RAIDFrom Wikipedia, the free encyclopediaThis article is about the data storage technology. For other uses, see Raid (disambiguation).RAID (redundant array of independent disks, originally redundant array of inexpensive disks) is astorage technology that combines multiple disk drive components into a logical unit. Data is distributedacross the drives in one of several ways called "RAID levels", depending on what level of redundancy andperformance (via parallel communication) is required. In October 1986, the IBM S/38 announced"checksum". Checksum was an implementation of RAID-5. The implementation was in the operatingsystem and was software only and had a minimum of 10% overhead. The S/38 "scatter loaded" all datafor performance. The downside was the loss of any single disk required a total system restore for alldisks. Under checksum, when a disk failed, the system halted and was then shutdown. Undermaintenance, the bad disk was replaced and then a parity-bit disk recovery was run. The system wasrestarted using a recovery procedure similar to the one run after a power failure. While difficult, therecovery from a drive failure was much shorter and easier than without checksum.RAID is an example of storage virtualization and was first defined by David Patterson, Garth A. Gibson,and Randy Katz at the University of California, Berkeley in 1987. Marketers representing industryRAID manufacturers later attempted to reinvent the term to describe a redundant array of independentdisks as a means of disassociating a low-cost expectation from RAID technology.RAID is now used as an umbrella term for computer data storage schemes that can divide and replicatedata among multiple physical drives. The physical drives are said to be "in a RAID", however the morecommon, incorrect parlance is to say that they are "in a RAID array". The array can then be accessedby the operating system as one single drive. The different schemes or architectures are named by theword RAID followed by a number (e.g., RAID 0, RAID 1). Each scheme provides a different balancebetween three key goals: resiliency, performance, and capacity.Contents [hide]1 Standard levels2 Nested (hybrid) RAID3 RAID parity4 RAID 6 replacing RAID 5 in enterprise environments5 RAID 10 versus RAID 5 in relational databases6 New RAID classification7 Non-standard levels
8 Data backup9 Implementations9.1 Software-based RAID9.1.1 Volume manager support9.1.2 File system support9.1.3 Other support9.2 Hardware-based RAID9.3 Firmware/driver-based RAID9.4 Hot spares9.5 Data scrubbing / Patrol read10 Reliability terms11 Problems with RAID11.1 Correlated failures11.2 Atomicity11.3 Write cache reliability11.4 Equipment compatibility11.5 Data recovery in the event of a failed array11.6 Drive error recovery algorithms11.7 Recovery time is increasing11.8 Operator skills, correct operation11.8.1 Hardware labeling issues11.9 Other problems12 History13 Non-RAID drive architectures14 See also
15 References16 External linksStandard levelsMain article: Standard RAID levelsA number of standard schemes have evolved which are referred to as levels. There were five RAID levelsoriginally conceived, but many more variations have evolved, notably several nested levels and manynon-standard levels (mostly proprietary). RAID levels and their associated data formats are standardisedby the Storage Networking Industry Association (SNIA) in the Common RAID Disk Drive Format (DDF)standard.Following is a brief textual summary of the most commonly used RAID levels.RAID 0 (block-level striping without parity or mirroring) has no (or zero) redundancy. It providesimproved performance and additional storage but no fault tolerance. Hence simple stripe sets arenormally referred to as RAID 0. Any drive failure destroys the array, and the likelihood of failureincreases with more drives in the array (at a minimum, potential for catastrophic data loss is double thatof isolated drives without RAID). A single drive failure destroys the entire array because when data iswritten to a RAID 0 volume, the data is broken into fragments called blocks. The number of blocks isdictated by the stripe size, which is a configuration parameter of the array. The blocks are written totheir respective drives simultaneously on the same sector. This allows smaller sections of the entirechunk of data to be read off each drive in parallel, increasing bandwidth. RAID 0 does not implementerror checking, so any error is uncorrectable. More drives in the array means higher bandwidth, butgreater risk of data loss.In RAID 1 (mirroring without parity or striping), data is written identically to two drives, therebyproducing a "mirrored set"; the read request is serviced by either of the two drives containing therequested data, whichever one involves least seek time plus rotational latency. Similarly, a write requestupdates the strips of both drives. The write performance depends on the slower of the two writes (i.e.,the one that involves larger seek time and rotational latency); at least two drives are required toconstitute such an array. While more constituent drives may be employed, many implementations dealwith a maximum of only two; of course, it might be possible to use such a limited level 1 RAID itself as aconstituent of a level 1 RAID, effectively masking the limitation. The array continues tooperate as long as at least one drive is functioning. With appropriate operating system support, therecan be increased read performance, and only a minimal write performance reduction; implementingRAID 1 with a separate controller for each drive in order to perform simultaneous reads (and writes) issometimes called "multiplexing" (or "duplexing" when there are only two drives).In RAID 10 (mirroring and striping), data is written in stripes across the primary disks and then mirroredto the secondary disks. A typical RAID 10 configuration consists of four drives. Two for striping and two
for mirroring. A RAID 10 configuration takes the best concepts of RAID 0 and RAID 1 and combines themto provide better performance along with the reliability of parity without actually having parity as withRAID 5 and RAID 6. RAID 10 is often referred to as RAID 1+0 (mirrored+striped).In RAID 2 (bit-level striping with dedicated Hamming-code parity), all disk spindle rotation issynchronized, and data is striped such that each sequential bit is on a different drive. Hamming-codeparity is calculated across corresponding bits and stored on at least one parity drive.In RAID 3 (byte-level striping with dedicated parity), all disk spindle rotation is synchronized, and data isstriped so each sequential byte is on a different drive. Parity is calculated across corresponding bytesand stored on a dedicated parity drive.RAID 4 (block-level striping with dedicated parity) is identical to RAID 5 (see below), but confines allparity data to a single drive. In this setup, files may be distributed between multiple drives. Each driveoperates independently, allowing I/O requests to be performed in parallel. However, the use of adedicated parity drive could create a performance bottleneck; because the parity data must be writtento a single, dedicated parity drive for each block of non-parity data, the overall write performance maydepend a great deal on the performance of this parity drive.RAID 5 (block-level striping with distributed parity) distributes parity along with the data and requires alldrives but one to be present to operate; the array is not destroyed by a single drive failure. Upon drivefailure, any subsequent reads can be calculated from the distributed parity such that the drive failure ismasked from the end user. However, a single drive failure results in reduced performance of the entirearray until the failed drive has been replaced and the associated data rebuilt. Additionally, there is thepotentially disastrous RAID 5 write hole. RAID 5 requires at least three disks.RAID 6 (block-level striping with double distributed parity) provides fault tolerance of two drive failures;the array continues to operate with up to two failed drives. This makes larger RAID groups morepractical, especially for high-availability systems. This becomes increasingly important as large-capacitydrives lengthen the time needed to recover from the failure of a single drive. Single-parity RAID levelsare as vulnerable to data loss as a RAID 0 array until the failed drive is replaced and its data rebuilt; thelarger the drive, the longer the rebuild takes. Double parity gives additional time to rebuild the arraywithout the data being at risk if a single additional drive fails before the rebuild is complete. Like RAID 5,a single drive failure results in reduced performance of the entire array until the failed drive has beenreplaced and the associated data rebuilt.The following table provides an overview of the most important parameters of standard RAID levels. Ineach case:Array space efficiency is given as an expression in terms of the number of drives, ; this expressiondesignates a value between 0 and 1, representing the fraction of the sum of the drives capacities that isavailable for use. For example, if three drives are arranged in RAID 3, this gives an array space efficiencyof (approximately 66%); thus, if each drive in this example has a capacity of 250 GB, then the array has atotal capacity of 750 GB but the capacity that is usable for data storage is only 500 GB.
Array failure rate is given as an expression in terms of the number of drives, , and the drive failure rate,(which is assumed to be identical and independent for each drive). For example, if each of three driveshas a failure rate of 5% over the next 3 years, and these drives are arranged in RAID 3, then this gives anarray failure rate of over the next 3 years.Level Description Minimum # of drives** Space Efficiency Fault Tolerance Array Failure Rate*** Read Benefit Write Benefit ImageRAID 0 Block-level striping without parity or mirroring. 2 1 0 (none) 1−(1−r)n nX nXRAID 1 Mirroring without parity or striping. 2 1/n n−1 drives rn nX 1XRAID 2 Bit-level striping with dedicated Hamming-code parity. 3 1 − 1/n ⋅ log2(n-1) RAID 2can recover from 1 drive failure or repair corrupt data or parity when a corrupted bits correspondingdata and parity are good. variablevariablevariableRAID 3 Byte-level striping with dedicated parity. 3 1 − 1/n 1 drive n(n−1)r2 (n−1)X (n−1)X*RAID 4 Block-level striping with dedicated parity. 3 1 − 1/n 1 drive n(n−1)r2 (n−1)X (n−1)X*RAID 5 Block-level striping with distributed parity. 3 1 − 1/n 1 drive n(n−1)r2 (n−1)X* (n−1)X*RAID 6 Block-level striping with double distributed parity. 4 1 − 2/n 2 drives n(n-1)(n-2)r3 (n−2)X* (n−2)X** Assumes hardware is fast enough to support; ** Assumes a nondegenerate minimum number ofdrives; *** Assumes independent, identical rate of failure amongst drivesNested (hybrid) RAIDMain article: Nested RAID levelsIn what was originally termed hybrid RAID, many storage controllers allow RAID levels to be nested.The elements of a RAID may be either individual drives or RAIDs themselves. However, if a RAID is itselfan element of a larger RAID, it is unusual for its elements to be themselves RAIDs.As there is no basic RAID level numbered larger than 9, nested RAIDs are usually clearly described byattaching the numbers indicating the RAID levels, sometimes with a "+" in between. The order of thedigits in a nested RAID designation is the order in which the nested array is built: For a RAID 1+0, drivesare first combined into multiple level 1 RAIDs that are themselves treated as single drives to becombined into a single RAID 0; the reverse structure is also possible (RAID 0+1).
The final RAID is known as the top array. When the top array is a RAID 0 (such as in RAID 1+0 and RAID5+0), most vendors omit the "+" (yielding RAID 10 and RAID 50, respectively).RAID 0+1: striped sets in a mirrored set (minimum four drives; even number of drives) provides faulttolerance and improved performance but increases complexity.The key difference from RAID 1+0 is that RAID 0+1 creates a second striped set to mirror a primarystriped set. The array continues to operate with one or more drives failed in the same mirror set, but ifdrives fail on both sides of the mirror the data on the RAID system is lost.RAID 1+0: (a.k.a. RAID 10) mirrored sets in a striped set (minimum four drives; even number of drives)provides fault tolerance and improved performance but increases complexity.The key difference from RAID 0+1 is that RAID 1+0 creates a striped set from a series of mirrored drives.The array can sustain multiple drive losses so long as no mirror loses all its drives.RAID 5+3: mirrored striped set with distributed parity (some manufacturers label this as RAID 53).Whether an array runs as RAID 0+1 or RAID 1+0 in practice is often determined by the evolution of thestorage system. A RAID controller might support upgrading a RAID 1 array to a RAID 1+0 array on the fly,but require a lengthy off-line rebuild to upgrade from RAID 1 to RAID 0+1. With nested arrays,sometimes the path of least disruption prevails over achieving the preferred configuration.[citationneeded]RAID parityFurther information: Parity bitMany RAID levels employ an error protection scheme called "parity", a widely used method ininformation technology to provide fault tolerance in a given set of data. Most use the simple XOR paritydescribed in this section, but RAID 6 uses two separate parities based respectively on addition andmultiplication in a particular Galois Field or Reed-Solomon error correction.In Boolean logic, there is an operation called exclusive or (XOR), meaning "one or the other, but notboth," that is:0 XOR 0 = 00 XOR 1 = 11 XOR 0 = 11 XOR 1 = 0
The XOR operator is central to how parity data is created and used within an array. It is used both forthe protection of data, as well as for the recovery of missing data.As an example, consider a simple RAID made up of 6 drives (4 for data, 1 for parity, and 1 for use as ahot spare), where each drive has only a single byte worth of storage (a - represents a bit, the value ofwhich doesnt matter at this point in the discussion):Drive #1: -------- (Data)Drive #2: -------- (Data)Drive #3: -------- (Data)Drive #4: -------- (Data)Drive #5: -------- (Hot Spare)Drive #6: -------- (Parity)Suppose the following data is written to the drives:Drive #1: 00101010 (Data)Drive #2: 10001110 (Data)Drive #3: 11110111 (Data)Drive #4: 10110101 (Data)Drive #5: -------- (Hot Spare)Drive #6: -------- (Parity)Every time data is written to the data drives, a parity value must be calculated in order for the array tobe able to recover in the event of a failure. To calculate the parity for this RAID, a bitwise XOR of eachdrives data is calculated as follows, the result of which is the parity data:00101010 XOR 10001110 XOR 11110111 XOR 10110101 = 11100110The parity data 11100110 is then written to the dedicated parity drive:Drive #1: 00101010 (Data)Drive #2: 10001110 (Data)Drive #3: 11110111 (Data)Drive #4: 10110101 (Data)
Drive #5: -------- (Hot Spare)Drive #6: 11100110 (Parity)Suppose Drive #3 fails. In order to restore the contents of Drive #3, the same XOR calculation isperformed against the data of all the remaining data drives and data on the parity drive (11100110)which was stored in Drive #6:00101010 XOR 10001110 XOR 11100110 XOR 10110101 = 11110111The XOR operation will yield the missing data. With the complete contents of Drive #3 recovered, thedata is written to the hot spare, which then acts as a member of the array and allows the group as awhole to continue operating.Drive #1: 00101010 (Data)Drive #2: 10001110 (Data)Drive #3: --Dead-- (Data)Drive #4: 10110101 (Data)Drive #5: 11110111 (Hot Spare)Drive #6: 11100110 (Parity)At this point the failed drive has to be replaced with a working one of the same size. Depending on theimplementation, the new drive becomes a new hot spare, and the old hot spare drive continues to actas a data drive of the array, or (as illustrated below) the original hot spares contents are automaticallycopied to the new drive by the array controller, allowing the original hot spare to return to its originalpurpose. The resulting array is identical to its pre-failure state:Drive #1: 00101010 (Data)Drive #2: 10001110 (Data)Drive #3: 11110111 (Data)Drive #4: 10110101 (Data)Drive #5: -------- (Hot Spare)Drive #6: 11100110 (Parity)This same basic XOR principle applies to parity within RAID groups regardless of capacity or number ofdrives. As long as there are enough drives present to allow for an XOR calculation to take place, paritycan be used to recover data from any single drive failure. (A minimum of three drives must be present in
order for parity to be used for fault tolerance, because the XOR operator requires two operands, and aplace to store the result).RAID 6 replacing RAID 5 in enterprise environmentsModern large drive capacities and the large RAID arrays used in modern servers create two problems(discussed below in Problems with RAID). First, in almost all arrays the drives are fitted at the time ofmanufacture and will therefore wear at similar rates and times. Therefore, the times of failure forindividual drives correlate more closely than they should for a truly random event. Second, it takes timeto replace the faulty drive and to rebuild the array.Rebuilding a RAID 5 array after a failure will add additional stress to all of the working drives becauseevery area on every disc marked as being "in use" must be read to rebuild the redundancy that has beenlost. If drives are close to failure, the stress of rebuilding the array can be enough to cause another driveto fail before the rebuild has been finished, and even more so if the server is still accessing the drives toprovide data to clients, users, applications, etc. It is during this rebuild of the "missing" drive that theentire raid array is at risk of a catastrophic failure. The rebuild of an array on a busy and large system cantake hours and sometimes days and therefore it is not surprising that when systems need to behighly available and highly reliable or fault tolerant RAID 6 is chosen.With a RAID 6 array using drives from multiple sources and manufacturers it is possible to mitigate mostof the problems associated with RAID 5. The larger the drive capacities and the larger the array size, themore important it becomes to choose RAID 6 instead of RAID 5.A disadvantage of RAID 6 is extra cost because two redundant drives are required. In small arrays, thiscan add significantly to the production cost and also to the ongoing cost because of the additionalpower consumption and additional physical space required. RAID 6 is a relatively new technologycompared to RAID 5, and therefore the hardware is more expensive to purchase and drivers will belimited to a smaller range of operating systems. In software implementations of RAID 6, the algorithmsrequire more CPU time when compared to RAID 5, because the algorithms are more complex and thereis more data to be processed. Therefore, RAID 6 in software implementations may require morepowerful CPUs than RAID 5.RAID 6 also suffers a greater write performance penalty than RAID 5. For small (non-full stripe in size)write operations, which are the dominant size in transaction processing systems, the spindle operationoverhead is 50% greater and latency will be slightly higher than with RAID 5. Providing the same writeperformance as a RAID 5 array requires that a RAID 6 array be built of approximately 50% more spindlesand this impacts the cost of performance.RAID 10 versus RAID 5 in relational databases
This section may contain original research. Please improve it by verifying the claims made and addingreferences. Statements consisting only of original research may be removed. (February 2011)A common opinion (and one which serves to illustrate the dynamics of proper RAID deployment) is thatRAID 10 is inherently better for relational databases than RAID 5, because RAID 5 requires therecalculation and redistribution of parity data on a per-write basis.While this may have been a hurdle in past RAID 5 implementations, the task of parity recalculation andredistribution within modern storage area network (SAN) appliances is performed as a back-end processtransparent to the host, not as an in-line process which competes with existing I/O. (i.e. the RAIDcontroller handles this as a housekeeping task to be performed during a particular spindles idletimeslices, so as not to disrupt any pending I/O from the host.) The "write penalty" inherent to RAID 5has been effectively masked since the late 1990s by a combination of improved controller design, largeramounts of cache, and faster drives. The effect of a write penalty when using RAID 5 is mostly a concernwhen the workload cannot be de-staged efficiently from the SAN controllers write cache.SAN appliances generally service multiple hosts that compete both for controller cache, potential SSDcache, and spindle time. In enterprise-level SAN hardware, any writes which are generated by the hostare simply stored in a small, mirrored NVRAM cache, acknowledged immediately, and later physicallywritten when the controller sees fit to do so from an efficiency standpoint. From thehosts perspective, an individual write to a RAID 10 volume is no faster than an individual write to a RAID5 volume, both are acknowledged immediately, and serviced on the back-end.The choice between RAID 10 and RAID 5 for the purpose of housing a relational database depends upona number of factors (spindle availability, cost, business risk, etc.) but, from a performance standpoint, itdepends mostly on the type of I/O expected for a particular database application. For databases that areexpected to be exclusively or strongly read-biased, RAID 10 is often chosen because it offers a slightspeed improvement over RAID 5 on sustained reads and sustained randomized writes. If a database isexpected to be strongly write-biased, RAID 5 becomes the more attractive option, since RAID 5 does notsuffer from the same write handicap inherent in RAID 10; all spindles in a RAID 5 can be utilized to writesimultaneously, whereas only half the members of a RAID 10 can be used. However, for reasons similarto what has eliminated the "read penalty" in RAID 5, the "write penalty" of RAID 10 has been largelymasked by improvements in controller cache efficiency and drive throughput.What causes RAID 5 to be slightly slower than RAID 10 on sustained reads is the fact that RAID 5 hasparity data interleaved within normal data. For every read pass in RAID 5, there is a probability that aread head may need to traverse a region of parity data. The cumulative effect of this is a slightperformance drop compared to RAID 10, which does not use parity, and therefore never encounters acircumstance where data underneath a head is of no use. For the vast majority of situations, however,most relational databases housed on RAID 10 perform equally well in RAID 5. The strengths andweaknesses of each type only become an issue in atypical deployments, or deployments onovercommitted hardware. Often, any measurable differences between the two formats are masked by
structural deficiencies at the host layer, such as poor database maintenance, or sub-optimal I/Oconfiguration settings.There are, however, other considerations which must be taken into account other than simply thoseregarding performance. RAID 5 and other non-mirror-based arrays offer a lower degree of resiliencythan RAID 10 by virtue of RAID 10s mirroring strategy. In a RAID 10, I/O can continue even in spite ofmultiple drive failures. By comparison, in a RAID 5 array, any failure involving more than one driverenders the array itself unusable by virtue of parity recalculation being impossible to perform. Thus,RAID 10 is frequently favored because it provides the lowest level of risk.Additionally, the time required to rebuild data on a hot spare in a RAID 10 is significantly less than in aRAID 5, because all the remaining spindles in a RAID 5 rebuild must participate in the process, whereasonly the hot spare and one surviving member of the broken mirror are required in a RAID 10. Thus, incomparison to a RAID 5, a RAID 10 has a smaller window of opportunity during which a second drivefailure could cause array failure.Modern SAN design largely masks any performance hit while a RAID is in a degraded state, by virtue ofbeing able to perform rebuild operations both in-band or out-of-band with respect to existing I/O traffic.Given the rare nature of drive failures in general, and the exceedingly low probability of multipleconcurrent drive failures occurring within the same RAID, the choice of RAID 5 over RAID 10 often comesdown to the preference of the storage administrator, particularly when weighed against other factorssuch as cost, throughput requirements, and physical spindle availability.In short, the choice between RAID 5 and RAID 10 involves a complicated mixture of factors. There is noone-size-fits-all solution, as the choice of one over the other must be dictated by everything from theI/O characteristics of the database, to business risk, to worst case degraded-state throughput, to thenumber and type of drives present in the array itself. Over the course of the life of a database, one mayeven see situations where RAID 5 is initially favored, but RAID 10 slowly becomes the better choice, andvice versa.New RAID classificationIn 1996, the RAID Advisory Board introduced an improved classification of RAID systems.[citationneeded] It divides RAID into three types:Failure-resistant (systems that protect against loss of data due to drive failure).Failure-tolerant (systems that protect against loss of data access due to failure of any singlecomponent).Disaster-tolerant (systems that consist of two or more independent zones, either of which providesaccess to stored data).
The original "Berkeley" RAID classifications are still kept as an important historical reference point andalso to recognize that RAID levels 0–6 successfully define all known data mapping and protectionschemes for disk-based storage systems. Unfortunately, the original classification caused someconfusion due to the assumption that higher RAID levels imply higher redundancy and performance; thisconfusion has been exploited by RAID system manufacturers, and it has given birth to the products withsuch names as RAID-7, RAID-10, RAID-30, RAID-S, etc. Consequently, the new classification describes thedata availability characteristics of a RAID system, leaving the details of its implementation to systemmanufacturers.Failure-resistant disk systems (FRDS) (meets a minimum of criteria 1–6)Protection against data loss and loss of access to data due to drive failureReconstruction of failed drive content to a replacement driveProtection against data loss due to a "write hole"Protection against data loss due to host and host I/O bus failureProtection against data loss due to replaceable unit failureReplaceable unit monitoring and failure indicationFailure-tolerant disk systems (FTDS) (meets a minimum of criteria 1–15)Disk automatic swap and hot swapProtection against data loss due to cache failureProtection against data loss due to external power failureProtection against data loss due to a temperature out of operating rangeReplaceable unit and environmental failure warningProtection against loss of access to data due to device channel failureProtection against loss of access to data due to controller module failureProtection against loss of access to data due to cache failureProtection against loss of access to data due to power supply failureDisaster-tolerant disk systems (DTDS) (meets a minimum of criteria 1–21)Protection against loss of access to data due to host and host I/O bus failureProtection against loss of access to data due to external power failure
Protection against loss of access to data due to component replacementProtection against loss of data and loss of access to data due to multiple drive failuresProtection against loss of access to data due to zone failureLong-distance protection against loss of data due to zone failureNon-standard levelsMain article: Non-standard RAID levelsMany configurations other than the basic numbered RAID levels are possible, and many companies,organizations, and groups have created their own non-standard configurations, in many cases designedto meet the specialised needs of a small niche group. Most of these non-standard RAID levels areproprietary.Himperia is using RAID 50EE in ZStore 3212L product. This is a RAID 0 of two pools with RAID 5EE(7+1+1). It is tolerant up to 2 disks failures at the same time, and up to 4 disks failures in degrade mode.Reconstruction time is set to a minimum thanks to RAID 5EE. And performance is increased, thanks toRAID 0.Storage Computer Corporation used to call a cached version of RAID 3 and 4, RAID 7. Storage ComputerCorporation is now defunct.EMC Corporation used to offer RAID S as an alternative to RAID 5 on their Symmetrix systems. Theirlatest generations of Symmetrix, the DMX and the V-Max series, do not support RAID S (instead theysupport RAID 1, RAID 5 and RAID 6.)The ZFS filesystem, available in Solaris, OpenSolaris and FreeBSD, offers RAID-Z, which solves RAID 5swrite hole problem.Hewlett-Packards Advanced Data Guarding (ADG) is a form of RAID 6.NetApps Data ONTAP uses RAID-DP (also referred to as "double", "dual", or "diagonal" parity), is a formof RAID 6, but unlike many RAID 6 implementations, it does not use distributed parity as in RAID 5.Instead, two unique parity drives with separate parity calculations are used. This is a modification ofRAID 4 with an extra parity drive.Accusys Triple Parity (RAID TP) implements three independent parities by extending RAID 6 algorithmson its FC-SATA and SCSI-SATA RAID controllers to tolerate a failure of 3 drives.Linux MD RAID10 (RAID 10) implements a general RAID driver that defaults to a standard RAID 1 with 2drives, and a standard RAID 1+0 with four drives, but can have any number of drives, including oddnumbers. MD RAID 10 can run striped and mirrored, even with only two drives with the f2 layout
(mirroring with striped reads, giving the read performance of RAID 0; normal Linux software RAID 1 doesnot stripe reads, but can read in parallel).Hewlett-Packards EVA series arrays implement vRAID - vRAID-0, vRAID-1, vRAID-5, and vRAID-6; vRAIDlevels are closely aligned to Nested RAID levels: vRAID-1 is actually a RAID 1+0 (or RAID 10), vRAID-5 isactually a RAID 5+0 (or RAID 50), etc.IBM (among others) has implemented a RAID 1E (Level 1 Enhanced). It requires a minimum of 3 drives. Itis similar to a RAID 1+0 array, but it can also be implemented with either an even or odd number ofdrives. The total available RAID storage is n/2.Hadoop has a RAID system that generates a parity file by xor-ing a stripe of blocks in a single HDFSfile.Data backupA RAID system used as secondary storage is not an alternative to backing up data. In parityconfigurations, a RAID protects from catastrophic data loss caused by physical damage or errors on asingle drive within the array (or two drives in, say, RAID 6). However, a true backup system has otherimportant features such as the ability to restore an earlier version of data, which is needed both toprotect against software errors that write unwanted data to secondary storage, and also to recover fromuser error and malicious data deletion. A RAID can be overwhelmed by catastrophic failure that exceedsits recovery capacity and, of course, the entire array is at risk of physical damage by fire, naturaldisaster, and human forces, while backups can be stored off-site. A RAID is also vulnerable to controllerfailure because it is not always possible to migrate a RAID to a new, different controller without dataloss.ImplementationsThe distribution of data across multiple drives can be managed either by dedicated computer hardwareor by software. A software solution may be part of the operating system, or it may be part of thefirmware and drivers supplied with a hardware RAID controller.Software-based RAIDSoftware RAID implementations are now provided by many operating systems. Software RAID can beimplemented as:a layer that abstracts multiple devices, thereby providing a single virtual device (e.g., Linuxs md).a more generic logical volume manager (provided with most server-class operating systems, e.g., Veritasor LVM).
a component of the file system (e.g., ZFS or Btrfs).Volume manager supportServer class operating systems typically provide logical volume management, which allows a system touse logical[jargon] volumes which can be resized or moved. Often, features like RAID or snapshots arealso supported.Vinum is a logical volume manager supporting RAID-0, RAID-1, and RAID-5. Vinum is part of the basedistribution of the FreeBSD operating system, and versions exist for NetBSD, OpenBSD, and DragonFlyBSD.Solaris SVM supports RAID 1 for the boot filesystem, and adds RAID 0 and RAID 5 support (and variousnested combinations) for data drives.Linux LVM supports RAID 0 and RAID 1.HPs OpenVMS provides a form of RAID 1 called "Volume shadowing", giving the possibility to mirrordata locally and at remote cluster systems.File system supportSome advanced file systems are designed to organize data across multiple storage devices directly(without needing the help of a third-party logical volume manager).ZFS supports equivalents of RAID 0, RAID 1, RAID 5 (RAID Z), RAID 6 (RAID Z2), and a triple parity versionRAID Z3, and any nested combination of those like 1+0. ZFS is the native file system on Solaris, and alsoavailable on FreeBSD.Btrfs supports RAID 0, RAID 1, and RAID 10 (RAID 5 and 6 are under development).Other supportMany operating systems provide basic RAID functionality independently of volume management.Apples Mac OS X Server and Mac OS X support RAID 0, RAID 1, and RAID 1+0.FreeBSD supports RAID 0, RAID 1, RAID 3, and RAID 5, and all nestings via GEOM modules andccd.Linuxs md supports RAID 0, RAID 1, RAID 4, RAID 5, RAID 6, and all nestings. Certainreshaping/resizing/expanding operations are also supported.Microsofts server operating systems support RAID 0, RAID 1, and RAID 5. Some of the Microsoft desktopoperating systems support RAID such as Windows XP Professional which supports RAID level 0 inaddition to spanning multiple drives but only if using dynamic disks and volumes. Windows XP can bemodified to support RAID 0, 1, and 5.
NetBSD supports RAID 0, RAID 1, RAID 4, and RAID 5, and all nestings via its software implementation,named RAIDframe.OpenBSD aims to support RAID 0, RAID 1, RAID 4, and RAID 5 via its software implementation softraid.FlexRAID (for Linux and Windows) is a snapshot RAID implementation.Software RAID has advantages and disadvantages compared to hardware RAID. The software must runon a host server attached to storage, and the servers processor must dedicate processing time to runthe RAID software; the additional processing capacity required for RAID 0 and RAID 1 is low, but parity-based arrays require more complex data processing during write or integrity-checking operations. As therate of data processing increases with the number of drives in the array, so does the processingrequirement. Furthermore, all the buses between the processor and the drive controller must carry theextra data required by RAID, which may cause congestion.Fortunately, over time, the increase in commodity CPU speed has been consistently greater than theincrease in drive throughput; the percentage of host CPU time required to saturate a given numberof drives has decreased. For instance, under 100% usage of a single core on a 2.1 GHz Intel "Core2" CPU,the Linux software RAID subsystem (md) as of version 2.6.26 is capable of calculating parity informationat 6 GB/s; however, a three-drive RAID 5 array using drives capable of sustaining a write operation at100 MB/s only requires parity to be calculated at the rate of 200 MB/s, which requires the resources ofjust over 3% of a single CPU core.Furthermore, software RAID implementations may employ more sophisticated algorithms thanhardware RAID implementations (e.g. drive scheduling and command queueing), and thus, may becapable of better performance.Another concern with software implementations is the process of booting the associated operatingsystem. For instance, consider a computer being booted from a RAID 1 (mirrored drives); if the first drivein the RAID 1 fails, then a first-stage boot loader might not be sophisticated enough to attempt loadingthe second-stage boot loader from the second drive as a fallback. In contrast, a RAID 1 hardwarecontroller typically has explicit programming to decide that a drive has malfunctioned and that the nextdrive should be used. At least the following second-stage boot loaders are capable of loading a kernelfrom a RAID 1:LILO (for Linux).Some configurations of the GRUB.The boot loader for FreeBSD.The boot loader for NetBSD.For data safety, the write-back cache of an operating system or individual drive might need to be turnedoff in order to ensure that as much data as possible is actually written to secondary storage before some
failure (such as a loss of power); unfortunately, turning off the write-back cache has a performancepenalty that can be significant depending on the workload and command queuing[jargon] support. Incontrast, a hardware RAID controller may carry a dedicated battery-powered write-back cache of itsown, thereby allowing for efficient operation that is also relatively safe. Fortunately, it is possible toavoid such problems with a software controller by constructing a RAID with safer components; forinstance, each drive could have its own battery or capacitor on its own write-back cache, and the drivecould implement atomicity in various ways, and the entire RAID or computing system could be poweredby a UPS, etc.Finally, a software RAID controller that is built into an operating system usually uses proprietary dataformats and RAID levels, so an associated RAID usually cannot be shared between operating systems aspart of a multi boot setup. However, such a RAID may be moved between computers that share thesame operating system; in contrast, such mobility is more difficult when using a hardware RAIDcontroller because both computers must provide compatible hardware controllers. Also, if the hardwarecontroller fails, data could become unrecoverable unless a hardware controller of the same type isobtained.Most software implementations allow a RAID to be created from partitions rather than entire physicaldrives. For instance, an administrator could divide each drive of an odd number of drives into twopartitions, and then mirror partitions across drives and stripe a volume across the mirrored partitions toemulate IBMs RAID 1E configuration. Using partitions in this way also allows for constructing multipleRAIDs in various RAID levels from the same set of drives. For example, one could have a very robustRAID 1 for important files, and a less robust RAID 5 or RAID 0 for less important data, all using the sameset of underlying drives. (Some BIOS-based controllers offer similar features, e.g. Intel Matrix RAID.)Using two partitions from the same drive in the same RAID puts data at risk if the drive fails; forinstance:A RAID 1 across partitions from the same drive makes all the data inaccessible if the single drive fails.Consider a RAID 5 composed of 4 drives, 3 of which are 250 GB and one of which is 500 GB; the 500 GBdrive is split into 2 partitions, each of which is 250 GB. Then, a failure of the 500 GB drive would remove2 underlying drives from the array, causing a failure of the entire array.Hardware-based RAIDHardware RAID controllers use proprietary data layouts, so it is not usually possible to span controllersfrom different manufacturers. They do not require processor resources, the BIOS can boot from them,and tighter integration with the device driver may offer better error handling.On a desktop system, a hardware RAID controller may be an expansion card connected to a bus (e.g., PCIor PCIe), a component integrated into the motherboard; there are controllers for supporting most typesof drive technology, such as IDE/ATA, SATA, SCSI, SSA, Fibre Channel, and sometimes even acombination. The controller and drives may be in a stand-alone enclosure, rather than inside acomputer, and the enclosure may be directly attached to a computer, or connected via a SAN.
Most hardware implementations provide a read/write cache, which, depending on the I/O workload,improves performance. In most systems, the write cache is non-volatile (i.e. battery-protected), sopending writes are not lost in the event of a power failure.Hardware implementations provide guaranteed performance, add no computational overhead to thehost computer, and can support many operating systems; the controller simply presents the RAID asanother logical drive.Firmware/driver-based RAIDA RAID implemented at the level of an operating system is not always compatible with the systems bootprocess, and it is generally impractical for desktop versions of Windows (as described above). However,hardware RAID controllers are expensive and proprietary. To fill this gap, cheap "RAID controllers" wereintroduced that do not contain a dedicated RAID controller chip, but simply a standard drive controllerchip with special firmware and drivers; during early stage bootup, the RAID is implemented by thefirmware, and once the operating system has been more completely loaded, then the drivers take overcontrol. Consequently, such controllers may not work when driver support is not available for the hostoperating system.Initially, the term "RAID controller" implied that the controller does the processing. However, while acontroller without a dedicated RAID chip is often described by a manufacturer as a "RAID controller", itis rarely made clear that the burden of RAID processing is borne by a host computers central processingunit rather than the RAID controller itself. Thus, this new type is sometimes called "fake" RAID; Adapteccalls it a "HostRAID".Moreover, a firmware controller can often only support certain types of hard drive to form the RAID thatit manages (e.g., SATA for an Intel Matrix RAID, as there is neither SCSI nor PATA support in modern IntelICH southbridges; however, motherboard makers implement RAID controllers outside of thesouthbridge on some motherboards).Hot sparesBoth hardware and software RAIDs with redundancy may support the use of a hot spare drive; this is adrive physically installed in the array which is inactive until an active drive fails, when the systemautomatically replaces the failed drive with the spare, rebuilding the array with the spare drive included.This reduces the mean time to recovery (MTTR), but does not completely eliminate it. As with non-hot-spare systems, subsequent additional failure(s) in the same RAID redundancy group before the array isfully rebuilt can cause data loss. Rebuilding can take several hours, especially on busy systems.RAID 6 without a spare uses the same number of drives as RAID 5 with a hot spare and protects dataagainst failure of up to two drives, but requires a more advanced RAID controller and may not performas well. Further, a hot spare can be shared by multiple RAID sets.Data scrubbing / Patrol read
Data scrubbing is periodic reading and checking by the RAID controller of all the blocks in a RAID,including those not otherwise accessed. This allows bad blocks to be detected before they are used.An alternate name for this is patrol read. This is defined as a check for bad blocks on each storage devicein an array, but which also uses the redundancy of the array to recover bad blocks on a single drive andreassign the recovered data to spare blocks elsewhere on the drive.Reliability termsFailure rateTwo different kinds of failure rates are applicable to RAID systems. Logical failure is defined as the loss ofa single drive and its rate is equal to the sum of individual drives failure rates. System failure is definedas loss of data and its rate will depend on the type of RAID. For RAID 0 this is equal to the logical failurerate, as there is no redundancy. For other types of RAID, it will be less than the logical failure rate,potentially very small, and its exact value will depend on the type of RAID, the number of drivesemployed, the vigilance and alacrity of its human administrators, and chance (improbable events dooccur, though infrequently).Mean time to data loss (MTTDL)In this context, the average time before a loss of data in a given array. Mean time to data loss of agiven RAID may be higher or lower than that of its constituent hard drives, depending upon what type ofRAID is employed. The referenced report assumes times to data loss are exponentially distributed, sothat 63.2% of all data loss will occur between time 0 and the MTTDL.Mean time to recovery (MTTR)In arrays that include redundancy for reliability, this is the time following a failure to restore an array toits normal failure-tolerant mode of operation. This includes time to replace a failed drive mechanismand time to re-build the array (to replicate data for redundancy).Unrecoverable bit error rate (UBE)This is the rate at which a drive will be unable to recover data after application of cyclic redundancycheck (CRC) codes and multiple retries.Write cache reliabilitySome RAID systems use RAM write cache to increase performance. A power failure can result in dataloss unless this sort of drive buffer has a supplementary battery to ensure that the buffer has time towrite from RAM to secondary storage before the drive powers down.Atomic write failure
Also known by various terms such as torn writes, torn pages, incomplete writes, interrupted writes, non-transactional, etc.Problems with RAIDCorrelated failuresThe theory behind the error correction in RAID assumes that failures of drives are independent. Giventhese assumptions, it is possible to calculate how often they can fail and to arrange the array to makedata loss arbitrarily improbable. There is also an assumption that motherboard failures wont damagethe hard drive and that hard drive failures occur more often than motherboard failures.In practice, the drives are often the same age (with similar wear) and subject to the same environment.Since many drive failures are due to mechanical issues (which are more likely on older drives), thisviolates those assumptions; failures are in fact statistically correlated. In practice, the chances of asecond failure before the first has been recovered (causing data loss) is not as unlikely as four randomfailures. In a study including about 100 thousand drives, the probability of two drives in the same clusterfailing within one hour was observed to be four times larger than was predicted by the exponentialstatistical distribution which characterizes processes in which events occur continuously andindependently at a constant average rate. The probability of two failures within the same 10-hourperiod was twice as large as that which was predicted by an exponential distribution.A common assumption is that "server-grade" drives fail less frequently than consumer-grade drives. Twoindependent studies (one by Carnegie Mellon University and the other by Google) have shown that the"grade" of a drive does not relate to the drives failure rate.In addition, there is no protection circuitry between the motherboard and hard drive electronics, so acatastrophic failure of the motherboard can cause the harddrive electronics to fail. Therefore, takingelaborate precautions via RAID setups ignores the equal risk of electronics failures elsewhere which cancascade to a hard drive failure. For a robust critical data system, no risk can outweigh another as theconsequence of any data loss is unacceptable.AtomicityThis is a little understood and rarely mentioned failure mode for redundant storage systems that do notutilize transactional features. Database researcher Jim Gray wrote "Update in Place is a PoisonApple" during the early days of relational database commercialization. However, this warning largelywent unheeded and fell by the wayside upon the advent of RAID, which many software engineersmistook as solving all data storage integrity and reliability problems. Many software programs update astorage object "in-place"; that is, they write a new version of the object on to the same secondarystorage addresses as the old version of the object. While the software may also log some deltainformation elsewhere, it expects the storage to present "atomic write semantics," meaning that thewrite of the data either occurred in its entirety or did not occur at all.
However, very few storage systems provide support for atomic writes, and even fewer specify their rateof failure in providing this semantic. Note that during the act of writing an object, a RAID storage devicewill usually be writing all redundant copies of the object in parallel, although overlapped or staggeredwrites are more common when a single RAID processor is responsible for multiple drives. Hence an errorthat occurs during the process of writing may leave the redundant copies in different states, andfurthermore may leave the copies in neither the old nor the new state. The little known failure mode isthat delta logging relies on the original data being either in the old or the new state so as to enablebacking out the logical change, yet few storage systems provide an atomic write semantic for a RAID.While the battery-backed write cache may partially solve the problem, it is applicable only to a powerfailure scenario.Since transactional support is not universally present in hardware RAID, many operating systems includetransactional support to protect against data loss during an interrupted write. Novell NetWare, startingwith version 3.x, included a transaction tracking system. Microsoft introduced transaction tracking viathe journaling feature in NTFS. ext4 has journaling with checksums; ext3 has journaling withoutchecksums but an "append-only" option, or ext3cow (Copy on Write). If the journal itself in a filesystemis corrupted though, this can be problematic. The journaling in NetApp WAFL file system gives atomicityby never updating the data in place, as does ZFS. An alternative method to journaling is soft updates,which are used in some BSD-derived systems implementation of UFS.This can present as a sector read failure. Some RAID implementations protect against this failure modeby remapping the bad sector, using the redundant data to retrieve a good copy of the data, andrewriting that good data to the newly mapped replacement sector. The UBE (Unrecoverable Bit Error)rate is typically specified at 1 bit in 1015 for enterprise class drives (SCSI, FC, SAS), and 1 bit in 1014 fordesktop class drives (IDE/ATA/PATA, SATA). Increasing drive capacities and large RAID 5 redundancygroups have led to an increasing inability to successfully rebuild a RAID group after a drive failurebecause an unrecoverable sector is found on the remaining drives. Double protection schemes such asRAID 6 are attempting to address this issue, but suffer from a very high write penalty.Write cache reliabilityThe drive system can acknowledge the write operation as soon as the data is in the cache, not waitingfor the data to be physically written. This typically occurs in old, non-journaled systems such as FAT32,or if the Linux/Unix "writeback" option is chosen without any protections like the "soft updates" option(to promote I/O speed whilst trading-away data reliability). A power outage or system hang such as aBSOD can mean a significant loss of any data queued in such a cache.Often a battery is protecting the write cache, mostly solving the problem. If a write fails because ofpower failure, the controller may complete the pending writes as soon as restarted. This solution stillhas potential failure cases: the battery may have worn out, the power may be off for too long, the drivescould be moved to another controller, and the controller itself could fail. Some systems provide thecapability of testing the battery periodically, however this leaves the system without a fully chargedbattery for several hours.
An additional concern about write cache reliability exists, specifically regarding devices equipped with awrite-back cache—a caching system which reports the data as written as soon as it is written to cache,as opposed to the non-volatile medium. The safer cache technique is write-through, which reportstransactions as written when they are written to the non-volatile medium.Equipment compatibilityThe methods used to store data by various RAID controllers are not necessarily compatible, so that itmay not be possible to read a RAID on different hardware, with the exception of RAID 1, which istypically represented as plain identical copies of the original data on each drive. Consequently a non-drive hardware failure may require the use of identical hardware to recover the data, and furthermorean identical configuration has to be reassembled without triggering a rebuild and overwriting the data.Software RAID however, such as implemented in the Linux kernel, alleviates this concern, as the setup isnot hardware dependent, but runs on ordinary drive controllers, and allows the reassembly of an array.Additionally, individual drives of a RAID 1 (software and most hardware implementations) can be readlike normal drives when removed from the array, so no RAID system is required to retrieve the data.Inexperienced data recovery firms typically have a difficult time recovering data from RAID drives, withthe exception of RAID1 drives with conventional data structure.Data recovery in the event of a failed arrayWith larger drive capacities the odds of a drive failure during rebuild are not negligible. In that event, thedifficulty of extracting data from a failed array must be considered. Only a RAID 1 (mirror) stores all dataon each drive in the array. Although it may depend on the controller, some individual drives in a RAID 1can be read as a single conventional drive; this means a damaged RAID 1 can often be easily recovered ifat least one component drive is in working condition. If the damage is more severe, some or all data canoften be recovered by professional data recovery specialists. However, other RAID levels (like RAID level5) present much more formidable obstacles to data recovery.Drive error recovery algorithmsMany modern drives have internal error recovery algorithms that can take upwards of a minute torecover and re-map data that the drive fails to read easily. Frequently, a RAID controller is configured todrop a component drive (that is, to assume a component drive has failed) if the drive has beenunresponsive for 8 seconds or so; this might cause the array controller to drop a good drive because thatdrive has not been given enough time to complete its internal error recovery procedure. Consequently,desktop drives can be quite risky when used in a RAID, and so-called enterprise class drives limit thiserror recovery time in order to obviate the problem.A fix specific to Western Digitals desktop drives used to be known: A utility called WDTLER.exe couldlimit a drives error recovery time; the utility enabled TLER (time limited error recovery), which limits theerror recovery time to 7 seconds. Around September 2009, Western Digital disabled this feature in theirdesktop drives (e.g., the Caviar Black line), making such drives unsuitable for use in a RAID.
However, Western Digital enterprise class drives are shipped from the factory with TLER enabled. Similartechnologies are used by Seagate, Samsung, and Hitachi. Of course, for non-RAID usage, an enterpriseclass drive with a short error recovery timeout that cannot be changed is therefore less suitable than adesktop drive.In late 2010, the Smartmontools program began supporting the configuration of ATA Error RecoveryControl, allowing the tool to configure many desktop class hard drives for use in a RAID.Recovery time is increasingDrive capacity has grown at a much faster rate than transfer speed, and error rates have only fallen alittle in comparison. Therefore, larger capacity drives may take hours, if not days, to rebuild. The re-buildtime is also limited if the entire array is still in operation at reduced capacity. Given a RAID with onlyone drive of redundancy (RAIDs 3, 4, and 5), a second failure would cause complete failure of the array.Even though individual drives mean time between failure (MTBF) have increased over time, thisincrease has not kept pace with the increased storage capacity of the drives. The time to rebuild thearray after a single drive failure, as well as the chance of a second failure during a rebuild, haveincreased over time.Operator skills, correct operationIn order to provide the desired protection against physical drive failure, a RAID must be properly set upand maintained by an operator with sufficient knowledge of the chosen RAID configuration, arraycontroller (hardware or software), failure detection and recovery. Unskilled handling of the array at anystage may exacerbate the consequences of a failure, and result in downtime and full or partial loss ofdata that might otherwise be recoverable.Particularly, the array must be monitored, and any failures detected and dealt with promptly. Failure todo so will result in the array continuing to run in a degraded state, vulnerable to further failures.Ultimately more failures may occur, until the entire array becomes inoperable, resulting in data loss anddowntime. In this case, any protection the array may provide merely delays this.The operator must know how to detect failures or verify healthy state of the array, identify which drivefailed, have replacement drives available, and know how to replace a drive and initiate a rebuild of thearray.In order to protect against such issues and reduce the need for direct onsite monitoring, some serverhardware includes remote management and monitoring capabilities referred to as BaseboardManagement, using the Intelligent Platform Management Interface. A server at a remote site which isnot monitored by an onsite technician can instead be remotely managed and monitored, using aseparate standalone communications channel that does not require the managed device to beoperating. The Baseboard Management Controller in the server functions independent of the installedoperating system, and may include the ability to manage and monitor a server even when it is in its"powered off / standby" state.
Hardware labeling issuesThe hardware itself can contribute to RAID array management challenges, depending on how the arraydrives are arranged and identified. If there is no clear indication of which drive is failed, an operator notfamiliar with the hardware might remove a non-failed drive in a running server, and destroy an alreadydegraded array.A controller may refer to drives by an internal numbering scheme such as 0, 1, 2... while an externaldrive mounting frame may be labeled 1, 2, 3...; in this situation drive #2 as identified by the controller isactually in mounting frame position #3.For large arrays spanning several external drive frames, each separate frame may restart the numberingat 1, 2, 3... but if the drive frames are cabled together, then the second row of a 12-drive frame mayactually be drive 13, 14, 15...SCSI IDs can be assigned directly on the drive rather than through the interface connector. For direct-cabled drives, it is possible for the drive IDs to be arranged in any order on the SCSI cable, and forcabled drives to swap position keeping their individually assigned ID, even if the servers external chassislabeling indicates otherwise. Someone unfamiliar with a servers management challenges could swapdrives around while the power is off without causing immediate damage to the RAID array, but whichmisleads other technicians at a later time that are assuming failed drives are in the original locations.Other problemsWhile RAID may protect against physical drive failure, the data is still exposed to operator, software,hardware and virus destruction. Many studies cite operator fault as the most common source ofmalfunction, such as a server operator replacing the incorrect drive in a faulty RAID, and disabling thesystem (even temporarily) in the process. Most well-designed systems include separate backupsystems that hold copies of the data, but do not allow much interaction with it. Most copy the data andremove the copy from the computer for safe storage.Hardware RAID controllers are really just small computers running specialized software. Although RAIDcontrollers tend to be very thoroughly tested for reliability, the controller software may still containbugs that cause damage to data in certain unforeseen situations. The controller software may also havetime-dependent bugs that dont manifest until a system has been operating continuously, beyond whatis a feasible time-frame for testing, before the controller product goes to market.HistoryNorman Ken Ouchi at IBM was awarded a 1978 U.S. patent 4,092,732 titled "System for recoveringdata stored in failed memory unit." The claims for this patent describe what would later be termed RAID5 with full stripe writes. This 1978 patent also mentions that drive mirroring or duplexing (what would
later be termed RAID 1) and protection with dedicated parity (that would later be termed RAID 4) wereprior art at that time.In October 1986, the IBM S/38 announced "checksum" - an operating system software levelimplementation of what became RAID-5. The S/38 "scatter-loaded" data over all disks for betterperformance and ease of use. As a result, a single disk failure forced the restore of the entire system.With S/38 checksum, when a disk failed, the system stopped and was powered off. Under maintenance,the bad disk was replaced and the new disk was fully recovered using RAID parity bits. While checksumhad 10%-30% overhead and was not concurrent recovery, non-concurrent recovery was still a far bettersolution than a reload of the entire system. With 30% overhead and the then high expense of extra disk,few customers implemented checksum.The term RAID was first defined by David A. Patterson, Garth A. Gibson and Randy Katz at the Universityof California, Berkeley, in 1987. They studied the possibility of using two or more drives to appear as asingle device to the host system and published a paper: "A Case for Redundant Arrays of InexpensiveDisks (RAID)" in June 1988 at the SIGMOD conference.This specification suggested a number of prototype RAID levels, or combinations of drives. Each hadtheoretical advantages and disadvantages. Over the years, different implementations of the RAIDconcept have appeared. Most differ substantially from the original idealized RAID levels, but thenumbered names have remained. This can be confusing, since one implementation of RAID 5, forexample, can differ substantially from another. RAID 3 and RAID 4 are often confused and even usedinterchangeably.One of the early uses of RAID 0 and 1 was the Crosfield Electronics Studio 9500 page layout systembased on the Python workstation. The Python workstation was a Crosfield managed internationaldevelopment using PERQ 3B electronics, benchMark Technologys Viper display system and Crosfieldsown RAID and fibre-optic network controllers. RAID 0 was particularly important to these workstationsas it dramatically sped up image manipulation for the pre-press markets. Volume production started inPeterborough, England in early 1987.Non-RAID drive architecturesMain article: Non-RAID drive architecturesNon-RAID drive architectures also exist, and are often referred to, similarly to RAID, by standardacronyms, several tongue-in-cheek. A single drive is referred to as a SLED (Single Large ExpensiveDisk/Drive), by contrast with RAID, while an array of drives without any additional control (accessedsimply as independent drives) is referred to, even in a formal context such as equipment specification,as a JBOD (Just a Bunch Of Disks). Simple concatenation is referred to as a "span".See also
Reliable array of independent nodes (RAIN)Redundant array of independent memory (RAIM)References^ Donald, L. (2003). MCSA/MCSE 2006 JumpStart Computer and Network Basics (2nd ed.). Glasgow:SYBEX.^ Howe, Denis, ed. Redundant Arrays of Independent Disks from FOLDOC. Imperial College Departmentof Computing. http://foldoc.org/RAID. Retrieved 2011-11-10.^ a b David A. Patterson, Garth Gibson, and Randy H. Katz: A Case for Redundant Arrays of InexpensiveDisks (RAID). University of California Berkeley. 1988.^ "Originally referred to as Redundant Array of Inexpensive Disks, the concept of RAID was firstdeveloped in the late 1980s by Patterson, Gibson, and Katz of the University of California at Berkeley.(The RAID Advisory Board has since substituted the term Inexpensive with Independent.)" StorageccArea Network Fundamentals; Meeta Gupta; Cisco Press; ISBN 978-1-58705-065-7; Appendix A.^ See RAS syndrome.^ "Common RAID Disk Drive Format (DDF) standard". Snia.org. Retrieved 2012-08-26.^ "SNIA Dictionary". Snia.org. Retrieved 2010-08-24.^ Vijayan, S.; Selvamani, S. ; Vijayan, S (1995). "Dual-Crosshatch Disk Array: A Highly Reliable Hybrid-RAID Architecture". Proceedings of the 1995 International Conference on Parallel Processing: Volume 1.CRC Press. pp. I–146ff. ISBN 0-8493-2615-X.^ a b Jeffrey B. Layton: "Intro to Nested-RAID: RAID-01 and RAID-10", Linux Magazine, January 6, 2011^ DAwkins, Bill and Jones, Arnold. "Common RAID Disk Data Format Specification" [Storage NetworkingIndustry Association] Colorado Springs, 28 July 2006. Retrieved on 22 February 2011.^ a b c d "Why RAID 6 stops working in 2019". ZDNet. 22 February 2010.^ "RAID Classifications". BytePile.com. 2012-04-10. Retrieved 2012-08-26.^ [dead link]^ a b "RAID Classifications". BytePile.com. 2012-04-10. Retrieved 2012-08-26.^ , question 4
^ "Main Page - Linux-raid". Linux-raid.osdl.org. 2010-08-20. Retrieved 2010-08-24.^ "Hdfs Raid". Hadoopblog.blogspot.com. 2009-08-28. Retrieved 2010-08-24.^ "The RAID Migration Adventure". Retrieved 2010-03-10.^ "Apple Mac OS X Server File Systems". Retrieved 2008-04-23.^ "Mac OS X: How to combine RAID sets in Disk Utility". Retrieved 2010-01-04.^ "FreeBSD System Managers Manual page for GEOM(8)". Retrieved 2009-03-19.^ "freebsd-geom mailing list - new class / geom_raid5". Retrieved 2009-03-19.^ "FreeBSD Kernel Interfaces Manual for CCD(4)". Retrieved 2009-03-19.^ "The Software-RAID HOWTO". Retrieved 2008-11-10.^ "RAID setup". Retrieved 2008-11-10.[dead link]^ "RAID setup". Retrieved 2010-09-30.^ "Using WindowsXP to Make RAID 5 Happen". Tomshardware.com. Retrieved 2010-08-24.^ "Rules of Thumb in Data Engineering". Retrieved 2010-01-14.^ "FreeBSD Handbook". Chapter 19 GEOM: Modular Disk Transformation Framework. Retrieved 2009-03-19.^ "SATA RAID FAQ - ata Wiki". Ata.wiki.kernel.org. 2011-04-08. Retrieved 2012-08-26.^ Ulf Troppens, Wolfgang Mueller-Friedt, Rainer Erkens, Rainer Wolafka, Nils Haustein. StorageNetworks Explained: Basics and Application of Fibre Channel SAN, NAS, ISCSI,InfiniBand and FCoE. JohnWiley and Sons, 2009. p.39^ Dell Computers, Background Patrol Read for Dell PowerEdge RAID Controllers, BY DREW HABAS ANDJOHN SIEBER, Reprinted from Dell Power Solutions, February 2006http://www.dell.com/downloads/global/power/ps1q06-20050212-Habas.pdf^ Jim Gray and Catharine van Ingen, "Empirical Measurements of Disk Failure Rates and Error Rates",MSTR-2005-166, December 2005^ Disk Failures in the Real World: What Does an MTTF of 1,000,000 Hours Mean to You? BiancaSchroeder and Garth A. Gibson^ "Everything You Know About Disks Is Wrong". Storagemojo.com. 2007-02-22. Retrieved 2010-08-24.^ Eduardo Pinheiro, Wolf-Dietrich Weber and Luiz André Barroso (February 2007). "Failure Trends in aLarge Disk Drive Population". Google Inc. Retrieved 2011-12-26.
^ Jim Gray: The Transaction Concept: Virtues and Limitations (Invited Paper) VLDB 1981: 144-154^ "Definition of write-back cache at SNIA dictionary".^ a b c "Error recovery control with smartmontools". Retrieved 2011.^ Patterson, D., Hennessy, J. (2009). Computer Organization and Design. New York: Morgan KaufmannPublishers. pp 604-605.^ Newman, Henry (2009-09-17). "RAIDs Days May Be Numbered". EnterpriseStorageForum. Retrieved2010-09-07.^ These studies are: Gray, J (1990), Murphy and Gent (1995), Kuhn (1997), and Enriquez P. (2003). Seefollowing source.^ Patterson, D., Hennessy, J. (2009), 574.^ US patent 4092732, Norman Ken Ouchi, "System for recovering data stored in failed memory unit",issued 1978-05-30External links Wikimedia Commons has media related to: Redundant array of independent disksRAID at the Open Directory ProjectCategories: RAIDFault-tolerant computer systemsCreate accountLog inArticleTalkReadEditView historyNavigationMain pageContentsFeatured contentCurrent eventsRandom articleDonate to WikipediaInteraction
HelpAbout WikipediaCommunity portalRecent changesContact WikipediaToolboxWhat links hereRelated changesUpload fileSpecial pagesPermanent linkPage informationCite this pagePrint/exportCreate a bookDownload as PDFPrintable versionLanguagesAfrikaansية ال عربБългарскиCatalàČeskyDanskDeutsch