JP4469822B2 - Disk array device and data management method for disk array device - Google Patents

Description

  The present invention relates to a data management technique suitable for application to, for example, a disk array device constituting a redundant array of inexpensive disks (RAID) 6.

  In recent years, with the spread of electronic commerce and the like, demands for data storage devices such as high responsiveness and fault tolerance have been greatly increased. As a technology that meets this requirement, there is a RAID function in which a plurality of disk devices are connected in parallel, and operate as if they were a single disk device.

In particular, in the latest RAID-6, the consistency check is strengthened by using two types of redundant data called parity data. Further, in the disk array device constituting RAID-6, various methods for specifying an illegal data location have been proposed (see, for example, Patent Document 1).
JP 2005-293263 A

  In the disk array device described in Patent Document 1, attention is given to the presence of two types of redundant data, and error data is identified by sequentially solving simultaneous equations with suspicious data as unknowns.

  However, this method is used when all of the plurality of disk devices included in the disk array device are operating normally. More strictly, when the area in the disk device in the failed state is not arranged in the stripe group. Only valid. In other words, if one disk unit is in a failed state, even if an inconsistency can be detected by the consistency check in the stripe group in which an area in the disk unit in the failed state is arranged, the data illegal location Cannot be specified.

  The present invention takes such circumstances into consideration, and for example, even if one of the plurality of disk devices constituting RAID-6 is in a failure state, inconsistency is detected by the consistency check. In such a case, it is an object of the present invention to provide a disk array device and a data management method for the disk array device that make it possible to specify an illegal portion of data.

In order to achieve the above-described object, the present invention provides a disk array device that records data and redundant data of the data in a plurality of disk devices in a distributed manner and is larger than the actual recording unit of the data and redundant data. The recording area of each of the plurality of disk devices is divided into sections by data length, and a redundant area for recording access history information related to writing of the data or the redundant data is secured for each of the plurality of disk devices. A redundant area securing means for the data, and at the time of data writing, for each stripe group using the same redundant data, the redundant area of one or more areas where the data is recorded and the area where the redundant data of the data is recorded in the redundant area, the data or in the entire area constituting the stripe group And the access history information recording means for serial redundant data records with the same access history information including bitmap data indicating the area to be synchronously recorded, the fault condition one disk apparatus in said plurality of disk devices When a mismatch is detected by the consistency check using the redundant data in the stripe group in which the area in the disk device in the failed state is arranged, the redundant area of the area in which the redundant data is recorded The access history information recorded in the redundant area of the area where the fact that the data has been recorded synchronously with the redundant data is read out by the access history information recorded in the bitmap and the bitmap data included in the access history information. by comparison, the characterized by comprising a data invalid portion identifying means for identifying the data invalid location, the That.

Further, the present invention provides a data management method for a disk array device for distributing and recording data and redundant data of the data on a plurality of disk devices, wherein the data is larger than the actual recording unit of the data and redundant data. The recording area of each of the plurality of disk devices is divided into sections by length, and a redundant area for recording access history information related to the writing of the data or the redundant data is secured for each section of the plurality of disk devices. , when writing data, each stripe group using the same redundant data, to the redundant area of zone 1 or 2 or more sections of the redundant area and the data redundancy data the data is recorded is recorded, the The data or the redundant data is recorded synchronously in all areas constituting the stripe group The same access history information including a bitmap data indicating the area to record that, when said plurality of one disk device in the disk device is in failure state, arranged areas in the disk device in this fault state Included in the access history information and the access history information recorded in the redundant area of the area where the redundant data is recorded when inconsistency is detected in the consistency check using the redundant data in the stripe group An illegal data location is identified by reading out and comparing access history information recorded in a redundant area of an area where it is indicated that the data is recorded synchronously with the redundant data by bitmap data. And

  According to the present invention, for example, if an inconsistency is detected in the consistency check even if one of the plurality of disk devices constituting RAID-6 is in a failure state, an illegal data location is detected. Can be specified.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  FIG. 1 shows a configuration example of a disk array device according to the present embodiment. This disk array device is composed of a RAID controller 2 and a plurality of hard disk drives (HDDs) 3 and behaves as if it is one disk device and accepts an access request from the host device 1. The RAID controller 2 configures RAID-4 in RAID-6 format using a plurality of HDDs 3. That is, the RAID controller 2 performs striping with a plurality of HDDs 3 connected in parallel to generate two types of parity data (redundant data) called P parity / Q parity for each stripe group 4A. Kinds of parity data are cyclically distributed and recorded in a plurality of HDDs 3.

  Here, the basic principle of RAID-6 will be described in order to facilitate understanding of the disk array device of the present embodiment. FIG. 2 shows a data arrangement example of RAID-6.

(1) Parity generation of RAID-6 (a) Calculation formula when generating P parity data RAID-6 P parity data, like RAID-5 parity data, is a bit of data stripes of all target stripe groups Calculated by exclusive OR (XOR) operation of units. For example, the P parity data of the stripe group 1 is obtained by the following equation (1).

(B) Q parity data is obtained by multiplying each data stripe of a target stripe group by a Galois field using a coefficient (weight) that is the only data in the stripe group, and XORing each product. To calculate. For example, the Q parity data of the stripe group 1 is obtained by the following equation (2).

(2) RAID-6 parity update RAID-6 P parity data and Q parity data must be updated to maintain consistency when data stripe data is updated with new data. .

(A) Calculation formula for updating P parity data For the calculation of updating P parity data, the same calculation method as that used when generating parity using new data, the old data to be written, the new data, and the old data of P parity are used. There is a method to calculate from. A calculation formula by the latter method when data 1 of the stripe group 1 is written is shown below (formula (3)).

(B) Calculation Formula for Updating Q Parity Data There are two types of calculation methods for Q parity data updating. When data 2 of stripe group 1 is written, the old data and new data to be written are A calculation formula for calculating from the Q parity old data is shown (Formula (4)).

(3) Consistency check of RAID-6 For example, the data and parity data stored in the HDD are read, the parity data is recalculated based on the read data, and whether or not it matches the previously read parity data Confirm.

(A) When the RAID is in a normal state When the RAID is in a normal state, the data is read from all the HDDs, the parity is recalculated, and it is confirmed whether or not it matches the read parity. When inconsistency is detected, recovery is performed based on the policy of the system, such as regenerating parity again.

(A) Consistency confirmation method using P parity data Confirm that the following equation (5) holds.

(A) Consistency confirmation method using Q parity data It is confirmed that the following equation (6) holds.

(B) When the RAID is in a degraded state (one HDD has failed) When the RAID is in a degraded state and the failed HDD is a data stripe, first, the failed HDD data is determined from the P parity and other normal HDD data. Then, the Q parity is recalculated from the restored data and the normal HDD data, and it is confirmed whether or not it matches the read Q parity.

For example, as shown in FIG. 3, assuming that the HDD in which the data 2 of the stripe group 1 is recorded is broken, the consistency check in this case is calculated in the order of the following formulas (7) and (8). Then, it is confirmed that the formula (8) is established.

  As described above, in RAID-6, data consistency can be confirmed even when one HDD is in a reduced state of failure. However, if an inconsistency is detected in this data consistency check, an illegal location with a data error cannot be specified. Therefore, the disk array device according to the present embodiment specifies an illegal data location when an inconsistency is detected in the data consistency check even when one HDD is in a reduced state of failure. Hereinafter, this point will be described in detail.

  As shown in FIG. 1, the RAID controller 2 includes a control unit 21, a consistency confirmation unit 22, a sector redundant area securing unit 23, an access history recording unit 24, an illegal data location specifying unit 25, and a simple consistency confirmation unit 26. is doing. The overall operation control of the RAID controller 2 is governed by the control unit 21, and includes a consistency check unit 22, a sector redundant area securing unit 23, an access history recording unit 24, an illegal data location specifying unit 25, and a simple consistency check. Each unit of the unit 26 operates under the control of the control unit 21.

  The consistency check unit 22 performs the above-described data consistency check using, for example, a free time when there is no data access from the host device 1. That is, the data and parity data stored in the HDD 3 are read, the parity data is recalculated based on the read data, and a process of confirming whether or not it matches the previously read parity data is executed.

  The sector redundant area reservation unit 23 sets the storage capacity (Byte par Sector) per sector of the HDD 3 so as to be larger than the capacity for actually storing valid data, and secures an area to be an unused part. To do. The area that becomes an unused part and can be arbitrarily used is referred to herein as a sector redundant area. For example, by setting the recording capacity per sector of the HDD 3 to a 520-byte format and an apparent format to be 512 bytes per sector, the sector redundant area securing unit 23 secures an 8-byte sector redundant area.

  The access history recording unit 24 records access history information and data protection information (CRC: cyclic redundancy check) in the sector redundancy area secured by the sector redundancy area securing unit 23. The access history information includes the following two types of contents.

(1) Time stamp In accordance with the write request, the same time stamp (time for writing data to the HDD 3) is recorded in all the HDDs 3 to which the target stripe group is simultaneously written.

(2) Bitmap data Bitmap data indicating the positions in the RAID of all HDDs 3 to which the target stripe group is simultaneously written is recorded. Even when data cannot be written because the write destination HDD 3 is out of order, it is reflected in the bitmap so as to indicate that it is the write target HDD 3 (the HDD 3 is replaced later, and when the data is restored, the access history of the sector redundancy area) Because it is used for control when copying information).

  FIG. 4 is a conceptual diagram showing how access history information and data protection information are recorded by the access history recording unit 24 in the sector redundant area secured by the sector redundancy area securing unit 23. The RAID initialization process is a process of generating parity from the data stripe and writing it to the HDD. During this initialization process, the initialization process is performed on the sector redundancy areas of all HDDs of the data stripe and the parity stripe. As a time stamp, the HDDs of all stripe groups record data to be written as bitmap data. Also, when there is a write request for a logical address that has not been initialized during the initialization process, the same access history information as in normal writing is recorded. As the initialization process progresses, the access history information is overwritten by the initialization of the stripe group.

  Based on the access history information recorded in the sector redundancy area, the data illegal location specifying unit 25 determines that the consistency checking unit 22 is not possible when one HDD has been reduced. Realize the location of illegal data when consistency is detected.

  As data is updated (written), P parity data and Q parity data must be updated simultaneously in order to maintain consistency between the data and parity of the target stripe group. Basically, in a data write in a normal RAID state, at least three data stripes, P parity stripes, and Q parity stripes should be updated simultaneously. That is, since the same information should be stored in the sector redundancy areas written in these HDDs 3, the data illegal location specifying unit 25 specifies the data illegal location by paying attention to this point.

  Next, an example of recording access history information (case 1 to case 3) to the sector redundant area associated with data writing will be described with reference to FIGS.

(Case 1: Fig. 5)
When receiving a data write request from the host device 1 (FIG. 5 (1)), the control unit 21 of the RAID controller 2 determines the HDD 3 to be written. Here, it is assumed that write data is written to the HDD (1), and accordingly, the P parity data of the HDD (4) and the Q parity data of the HDD (5) are updated (written).

  In this case, since the access history recording unit 24 determines that the HDD (1), the HDD (4), and the HDD (5) are write target HDDs, the access history recording unit 24 generates bitmap data “10011”. Further, the access history recording unit 24 acquires time from hardware resources such as a built-in clock module, and prepares a time stamp.

  In parallel with this, the RAID controller 2 reads the old data for updating the P parity data / Q parity data (FIG. 5 (2)) and recalculates the P parity data / Q parity data (FIG. 5 ( 3) and (4)) are performed. In the example of FIG. 5, the calculation method using the write target old data and new data among the two calculation methods described above is used.

  When the write data and P parity data / Q parity data are written to the HDD (FIG. 5 (5)), the access history recording unit 24 writes the time stamp and bitmap data to the sector redundant area. Do it at the same time. As a result, the sector redundancy area of each sector of HDD (1), HDD (4), and HDD (5) to which the write data and P parity data / Q parity data are written is as shown in FIG. The same access history information is recorded. When writing the time stamp and bitmap data to the sector redundancy area, the access history recording unit 24 calculates the CRC of the write data and P parity data / Q parity data and writes the CRC to each sector redundancy area. (The same applies hereinafter).

(Case 2: Fig. 6)
In this case, write data is written to the HDD (1) and HDD (2), and accordingly, the P parity data of the HDD (4) and the Q parity data of the HDD (5) are updated (written). Therefore, the access history recording unit 24 generates bitmap data “11011”.

  In the example of FIG. 6, P parity data / Q parity data is recalculated by using the same calculation method as that for parity generation using new data. For this reason, data reading from the HDD (3) is performed, but data writing to the HDD (3) is not performed, so that access history information is not recorded. Therefore, sector redundancy of each sector of HDD (1), HDD (2), HDD (4), and HDD (5) to which write data and P parity data / Q parity data are written, excluding this HDD (3). The same access history information as shown in FIG. 6B is recorded in the area.

(Case 3: Fig. 7)
In this case, as in the case 1 shown in FIG. 5, the write data is written to the HDD (1), and accordingly, the P parity data of the HDD (4) and the Q parity data of the HDD (5) are updated ( Is written). However, since the HDD (1) is in a failure state, the HDD (1) cannot be written.

  For this purpose, as shown in FIG. 7, the P parity data / Q is not calculated by using the old data and new data to be written, but by the same calculation as that at the time of parity generation using new data. Parity data is recalculated, and writing of only the P parity data / Q parity data is performed on the HDD (4) and the HDD (5). Therefore, the writing to the HDD (1) is not actually performed.

  Even in this case, the access history recording unit 24 creates bitmap data “10011” on the assumption that the HDD (1) should be written. Since the HDD (1) is in a failure state, the same access history information as shown in FIG. 7B is recorded in the HDD (4) and the HDD (5).

  Next, referring to FIG. 8, in the disk array device of this embodiment in which the access history information is recorded in this way, the data illegal location specifying unit 25 is a consistency checking unit when one HDD is reduced. A description will be given of how data illegal portions are identified when 22 detects inconsistencies.

  The illegal data location specifying unit 25 first reads the access history information of the P parity and the Q parity (step A1) and checks whether they match (step A2). If they match (Yes in step A2), the illegal data location specifying unit 25 refers to the parity stripe bitmap and reads the access history information from the target simultaneously written HDD 3 (step A3). . Then, the illegal data location specifying unit 25 compares the access history information of the data HDD and the access history information of the parity HDD (step A4), finds the HDD 3 having a contradiction where the information does not match, and corrects the data using the RAID function ( Step A5).

  On the other hand, if the access history information of the P parity and the Q parity do not match (No in step A2), the data illegal location specifying unit 25 firstly based on the bitmap data of the access history information of the P parity. Access history information is read from the simultaneously written HDD 3 to be compared and compared (step A6). Here, the illegal data location specifying unit 25 classifies the comparison result A if “no comparison target HDD”, the comparison result B if “all match”, and the comparison result C if “no match”. (Step A7).

  Secondly, the illegal data location specifying unit 25 reads and compares the access history information from the simultaneously written HDD 3 based on the bitmap data of the Q parity access history information (step A8). Here again, the data illegal location specifying unit 25 classifies the comparison result D if “no comparison target HDD”, comparison result E if “all match”, and comparison result F if “no match”. (Step A9).

  Then, based on the combination of the classification of the first comparison result and the classification of the second comparison result, the data illegal location specifying unit 25 determines and corrects the data illegal location as shown in FIG. 9 ( Step A10).

  As described above, the disk array device according to the present embodiment secures the sector redundancy area by the sector redundancy area securing unit 23 and records the access history information by the access history recording unit 24 (in this secured sector redundancy area). The data illegal location specifying unit 25 realizes specification of the data illegal location when the consistency checking unit 22 detects inconsistency when the HDD is in a reduced state of failure, which was impossible until now.

  By the way, as described above, as an application of the access history information, an example has been described in which it is used for specifying an illegal data portion when a data inconsistency is detected by the consistency check unit 22. However, the access history information is not limited to this, and can be used for verification at the time of data write, for example. For this purpose, the disk array device of this embodiment includes a simple consistency confirmation unit 26.

  The simple consistency check unit 26 operates in the process of write processing, and verifies whether or not the write processing has been performed normally without executing high load calculation using P parity or Q parity. . With reference to FIG. 10, the procedure of the simple consistency confirmation which this simple consistency confirmation part 26 performs is demonstrated.

  The simple consistency confirmation unit 26 first reads both parity stripes and confirms whether the access history information matches (Yes in Step B1, Step B2). If they do not match (No in step B3), it is determined that a mismatch has occurred at this point (step B8).

  On the other hand, if the access history information matches (Yes in step B3), the simple consistency check unit 26 refers to the bitmap data, reads the access history information from the HDD to be simultaneously written, and information on the parity stripe. (Step B4, step B5). If they do not match (No in step B5), it is determined that a mismatch has occurred (step B8).

  If they match (Yes in step B5), the simple consistency check unit 26 checks the data itself using the data protection information (CRC) (step B6), and if an error is detected (step B7). No), it is determined that a mismatch has occurred (step B8).

  Therefore, the data write procedure executed by the disk array device of this embodiment is as shown in FIG.

  When a data write request is received from the host device 1 (step C1), it is determined which address (sector) of which HDD should be written, and a bitmap indicating the positions of all HDDs to be written in the target stripe group is generated. (Step C2) The data writing time (time stamp) is obtained from the built-in hardware or the like (Step C3).

  Next, if necessary, old data is read from each HDD (step C4), parity update data is calculated (step C5), and CRC of write data and parity update data is calculated (step C6). Then, data is set in the buffer area so that the host request write data is stored in the data area of the target sector, and the time stamp, bitmap data, and CRC are stored in the redundant area (step C7), and a write request is issued to the HDD. (Step C8).

  When data writing to all write target HDDs is completed (YES in step C9), a simple consistency check using the access history information shown in the detailed procedure in FIG. 10 is performed (step C10). If it is determined that it is inconsistent (YES in step C11), the write process is re-executed (step C12). If it is determined that it is consistent (NO in step C11), a data write completion response is returned to the host device 1 (step S11). C13).

  If inconsistency can be detected by simple consistency check at the time of data writing, the data to be written exists in the RAID controller 2, so that recovery is possible in a short time. In addition, the simple consistency check itself does not require a parity calculation, so that the responsiveness is not greatly reduced. Furthermore, if this access history information match confirmation is performed by hardware, high-speed processing is possible.

  Next, a description will be given of a procedure for copying access history information when a single HDD is restored from a reduced state of failure to a normal state.

  When the HDD fails, it is possible to recover from the failed state by restoring the failed HDD data by the RAID function by replacing the HDD or using a hot spare HDD and storing the data in the replaced HDD. Therefore, in the disk array device of this embodiment, the access history information of the sector redundant area is also copied at the time of this data restoration. FIG. 12 is a flowchart showing this copy procedure.

  First, the P or Q accessible parity stripe is read, and the bitmap data of the access history information is copied (Yes in step D1, step D2). If the data restoration destination HDD is the target of the write destination HDD in the bitmap information (Yes in step D3), the time stamp of the parity stripe is also copied (step D4). On the other hand, if it is not the target of the write destination HDD (No in step D3), the time stamp is stored as zero (step D5).

  If both parity stripes cannot be read due to a failure or the like (No in step D1), the latest data stamp access history information time stamp is found and the information is copied (step D6). If the data restoration destination HDD is the target of the write destination HDD in the bitmap information (Yes in step D7), the time stamp also copies the time stamp of the data stripe (step D8), If it is not the target of the write destination HDD (No in step D7), the time stamp is stored as zero (step D9).

  As a result, the access history information of the sector redundancy area is also restored at the time of restoring the valid data. The data protection information (CRC) is recalculated from each restored valid data.

  As described above, according to the disk array device of the present embodiment, data fraud due to data write omission can be detected using the time stamp and bitmap data recorded in the sector redundancy area. Even if one of the disk devices is in a failed state, if an inconsistency is detected by the consistency check, it is possible to specify an illegal data portion. Furthermore, it is possible to detect data fraud due to garbled bits using the CRC recorded in the sector redundancy area.

  In the above, RAID-6 format RAID has been described. However, the technique of securing sector redundancy areas and recording access history information can also be applied to RAIDs other than RAID-6 format.

  In addition, the consistency check unit 22 that periodically performs data consistency check using, for example, free time when there is no data access from the host device 1 re-creates parity data from the read data as described above. In general, the consistency is confirmed by calculating and comparing, but in the disk array device of the present embodiment, the time stamp and the bit recorded in the sector redundancy area as in the simple consistency confirmation unit 26. It is also possible to efficiently check the consistency by using the map data and the CRC (that is, the consistency checker 22 is not provided, and the simple consistency checker 26 can also periodically check the consistency. Possible). In particular, the bitmap data is more useful for checking the periodic consistency than the verification at the time of data writing. Which HDD 3 has been written is stored in the memory of the RAID controller 2 immediately after the data write, and can be determined using this, but (the data write has already occurred in the past. This is because it can be determined because the bitmap data exists in the periodic consistency check.

  As described above, the present invention is not limited to the above-described embodiment as it is, and can be embodied by modifying the constituent elements without departing from the scope of the invention in the implementation stage. In addition, various inventions can be formed by appropriately combining a plurality of components disclosed in the embodiment. For example, some components may be deleted from all the components shown in the embodiment. Furthermore, you may combine a component suitably in different embodiment.

1 is a diagram showing a configuration example of a disk array device according to an embodiment of the present invention. The figure which shows the data arrangement example of RAID-6 The figure for demonstrating the data consistency check in case a RAID is in a degenerate state (one HDD has failed) The conceptual diagram which shows a mode that access history information is recorded by the access history recording part in the sector redundant area ensured by the sector redundant area securing part in the disk array apparatus of the embodiment. A diagram showing a recording example (case 1) of access history information to a sector redundant area accompanying a data write executed by the disk array device of the same embodiment A diagram showing a recording example (case 2) of access history information to a sector redundant area accompanying a data write executed by the disk array device of the same embodiment A diagram showing a recording example (case 3) of access history information to a sector redundant area accompanying a data write executed by the disk array device of the same embodiment A flowchart executed by the disk array device according to the embodiment, showing a procedure for identifying an illegal data portion when an inconsistency is detected when one HDD is in a reduced state of failure. The figure for demonstrating the determination and correction of a data illegal location by the disk array apparatus of the embodiment A flowchart showing a procedure of simple consistency check executed by the disk array device of the same embodiment A flowchart showing a data write procedure executed by the disk array device of this embodiment A flowchart showing a procedure of copying access history information of a sector redundant area at the time of data restoration executed by the disk array device of the same embodiment

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Host device, 2 ... RAID controller, 3 ... HDD, 4 ... RAID, 21 ... Control part, 22 ... Consistency confirmation part, 23 ... Sector redundant area reservation part, 24 ... Access history recording part, 25 ... Data illegal location Specific part, 26... Simple consistency confirmation part.

Claims (11)

In a disk array device that records data and redundant data of the data in a plurality of disk devices in a distributed manner,
Redundancy for recording access history information related to writing of the data or the redundant data by dividing the recording area of each of the plurality of disk devices with a data length larger than the actual recording unit of the data and the redundant data Redundant area securing means for securing an area for each area of each of the plurality of disk devices;
When writing data, each stripe group using the same redundant data, to the redundant area of zone 1 or 2 or more sections of the redundant area and the data redundancy data the data is recorded is recorded, the stripe Access history information recording means for recording the same access history information including bitmap data indicating an area in which the data or the redundant data is recorded synchronously among all the areas constituting the group ;
When one of the plurality of disk devices is in a failed state, inconsistency is detected in the consistency check using the redundant data in the stripe group in which the area in the failed disk device is arranged. When detected, the fact that the data was recorded synchronously with the redundant data by the access history information recorded in the redundant area of the area where the redundant data is recorded and the bitmap data included in the access history information By reading and comparing the access history information recorded in the redundant area of the area indicated by, the data illegal location specifying means for specifying the data illegal location,
A disk array device comprising:   2. The disk array device according to claim 1, wherein the access history information recording means records a time stamp as access history information in a redundant area of each section.   2. The disk array device according to claim 1, wherein the access history information recording means further records data protection information for detecting an error in the data or the redundant data in a redundant area of each section. 4. The disk array device according to claim 3, wherein the access history information recording means records CRC (cyclic redundancy check) as data protection information in a redundant area of each section. RAID (redundant array of inexpensive disks) disk array device according to claim 1, wherein that it is an shall make up the 6.   Immediately after the data is written, it comprises simple consistency confirmation means for confirming whether or not all the access history information of the redundant area of the area where the data and the redundant data of the data are recorded is the same. The disk array device according to claim 1. 7. The disk array device according to claim 6 , further comprising a control unit that re-executes data writing when a mismatch in access history information is detected by the simple consistency checking unit. A data management method for a disk array device that records data and redundant data of the data in a distributed manner on a plurality of disk devices,
Redundancy for recording access history information related to writing of the data or the redundant data by dividing the recording area of each of the plurality of disk devices with a data length larger than the actual recording unit of the data and the redundant data An area is secured for each area of each of the plurality of disk devices,
When writing data, each stripe group using the same redundant data, to the redundant area of zone 1 or 2 or more sections of the redundant area and the data redundancy data the data is recorded is recorded, the stripe Record the same access history information including bitmap data indicating areas where the data or the redundant data is recorded synchronously among all the areas constituting the group ,
When one of the plurality of disk devices is in a failed state, inconsistency is detected in the consistency check using the redundant data in the stripe group in which the area in the failed disk device is arranged. When detected, the fact that the data was recorded synchronously with the redundant data by the access history information recorded in the redundant area of the area where the redundant data is recorded and the bitmap data included in the access history information By identifying and comparing the access history information recorded in the redundant area in the area indicated by
A data management method for a disk array device. 9. The data management method for a disk array device according to claim 8 , wherein data protection information for detecting an error in the data or the redundant data is further recorded in a redundant area of each section. Said disk array device, a data management method for a disk array device according to claim 8, wherein a constitutes a RAID6. Immediately after the data is written, it is checked whether all access history information in the redundant area of the area where the data and the redundant data of the data are recorded matches, and if a mismatch of the access history information is detected, the data To re-write
9. A data management method for a disk array device according to claim 8, wherein:

Images