RAID - DOC by fdjerue7eeu


?raid0 Overview

RAID is "Redundant Array of Independent Disk" stands for the
Chinese meaning is independent of redundant disk array. Redundant disk array
technology was born in 1987, from the University of California at Berkeley made.

RAID disk arrays (Redundant Array of Independent Disks)

Simply explained, is to drive through the N-Taiwan RAID Controller (sub-Hardware,
Software) combined into a single virtual hard disk to use large capacity, which is
characterized by N of drives to read faster and provide fault tolerance Fault Tolerant,
so RAID is as usual is not the main access to data Storage Backup Solution.

A basic concept in the RAID as EDAP (Extended Data Availability and Protection),
which emphasizes scalability and fault tolerance mechanisms, but also various
manufacturers such as: Mylex, IBM, HP, Compaq, Adaptec, Infortrend and other key
demands, including without need downtime can be handled under the following

RAID disk array support for automatic detection of failed hard;

RAID disk arrays support the reconstruction of the hard disk bad track information;

RAID disk array support for support do not have to stop the hard drive backup Hot

RAID disk array support for support to stop the hard drive does not replace Hot Swap;

RAID disk array support for the expanded hard drive capacity and so on.

Technical Specifications

(1) RAID specification Introduction

Redundant disk array technology developed originally intended to mix in place of
small, inexpensive disk large expensive disks to reduce the cost of mass data storage,
but also hope that the way information is redundant, so disk failure will not lead to the
access to data loss, and thus develop a certain level of data protection technologies,
and can raise the appropriate data transmission speed.

RAID has been in the past was destined to enjoy high-end server has SCSI hard drive
as a supporting technology for high-end applications. Recently, as technology
development and product costs continue to decline, IDE hard disk performance has
greatly enhanced the popularity of plus RAID chip, making RAID also gradually get
PC applications.

Why then is called redundant disk array then? Redundancy means that Chinese
surplus, redundant. The disk shows more than just a disk array, but a group of disks.
Then you should understand that it is to use the disk to handle duplicate data, making
data stability is improved.

(2) RAID works

How RAID data storage and high stability it? Let us look at how it works. RAID in
accordance with the realization of the principle of divided into different levels,
between different levels of work pattern is different. The entire structure is some disk
RAID structure on the disk combination to improve efficiency and reduce errors
purpose, not because of so many terms have been frightened, and their philosophy
behind it is simple. For illustration, the following diagram of each box represents a
disk block is called vertical or disk array, called horizontal bands.

(3) RAID specification

Consists mainly of RAID 0 ~ RAID 7 and a few specifications, and their different
focus, common standard the following categories:

RAID 0: No error control with block

To achieve RAID0 must have two or more hard drives, RAID0 achieved with a block
of data is not stored in a hard disk, but is divided into blocks of data stored in different
drives. Because the data in different drives, so the data throughput significantly
increased load on the drive more balanced. If you just need the data in a different
drive the best efficiency. It does not calculate checksum and easy to implement. The
disadvantage is that it does not have the data error control, if the data in a drive error,
even if other data on the correct disk also serves no purpose. It should not be used for
occasions that require high data stability. If the user images (including animation)
editor and other occasions, the transmission of larger more appropriate to use RAID0.
At the same time, RAID can improve data transfer rates, such as distribution of the
required documents to read in two hard disk, two hard drives can also read. Then read
the same original file is reduced to 1 / 2. At all levels of, RAID 0 speed is the fastest.
However, there is no redundancy RAID 0, if a disk (physical) damage, all of the data
are not available.

RAID 1: mirroring structure

raid1 structure for the use of such equipment for RAID1, RAID controller must be
able to read the two disks and two mirror disks for writing. By the following chart you
can see that there must be two drives. Mirror the structure because it is a problem in a
set, you can use the mirror to improve the system's fault tolerance. It is
relatively easy to design and implementation. Each time a disk is only allowed to
deliver a data transfer rate and data block that is a separate disk to read the same rate.
RAID1 is complete because validation, processing power, therefore the system has a
great influence, usually by the software RAID functionality, but this realization
heavier load on the server when the server will greatly affect the efficiency. When
your system needs a high level of reliability, such as statistical data, it is appropriate
to use RAID1. And RAID1 support "hot replacement", that is
the case of UPS replacement of the failed disk, replace the complete recovery as long
as the data from the mirrored disk can. When the main hard disk is damaged, the
mirror can replace the primary hard drive work. Equivalent to a backup disk image
hard disk, can be imagined, this hard drive model of security is very high, RAID 1 for
data security of RAID at all levels is the best. But the disk utilization is only 50%,
which is the lowest of all RAID levels.

RAID2: Hamming code with a checksum

Conceptually, RAID 2 RAID 3 with similar compartmentalization of both the
distribution of data on different hard disk, slice in bits or bytes. RAID 2, however, use
some coding techniques to provide error checking and recovery. This encoding
requires more disk storage inspection and recovery information, making the technical
implementation of more complex RAID 2. Therefore, rarely used in a commercial
environment. The figure on the left of each disk on each bit of data from a data-bit
computing are different Hamming checksum code can be saved on the disk to another
group, under the circumstances, see Fig. Because the characteristics of Hamming
codes, it can be an error in the data where the error correction to ensure the correct
output. It is a very high data transfer rate, ideal if you want to achieve the speed, it is
best to improve check code ECC code stored hard disk, the controller design for it
than RAID3, 4 or 5 should be simple. No free lunch, here too, to make use of
Hamming codes, have to pay the cost of data redundancy. Output data rate and drive
the group equal to the slowest.

RAID3: parity codes with parallel transmission

raid3 this verification code and RAID2 different, not only detecting errors correction.
Access to data once it is dealing with a band, this can increase the read and write
speed, it is like RAID 0 as a way to store in parallel a few, but the speed is not fast
RAID 0. Check code generated in the write data and save it in another disk. Need to
achieve the user must have three or more drives, write speed and read out rate is high,
because the parity bit less, so the computation time is relatively small. Using software
RAID control will be very difficult to achieve the controller is not easy. It is mainly
used for graphics (including animation) and so on the occasion demands a relatively
high throughput. Unlike RAID 2, RAID 3 uses a single block of disk storage of parity
information. If a disk failure, parity disks and other data disk can be re-generated data.
If the parity disk failure, did not affect the use of data. RAID 3 for large numbers of
continuous data transfer rate can provide very good, but for random data, parity disk
will become a write bottleneck. Use a separate check disk to protect data while not
mirroring the safe, but the hard drive has been greatly improved efficiency for the n-1.

RAID4: parity check codes with the structure of independent disks

raid4RAID4 and RAID3 much like the difference is, its access to data is based on data
blocks, that is conducted according to the disk each time a plate. It can be see in the
chart, RAID3 is a one bar, while a vertical bar RAID4 time. It features the RAID3
also Tingxiang, but in the failure to resume, it's difficult to be much larger
than RAID3, and controller design difficult to have a large number of, and access to
data the efficiency of the not very good.

RAID5: the independent distributed parity disk structure

raid5 from its diagram can be seen, it exists in all the parity bits on the disk, where the
p0 band on behalf of 0 parity value, the other meaning is the same. RAID5 efficient
read out, write the general efficiency, block access to the efficiency of the collective
good. Because parity code in a different disk, so improving the reliability, allowing a
single disk errors. RAID 5 is also based on the data parity bit to ensure data security,
but it is not a separate hard drive to store data parity bit, but the data section of the
parity bit stored in each disk interaction. In this way, any damage to a hard disk, can
check under the other places on the hard disk to reconstruct damaged data. Hard disk
utilization for the n-1. But it's parallel data transfer solution is not good,
and the controller design is also very difficult. Compared to RAID 3 and RAID 5,
RAID 3 important difference is that every once data transfer, to cover all of the array
disk. For RAID 5, the most of the data transmission is only one disk operation to
allow parallel operation. In RAID 5 in the "write losses", that is,
every write operation, will have four actual read / write operations, two of which read
the old data and parity information, the two write new data and parity information.
RAID-5, then the advantage of providing redundancy (to support a normal operation
even after the disc dropped), disk space utilization rate is higher (N-1 / N), read and
write speed (N-1 times). But off set, and operating efficiency has dropped

RAID6: the distribution with two parity bits stored in the structure of independent

raid6 name is long, but if you see diagram, you will immediately understand why,
note that the representative p0 0 parity with district value, representative of data
blocks A, pA parity value. It is an extension of RAID5 is mainly used for required
data must not mistake the situation. Of course, the introduction of a second parity
value, so they need N +2 disks, while the controller design becomes very complicated,
write speed is not good, used to calculate the parity value and validate data accuracy
of more time spent, resulting in a non-essential load. I think apart from the military
that no one can afford this kind of thing.

RAID7: Optimization of high-speed data transfer disk structure

RAID7 all the I / O transmission are synchronized, can be controlled separately, so
that the parallel system is improved, and improving the speed of the system to access
the data; each disk comes with high-speed buffer memory, real-time operating system
can use any real-time operation chip, to achieve real-time systems with different needs.
Allows the use of SNMP protocol to manage and monitor the designated area can
check independent transmission channels to improve efficiency. Can connect multiple
hosts, because adding high-speed buffer memory, when multiple users access the
system, the access time is almost close to 0. As a result of parallel structure, so data
access efficiency is improved. Note that it introduces a high-speed buffer memory, the
pros and cons, because once the system is powered down, in the data cache will be
lost and the UPS needed to work together. Of course, things so quickly, the price is
also very expensive.

RAID10: disk structure, high reliability and high efficiency

This structure is nothing but a band structure, plus a mirror structure, because the two
structures have their advantages and disadvantages, so you can complement each
other, to achieve both high speed and efficient they can be objective. We can combine
the two structures to understand the strengths and weaknesses of this new structure.
The new structure of high prices, poor scalability. Capacity is not used, but require
speed and error control of the database.

RAID53: Efficient data transfer disk structure

More to the back of the structure is a repetition of the previous structure and re-use,
this structure is RAID3 and bands with structure, so it faster, but also fault tolerance.
But the price is very high, not easy to achieve. This is yes because all the data must be
stored with two zones and by-bit approach, taking into account efficiency, asking that
it is not easy these disks Tongbu.

RAID0 +1:

RAID0 and RAID1 technology to combine the RAID0 +1. In addition to the
distribution of data across multiple disks, each disk has its own physical disk
mirroring to provide full redundancy, allowing a less disk failure, without affecting
data availability and fast read / write capability. Requires at least four hard drives can
Zuocheng RAID0 +1.

(3) JBOD mode

JBOD is usually known as Span. It is logically one after several physical disks
connected together to form one large logical disk. JBOD does not provide fault
tolerance, the capacity of the array is equal to the composition of the capacity of all
disks Span combined. JBOD, strictly speaking, not part of RAID range. But now
many IDE RAID control chip with a kind of model, JBOD hard drive capacity is a
simple superposition, but the system does not use parallel processing way, write data,
it is the first to write a hard drive, filled with the write article two hard drives ... ...

The most common practice is RAID0 RAID1 RAID5 and RAID10 because in most
cases, RAID5 contains RAID2-4 advantage, so the basic out of the market RAID2-4

Now, generally considered only for RAID RAID2-4 Development

(4) we can actually be put to the IDE RAID

The above is a description of RAID principle, and we are most concerned about is the
RAID Pcfans application. We daily use IDE hard drive, and very easy to get IDE
RAID cards and RAID chips integrated motherboard. So now we are most close to the
IDE RAID. Limited to the application level is very low, IDE RAID majority support
only RAID 0, RAID 1, RAID 0 +1, JBOD mode.

RAID applications

Program targeted at the beginning of SCSI hard disk RAID systems, the system costs
more expensive. 1993, HighPoint company introduced the first IDE-RAID controller
chip, to take advantage of relatively inexpensive IDE hard drives to form RAID
system, which greatly reduces the RAID "threshold." Since then,
individuals begin to focus on the technology, because the hard drive is the most
modern personal computers in the development of the "slow"
and the most lack of safety equipment, and user data stored in them often far exceeds
the price of the computer itself. In the case of relatively low cost, RAID technology
allows individual users to enjoy double the disk speed boost and higher data security,
personal computer market, IDE-RAID controller chip companies mainly from
HighPoint and Promise, There are also part of the company from AMI (Table 2).

For the individual user's IDE-RAID chip generally provides RAID 0,
RAID 1 and RAID 0 +1 (RAID 10) and other standard support for RAID, although
they technically can not be compared with commercial systems, but for ordinary users
provided to enhance the speed and security guarantees have been sufficient. With the
hard disk transfer rate interface, the continuous improvement, IDE-RAID Xinpian
also continuously upgrading, chip, chip Yijing the mainstream market, all support
ATA 100 standard and HighPoint's new HPT 372 chip and the latest
Promise PDC20276 chip, even have support for ATA 133 IDE hard drive standard.
Increased competition in the motherboard manufacturers, gradually increasing
requirements for PC users today, onboard RAID chip on the motherboard
manufacturers have been few, users can purchase RAID cards do not directly set up
their own disk array, feel the speed of disk winds.

Matrix RAID:

Matrix RAID so-called "Matrix RAID", is supported by the
Southbridge ICH6R a cheap disk redundancy, is a highly innovative economy RAID
solution. Matrix RAID technology principle is very simple and requires only two hard
drives can achieve a RAID 0 and RAID 1 disk array, and do not need to add
additional RAID controller, which is what we expect ordinary users. Matrix RAID
needs hardware layer and software layer can be achieved at the same time support,
hardware, now is the ICH6R Southbridge and higher order ICH6RW South Bridge,
the Intel Application Acclerator software and Windows operating systems are
supported on the software layer.

Matrix RAID is the principle of the capacity of each disk is divided into two parts
each (ie: a virtual hard drive into two sub-disk, then the total number of sub-Hard 4),
in which children with two virtual hard disk to create a RAID0 mode to improve
performance , while the other two virtual sub-disk is formed through RAID 1
mirroring backup to backup data. Matrix RAID mode in the data storage mode as
follows: two disk drives are used to create the first part of the RAID 0 array, is mainly
used to store the operating system, applications, and exchange files, this is the
beginning of the region because of the disk have a higher survival access speed,
Matrix RAID RAID 0 logical partitions will be placed on the hard disk front (outer)
of the reasons is that the module can require the best performance performance; and
two disk drives of the second part of the model used to create RAID1 mainly used to
store user personal files and data.

For example, two 120GB hard drives, can be composed of two 120GB hard drives
before the 60GB logical partition, and then the remaining two blocks, a 60GB 60GB
data backup partition. As the need to high-performance, do not require the application
of security, you can install the RAID 0 partition, while the need for security Backup
data can be installed in RAID 1 partition. In other words, users get the total hard disk
space is 180GB, and the traditional RAID 0 +1 compared to the benefits of capacity
use is very high, and the allocation of capacity will have more flexibility. In the event
of damage to the hard disk, RAID 0 partition data naturally can not be undone, but the
data partition RAID 1 but will be preserved.

Can be said that the use of Matrix RAID technology, we only need two hard drives
you can access to efficient data access while ensuring data security. This means that
ordinary users can also enjoy low-cost RAID 0 +1 application mode.


NV RAID is the RAID technology developed by nVidia, with the nForce series of
chipsets, the development of an ongoing basis. Compared to other RAID technology,
the latest of the nForce4-series chipsets, NV RAID has its own distinctive features,
mainly the following:

(1) interlaced RAID (Cross-Controller RAID): staggered, commonly known as hybrid
RAID RAID, that is, the SATA interface hard disk drives with IDE interfaces together
to form a RAID mode. Staggered series of RAID in the nForce3 250 chipset it has
emerged in the nForce 4 series chipset, the body of the function will be renewed and

(2) redundant hot backup feature: in the nForce 4 family of chipsets, supporting the
Serial ATA 2.0 hot swap functionality, users can use the process to replace the
damaged hard drive, and run the state in the re-establishment of a new image to
ensure the security of important data. Even more gratifying is that, nForce 4 of
nVIDIA RAID controller also allows the user to run in a RAID system to add a
redundant backup feature, without regard to what kind of RAID system uses pattern,
the user can provide the driver a "management tool "assign any
extra hard disk for RAID system hot backup. The heat allows multiple RAID
redundant disk system (such as a RAID 0 and a RAID1) share, one can also hold its
own RAID system, functions like high-end RAID system nowadays.

(3) a simple RAID mode migration: nForce 4 series chipset NV RAID module added
called "Morphing" the new feature, users only need to select the
conversion     after  the   RAID    mode,     then    the   implementation      of
"Morphing" operation, RAID Remove and patterns of work can
be reset automatically, without human intervention, ease of use has improved

RAID Fault:

Now select the IDE RAID card (IDE RAID card) to ensure data security, more and
more people, how to properly use the IDE RAID card is also learning. Here we have
to use HPT370A/372 control chip Rocket100 RAID card as an example to explain to
everyone to do common faults and techniques.

Installation Tips

Go first to a free PCI slot and the card inserted to the hard disk drive cable with the
card installed and connected, adapter card installed in your computer in the process,
you will see that the adapter has been successfully installation was system
identification. In the system boot, the BIOS will display the hard disk controller status
information, press CTRL + H to enter the settings menu structure is very clear, where
you can set the disk array: Two hard drives can choose to strip mode ( RAID 0) and
mirror mode (RAID 1), there are three hard drives can only be selected if expansion
or stripe pattern across, and 4 can be selected across the model, strip or stripe pattern
combined with the mirror mode (RAID 0 +1), If the choice of RAID 1 hard disk must
be synchronized.

Troubleshooting Common Installation

When Rocket100 RAID card is identified, the board began examining the connection
device BIOS. Please note that the device appears on the screen, if all the connected
device is correctly scanned, the description of equipment has been properly connected
and are system identification, and then after you installed the drivers you can use
RAID function. And if some of them have not been identified device, you can open
the case, will the connected device is plugged in the power cord firmly, when
necessary, change a plug to try; the connected device data cable is correctly connected
and fit snugly necessary for a data line to try; If a data line connected with two
devices, make sure that both devices are master-slave jumper settings conflicts (one
data line of the two devices must be a primary one from).

The choice of hard disk capacity

Taking into account the system's operational performance and disk
utilization, we recommend that you better make the same hard drive capacity.
However, if you must use disks of different capacity, to note that the capacity of the
array by the array to the smallest-capacity hard drive decisions, such as in the three
disk RAID 0 array consisting of a total capacity equal to the minimum disk capacity 3
times. In RAID 1 array, the target disk capacity can not be less than the source disk
capacity. The total capacity of the array is equivalent to the smallest disk capacity.
JBOD is an exception, however, two or more different capacity hard drives can be
combined to form a single logical disk.

BIOS Settings Guide

IDE RAID card is Plug and Play devices, so you do not have to change the system
CMOS. That will automatically interrupt and port address. In the CMOS settings, all
equipment is in none or unstalled state, IDE RAID cards can be set to start the card; or
SCSI device boot sequence of notes of the first one, RAID card can also be set to start
the card.

RAID damaged how to do?
If a RAID level to be destroyed, can be checked the following: First turn off the
computer's Dianyuan, check the hard disk power of Lian Jie, and data cable
is hard He RAID 控制 Ka Yu Lian Jie Zai Zhong Xin normal boot; If the above
action does not resolve the problem may be hard drive problems, for RAID 1 and
RAID 0 +1, you can use a new hard drive will have been damaged hard disk data
backup, all data are not lost. For RAID 0 and JBOD, you must first delete the original
RAID levels, further RAID created, but note that this action will drive all data is lost,
so, on the RAID 0 and JBOD, please be sure to regularly back up data.

RAID disk array data recovery guide

RAID disk array device, during use, often encounter some common faults, which also
makes RAID in the mass storage space to us the application, but also brings a lot of
data is difficult to assess the risk. This article will focus on common faults, and RAID
related approach.

Note RAID failure

1, data is lost, the user must not carry out any operation on the hard disk, remove the
hard disk in order to use imaging software to the hard disk made of each block image
file can also be handed over to professional data recovery center.

2, Do not Rebuild Raid card operation, otherwise it will increase the difficulty of
recovering data.

3, marking a good hard drive in Raid card above order.

4, if there are problems, you can call the professional data recovery center looking for
professional engineers consultation phone consultation, should not attempt to repair
itself, unless you are sure they have sufficient skills and experience to handle the data

Common Faults and recoverability of Raid

1, software failure:

a. RAID disk array, a sudden power outage caused the loss of card data recovery

b. Reconfigure the RAID array information, resulting in data loss recovery.

c. If the disk sequence error, will cause the system to not recognize the data.
d. Deletions, incorrect formatting, partition error, error cloning, file decryption, data
on life toxic damage, etc. back to work.

2, hardware damage:

a. usually have a few hard drives raid, in which a piece of the hard disk are damaged,
the data will not be able to read.

b. raid appears bad sectors, resulting in data loss, this recovery success rate is
relatively large.

c. If two or more simultaneous disk damage, restoration work is very complex and
relatively low success rate.

Case 1

February 2009, a machinery manufacturing group had the only real network data
recovery call center services, telephone, a server of the company, I do not know why
suddenly can not start, the data can not be read. The server is using RAID 5 works.
The only real data recovery center engineers determine the user's initial
statement is the order of the hard drive problem. This is a 5 75G array of servers, to
do this piece of hard data, at least 400G space.

Recovery steps are as follows:

* Carried out on each hard drive image, do not operate in the original disc, so as to
avoid secondary damage.

* By calculating the first judged the order of the hard disk, the data will only order,
otherwise data recovery is impossible with this will come out of.

* After the judge finished it, these five image files to be restored. Three days later,
retrieve the data successfully.

Note raid data recovery

A basic concept in the RAID as EDAP (Extended Data Availability and Protection),
which emphasizes scalability and fault tolerance mechanisms, but also various
manufacturers such as: Mylex, IBM, HP, Compaq, Adaptec, Infortrend and other key
demands, including without need downtime can be handled under the following

RAID disk array support to automatically detect hard disk failure.
RAID disk arrays support the reconstruction of the hard disk bad track information.

RAID disk array support for support do not have to stop the hard drive backup Hot

RAID disk array support for support to stop the hard drive does not replace Hot Swap.

Hard disk arrays to support RAID Capacity Expansion, etc..

Once the RAID array fails, the hardware service provider only to the customer to
re-initialize or REBUILD, so that the customer data is irretrievable. After failure to
initialize the array as long as the operation does not have the opportunity to restore the
fault RAID disk array data.


MMORPG in the definition: a group of players in a region of large-scale operations
are called Raid.

For example: a copy of World of Warcraft.

Burning Crusade more than 40 people before the raid, 25 people after Burning
Crusade raid

MMORPG games by a group of players form a team (team format is not RAID) in an
area of organized large-scale operations, from the D & D game is first

Neither Tezhi is in union activities, not a single copy of the play Tezhi, more does not
specify that the activities of World of Warcraft.

MMORPG game as long as a team to meet the conditions of the form of two
organized a purpose both for RAID 3
D 3

To top