Introduction; Raid Basics; Raid Overview; Software Raid In Linux - HP xw3400 User Manual

Introduction

This document provides an overview of software RAID solutions for HP workstations running the Linux
operating system. It assumes a basic understanding of computer hardware, filesystems, and the Linux
operating system. It covers the various RAID levels, the differences between hardware RAID and
software RAID, steps to configure software RAID on Linux workstations, and information regarding
disk failure, recovery, and other information pertaining to running software RAID on a Linux system.
Please note that while BIOS messages may refer to hardware RAID, hardware RAID is not supported
by HP Linux workstations.

RAID Basics

RAID Overview

Standing for Redundant Array of Independent (or Inexpensive) Disks, RAID is a technology that allows
a computer system to increase capacity, performance, and/or reliability of storage solutions using
multiple physical hard disks. There are several kinds of RAID; each RAID configuration is assigned to
its own RAID "level." There are four basic RAID levels: 0, 1, 5, and 10. Over time, a number of other
RAID levels have been defined as well; RAID levels 1-5 were those defined by the original
RAID paper. These levels have been implemented with dedicated RAID disk controllers, hardware
RAID, and with kernel-level code controlling RAID operations, software RAID.
RAID was originally developed as a means to provide a degree of redundancy for disk data. It first
became popular in servers with SCSI drives where data reliability was key. Today, RAID can be
found across a number of different systems, including workstations, and a variety of drive types,
including IDE, SATA, and SAS, as well as SCSI.

Software RAID in Linux

The Linux kernel, as of revision 2.4, offers integrated software RAID without the need for additional
hardware disk controllers or kernel patches. All that is required is multiple hard disks and a small
amount of setup. Unlike most hardware RAID solutions, software RAID can be used with IDE disks as
well as SATA, SAS, and SCSI.
Software RAID does not come entirely without cost: in order to manage the disks, break up data as
necessary, and manage parity data, the CPU must take on some extra loading. It has been found that
heavily disk-intensive workloads result in roughly double the CPU overhead (for example, from 15% to
30%) when software RAID is in use. For most applications, this overhead is easily handled by excess
headroom in the processors, but for some applications where disk performance and CPU
performance are very well balanced and already near-bottleneck, this additional CPU overhead may
become burdensome and hardware RAID may look attractive. Hardware RAID also offers advantages
due to its large hardware cache, and the capability for better scheduling of operations in parallel;
however, software RAID offers much more flexibility in terms of disk and disk controller setup.
Additionally, hardware RAID requires that a failed RAID controller must be replaced with an identical
model in order to avoid data loss, whereas software RAID imposes no such requirements.
3