CN104714758A - Method for building array by adding mirror image structure to check-based RAID and read-write system - Google Patents
Method for building array by adding mirror image structure to check-based RAID and read-write system Download PDFInfo
- Publication number
- CN104714758A CN104714758A CN201510025251.XA CN201510025251A CN104714758A CN 104714758 A CN104714758 A CN 104714758A CN 201510025251 A CN201510025251 A CN 201510025251A CN 104714758 A CN104714758 A CN 104714758A
- Authority
- CN
- China
- Prior art keywords
- data
- mirror image
- disk
- section
- read
- 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.)
- Granted
Links
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Signal Processing For Digital Recording And Reproducing (AREA)
Abstract
The invention discloses a method for building an array by adding a mirror image structure to a check-based RAID and read-write system. The method comprises the four steps of address layout, data layout, data access and data reconstruction. The read-write system comprises an I/O module, a mirror image management module, an address mapping module and the check-based RAID module. On the basis of the check-based RAID, the mirror image structure is added, so that the I/O performance of the array is improved; the reconstruction time is shortened by using mirror image data reconstruction; the novel address layout mode puts original data and mirror image data at the adjacent positions as much as possible, and the movement distance of a head is shortened; checking information is updated in the background, the write amplification problem of the check-based RAID is relieved, and the write performance of the array is improved; as the mirror image data exist, the fault-tolerant ability of the array is improved, and the availability and reliability of the array are greatly improved.
Description
Technical field
The invention belongs to data storing technical field, more specifically, relate to one and add enantiomorphous array construction method and read-write system based on verification RAID.
Background technology
Redundant Array of Independent Disks (RAID) (Redundant Arrays of Independent Disks, RAID) is widely used because of its performance and reliability in large-scale distributed parallel memory system.In performance, RAID realizes high-throughput by striping, and in reliability, RAID utilizes the redundant data such as mirror image or check code to improve reliability.
RAID can be divided three classes by the difference according to its redundancy structure.One, do not have redundancy structure, representational array is RAID0, RAID0 usage data disk-striping strategy (Data stripping), by data scatter in all disks, making between disk can parallel work-flow, improves read or write speed, owing to not having redundancy structure, its space availability ratio is the highest in array structure, but also therefore cause its reliability low, in array, any one piece of disk breaks down, and whole system all can be caused to be damaged.They are two years old, based on the RAID of verification, representational array is the technology that RAID5, RAID5 have employed date classification equally, data strip blocking is stored in all disks, make disk can parallel work-flow, with RAID0 unlike, each band has a parity values, all parity values are evenly distributed in all disks, its advantage improves reliability, and in array, any one piece of disk breaks down, and system can calculate the data of faulty disk according to remaining disk; Owing to adding redundancy structure, space availability ratio declines to some extent, and its redundancy structure needs the capacity of a disk altogether, meanwhile, due to its redundancy structure, result in the problem writing amplification, an i.e. write operation, by I/O operation actual for generation four, comprises and reads for twice to write with twice; Under reconfiguration mode, the data reading remaining all disks are needed to calculate the data of faulty disk; Under degraded mode, the request of access fault dish is broken down into multiple I/O asks, and this will cause reconstitution time long, and under degraded mode and reconfiguration mode, I/O performance is low.They are three years old, based on the RAID of mirror image, representational array is RAID1, RAID1 is also referred to as disk mirroring, it by the data image of a disk on another disk, because data exist copy in different disk, I/O shunting can be realized, make it access can be evenly distributed in mirror image, and any one piece of disk breaks down in array, data all can not be lost, the request of access fault dish directly can be converted to the request of its mirror drive of access, multiple I/O can not be resolved into, and reconstruct time can directly read failure dish correspondence mirror drive data, reconstitution time is short, I/O performance is high, but its space utilization rate variance, only can use the half of array capacity.
Summary of the invention
For above defect or the Improvement requirement of prior art, the invention provides a kind of elastic data mirror-image system towards check disk array and method, its object is to: on the basis based on verification RAID, add mirror-image structure, improve the fault-tolerant ability of array.
For achieving the above object, according to one aspect of the present invention, provide a kind of and add enantiomorphous array construction method based on verification RAID, it is characterized in that, described array construction method comprises address layout, data layout, data access and data reconstruction four steps, specific as follows:
(1) address layout:
(1.1) set band number in section according to disk number M, in section, band number is the integral multiple of M;
(1.2) according to band number determination hop count K in the number of data block in disk and section, in hop count K=disk data block number/section in band number;
(1.3) N=K/2 is got; The section be section and the segment number of 1 to N by segment number being N+1 to K is intersected and is deposited: deposit the section being numbered N+1 after being numbered the section of 1, deposit the section being numbered 2 afterwards, deposit the section being numbered N+2 after being numbered 2, so intersect and deposit; If K/2 is non-integer, then round;
(2) data layout:
(2.1) data in raw data section section with based on the RAID identical topology verified;
(2.2) mirror image data section is on the basis of raw data segment layout, changes the disk number that in same band, data block is deposited, and disk number is unified to be offset to the right or left; Unification offsets to the right, and namely disk number adds i, if add the disk number after i to be greater than M, then subtracts M; Unification offsets left, and namely disk number subtracts i, if subtract the disk number after i to be less than 0, then adds M;
(3) reading and writing data:
(3.1) physical address intending visit data is calculated; If write request, also want calculation check information physical address;
(3.2) judge whether the data intending access have mirror image; If there is no mirror image, then enter step (3.3), if there is mirror image, then enter step (3.4);
(3.3) disk that read-write requests is pointed to the physical address of data is issued; If write request, also upgrade the check information of raw data;
(3.4) physical address of mirror image data is calculated; If write request, also calculate the physical address of mirror image data checking data, issue the disk of disk pointed by write request to corresponding data physical address and the sensing of mirror image physical address, and upgrade the check information of raw data and mirror image data; If read, the load of balanced raw data place disk and mirror image data place disk, selects load smaller to issue read request;
(4) data reconstruction:
If there is faulty disk, whether the data on failure judgement dish have mirror image, if there is mirror image, read corresponding data write back-up disk from mirror image; If there is no mirror image, then according to the reconstructing method based on verification RAID, read the data write back-up disk that other data block in band calculates faulty disk.
Preferably, when writing data, if array manifold is not enough, then according to the priority elasticity release mirror image data space that user specifies, described elasticity release is specially: when user data is less than or equal to a half of array capacity by the space required for the RAID tissue based on verification, all data have mirrored-redundant; When user data is greater than a half of array capacity by the space required for the RAID tissue based on verification, user can the priority of specific data, the mirror image of the data segment that first Release priority level is low; Along with increasing progressively of user data reaches capacity, array changes in quality into the RAID based on verification.
Preferably, judge whether that the method for mirror image is specially: the content according to the band section figure record in mirror image administration module judges, each section of use 2 bit carrys out recording status, for " 00 " represents idle; " 01 " indicates that data do not have mirror image; " 10 " indicate data and its mirror image data in position adjacent downwards; " 11 " indicate data and these data have mirror image, and mirror position need be searched by Mirroring Mapping table.
Preferably, the mode of described renewal check information has two kinds: one, if user does not require checking data synchronized update, then utilizes the verification delay queue background process in I/O module; Its two, synchronously process, the check information of synchronized update raw data and the check information of mirror image data when writing raw data and mirror image data.
Preferably, when not having mirror image, in described read request implementation, if raw data place dish breaks down, then the method following verification RAID performs to be read; This method following verification RAID read operation is degradation and reads.
Preferably, when there being mirror image, in described read request implementation, if raw data place dish breaks down, directly user's read request is issued to mirror image data place disk.
For realizing object of the present invention, according to another aspect of the present invention, provide and a kind of add enantiomorphous array read-write system based on verification RAID, it is characterized in that, described system comprise I/O module, mirror image administration module, address changing module, based on verification RAID module;
Described I/O module receives upper strata read-write requests, exports the logical address that read-write requests comprises; Mirror image administration module is according to described logical address, and judge the mirror image data of the raw data of searching pointed by described address, outgoing mirror is as data logical address; Address changing module receives the logical address of raw data and mirror image data, exports the physical address of raw data and mirror image data; If be write request, also export the physical address of raw data and mirror image data proof mass; I/O module, according to described physical address, issues read-write requests to corresponding disk; Based on verification RAID module then for when there is no mirror image, when raw data place dish breaks down, perform read request.
Preferably, described I/O module comprises verification delay queue, for the check information of background update data segment.
Preferred further, described verification delay queue easily extensible 8 bits or 16 bits; Queue after expansion, for recording the data block place segment number and corresponding skew bar reel number that need to upgrade verification, realizes the renewal of small grain size.
Preferably, described mirror image administration module comprises band section figure and Mirroring Mapping table, the state of described band section figure record band section, the position of the non-conterminous mirror image of described Mirroring Mapping table record; Search idle section by traversal band section figure and distribute to raw data and mirror image data; If the space that mirror image data distributes is not closed on, then the position of non-conterminous mirror image is recorded in Mirroring Mapping table downwards; Mirror image administration module is for managing space and mirror image data: for raw data allocation space, distribute mirror image; The priority of search corresponding mirror image according to raw data, specifying when insufficient space according to user and array capacity elasticity discharge mirror image data space.
In general, the above technical scheme conceived by the present invention compared with prior art, can obtain following beneficial effect:
(1) owing to adding mirror-image structure on the basis based on verification RAID, utilize the mirror image of free space store data, in read request implementation, mirror image can be utilized to do equilibrium between disk, when degradation is read, directly return the data of mirror image; Can postpone in write request implementation upgrade check information, alleviate based on verification RAID write scale-up problem, improve the readwrite performance of array;
(2) owing to adding mirror-image structure, utilize mirror image data to carry out data reconstruction, shorten the time of data reconstruction;
(3) owing to present invention employs address layout's mode of intersecting and depositing, raw data and mirror image data are placed on contiguous position as much as possible, shorten the distance of magnetic head movement, improve the read or write speed of array;
(4) because I/O module of the present invention introduces delay queue, utilize delay queue to realize the background update of check information, alleviate and write scale-up problem based on verification RAID, improve the write performance of array;
(5) owing to adding mirror-image structure, the existence of mirror image data, the fault-tolerant of array can reach 1 to M/2 disk (M is the sum of disk in array), improves availability and the reliability of array, and improves the performance of system.
Accompanying drawing explanation
What Fig. 1 embodiment of the present invention 1 provided adds enantiomorphous array read-write system block diagram based on verification RAID;
Fig. 2 is the data structure schematic diagram that the embodiment of the present invention 1 provides;
Fig. 3 is RAID5 data layout schematic diagram in the embodiment of the present invention 2;
Fig. 4 be the embodiment of the present invention 2 provide based on verification RAID add data layout's schematic diagram that enantiomorphous array construction method is applied to RAID5;
Fig. 5 is the embodiment of the present invention 2 stage casing address layout schematic diagram;
Fig. 6 is method for writing data process flow diagram in the embodiment of the present invention 2;
Fig. 7 is method for reading data process flow diagram in the embodiment of the present invention 2;
Fig. 8 is data reconstruction method process flow diagram in the embodiment of the present invention 2.
Embodiment
In order to make object of the present invention, technical scheme and advantage clearly understand, below in conjunction with drawings and Examples, the present invention is further elaborated.Should be appreciated that specific embodiment described herein only in order to explain the present invention, be not intended to limit the present invention.In addition, if below in described each embodiment of the present invention involved technical characteristic do not form conflict each other and just can mutually combine.
The one that Fig. 1 shows the embodiment of the present invention 1 to be provided adds enantiomorphous array read-write system based on verification RAID, illustrate only part related to the present invention for convenience of explanation.
The system that embodiment 1 provides, comprises I/O module, mirror image administration module, address changing module and based on verification RAID module;
Described I/O module accepts upper strata read-write requests; Mirror image administration module judges the mirror image data section of the raw data section of searching pointed by address according to the address information in read-write requests, return mirror image data sector address; Logical address is mapped to physical address on disk according to sector address layout and based on the data layout of verification RAID by address changing module, if be write request, and the physical address of the proof mass also needing calculating raw data corresponding with mirror image data; I/O decomposition module read-write requests, is issued to the disk corresponding to physical address; Based on verification RAID module then for when there is no mirror image, when raw data place dish breaks down, perform read request.
The I/O module of embodiment 1 has verification delay queue, as verified shown in delay queue 11 in Fig. 2, for the check information of background update data segment; In ablation process, if user does not require checking data synchronized update, then segment number is recorded in verification delay queue, when treating that system idles or load are not high, the check information of the data block of background update record; Verification delay queue easily extensible 8 bits or 16 bits, the queue after expansion can record the data block place segment number and corresponding skew bar reel number that need to upgrade verification, realizes the renewal of small grain size.
Described mirror image administration module is used for managing space and mirror image data, comprise for raw data allocation space, for mirror image data allocation space, search corresponding mirror image, insufficient space according to raw data time, the priority of specifying according to user and array capacity elasticity discharge mirror image data space, be specially: when user data is less than or equal to a half of array capacity by the space required for the RAID tissue based on verification, all data have mirrored-redundant; When user data is greater than a half of array capacity by the space required for the RAID tissue based on verification, user can the priority of specific data, the mirror image of the data segment that first Release priority level is low; Along with increasing progressively of user data reaches capacity, array changes in quality into the RAID based on verification.
Described mirror image administration module has Mirroring Mapping table and band section figure, as shown in Mirroring Mapping table 21 in Fig. 2 and band section Figure 22, the state of band section figure record segment, a section needs 2 bit recording status, always have four kinds of states, be respectively: " 00 " represents idle; " 01 " indicates data; " 10 " indicate data and its mirror image data in position close downwards, such as band section 2 is adjacent downwards with band section N+2; " 11 " indicate data and these data have mirror image, and mirror position need be searched by Mirroring Mapping table; The described Mirroring Mapping table record position of non-conterminous mirror image, last item is the segment number of raw data band section, and latter one is the segment number of mirror image section.
Described address changing module is for being mapped to physical address on disk by logical address according to sector address layout and based on the data layout of verification RAID.
Described based on the data layout in the verification RAID module section of completing, check code calculating, and when not having mirror image data, degradation is read and data reconstruction.
Add enantiomorphous array construction method based on verification RAID in embodiment 2 to be applied on array read-write system of the present invention, comprise address layout, data layout, data access and data reconstruction four steps, specific as follows:
(1) address layout:
(1.1) set band number in section according to disk number M, in section, band number is the integral multiple of M;
(1.2) according to band number determination hop count K in the number of data block in disk and section, in hop count K=disk data block number/section in band number;
(1.3) N=K/2 is got; The section be section and the segment number of 1 to N by segment number being N+1 to K is intersected and is deposited: deposit the section being numbered N+1 after being numbered the section of 1, deposit the section being numbered 2 afterwards, deposit the section being numbered N+2 after being numbered 2, so intersect and deposit; If K/2 is non-integer, then round;
(2) data layout:
(2.1) data in raw data section section with based on the RAID identical topology verified;
(2.2) mirror image data section is on the basis of raw data segment layout, changes the disk number that in same band, data block is deposited, and disk number is unified to be offset to the right or left; Unification offsets to the right, and namely disk number adds i, if add the disk number after i to be greater than M, then subtracts M; Unification offsets left, namely moves disk number and subtracts i, if subtract the disk number after i to be less than 0, then add M;
(3) reading and writing data:
(3.1) physical address intending visit data is calculated; If write request, also want calculation check information physical address;
(3.2) judge whether the data intending access have mirror image; If there is no mirror image, then enter step (3.3), if there is mirror image, then enter step (3.4);
(3.3) disk that read-write requests is pointed to the physical address of data is issued; If write request, also upgrade the check information of raw data;
(3.4) physical address of mirror image data is calculated; If write request, also calculate the physical address of mirror image data checking data, issue the disk of disk pointed by write request to corresponding data physical address and the sensing of mirror image physical address, and upgrade the check information of raw data and mirror image data; If read, the load of balanced raw data place disk and mirror image data place disk, selects load smaller to issue read request;
(4) data reconstruction:
If there is faulty disk, whether the data on failure judgement dish have mirror image, if there is mirror image, read corresponding data write back-up disk from mirror image; If there is no mirror image, then according to the reconstructing method based on verification RAID, read the data write back-up disk that other data block in band calculates faulty disk.
Verification RAID in the array construction method of embodiment 2 be the data layout of RAID5, RAID5 as shown in Figure 3, Ai, Bi, Ci are data block, and Pi is the check information of the i-th band data, and such as P1 is the check information of A1, B1, C1.
Data layout's method in embodiment 2 is as shown in Figure 4: Ai, Bi, Ci are data block, and Pi is the check information of the i-th band data, and Ai ', Bi ', Ci ', Pi ' are respectively the mirror image of Ai, Bi, Ci, Pi.The Data Placement section of being (section), a section is divided into several band (stripe), and in section, the number of band is specified by user, and in the present embodiment, a section comprises 4 bands.Data layout in raw data section section is identical with RAID5 layout, and mirror image data section, on the basis of raw data segment layout, changes the disk number deposited, and unification offsets 1 to the right, as shown in (a) part in Fig. 4; Or unification offsets 1 left, as shown in (b) part in Fig. 4.Such as, originally leave the data be numbered on m disk in, leave in mirror image and be numbered on the disk of m+1, or leave in and be numbered on the disk of m-1.
In embodiment 2, sector address layout method is specially: suppose there be 2N section, and the section be section and the segment number of 1 to N by segment number being N+1 to 2*N is intersected and deposited.Concrete after being numbered the section of 1, arrangement is numbered the section of N+1 as shown in Figure 5, and arrangement is afterwards numbered the section of 2, and arrangement is afterwards numbered the section of N+2, so intersects addressing.In mirror image administration module, for raw data allocation space was distributed from segment labeling 1 to the 2N sequential search free time, for raw data distributes mirror image space, the band section that priority allocation raw data physical address is close to downwards, such as in Figure 5, band section k is raw data, if write new file again, if band section k+1 is idle, priority allocation band section k+1, if band section k+1 has data, then check that whether band section k+2 is idle, sequential search free space distributes, if be the raw data distribution mirror image of band section k, priority allocation band section N+k, if band section N+k has data, then search idle band section first forward to distribute, if do not had, then search idle band section backward.
In embodiment 2, described user data wiring method is specially: for upper-layer user's write request, if the data that write request is pointed to do not have mirror image, calculates raw data physical address, issues write request and upgrade check information; If the data that write request is pointed to have mirror image, according to data layout and mirror image management, find raw data physical address and mirror image data physical address, issue write request and upgrade check information.The method upgrading check information has two kinds, one, background process, when waiting system idles or load not high, upgrades original checksums information and its mirror image check information; Its two, process simultaneously, upgrade when writing raw data and mirror image data simultaneously original checksums and mirror image verification.
More specifically, as shown in Figure 6, described user data write specifically comprises the steps:
Step S401 is according to data layout and sector address layout calculation raw data physical address and check information physical address thereof;
Step S402 has judged whether mirror image, if there is mirror image, proceeds to step S403; Otherwise, proceed to step S408; It is judge according to the numerical value of the band section figure record in mirror image administration module that described mirror image judges, if be " 10 " or " 11 ", indicates mirror image, if be " 01 ", represents do not have mirror image;
Step S403 calculates the data physical address of mirror image data and the physical address of checking data thereof;
Step S404 issues request to raw data and mirror image data place disk;
Step S405 check information whether synchronized update, if so, proceeds to step S406, otherwise, proceed to step S407;
Step S406 upgrades the check information of raw data and mirror image data;
The system idles such as step S407 or disk load not high time, the check information of background update raw data and mirror image data; If background update, the skew bar reel number in corresponding segment number and section can add verification delay queue 11;
Step S408 issues write request to the disk at raw data place and upgrades the check information of raw data.
Described user data reads process and is specially: for upper-layer user's read request, calculates raw data physical address; If there is mirror image, calculate mirror image data physical address, the disk getting load less issues read request; If there is no mirror image, be directly issued to raw data place disk; When data place dish breaks down, if data have mirror image, calculate the physical address of mirror image data, user's read request is issued to mirror image data place disk; If data do not have mirror image, then basis is based on verification RAID module, and degradation is read.
More specifically, as shown in Figure 7, described user data reads and specifically comprises the following steps:
Step S501 is according to data layout and sector address layout calculation raw data physical address;
Step S502 has judged whether mirror image, if there is mirror image, proceeds to step S503, otherwise, proceed to step S507;
Step S503 calculates the data physical address of mirror image data;
Step S504 has judged whether low-quality disk, if having low-quality disk in data and mirror drive, then proceeds to step S505, otherwise, proceed to step S506;
Step S505 issues the dish of request to mirror image data place;
The load that the balanced raw data of step S506 and mirror image data place are coiled, the dish selecting load lighter issues request;
Step S507 has judged whether low-quality disk, if raw data place dish breaks down, then proceeds to step S509, otherwise, proceed to step S508;
Step S508 issues the disk of request to raw data place;
Step S509 degradation is read, and performs degradation and reads, follow the method for RAID5 in the present embodiment, such as in the diagram, if read block C5, C dish fault, now read block A5, B5, P5, calculate the data of C5 according to the method based on verification RAID.
In embodiment 2, user data reconstructing method is specially: if there is faulty disk, whether the data on failure judgement dish have mirror image, if there is mirror image, corresponding data write back-up disk is read from mirror image, if there is no mirror image, then according to the reconstructing method based on verification RAID, read the data write back-up disk that other data block in band calculates faulty disk.
More specifically, as shown in Figure 8, described reconstructing method specifically comprises the following steps:
Step S601 judges whether data have mirror image, if there is mirror image, proceeds to step S603, otherwise, proceed to step S602;
Step S602 reads the data on the information calculating faulty disk of check information and other normal disks;
Step S603 is from the data of mirror image read failure dish;
Data are write back-up disk by step S604.
The array construction method of embodiment 2, by adding mirror-image structure on the basis based on verification RAID, utilizing the mirror image of free space store data, improve the readwrite performance of array, alleviation writes scale-up problem based on verification RAID's, improves availability and the reliability of array.
Those skilled in the art will readily understand; the foregoing is only preferred embodiment of the present invention; not in order to limit the present invention, all any amendments done within the spirit and principles in the present invention, equivalent replacement and improvement etc., all should be included within protection scope of the present invention.
Claims (10)
1. add an enantiomorphous array construction method based on verification RAID, it is characterized in that, described array construction method comprises address layout, data layout, data access and data reconstruction four steps, specific as follows:
(1) address layout:
(1.1) set band number in section according to disk number M, in section, band number is the integral multiple of M;
(1.2) according to band number determination hop count K in the number of data block in disk and section, in hop count K=disk data block number/section in band number;
(1.3) N=K/2 is got; The section be section and the segment number of 1 to N by segment number being N+1 to K is intersected and is deposited: deposit the section being numbered N+1 after being numbered the section of 1, deposit the section being numbered 2 afterwards, deposit the section being numbered N+2 after being numbered 2, so intersect and deposit; If K/2 is non-integer, then round;
(2) data layout:
(2.1) data in raw data section section with based on the RAID identical topology verified;
(2.2) mirror image data section is on the basis of raw data segment layout, changes the disk number that in same band, data block is deposited, and disk number is unified to be offset to the right or left; Unification offsets to the right, and namely disk number adds i, if add the disk number after i to be greater than M, then subtracts M; Unification offsets left, namely moves disk number and subtracts i, if subtract the disk number after i to be less than 0, then add M;
(3) reading and writing data:
(3.1) physical address intending visit data is calculated; If write request, also want calculation check information physical address;
(3.2) judge whether the data intending access have mirror image; If there is no mirror image, then enter step (3.3), if there is mirror image, then enter step (3.4);
(3.3) disk that read-write requests is pointed to the physical address of data is issued; If write request, also upgrade the check information of raw data;
(3.4) physical address of mirror image data is calculated; If write request, also calculate the physical address of mirror image data checking data, issue the disk of disk pointed by write request to corresponding data physical address and the sensing of mirror image physical address, and upgrade the check information of raw data and mirror image data; If read, the load of balanced raw data place disk and mirror image data place disk, selects load smaller to issue read request;
(4) data reconstruction:
If there is faulty disk, whether the data on failure judgement dish have mirror image, if there is mirror image, read corresponding data write back-up disk from mirror image; If there is no mirror image, then according to the reconstructing method based on verification RAID, read the data write back-up disk that other data block in band calculates faulty disk.
2. array construction method as claimed in claim 1, it is characterized in that, in described step (3), when writing data, if array manifold is not enough, then according to the priority elasticity release mirror image data space that user specifies: when user data is less than or equal to a half of array capacity by the space required for the RAID tissue based on verification, all data have mirrored-redundant; When user data is greater than a half of array capacity by the space required for the RAID tissue based on verification, according to the priority of user's specific data, the mirror image of the data segment that first Release priority level is low; Along with increasing progressively of user data reaches capacity, array changes in quality into the RAID based on verification.
3. array construction method as claimed in claim 1 or 2, is characterized in that, in described step (3), judge whether that the method for mirror image is specially: the content according to band section figure record in mirror image administration module judges, for " 00 " represents idle; " 01 " indicates that data do not have mirror image; " 10 " indicate data and its mirror image data in position adjacent downwards; " 11 " indicate data and these data have mirror image, and mirror position need be searched by Mirroring Mapping table.
4. the array construction method as described in any one of claims 1 to 3, it is characterized in that, upgrade check information in described step (3) and have two kinds of modes: one, if user does not require checking data synchronized update, then utilizes the verification delay queue background process in I/O module; Its two, synchronously process, the check information of synchronized update raw data and the check information of mirror image data when writing raw data and mirror image data.
5. the array construction method as described in any one of Claims 1-4, is characterized in that, in described step (3), performs in read request process, if raw data place dish breaks down, then the method following verification RAID performs to be read.
6. the array construction method as described in any one of claim 1 to 5, is characterized in that, in described step (3), in read request implementation, if raw data place dish breaks down, directly user's read request is issued to mirror image data place disk.
7. add an enantiomorphous array read-write system based on verification RAID, it is characterized in that, described system comprise I/O module, mirror image administration module, address changing module, based on verification RAID module;
Described I/O module receives upper strata read-write requests, output logic address information; Mirror image administration module is according to described logical address, and judge the mirror image data of the raw data of searching pointed by described address, outgoing mirror is as data logical address; Address changing module receives the logical address of raw data and mirror image data, exports the physical address of raw data and mirror image data; If be write request, also export the physical address of raw data and mirror image data proof mass; I/O module, according to described physical address, issues read-write requests to corresponding disk; Based on verification RAID module then for when there is no mirror image, when raw data place dish breaks down, perform read request.
8. array read-write system as claimed in claim 7, it is characterized in that, described I/O module comprises verification delay queue, for the check information of background update data segment.
9. array read-write system as claimed in claim 8, is characterized in that, described verification delay queue easily extensible 8 bits or 16 bits; Queue after expansion, for recording the data block place segment number and corresponding skew bar reel number that need to upgrade verification, realizes the renewal of small grain size.
10. the array read-write system as described in any one of claim 7 to 9, is characterized in that, described mirror image administration module is used for for raw data allocation space and distributes mirror image, comprises band section figure and Mirroring Mapping table; Search idle section by traversal band section figure and distribute to raw data and mirror image data; If the space that mirror image data distributes is not closed on downwards, then the position of non-conterminous mirror image is recorded in Mirroring Mapping table.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510025251.XA CN104714758B (en) | 2015-01-19 | 2015-01-19 | A kind of array construction method and read-write system based on verification RAID addition mirror-image structures |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510025251.XA CN104714758B (en) | 2015-01-19 | 2015-01-19 | A kind of array construction method and read-write system based on verification RAID addition mirror-image structures |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104714758A true CN104714758A (en) | 2015-06-17 |
CN104714758B CN104714758B (en) | 2017-07-07 |
Family
ID=53414143
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510025251.XA Active CN104714758B (en) | 2015-01-19 | 2015-01-19 | A kind of array construction method and read-write system based on verification RAID addition mirror-image structures |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104714758B (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106339276A (en) * | 2016-08-16 | 2017-01-18 | 浪潮(北京)电子信息产业有限公司 | Data backup state-based data recovery method and system |
CN105022673B (en) * | 2015-07-15 | 2018-07-20 | 南京师范大学 | A kind of fault-tolerant fast parallel double calculation method of data-oriented parallel computation |
CN109032513A (en) * | 2018-07-16 | 2018-12-18 | 山东大学 | Based on the RAID framework of SSD and HDD and its backup, method for reconstructing |
CN113407125A (en) * | 2021-08-20 | 2021-09-17 | 苏州浪潮智能科技有限公司 | Method, system and related device for determining block number in RAID6 array |
CN114510379A (en) * | 2022-04-21 | 2022-05-17 | 山东百盟信息技术有限公司 | Distributed array video data storage device |
CN115562594A (en) * | 2022-12-06 | 2023-01-03 | 苏州浪潮智能科技有限公司 | Method, system and related device for constructing RAID card |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1247608A (en) * | 1997-02-27 | 2000-03-15 | 国际商业机器公司 | Transformational raid for hierarchical storage management system |
CN1518697A (en) * | 2001-04-19 | 2004-08-04 | �Ҵ���˾ | Method, apparatus and program for providing hybrid disk mirroring and striping |
KR20050060804A (en) * | 2003-12-17 | 2005-06-22 | 한국전자통신연구원 | Data mirroring system to improve the performance of read operation for large data |
US20100191907A1 (en) * | 2009-01-26 | 2010-07-29 | Lsi Corporation | RAID Converter and Methods for Transforming a First RAID Array to a Second RAID Array Without Creating a Backup Copy |
CN101866307A (en) * | 2010-06-24 | 2010-10-20 | 杭州华三通信技术有限公司 | Data storage method and device based on mirror image technology |
CN102662607A (en) * | 2012-03-29 | 2012-09-12 | 华中科技大学 | RAID6 level mixed disk array, and method for accelerating performance and improving reliability |
CN103761058A (en) * | 2014-01-23 | 2014-04-30 | 天津中科蓝鲸信息技术有限公司 | RAID1 and RAID4 hybrid structure network storage system and method |
US20140304470A1 (en) * | 2013-04-04 | 2014-10-09 | Lsi Corporation | Reverse mirroring in raid level 1 |
CN104281499A (en) * | 2014-10-28 | 2015-01-14 | 苏州工业职业技术学院 | Odd-even check-based RAID (redundant arrays of inexpensive disks) striped mirror data distribution method |
-
2015
- 2015-01-19 CN CN201510025251.XA patent/CN104714758B/en active Active
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1247608A (en) * | 1997-02-27 | 2000-03-15 | 国际商业机器公司 | Transformational raid for hierarchical storage management system |
CN1518697A (en) * | 2001-04-19 | 2004-08-04 | �Ҵ���˾ | Method, apparatus and program for providing hybrid disk mirroring and striping |
KR20050060804A (en) * | 2003-12-17 | 2005-06-22 | 한국전자통신연구원 | Data mirroring system to improve the performance of read operation for large data |
US20100191907A1 (en) * | 2009-01-26 | 2010-07-29 | Lsi Corporation | RAID Converter and Methods for Transforming a First RAID Array to a Second RAID Array Without Creating a Backup Copy |
CN101866307A (en) * | 2010-06-24 | 2010-10-20 | 杭州华三通信技术有限公司 | Data storage method and device based on mirror image technology |
CN102662607A (en) * | 2012-03-29 | 2012-09-12 | 华中科技大学 | RAID6 level mixed disk array, and method for accelerating performance and improving reliability |
US20140304470A1 (en) * | 2013-04-04 | 2014-10-09 | Lsi Corporation | Reverse mirroring in raid level 1 |
CN103761058A (en) * | 2014-01-23 | 2014-04-30 | 天津中科蓝鲸信息技术有限公司 | RAID1 and RAID4 hybrid structure network storage system and method |
CN104281499A (en) * | 2014-10-28 | 2015-01-14 | 苏州工业职业技术学院 | Odd-even check-based RAID (redundant arrays of inexpensive disks) striped mirror data distribution method |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105022673B (en) * | 2015-07-15 | 2018-07-20 | 南京师范大学 | A kind of fault-tolerant fast parallel double calculation method of data-oriented parallel computation |
CN106339276A (en) * | 2016-08-16 | 2017-01-18 | 浪潮(北京)电子信息产业有限公司 | Data backup state-based data recovery method and system |
CN106339276B (en) * | 2016-08-16 | 2019-10-18 | 浪潮(北京)电子信息产业有限公司 | A kind of data reconstruction method and system based on data backup status |
CN109032513A (en) * | 2018-07-16 | 2018-12-18 | 山东大学 | Based on the RAID framework of SSD and HDD and its backup, method for reconstructing |
CN113407125A (en) * | 2021-08-20 | 2021-09-17 | 苏州浪潮智能科技有限公司 | Method, system and related device for determining block number in RAID6 array |
CN113407125B (en) * | 2021-08-20 | 2021-11-09 | 苏州浪潮智能科技有限公司 | Method, system and related device for determining block number in RAID6 array |
CN114510379A (en) * | 2022-04-21 | 2022-05-17 | 山东百盟信息技术有限公司 | Distributed array video data storage device |
CN114510379B (en) * | 2022-04-21 | 2022-11-01 | 山东百盟信息技术有限公司 | Distributed array video data storage device |
CN115562594A (en) * | 2022-12-06 | 2023-01-03 | 苏州浪潮智能科技有限公司 | Method, system and related device for constructing RAID card |
CN115562594B (en) * | 2022-12-06 | 2023-03-24 | 苏州浪潮智能科技有限公司 | Method, system and related device for constructing RAID card |
Also Published As
Publication number | Publication date |
---|---|
CN104714758B (en) | 2017-07-07 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104714758A (en) | Method for building array by adding mirror image structure to check-based RAID and read-write system | |
US9329957B2 (en) | Method for managing storage system using flash memory, and computer | |
US9448886B2 (en) | Flexible data storage system | |
US8677063B2 (en) | Parity declustered storage device array with partition groups | |
US10365983B1 (en) | Repairing raid systems at per-stripe granularity | |
KR101502519B1 (en) | Memory Management System and Method | |
US5301297A (en) | Method and means for managing RAID 5 DASD arrays having RAID DASD arrays as logical devices thereof | |
JP3505093B2 (en) | File management system | |
US20140365726A1 (en) | Memory system management | |
CN102147713B (en) | Method and device for managing network storage system | |
CN107924291B (en) | Storage system | |
US9251059B2 (en) | Storage system employing MRAM and redundant array of solid state disk | |
CN104956309A (en) | Storage system and data management method | |
CN103049222A (en) | RAID5 (redundant array of independent disk 5) write IO optimization processing method | |
US10678470B2 (en) | Computer system,control method for physical storage device,and recording medium | |
CN102164165B (en) | Management method and device for network storage system | |
CN102122235A (en) | RAID4 (redundant array of independent disks) system and data reading and writing method thereof | |
US20200125447A1 (en) | Memory reallocation during raid rebuild | |
CN103207840A (en) | Demoting partial tracks from a first cache to a second cache | |
CN101976181A (en) | Management method and device of storage resources | |
US20190196911A1 (en) | Computer system | |
CN109445681A (en) | Storage method, device and the storage system of data | |
CN101997919B (en) | Storage resource management method and device | |
KR102133316B1 (en) | Memory system management | |
CN102158538B (en) | Management method and device of network storage system |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |