent
• PXE Network Boot Server

Options for Performing a Bare-Metal Restore

There are different kinds of disasters you may need to recover from. At R1Soft, we consider any situation requiring you to take your server off-line to Restore as Disaster Recovery, and we call this Bare-Metal Restore. You typically want to perform a Bare-Metal Restore any time you need to Restore more than just a handful of files.

There are two options for performing a Bare-Metal Restore.

Note Disaster Recovery Boot Media supports the following two recovery Methods for both Windows and Linux Bare-Metal Restores.

• Live CD (Disaster Recovery Boot Media Version 3.18.1 and higher) - Download the CD-ROM ISO image, also referred to as Live CD. This ISO image should be burned to a CD-ROM using a CD-writer drive and any common CD burning software (see Write ISO file CDs). This method requires physical access to the computer to insert the CD-ROM. It cannot be easily automated or remotely initiated. See Use a boot CD-ROM.

• PXE Network Boot (Disaster Recovery Boot Media Version 3.18.1 and higher) - With PXE Boot and an installed Server Backup it is possible to provide a hands-free, fully automated Disaster Recovery and re-deployment system for the entire data center. See Use PXE network boot. 

Note To use this method, you need a special "boot" server running the services required for PXE Environment to work - DHCP and TFTP, as well as NFS service. Having this "boot" server and Server Backup, you will be able to carry out a hands-free, fully automated Disaster Recovery and re-deployment system. For servers which provide remote access to the server console, it can be done via IP-KVM solutions. This method does not require physical access to the server.

Each of the listed options has two sub-options: 

• Booting into Agent environment - To use this option, you need a running Backup Manager installed on a machine different from one to which you are going to restore.
• Booting into Server environment - This option allows to restore the current machine from a Disk Safe located on a network drive or on the secondary physical disk. When you boot from Live CD Server or PXE Network Boot Server, then a full-fledged Backup Manager runs. In the Backup Manager Web Interface you will be able to add the Disk Safe to restore from.

Choose the option that fits your specific needs.

Test a Linux Bare-Metal Restore Before an Emergency Occurs

It is highly recommended to test a Bare-Metal Restore for each hardware platform before you need it. For example, most Data Centers have many different server platforms in use. Test the Bare-Metal Restore on each of the platforms.

Understand How Bare-Metal Restore Works

Make sure you understand how the Bare-Metal Restore works and how to initiate it. If you have an emergency, you would not want to start reading the User Guide or be waiting for technical support to respond.

Test Bare-Metal Restore on Each Configuration

Try the Disaster Recovery Boot on the different hardware flavors you use and make sure you have worked out any unexpected problems.

Both Boot Methods depend on booting a "silver bullet" Linux Kernel. The Linux Kernel works on 95% of hardware without any additional settings. Sometimes, as is common for any Linux Kernel, certain hardware requires you to disable APIC (noapic) and boot with special PCI Bus options for broken hardware adapters.

If you rent a server at a remote data center, they may have already worked through problems with broken hardware or Linux Device Drivers before you get your server. Do not assume that because you rent a Centos dedicated server that it is "vanilla" Centos. It is likely it may have some special boot options to get it to run on some problematic hardware, or a tweak for some init scripts to load a Device Driver that would not play properly. The Boot CD does not know about all of these tweaks and may need additional settings.

If our Boot CD does not support your hardware, you may need to install additional drivers supplied by your Linux hardware or software vendor when Linux is already booted from CD. Live CD is based on Debian, so you can use use Debian apt-get command to install prepackaged drivers or other software. Make sure you know how to connect to your network. Also, note special accommodations used for a particular NIC adapter such as forcing a switch port to full duplex to avoid SLOW transfers.

If you test each detail before a Disaster Recovery, the process will be much smoother.