CN108228382B - Data recovery method for single-disk fault of EVENODD code - Google Patents
Data recovery method for single-disk fault of EVENODD code Download PDFInfo
- Publication number
- CN108228382B CN108228382B CN201810027966.2A CN201810027966A CN108228382B CN 108228382 B CN108228382 B CN 108228382B CN 201810027966 A CN201810027966 A CN 201810027966A CN 108228382 B CN108228382 B CN 108228382B
- Authority
- CN
- China
- Prior art keywords
- data
- disk
- redundant
- failure
- disks
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/07—Responding to the occurrence of a fault, e.g. fault tolerance
- G06F11/08—Error detection or correction by redundancy in data representation, e.g. by using checking codes
- G06F11/10—Adding special bits or symbols to the coded information, e.g. parity check, casting out 9's or 11's
- G06F11/1076—Parity data used in redundant arrays of independent storages, e.g. in RAID systems
Abstract
The invention discloses a data recovery method for single-disk faults of EVENODD codes, and belongs to the technical field of computer information storage and recovery. The method is based on the existing EVENODD coding, and the EVENODD comprising p data disks is selected as the former
Carrying out XOR calculation on the data disks, and storing the XOR result into a new redundant disk, wherein the redundant disk is defined as a local redundant disk; when the data disk fails, corresponding data recovery is carried out according to the position of the data disk; and when the redundant disk fails, performing corresponding data recovery according to the type of the redundant disk. The method utilizes the newly added local redundant disk to reduce the number of disks related to a single data disk. The number of disk reading is reduced during data recovery, the data reading overhead is low, the data recovery efficiency is high, and the single-disk failure recovery algorithm is superior to the traditional EVENODD code single-disk failure recovery algorithm.
Description
Technical Field
The invention relates to the technical field of computer information storage and recovery, in particular to a data recovery algorithm for an EVENODD code single-disk fault.
Background
In recent years, with the rapid development of society and technology, the amount of data in databases is increasing at an explosive rate, and thus the scale of storage systems is also increasing. As the scale increases, the problems faced with the stability of storage systems also increase. The main expression is that as the number of disks increases, the probability of data loss caused by disk failure also increases. Data recovery is required to ensure that the data remains intact even after a disk failure. Therefore, the time for data recovery and the amount of data that needs to be read for the recovery process become critical to system stability.
EVENODD codes proposed in the document (EVENODD: an effective scheme for completing double disk failures in RAID architecture) are a type of array codes which are commonly used in storage systems and are tolerant to two errors. The principle is mainly to ensure the integrity of data by adding two extra redundant disks. The EVENODD code can correctly recover the original data after any two data disks fail. For EVENODD coding, the advantages are simple structure, and fast speed because the coding and decoding process is based on XOR operation. However, one of the main disadvantages is that the single-disk failure recovery requires too much data to be read, all the remaining original data needs to be read, the data reading time is long, and the data recovery efficiency is low.
Disclosure of Invention
Based on the above, the technical problem to be solved by the present invention is to provide a new data recovery method for single disc failure of an event odd code, aiming at the defects of long data recovery time and low recovery efficiency of the single disc failure of the existing event odd code, wherein the storage overhead of the algorithm is slightly larger than that of the traditional event odd code, but the reading overhead of the single disc failure recovery is only about 50% of that of the existing event odd code, the data recovery time is short, and the recovery efficiency is high.
The overall principle of the data recovery method provided by the invention is as follows: taking the front of the EVENODD code
And carrying out XOR calculation on the data disks, and storing the XOR result into a new redundant disk, wherein the new redundant disk is defined as a local redundant disk. When a single disk has data failure, corresponding data recovery is carried out according to the position or the type of the failed disk. If the data disk fails, corresponding data recovery is carried out according to the position of the data disk; and if the redundant disk fails, performing corresponding data recovery according to the type of the redundant disk.
The data recovery method comprises the following steps:
(1) firstly, encoding the EVENODD comprising p data disks and two redundant disks, and taking the encoded EVENODDFront side
Performing exclusive-or calculation on the data disks, wherein p is the number of the data disks, the value of p is a prime number, and the two redundant disks are a horizontal redundant disk and a diagonal redundant disk respectively;
(2) storing the result of the exclusive or into a new redundant disk, wherein the redundant disk is defined as a local redundant disk, and the calculation formula is as follows:
ai,jfor a data block in the disk array, i, j is the row-column coordinate of the data block;
(3) when a single data disk has data failure (data loss), performing failure recovery according to the position of the failed data disk;
(4) when a data failure (data loss) occurs in a single redundant disk, failure recovery is performed according to the type of the failed redundant disk.
Specifically, the step (3) includes the following steps:
firstly, judging the position of a data disk with data failure, and dividing the situation into the situation that the failure occurs before
A data disk and a back disk
Two cases of a data disk;
② if the data disk with data failure is positioned in front
In a data disk, before reading
Carrying out the difference on the data of the intact disk and the local redundant disk in each diskOr calculating to obtain data, namely the data of the disk with the data failure, setting the data disk with the data failure as e, and setting a data recovery formula as follows;
ai,jfor a data block in the disk array, i, j is the row-column coordinate of the data block;
third, if the data disk with data failure is behind
In one disk, then read later
Performing exclusive-or calculation on data of a complete disk, a horizontal redundant disk and a local redundant disk in each data disk to obtain data, namely the data disk of the disk with the data fault, setting the data disk with the data fault as e, wherein a data recovery formula is as follows:
ai,jfor a data block in the disk array, i, j is the row and column coordinates of the data block
Specifically, the step (4) includes the following steps:
firstly, judging the type of a failed redundant disk, and dividing the type into three conditions:
if the disk with data failure is horizontal redundant disk, then after reading
Performing exclusive-or calculation on the data of the data disks and the local redundant disk to obtain data, namely the horizontal redundant disk data with data failure, wherein the data recovery formula is as follows:
ai,jfor a data block in the disk array, i, j is the row-column coordinate of the data block;
if the data failure occurs in the diagonal redundant disk, the data in all the data disks needs to be read for exclusive-or calculation, and the calculation formula is as follows:
ai,jfor a data block in the disk array, i, j is the row-column coordinate of the data block;
if the data failure occurs to the local redundant disk, before reading
Performing exclusive-or calculation on the data disks to obtain data, namely the data of the local redundant disk with the data failure, wherein the calculation formula is the same as the formula in the step (2), and specifically is as follows:
ai,jfor a data block in the disk array, i, j is the row and column coordinates of the data block
The method utilizes the newly added local redundant disk to reduce the number of disks related to a single data disk. The recovery algorithm reduces the number of disk reads, and is superior to the traditional EVENODD single-disk failure recovery algorithm in time and space.
Generally, the data recovery method can comprehensively evaluate the storage overhead, the recovery reading overhead, the system stability and the correctness. The storage overhead can be reduced by comparing the size required for encodingCompared with the existing EVENODD, the algorithm adds one more redundant column (local redundant disk), but in practice, the value of p is often large, and the added storage overhead can be almost ignored; the recovery reading overhead index can be observed by the number of file blocks which need to be read for recovering the same type of errors, because the probability of the data disk failing is greater than that of the redundant disk, so to speak, the probability is far greater, and when p tends to be infinite, the single-disk failure overhead approaches to the infinite
Therefore, the reading overhead can be reduced by 50% theoretically, and the recovery efficiency is high. The stability of a system in a disk array is often determined by the failure recovery time of a single disk, and the shorter the data recovery time is, the lower the probability of generating multiple errors is; because the number of the read data blocks is reduced, the recovery time of the algorithm is also reduced, and the system stability is improved; finally, because the first p +2 columns of the algorithm reserve the EVENODD codes, any two errors can be recovered, and the decoding correctness is ensured. Therefore, the data failure recovery method provided by the invention is excellent.
According to the method, the beneficial effects of the invention are as follows:
1. the problem that the reading overhead of the single-disk fault recovery of the existing EVENODD code is too high is solved; the invention is based on the improvement of the EVENODD single disk fault recovery, and the EVENODD coding method is a coding mode with simple algorithm and high efficiency based on XOR operation. However, the EVENODD has a significant disadvantage that, when a single disk fails, all data is needed to be used in recovery, and the reading overhead is very large. According to the data recovery method provided by the invention, as the local redundant disk is added, the number of disks related to a single data disk is reduced to half, and the reading overhead is effectively reduced.
2. The overall stability of the system is improved; most disk failures in array storage systems are single disk failures, and system stability depends on recovery time and read overhead of the single disk failure. The data recovery method provided by the invention reduces the number of the disks participating in the operation, thereby not only reducing the reading overhead, but also saving the operation time, further improving the recovery efficiency and further indirectly improving the overall stability of the system.
Drawings
FIG. 1 is a schematic diagram of a prior art EVENODD (5, 7) encoded disk array model;
FIG. 2 is a schematic diagram of the data algorithm for the horizontal redundant disk and the diagonal redundant disk in FIG. 1:
FIG. 3 is a schematic diagram of an EVENODD (5, 7) encoded disk array model after modification in accordance with the present invention;
FIG. 4 is a schematic structural diagram of the EVENODD (5, 7) encoded disk array model when a data failure occurs in the first data disk;
FIG. 5 is a schematic structural diagram of the EVENODD (5, 7) encoded disk array model when a data failure occurs in the third data disk;
FIG. 6 is a schematic structural diagram of a horizontal redundant disk in an EVENODD (5, 7) encoded disk array model when a failure occurs;
Detailed Description
Embodiments of the present invention will be described in detail below with reference to the accompanying drawings.
The described embodiments of the invention are only some embodiments of the invention, and not all embodiments. For the purpose of illustration, only some, but not all, aspects relevant to the present invention are shown in the drawings.
The data failure described in the following embodiments may also be referred to as data loss, and all data failures and data losses occurring in this specification are interpreted as the same; similarly, the same explanations apply to failure and loss.
The invention is explained by taking the existing EVENODD (5, 7) structure as a reference, wherein the EVENODD (5, 7) structure comprises 5 data disks, namely p is 5; and 2 redundant disks are respectively a horizontal redundant disk and a diagonal redundant disk.
In general, the value of p is large, often greater than 100, and for convenience of describing the operation mode of the present invention, in this embodiment, only the example where p is 5 is taken as an example, and when p is another value, the operation mode may refer to the example where p is 5.
Referring to fig. 1-2, fig. 1 is a schematic diagram of a conventional EVENODD (5, 7) disk array structure, and fig. 2 is a schematic diagram of a corresponding algorithm; disk 0-Disk 4 are 5 data disks grouped together in an array, Disk 5-Disk 6 are 2 redundant disks grouped together, wherein Disk5 is a horizontal redundant Disk, and Disk6 is a diagonal redundant Disk.
The data algorithm of the horizontal redundant disk is
The data algorithm of the diagonal redundant disk is
Wherein the content of the first and second substances,
i＝0,1...3；
ai,jfor a data block in the disk array, i, j is the row and column coordinates of the data block.
Fig. 3 is a schematic diagram of a modified EVENODD (5, 7) Disk array structure of the present invention, and compared with fig. 1, it can be known that a redundant Disk is added to the modified EVENODD (5, 7) Disk array, that is, a local redundant Disk, that is, a Disk7 shown in fig. 3, is added on the basis of the existing horizontal redundant Disk and diagonal redundant Disk.
The data of the local redundant Disk is obtained by performing exclusive-or calculation on the data in the first 2 data disks, namely performing exclusive-or by using the data in Disk0 and Disk 1;
the calculation formula is as follows:
i＝0,1...3。
referring to fig. 4, in the drawing, x represents occurrence of data failure, and o represents data integrity. When the Disk0 fails to be found in the first data Disk of the first 2 data disks in the Disk array, reading the data of the second data Disk and the local redundant Disk, namely reading the data in Disk1 and Disk7 for exclusive or, thereby recovering the lost data in Disk 0;
the data recovery formula is:
referring to fig. 5, in the drawing, x represents occurrence of data failure, and o represents data integrity. When the data failure occurs in the third data Disk, i.e. Disk2, of the last 3 data disks in the Disk array, the data in Disk3 and Disk4 are read first, then the data in the horizontal redundant Disk and the local redundant Disk, i.e. the data in Disk5 and Disk7, are read, and the read data are subjected to exclusive-or calculation, so that the lost data in Disk2 is recovered;
the data recovery formula is:
referring to fig. 6, in the drawing, x represents occurrence of a data failure, o represents perfect data, and when a data failure occurs in a redundant disk, the type of the redundant disk with the data failure is determined first;
if the data failure occurs in the horizontal redundant Disk, namely the data failure occurs in Disk5, the data in the next 3 data disks, namely Disk2, Disk3 and Disk4, are read first, then the data in the local redundant Disk, namely the data in Disk7, are read, and the read data are subjected to exclusive-or calculation. Thereby recovering the lost data in Disk 5;
the data recovery formula is:
if the data failure occurs in the diagonal redundant Disk, namely the data failure occurs in Disk6, the exclusive-or calculation is performed according to the existing EVENODD (5, 7) array structure model, and the calculation is not different from the existing calculation. The formula is omitted here. Generally, the probability of data failure of the diagonal redundant disk is very low due to the large p value, so that the data recovery situation is rarely applied, and once the data recovery situation occurs, the exclusive or calculation is performed according to the existing EVENODD (5, 7) model.
If the data failure occurs in the local redundant Disk, namely the data failure occurs in Disk7, reading the data in the first 2 data disks, namely the data in Disk0 and Disk1, and performing exclusive-or calculation;
the data recovery formula is:
i＝0,1...3
the above embodiments only show the case where p is 5, and since the value of p is usually large, a plurality of typical embodiments cannot be listed, and it can be considered that the above embodiments are preferred embodiments of the present invention rather than all embodiments, and any technical solutions that can be realized on the basis of the above embodiments without creative work should be considered to fall within the scope of protection of the patent of the present invention.
Claims (2)
1. A data recovery method for single disk failure of EVENODD code is characterized by comprising the following steps:
the method comprises the steps of encoding EVENODD comprising p data disks and two redundant disks, and taking the first one of the data disks
Performing exclusive-or calculation on the data disks, wherein p is a prime number, and the two redundant disks are a horizontal redundant disk and a diagonal redundant disk respectively;
secondly, storing the results of the XOR in the steps into a new redundant disk, wherein the redundant disk is defined as a local redundant disk;
thirdly, when a single data disk has data failure, performing data failure recovery according to the position of the failed data disk;
fourthly, when a single redundant disk has a data fault, performing data fault recovery according to the type of the fault redundant disk;
wherein the step three comprises the following substeps:
firstly, judging the position of a data disk with data failure, and dividing the situation into the situation that the data failure occurs before
A data disk and a back disk
Two cases of a data disk;
② if the data disk with data failure is positioned in front
In a data disk, before reading
Carrying out exclusive-or calculation on data of an intact data disk and a local redundant disk in each data disk to obtain data which is data of a disk with data failure;
third, if the data disk with data failure is located at the back
In a data disk, then read
Carrying out exclusive or calculation on the data of the intact data disk, the horizontal redundant disk and the local redundant disk in each data disk to obtain data, namely the data of the disk with data failure;
the step four includes the following substeps:
firstly, judging the type of a failed redundant disk, and dividing the type into three conditions;
if the disk with data failure is horizontal redundant disk, then after reading
Performing exclusive-or calculation on data in the data disks and the local redundant disk to obtain data which is the horizontal redundant disk data with data failure;
if the disk with the data failure is the diagonal redundant disk, reading the data in all the data disks to perform exclusive-or calculation, and obtaining the data which is the diagonal redundant disk data with the data failure;
if the data failure occurs to the local redundant disk, before reading
And carrying out XOR calculation on the data disks to obtain data which is the data of the local redundant disk with the fault.
2. The method for data recovery against an EVENODD code single-disk failure of claim 1, wherein the probability of the data disk failing is greater than the probability of the redundant disk failing.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810027966.2A CN108228382B (en) | 2018-01-11 | 2018-01-11 | Data recovery method for single-disk fault of EVENODD code |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810027966.2A CN108228382B (en) | 2018-01-11 | 2018-01-11 | Data recovery method for single-disk fault of EVENODD code |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108228382A CN108228382A (en) | 2018-06-29 |
| CN108228382B true CN108228382B (en) | 2021-08-10 |
Family
ID=62640197
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810027966.2A Active CN108228382B (en) | 2018-01-11 | 2018-01-11 | Data recovery method for single-disk fault of EVENODD code |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108228382B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11038533B2 (en) | 2019-04-25 | 2021-06-15 | International Business Machines Corporation | Expansion for generalized EVENODD codes |
| CN111143108B (en) * | 2019-12-09 | 2023-05-02 | 成都信息工程大学 | Coding and decoding method and device for reducing array code Xcode repair |
| CN111475329B (en) * | 2020-02-25 | 2023-07-18 | 成都信息工程大学 | Method and device for reducing predictive erasure code repair under big data application platform |
| CN111679793B (en) * | 2020-06-16 | 2023-03-14 | 成都信息工程大学 | Single-disk fault rapid recovery method based on STAR code |
| CN113094213B (en) * | 2021-05-07 | 2022-07-05 | 河北工业大学 | GPyramid-Code layout and data recovery and storage method for single-disk failure of layout |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1670682A (en) * | 2004-03-15 | 2005-09-21 | 华为技术有限公司 | A data reintegration method |
| CN1906592A (en) * | 2003-07-14 | 2007-01-31 | 国际商业机器公司 | Redundancy in array storage system |
| CN101770413A (en) * | 2010-01-07 | 2010-07-07 | 杭州华三通信技术有限公司 | Method and equipment for rebuilding redundant disk array |
| CN103744750A (en) * | 2014-01-13 | 2014-04-23 | 杭州华为数字技术有限公司 | Method and device for recovering data |
| CN104765660A (en) * | 2015-04-24 | 2015-07-08 | 中国人民解放军国防科学技术大学 | Single-disc quick recovery method machine based on RAID6 of SSD |
| KR101756136B1 (en) * | 2016-01-22 | 2017-07-26 | 계명대학교 산학협력단 | A fault tolerant method for a hierarchical system |
Family Cites Families (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR100390853B1 (en) * | 2000-06-07 | 2003-07-10 | 차상균 | A Logging Method and System for Highly Parallel Recovery Operation in Main-Memory Transaction Processing Systems |
| TWI310497B (en) * | 2004-02-06 | 2009-06-01 | Hon Hai Prec Ind Co Ltd | System and method for disk fault tolerance |
| US8250427B2 (en) * | 2008-06-18 | 2012-08-21 | Hewlett-Packard Development Company | Selecting erasure codes for a fault tolerant system |
| CN101387975B (en) * | 2008-10-20 | 2010-09-15 | 中科院成都信息技术有限公司 | Magnetic disc array system |
| CN101980137B (en) * | 2010-10-19 | 2012-05-30 | 成都市华为赛门铁克科技有限公司 | Method, device and system for reconstructing redundant array of inexpensive disks |
| CN102023819B (en) * | 2010-12-01 | 2012-08-22 | 北京同有飞骥科技股份有限公司 | Method for constructing double-disk fault tolerance horizontal grouping and parallel access disk array |
| CN102081559B (en) * | 2011-01-11 | 2013-06-05 | 华为数字技术（成都）有限公司 | Data recovery method and device for redundant array of independent disks |
| CN103365739B (en) * | 2013-08-02 | 2016-03-02 | 深圳市瑞耐斯技术有限公司 | A kind of NAND flash memory storage equipment and data reconstruction method thereof |
| CN106155847A (en) * | 2016-08-19 | 2016-11-23 | 浪潮(北京)电子信息产业有限公司 | A kind of disk array |
-
2018
- 2018-01-11 CN CN201810027966.2A patent/CN108228382B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1906592A (en) * | 2003-07-14 | 2007-01-31 | 国际商业机器公司 | Redundancy in array storage system |
| CN1670682A (en) * | 2004-03-15 | 2005-09-21 | 华为技术有限公司 | A data reintegration method |
| CN101770413A (en) * | 2010-01-07 | 2010-07-07 | 杭州华三通信技术有限公司 | Method and equipment for rebuilding redundant disk array |
| CN103744750A (en) * | 2014-01-13 | 2014-04-23 | 杭州华为数字技术有限公司 | Method and device for recovering data |
| CN104765660A (en) * | 2015-04-24 | 2015-07-08 | 中国人民解放军国防科学技术大学 | Single-disc quick recovery method machine based on RAID6 of SSD |
| KR101756136B1 (en) * | 2016-01-22 | 2017-07-26 | 계명대학교 산학협력단 | A fault tolerant method for a hierarchical system |
Non-Patent Citations (1)
| Title |
|---|
| RAID-6 存储系统中故障修复问题研究;项利萍;《中国博士学位论文全文数据库 信息科技辑》;20131015(第10期);I137-14页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108228382A (en) | 2018-06-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108228382B (en) | Data recovery method for single-disk fault of EVENODD code | |
| US9229810B2 (en) | Enabling efficient recovery from multiple failures together with one latent error in a storage array | |
| US7640452B2 (en) | Method for reconstructing data in case of two disk drives of RAID failure and system therefor | |
| US7254754B2 (en) | Raid 3+3 | |
| CN100337209C (en) | Method and means for tolerating multiple dependent or arbitrary double disk failures in a disk array | |
| US8489916B2 (en) | Multi-disk fault-tolerant system, method for generating a check block, and method for recovering a data block | |
| CN109086000B (en) | Three-fault-tolerant data layout method in RAID storage system | |
| US20140047265A1 (en) | Enhanced storage of metadata utilizing improved error detection and correction in computer memory | |
| US7818524B2 (en) | Data migration systems and methods for independent storage device expansion and adaptation | |
| US9063869B2 (en) | Method and system for storing and rebuilding data | |
| CN106874140B (en) | Data storage method and device | |
| CN112000512B (en) | Data restoration method and related device | |
| CN109358980B (en) | RAID6 encoding method friendly to data updating and single-disk error recovery | |
| US20060107091A1 (en) | System and method for recovery of data for a lost sector in a storage system | |
| CN110555055A (en) | data mining method for redo log file of Oracle database | |
| Iliadis | Reliability evaluation of erasure-coded storage systems with latent errors | |
| CN105630417A (en) | RAID5 (Redundant Array Of Independent Disks) system and method for continuously writing data after failure of RAID5 system | |
| CN110600070B (en) | Coding and repairing method for improving repairing performance of solid state disk array system | |
| CN114064346B (en) | Erasure code data consistency guaranteeing method and system | |
| US20050283652A1 (en) | Fault tolerance system and method for one or two failed disks in a disk array | |
| CN104932836A (en) | Three-disk fault-tolerance coding and decoding methods for improving single-writing performance | |
| CN102929736B (en) | A kind of flash memory interweave verification error correction method and flash controller | |
| CN106933707B (en) | Data recovery method and system of data storage device based on raid technology | |
| CN115269258A (en) | Data recovery method and system | |
| CN104156276A (en) | RAID method for preventing two disks from being damaged |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |