RAID

1
RAID

• Rithy Chhay
• Shari Holstege
• CMSC 691X UNIX Systems Administration

2
What is RAID?

• Redundant Array of Inexpensive/Independent Disks
• RAID can improve availability and throughput
  (although actually reliability whether anything
  is broken suffers because of the larger number
  of disks)
• Data is stored on several disks instead of a
  single disk

3
RAID Levels
RAID Level RAID Level Failures Survived Data Disks Check Disks
0 Nonredundant 0 8 0
1 Mirrored 1 8 8
2 Memory-style ECC 1 8 4
3 Bit-interleaved parity 1 8 1
4 Block-interleaved parity 1 8 1
5 Block-interleaved distributed parity 1 8 1
6 PQ redundancy 2 8 2
4
RAID 0 Striping

• This level offers no redundancy no extra data
  is kept.
• The performance is the best of any level.
  Throughput is increased by striping data across
  several disks.

5
RAID 1 Mirroring

• Uses twice as many disks
• Whenever data is written to one disk, that data
  is also written to a redundant disk so that there
  are always two copies of the information
• When a disk fails, the system merely goes to its
  mirror for the data

6
RAID 3 Bit-Interleaved Parity

• Reads and writes go to all disks in a group, with
  one extra disk to hold the check information in
  case there is a failure.
• Parity is simply the sum of the data in all the
  disks modulo 2. Lost data can be reconstructed
  by examining the parity.
• Every access goes to all disks.

7
RAID 4 Block-Interleaved Parity

• Allows applications to do smaller accesses than
  RAID 3, allowing independent accesses to occur in
  parallel.
• Small, independent reads are easy simply read
  the data and then check for error detection.
• Writes are harder old data is read, new data is
  compared, and only those parity bits whose values
  change are updated.
• The parity disk becomes a bottleneck, since the
  parity disk must be updated on every write.

8
RAID 5 Block-Interleaved Distributed Parity

• This is a way to get rid of the bottleneck of
  RAID 4 distribute the parity information across
  all disks.

0
1
2
P0
0
1
2
P0
3
4
5
P1
3
4
P1
5
6
7
8
P2
6
P2
7
8
9
10
11
P3
P3
9
10
11
RAID 4
RAID 5
9
Higher Levels

• RAID 6 Adds a second parity scheme that is
  distributed across different drives and thus
  offers extremely high fault- and drive-failure
  tolerance
• RAID 7 Includes a real-time embedded operating
  system as a controller, caching via a high-speed
  bus, and other characteristics of a stand-alone
  computer.

10
Higher Levels (cont.)

• RAID 10 Offers an array of stripes in which
  each stripe is a RAID-1 array of drives. This
  offers higher performance than RAID-1 but at much
  higher cost.
• RAID 53 Offers an array of stripes in which
  each stripe is a RAID 3 array of disks. This
  offers higher performance than RAID 3 but at much
  higher cost.

11
RAID in Linux

• Linux offers built-in software RAID capabilities
• Advantages of Linux Software RAID
• Threaded rebuild process
• Fully kernel-based configuration
• Backgrounded array reconstruction
• Hot-swappable drive support
• Automatic CPU detection to use CPU Optimizations

12
Options for Software RAID

• You can create a software RAID in Linux using
  only one hard disk!
• You can create a software RAID in Linux using
  multiple drives.
• Specifying the type of RAID you wish to install
  on your systems, depends on how they are used
  refer back to RAID Levels.

13
Creating RAID Partitions

• Using Disk Druid, Select software RAID from the
  Filesystem Type
• Select the drive on which the RAID is to be
  created
• Enter the size of the RAID partition
• Choose other options as needed for your RAID

14
Software RAID Configuration

• Once RAID partitions have been created, select
  the Make RAID option on the Disk Druid main
  partitioning screen.
• Enter a mount point, select a filesystem, and
  choose your RAID Level.
• A spare partition can be specified for RAID 1 and
  RAID 5

15
RAID Drive Summary

• You have now created a software RAID in Linux.
• Try it at home!

16
Rules of Thumb

• When setting up a RAID with different sized hard
  drives, configure partitions on each drive to be
  the same size.
• Whenever possible, use the same hardware
  specifications for multiple hard drives
• Creating a software RAID on a single hard drive
  will slow performance because data must be
  written twice using only one head.

17
Do-It-At-Home RAID

• Build your own home RAID using
• Soyo KT-SY333 Dragon Ultra Motherboard
• AMD Athlon XP 2200 CPU
• 1024MB DDR PC2700 RAM
• (4) 160GB Maxtor DiamondMax DX540 Hard drives
• Multiple cooling fans
• Large Tower Case with a 400W power supply

18
Soyo KT-SY333 Motherboard

• Embedded Hipoint IDE-RAID chip, providing ATA-133
  IDE-RAID 0,1,01
• Setup both Hardware and Software RAIDs
  simultaneously!

19
Setting up your RAID

• Enter the BIOS set-up of your Soyo motherboard.
• Enable the desired hardware RAID Level.
• Install Linux on your machine with or without the
  software RAID option.
• Selecting hardware RAID Level 0 and software RAID
  Level 0 will provide the best level of
  performance.
• This setup provides for a fast 640GB RAID.

20
Additional Resources

• Patterson, David A. and John L. Hennessy.
  Computer Architecture A Quantitative Approach.
  San Francisco, CA Morgan Kaufmann Publishers,
  Inc., 1996.
• http//whatis.techtarget.com/definition/0,289893,s
  id9_gci214332,00.html

21
Additional Resources

• http//www.redhat.com/docs/manuals/linux/RHL-7.3-M
  anual/custom-guide/
• http//www.soyotek.com/products/proddesc.php?id46
  
• http//www.maxtor.com/
• http//www.amd.com/