JPH11259991A - Device and method for recording/reproducing data, and av server - Google Patents

Abstract

PROBLEM TO BE SOLVED: To detect and correct a data error even during reallocation, performing reallocation without interrupting the operation of data recording/ reproducing, and to perform data recording/reproducing suited to real time processing. SOLUTION: This device 11 generates a plurality of divided data by dividing input data D1 by a data distributor 22, generates redundant code data P1 and P2 based on Reed Solomon Code by a redundant code generator 26, and these are recorded by HDDs 211 to 21k , 281 and 282 . In reallocation, a CPU 33 excludes HDD for performing reallocation from the recording/reproducing targets of the divided data, and performs reallocation for the HDD for reallocation.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data recording / reproducing apparatus and method for recording and reproducing video data and audio data, and an AV (audio / video) server.

[0002]

2. Description of the Related Art In recent years, with the spread of information channels due to the spread of CATV (cable television) and the like, unlike conventional VTRs (video tape recorders),
Multiple video / audio data from one video / audio data recording / reproducing device
There is an increasing demand for simultaneously reproducing audio data. In order to satisfy this demand, a device called a video server that records and reproduces video and audio data using a random access recording and reproduction medium such as a hard disk is becoming widespread.

In general, for example, in a video server in a broadcasting station, a required data transfer rate is high and a large capacity is required for recording long-time data due to requirements for image quality and sound quality. There is. Therefore,
Attempt to increase the data transfer rate and increase the capacity by using a data recording / reproducing device that includes a plurality of hard disk devices that can store video and audio data and operate in parallel, and record parity data Thus, an attempt has been made to ensure reliability even if any hard disk device fails. As a result, even when the number of channels required varies depending on the content of the program or the broadcasting format that the broadcasting station intends to provide, a plurality of material data are recorded in a distributed manner, and simultaneous transmission of multiple channels is performed. Or a multi-channel video server capable of supporting various usage modes, such as constructing a near video on demand (NVOD) system by reproducing the same material data on multiple channels with different reproduction times. be able to.

A data recording / reproducing apparatus used in such a multi-channel video server includes Patt in 1988.
R proposed in a paper published by erson et al.
AID (Redundant Arrays of Inexpensive Disks) technology is used. In the above paper, RAID is R
It is classified into five from AID-1 to RAID-5. Among them, the typical ones are RAID-1, RAID
-3 and RAID-5. RAID-1 is a method of writing the same contents to two hard disks.

[0005] RAID-3 is a system in which input data is divided into fixed lengths, recorded on a plurality of hard disk drives, and parity data is generated and written into another hard disk drive.

FIG. 11 is a block diagram showing an example of the configuration of a data recording / reproducing apparatus using RAID-3. The data recording / reproducing apparatus 101 includes a plurality of hard disk devices (hereinafter, referred to as HDDs) 102 _{1 to} 102 _N for recording input data (N is an integer value of 2 or more).
When, the HDD109 for recording parity data P as a redundant code, and divides the input data D _I to length to generate a plurality of divided data, each HDD each divided data
A data distributor 106 for distributing to 102 _{1 to} 102 _N ;
Parity generator 107 for generating parity data P from the divided data output from data distributor 106
An input memory 104 _{1 to} 104 _N for temporarily storing each divided data output from the data distributor 106, an input memory 108 for temporarily storing parity data P output from the parity generator 107, Each HDD
102 _{1 to} 102 _N , connected to the input memory 1
Recording the data held by the HDDs 104 _{1 to} 104 _N and 108 in the HDDs 102 ₁ to 102 _N and 109 and the HDD 1
Controllers 103 _{1 to} 103 _N , 110 for controlling reproduction of data from 02 _{1 to} 102 _N , 109;
Output memories 105 _{1 to} 105 _N for temporarily holding data read from the DDs 102 _{1 to} 102 _N , 109;
111, an error corrector 112 that performs error detection and error correction based on the data held by the output memories 105 _{1 to} 105 _N and 111 and error information to be described later to repair the divided data, and an error corrector 112 output data by multiplexing of the data multiplexer 113 to output as the output data D _o, C which controls the entire these devices
PU114.

Next, a data write operation of the data recording / reproducing apparatus 101 will be described. Input data D _I is,
The divided data is input to the data distributor 106 to generate a plurality of divided data. Each divided data is stored in the input memories 104 _{1 to} 104 _{1
The data} is distributed to _N , recorded once, and input to the parity generator 107. At this time, as a data distribution method, for example, the data D1 is distributed to the _first HDD 102 Then, the data D2 is distributed to the _second HDD 1022, and so on.
Is distributed to the last Nth HDD 102 _N , there is a method of again distributing data sequentially from the first HDD 102 ₁ .

[0008] The parity generator 107 is a data distributor 1
The parity data P is generated and output based on the divided data output from the data 06. The input memory 108 records the parity data P once. After that, each HDD 1
02 _{1 to} 102 _N , 109 controllers 103 ₁ to 1
03 _N, 110, under control of the CPU 114, reads the divided data and the parity data P from the input memory 104 ₁ -104 _N, 108, each HDD 102 ₁
Write to 102 _N , 109.

Next, the data reading operation of the data recording / reproducing apparatus 101 will be described. Each controller 10
31 _{1 to} 103 _N , 110 are HDDs 102 _{1 to} 102, respectively.
_N , 109, the divided data and the parity data P are read out, and the output memories 105 _{1 to} 105 _N , 11 are respectively read.
Write to 1. At this time, the HDDs 102 _{1 to} 102 _N ,
An error occurred in the data read operation at 109 (hereinafter, referred to as
This is called a read error. ) Occurs, error information indicating that fact is stored in the HDDs 102 _{1 to} 102 _N , 10 _N.
9 is sent to the controllers 103 _{1 to} 103 _N and 110 as status data from the controller 103 ₁ , and further sent from the controllers 103 _{1 to} 103 _N and 110 to the CPU 114.
To, error information Er ₁ to Er _N, sent as Er _p.

The data recorded in the output memories 105 _{1 to} 105 _N and 111 are synchronized with each other, and
12 is output. At this time, if a read error has occurred, error information indicating that
14 to the error corrector 112. This error information includes information for specifying the HDD in which the read error has occurred. The error corrector 112 restores the divided data using the error information and the parity data P, and outputs the restored data to the data multiplexer 113. Note that the error corrector 11
2 can recover data only when there is a read error in one HDD. If there is a read error in a plurality of HDDs, the error corrector 112
Can detect errors, but cannot repair data. The data multiplexer 113 rearranges the divided data output from the error corrector 112 into the original data sequence, and outputs it as output data _Do to the outside.

On the other hand, in RAID-5, the unit of data division (block) is increased, one divided data is recorded as a data block in one HDD, and each H
In this system, the result (parity data) of the exclusive OR of the DD corresponding data blocks is recorded as a parity block in another HDD, and the parity block is distributed to all HDDs.

FIG. 12 is a block diagram showing an example of the configuration of a data recording / reproducing apparatus using RAID-5. The data recording and reproducing apparatus 201 includes a plurality of HDD 202 ₁ to 202 _N for recording input data (N is 2 or more integer), the input data D _I and HDD 202 ₁ to 202 _N
Parity generation and error correction for generating parity data based on the data recorded in the HDD and performing error correction based on the data read from each of the HDDs 202 _{1 to} 202 _N and error information to restore the data. Vessel 21
2, input memories 204 _{1 to} 204 _N for temporarily holding output data of the parity generation / error corrector 212,
Connected to each HDD 202 ₁ to 202 _N, the data held by the input memory 204 ₁ to 204 _N H
Recording on the DD 202 _{1 to} 202 _N and the HDD 202 ₁ to 2
Controlling the reproduction of data from the 02 _N controller 2
03 _{1 to} 203 _N and an output memory 20 for temporarily holding data read from each of the HDDs 202 _{1 to} 202 _N.
And 5 ₁ ~205 _N, CPU2 for controlling the whole of these devices
14 is provided. Note that the HDDs 202 _{1 to} 202 _N
When a read error occurs in the HDD 202 _{1 to} 202 _N , the control unit in the HDDs 202 _{1 to} 202 _N sends the error information as status data to the controller 203 _1.
To 203 _N , and further to controllers 203 ₁ to 203 _N
03 _N to the CPU 214, the error information Er ₁ to Er _N
Sent as

Next, a data write operation of the data recording / reproducing apparatus 201 will be described. For example, HDD 202
When writing data D in the address A in the _1, the parity data P corresponding to the data D is to have been recorded in the HDD 202 _2, CPU 214 is HDD 202 ₁
Data D ₁ already read from the HDD
The controller 203 ₁ , 203 ₂ is controlled so that the parity data P is read from 202 ₂ . At this time, the parity generation / error corrector 212 calculates the exclusive OR of the data D ₁ and the parity data P, and restores the parity data P ₁ without the data D ₁ . Thereafter, the parity generation / error corrector 212 calculates an exclusive OR of the data D and the parity data P ₁ to obtain new parity data P ₂ . The CPU 214 sets the data D to H
DD202 ₁ and parity data P ₂ is stored in HDD
Controller 203 ₁ , 20 so as to write to 202 ₂
To control the 3 _2.

Next, the data reading operation of the data recording / reproducing apparatus 201 will be described. For example, HDD 202
To read data D from address A at ₁
PU214 controls the controller 203 ₁ to read the data D from the HDD 202 _1, this time, if the read error occurs, the data D read out from the HDD 202 _1, output memory 205 ₁ and a parity generation and error corrector The parity generator / error corrector 212 is controlled so as to output as the output data _Do via the 212. At this time, the parity generation / error correction unit 212
No particular processing is performed.

On the other hand, when the data reading of the data recording / reproducing apparatus 201 is not performed normally,
202 was attempting to read the data D from an address A to _1, if it can not be read data D by the defective sector or the like, CPU 214 receives the error information Er ₁ from the controller 203 _1. In this case, C
The PU 214 reads data from the corresponding addresses of the other HDDs 202 _{2 to} 202 _N and sends the data to the parity generation / error correction unit 212. Based on these data, the parity generation / error correction unit 212 reproduces the data D, The parity generator / error corrector 212 is controlled so as to output as the output data _Do.

As described above, in the data recording / reproducing apparatus 201 using RAID-5, when input data is written, it is necessary to read and write a data block and read and write a parity block. The number increases. Further, when an error occurs during data reading, data is read from another HDD and data is restored, so that the number of accesses also increases. Therefore, the data recording / reproducing apparatus 201 using RAID-5 is suitable for a process of randomly accessing a logical block of a predetermined size, but is not suitable for a process requiring real-time properties.

On the other hand, in the data recording / reproducing apparatus 101 using RAID-3, writing of input data is possible by one access, and error correction after data reading can be performed immediately. Therefore, the data recording / reproducing apparatus 101 using RAID-3 is suitable for processing for recording and reproducing data at high speed. Therefore, a data recording / reproducing device using RAID-3 is suitable for a device requiring real-time properties such as a multi-channel video server.

[0018]

On the hard disk, a plurality of tracks are set concentrically, and the tracks are radially divided to set a plurality of sectors as data recording units. In some of these sectors, there is a sector in which an error always occurs when writing or reading data. Such a sector is called a defective sector, and it is considered that data cannot be read or written correctly due to physical damage or the like. In preparation for such a defective sector, there is a method in which a spare sector is reserved in advance on the hard disk, and data is recorded in the spare sector instead of the defective sector as necessary. Such a spare sector is called a replacement sector. In the HDD having the replacement sector, the control unit in the HDD is configured to provide a logical sector number (LBA) such that even if the replacement sector is used instead of the defective sector, the HDD can be referred to by the same sector number from a higher rank.
Has a correspondence table for associating the LBA with the physical sector number in the recording area on the hard disk when a defective sector occurs. ).

The re-allocation process as described above requires a relatively long time (several seconds), and has not been generally performed during the operation of the data recording / reproducing apparatus. However,
In a device such as a multi-channel video server that requires real-time processing, it is not preferable that the operation of the device is interrupted by the re-allocation process.

Therefore, as described below, it is conceivable to perform the reallocation processing during the operation of the data recording / reproducing apparatus. Hereinafter, the operation of the reallocation processing when the data recording / reproducing apparatus 101 using RAID-3 shown in FIG. 11 performs the reallocation processing during operation will be described.

Here, it is assumed that the data recording / reproducing apparatus 101 has been instructed by a higher-level apparatus to perform a re-allocation process. Upon receiving this instruction, the CPU 114
The write operation and read operation in the HDD to be reallocated are stopped. Next, the CPU 114 specifies the sector to be reallocated to the HDD to be reallocated, and starts the reallocation process. As described above, the reallocation processing is to change the correspondence between the LBA and the physical sector number. When the CPU 114 receives a command instructing writing during the reallocation processing, the CPU 114 causes the HDD other than the HDD to perform reallocation to perform a write operation and performs H reallocation.
The DD and the LBA to which writing has been instructed are stored.
When receiving a command instructing reading during the reallocation processing, the CPU 114 performs reallocation.
A read operation is performed in an HDD other than the DD, and reallocation is performed to the error corrector 112.
An instruction is issued to invalidate data from the DD and perform error correction. When the reallocation processing is completed, the CPU 114 releases the suspension of the write operation and the read operation in the HDD to be reallocated.

As described above, after performing the reallocation processing, the HDD and the address for which the reallocation processing has been performed,
The data needs to be restored for the HDD and the address stored during the stop of the write operation in the reallocation processing. Such a recording medium (hard disk)
Is referred to as a partial rebuilding process (partial Rebuild) in the present application.

Hereinafter, the operation of the partial reconstruction process will be described. Here, it is assumed that the CPU 114 stores the HDD and the address (sector) to be subjected to the partial rebuilding process, and performs the partial rebuilding process on the HDD and the address. In this partial reconstruction process, first, C
The PU 114 specifies an address at which the partial reconstruction process is to be performed, and causes each of the controllers 103 _{1 to} 103 _N and 110 to perform a read operation. In response to this, each of the controllers 103 _{1 to} 103 _N and 110 communicates with each HDD 10
The data at the designated address is read out from 2 _{1 to} 102 _N , 109. Each of the read data passes through the output memories 105 _{1 to} 105 _N and 111, and passes through the error corrector 11.
2 is input. At this time, the CPU 114 requests the error corrector 112 to perform the partial rebuilding process on the HDD.
Is instructed not to use the data read from the. The error corrector 112 includes output memories 105 ₁ to 105 ₁ .
Of the data output from the 105 _N and 111, data other than the data read from the HDD to be subjected to the partial rebuilding process is used to restore the divided data, and the restored divided data is copied to the data multiplexer 113. Output to The data multiplexer 113 rearranges the divided data output from the error corrector 112 into the original data sequence, and outputs it as output data _Do. Next, the control of the CPU 114, data distributor output data D _o from the data multiplexer 113 1
Type 06, by performing the same writing operation and writing the input data D _I, to perform the partial reconstruction process HDD
Then, the restored divided data is written to the sub-program, and the partial reconstruction process ends.

However, in the data recording / reproducing apparatus 101 using RAID-3, the apparatus becomes incapable of detecting and correcting data errors during the reallocation processing. Has been difficult to realize because the reliability is significantly reduced.

In the partial rebuilding process, data can be restored during the rebuilding process only when correct data can be read from all HDDs other than the HDD to be rebuilt. If at least one of the HDDs other than the HDD to be reconstructed has a read error, the error corrector 112 can detect the error but cannot recover the data. That is, RAI
In the data recording / reproducing apparatus 101 using the D-3, the apparatus becomes incapable of detecting and correcting data errors during the reconstruction process. Therefore, if a new failure occurs, the data cannot be restored. There was a problem.

The document "Kitsuregawa: Recent Secondary Storage: Disarray" (Information Processing, Vol. 34, No. 5, pp. 642-651)
(May 1993), a system that extends RAID-5 has also been proposed. This method prepares two parity blocks based on Reed-Solomon coding,
It is possible to cope with a failure of up to two HDDs in a parity group.

However, according to the method of extending the RAID-5, when an error occurs at the time of input data writing or data reading, the number of accesses is larger than that of the RAID-5. It is not suitable for processing that requires sexuality. Therefore,
It is difficult for a device such as a multi-channel video server that requires real-time properties to use a method that is an extension of RAID-5 described above.

The present invention has been made in view of such a problem, and a first object of the present invention is to enable data error detection and correction even during re-allocation processing, thereby realizing re-allocation without interrupting data recording / reproduction operation. It is an object of the present invention to provide a data recording / reproducing apparatus and method, and an AV server that enable processing and are suitable for processing that requires real-time processing.

A second object of the present invention is, in addition to the first object, a data recording / reproducing apparatus and method and an AV server which can detect and correct data errors even in data restoration processing after reallocation processing. Is to provide.

[0030]

According to a first aspect of the present invention, there is provided a data recording / reproducing apparatus for recording a plurality of divided data obtained by dividing input data into predetermined units on a plurality of first non-linearly accessible recording media. And a divided data recording / reproducing means for reproducing the divided data recorded on the first recording medium, and generating error correction code data for the divided data, and providing a plurality of error correction codes on a second non-linearly accessible recording medium. An error correction code data recording / reproducing means for recording the code data and reproducing the error correction code data recorded on the second recording medium; and a correspondence relationship between a logical address and a physical address in a recording area of the first recording medium. For the first recording medium on which the reallocation processing for changing the data is to be performed, the other first recording medium on which recording and reproduction are performed by the divided data recording / reproducing means And et excluded, in which a re-allocation process control means for controlling the divided data recording and reproduction means so as to perform the re-allocation process.

According to a ninth aspect of the present invention, the AV server converts data including video and audio data input from the outside into data recordable on a non-linearly accessible recording medium, and outputs the converted data. A plurality of input / output processing means for converting the output data into a format which can be output to the outside, and a plurality of divided data obtained by dividing the data output from each of the plurality of input / output processing means into predetermined units. Division data recording / reproduction means for recording on a first non-linearly accessible recording medium and reproducing the division data recorded on the first recording medium;
While generating error correction code data of the divided data and recording it on a second non-linearly accessible recording medium,
Error correction code data recording / reproducing means for reproducing error correction code data recorded on the second recording medium; and reallocation processing for changing the correspondence between logical addresses and physical addresses in the recording area of the first recording medium. Re-allocation processing control for controlling the divided data recording / reproducing means to perform re-allocation processing by excluding the first recording medium to be performed from the other first recording media for performing recording / reproduction in the divided data recording / reproducing means. Means.

According to a tenth aspect of the present invention, there is provided a data recording / reproducing method,
A plurality of divided data obtained by dividing the input data into predetermined units is recorded on a first non-linearly accessible recording medium, and a plurality of error correction code data for the divided data is generated to perform a second non-linearly accessible recording. A first step of recording on the recording medium, and a second step of reproducing the divided data recorded on the first recording medium in the first step and reproducing the error correction code data recorded on the second recording medium. When performing the reallocation process for changing the correspondence between the logical address and the physical address in the recording area of the first recording medium, the first and second steps for performing recording and reproduction in the first and second steps are performed. And a third step of performing a reallocation process by excluding from the recording medium.

In the data recording / reproducing apparatus according to the first aspect,
The divided data recording / reproducing means records a plurality of divided data obtained by dividing the input data into predetermined units on a plurality of first non-linearly accessible recording media, respectively, and records the plurality of divided data on the first recording medium. The divided data is reproduced. Further, the error correction code data recording / reproducing means generates error correction code data of the divided data,
A plurality of error correction code data are recorded on a second non-linear accessible recording medium, and the error correction code data recorded on the second recording medium is reproduced. For the first recording medium on which the reallocation processing for changing the correspondence between the logical address and the physical address in the recording area of the first recording medium is to be performed by the reallocation processing control means,
The divided data recording / reproducing means is controlled so that the divided data recording / reproducing means is excluded from other first recording media on which recording / reproducing is performed, and is re-allocated.

In the AV server according to the ninth aspect, data including video and audio data input from outside is converted into data recordable on a non-linearly accessible recording medium and output by a plurality of input / output processing means. At the same time, the data output from the recording medium is converted into a format that can be output to the outside and output. Further, the divided data recording / reproducing means records a plurality of divided data obtained by dividing the data output from each of the plurality of input / output processing means into a predetermined unit on the first non-linearly accessible recording medium. The divided data recorded on the first recording medium is reproduced. The error correction code data recording / reproducing means generates the error correction code data of the divided data and records it on the second non-linearly accessible recording medium, and the error correction code data recorded on the second recording medium. Is played. The reallocation processing control means may perform the reallocation processing for changing the correspondence between the logical address and the physical address in the recording area of the first recording medium.
The divided data recording / reproducing means is controlled so that the divided data recording / reproducing means is excluded from other first recording media on which recording / reproducing is performed, and is re-allocated.

In the data recording / reproducing method according to the tenth aspect, in the first step, a plurality of divided data obtained by dividing the input data into predetermined units are recorded on the first non-linearly accessible recording medium. A plurality of error correction code data for the divided data is generated and recorded on a second non-linearly accessible recording medium.
Also, the first step is performed in the first step by the second step.
The divided data recorded on the second recording medium is reproduced, and the error correction code data recorded on the second recording medium is reproduced. In the third step, when performing the reallocation process for changing the correspondence between the logical address and the physical address in the recording area of the first recording medium,
The re-allocation process is performed excluding from the other first recording media for performing recording and reproduction in the step and the second step.

[0036]

Embodiments of the present invention will be described below in detail with reference to the drawings.

FIG. 1 is a block diagram showing an example of the configuration of a video server including a data recording / reproducing apparatus according to one embodiment of the present invention. This video server 10 is used as a multi-channel video server. The video server 10 records and reproduces data including video data, and a plurality of data recording / reproducing devices (referred to as RAID in the figure) 11 _{1 to} 11 according to the present embodiment.
_n (n is an integer value of 2 or more) and each data recording / reproducing device 1
And an input-output processor unit 12 for outputting data reproduced from 1 ₁ to 11 _n of the data to be recorded to the input and the data recording and reproducing apparatus 11 ₁ to 11 _n.

The input-output processor unit 12, when the access for the data recording and reproducing apparatus 11 ₁ to 11 of data to _n recorded or reproduced in a time slot unit is a time-separated by dividing a predetermined time interval into a plurality split I / O processor devices (referred to as IOP in the figure) 13 _{1 to} 13 _m (m is an integer value of 2 or more), a management device 14 for managing material information and the like, and an input / output processor device 13 up data bus 1 for connecting a ₁ to 13 _m and the management apparatus 14 and the data recording and reproducing apparatus 11 ₁ to 11 _n
5 and a downstream data bus 16. The number of data recording / reproducing devices 11 _{1 to} 11 _n is _{equal to} the number of data recording / reproducing devices 11 _{1 to} 11 _n. One uplink data bus 15 and one downlink data bus 16 have one data recording / reproducing device. 11 _i (i is an arbitrary integer value of 1 or more and n or less) and the input / output processor devices 13 ₁ to 13 ₁
3 _m and the management device 14 are connected. That is,
One input / output processor device 13 _j (j is an arbitrary integer value of 1 or more and m or less) and the management device 14 are connected to a plurality of upstream data buses 15 and a plurality of downstream data buses 16. In this embodiment, the downstream data bus 1
6, a bus for transferring the direction of the data directed from the input-output processor apparatus 13 ₁ to 13 _m and the management apparatus 14 to the data recording and reproducing apparatus 11 ₁ to 11 _n, a bus for transferring data in the opposite direction Is an up data bus 15.

When recording data, each of the input / output processor units 13 _j converts an input signal SI _j such as a video signal into data of a predetermined format, and transmits this data and a command for recording this data. , via the down data bus 16, and transmits to each of the data recording and reproducing apparatus 11 ₁ to 11 _n. Further, each input / output processor device 13
_j, when the data reproduction, a command for instructing reproduction of the data, via the down data bus 16, and transmitted to the data recording and reproducing apparatus 11 ₁ to 11 _n, the data recording and reproducing apparatus 11 ₁ It reproduced from to 11 _n, the uplink data bus 1
5 is converted into a predetermined signal,
And outputs to the outside as an output signal SO _j.

Each data recording / reproducing device 11 _i has a plurality of HDDs for recording input data and redundant code data.
It has. When recording data, each data recording / reproducing device 11 _i receives input data and a command from the input / output processor device 13 _j via the downlink data bus 16, divides the input data into predetermined units, and , Redundant code data is generated based on the input data, and the divided data and the redundant code data are recorded in a plurality of HDDs in accordance with a command. When data is reproduced, each data recording / reproducing device 11 _i receives a command from the input / output processor device 13 _j via the downlink data bus 16 and controls a plurality of HDDs according to the command. The divided data and the redundant code data are reproduced, the error correction processing is performed on the divided data using the redundant code data, and the divided data after the error correction processing is multiplexed and output as data through the upstream data bus 15. Output to the input / output processor _13j .

FIG. 2 is a block diagram showing a configuration of the data recording / reproducing apparatus according to the present embodiment. The data recording and reproducing apparatus 11 (representing 11 ₁ to 11 _n.) Includes a plurality of HDD 21 ₁ through 21 _K for recording input data and (K is an integer of 2 or more values), as an error correction code data HDDs 2 for recording redundant code data of
8 _1, 28 and _2, by dividing the input data D _I to length to generate a plurality of divided data, the divided data each HDD
A data distributor 22 for distributing the data to 21 _{1 to} 21 _K; a redundant code generator 26 for generating and outputting redundant code data P1 and P2 from the divided data output from the data distributor 22; an input memory 23 ₁ ~ 23 _K for temporarily holding each of the divided data are more output, redundant code data output from the redundant code generator 26 P1, P
An input memory 27 _1, 27 ₂ for temporarily holding the 2, are respectively connected to the HDD 21 ₁ through 21 _K, 28 _1, 28 _2, data held by the input memory 23 ₁ ~ 23 _K, 27 _1, 27 ₂ HDDs 21 _{1 to} 21 _K , 28 ₁ ,
28 record to ₂ (hereinafter also referred to as writing.) And HDD
21 ₁ through 21 _K, 28 _1, data from 28 ₂ reproduction (hereinafter, also referred to as read.) And a controller _{24 1 ~24 K, 29 1,} 29 2 for controlling the respective HDD 21 ₁
Through 21 _K, 28 _1, 28 output memory 25 ₁ to 25 _K for temporarily holding data read from the _2, 30 _1, 3
0 ₂ and an error corrector 31 that performs error detection and error correction using the data held by the output memories 25 _{1 to} 25 _K , 30 ₁ , and 30 ₂ and error information to be described later to repair the divided data. multiplexes the output data of the error corrector 31, a data multiplexer 32 for outputting as output data D _o, and controls the entire these devices CPU3
3 is provided. The output data of the data multiplexer 32 can be input to the data distributor 22 under the control of the CPU 33. Also, the CPU 33
It includes a ROM (Read Only Memory) storing a program and a RAM (Random Access Memory) serving as a working area.

HDDs 21 _{1 to} 21 _k , 28 ₁ , 28
No. ₂ accesses an arbitrary recording area on a hard disk (magnetic disk) which is a recording medium that can be accessed randomly, and performs a data recording operation or a data reproducing operation on the hard disk. Also,
_{HDD21 1 ~21 k, 28 1,} 28 control unit, not shown in the _2, when a read error of data occurs, the controller 24 ₁ ~24 _K, 29 ₁ error information indicating to that effect, as status data , which is to send in 29 _2. This error information further from the controller _{24 1 ~24 K, 29 1,} 29 2 CPU33
To, and adapted to be sent as error information _{ER 1 ~ER K, ER P1,} ER P2.

Here, the format on the hard disk will be described with reference to FIG. A plurality of tracks are set concentrically on the hard disk. FIG. 8 shows only one track. The track is radially divided, and a plurality of sectors as data recording units are set. In some of these sectors, there is a sector in which an error always occurs when writing or reading data. Such a sector is called a defective sector, and it is considered that data cannot be read or written correctly due to physical damage or the like. In preparation for such a defective sector, there is a method in which a spare sector is reserved in advance on the hard disk, and data is recorded in the spare sector instead of the defective sector as necessary. Such a spare sector is called a replacement sector. In an HDD having this alternative sector,
The control unit in the HDD associates a logical sector number (LBA) with a physical sector number so that the upper sector can refer to the same sector number even when an alternative sector is used instead of a defective sector. Have inside.

In the example shown in FIG. 8, N sectors (Sector-1 to Sector-N) are physically allocated to one track, and two alternative sectors (Spare-1 and Spare-1) are further allocated.
are-2) are available. FIG. 9 shows the correspondence between LBAs and physical sector numbers represented in the above-mentioned correspondence table. FIG. 9 shows a normal case with no defective sectors.
As shown in (a), the LBA and the physical sector number match. If normal, the control unit in the HDD
The recording area on the hard disk is managed based on the correspondence table shown in FIG. 9A. On the other hand, for example, if the sector Sector-5 is defective,
The alternative sector Spare-1 is used instead of ctor-5.
In this case, as shown in FIG. 9B, the alternative sector Spare-1 is associated with “5” in the LBA,
The control unit in the HDD manages the recording area on the hard disk based on the correspondence table shown in FIG. 9B. Thus, according to the occurrence of defective sectors,
Changing the correspondence between LBAs and physical sector numbers is referred to as reassignment processing (Reassign).

The CPU33 via the down data bus 16, receives a command from the input-output processor apparatus 13 _j, in accordance with this command, the controller 24 _1-24
k, 29 _1, 29 ₂ to give the instruction, each HDD21 ₁ ~
21 _k , 28 ₁ , and 28 ₂ are controlled.
If a data reading error has occurred, the CPU 33 sends error information indicating the fact to the error corrector 31. This error information
It contains information for identifying the HDD in which a read error has occurred.

The error corrector 31 outputs a CP
Error information from U33 and output memories 25 _{1 to} 25 _K ,
By using the held data by the 30 _1, 30 _2,
The divided data is restored by performing error detection and error correction, and the restored divided data is output to the data multiplexer 32. Data multiplexer 32 rearranges the divided data output from the error corrector 31 based on the data string, and outputs as output data D _O.

The HDDs 21 _{1 to} 21 _K correspond to the divided data recording / reproducing means in the present invention, and the HDDs 28 ₁ , 28 _K
2 corresponds to the error correction code data recording / reproducing means of the present invention, the data distributor 22 corresponds to the dividing means of the present invention, and the redundant code generator 26 corresponds to the error correction code data generating means of the present invention. And the error corrector 31
Corresponds to the error correcting means of the present invention, and the data multiplexer 32 corresponds to the multiplexing means of the present invention.

In this embodiment, the error correction code is
A code capable of correcting an error of a plurality of divided data is used. There is a BCH code as such a code. In the present embodiment, a Reed-Solomon code, which is a type of BCH code, is used as an error correction code. Here, the Reed-Solomon code will be briefly described.

First, in the Reed-Solomon code, one byte, that is, eight bits, is treated as one number, and from 00h to F
256 numbers up to Fh (h represents a hexadecimal number) are used, and an arithmetic rule different from that of the integer world is used for the four arithmetic operations.

In the Reed-Solomon code, a polynomial called a code polynomial or a generator polynomial is used. Here, 1
An example in which two parity bytes are used for four data bytes will be described. in this case,
A total of 16 bytes of 14 bytes of data bytes and 2 bytes of parity bytes are processed together. Therefore, the data byte of 14 bytes, W _15, respectively, W _14, ..., W ₂
And two parity bytes are W ₁ and W, respectively.
_Set to ₀ . W _i (i is an integer from 0 to 15)
Is any number from 00h to FFh in the format of all bytes.

The code polynomial W (X) is defined as the following equation (1).

W (X) = W ₁₅ × X ¹⁵ + W ₁₄ × X ¹⁴ + ... + W ₂ × X ² + W ₁ × X + W ₀ (1)

Note that X is any number from 00h to FFh in a byte format.

Here, in W (X) of equation (1), W
₁ × X + W ₀ is W ₁₅ × X ¹⁵ + W ₁₄ × X ¹⁴ + ... + W ₂ ×
X ² is obtained as a remainder when divided by a generator polynomial G (X) represented by the following equation (2).

G (X) = (X-02h) × (X-01h) (2)

Therefore, the values of the parity bytes W ₁ and W ₀ are determined so that the following equations (3) and (4) hold.

W (02h) = 00h (3) W (01h) = 00h (4)

In the coding in the Reed-Solomon code,
A parity byte is obtained from the data bytes based on such a relationship.

Next, error correction in the Reed-Solomon code will be described. First, the data bytes W ₁₅ , W ₁₄ ,
..., data, respectively HDD (15) corresponding to W ₂ and parity bytes _{W 1, W 0, HDD (} 14),
..., HDD (1), V 15 from the _{HDD (0), V 14,} ...,
It is assumed that they are read as V ₁ and V ₀ . Where HDD
The presence of an error in the read data V _i of (i), the read data V _i is the following equation (5), represented by (6). E _i represents the magnitude of the error.

When there is no error in HDD (i): V _i = W _i (5) When there is an error in HDD (i): V _i = W _i + E _i (6)

Next, for the code polynomial W (X), V
(X) is defined as in the following equation (7).

V (X) = V ₁₅ × X ¹⁵ + V ₁₄ × X ¹⁴ +... + V ₂ × X ² + V ₁ × X + V ₀ (7)

Further, values S ₁ , S called syndromes
₀ is defined as in the following equations (8) and (9).

S ₁ = V (02h) (8) S ₀ = V (01h) (9)

Here, if there is no error in the read data of any HDD, since V (X) = W (X), the syndrome is expressed by the following equations (10) and (11).

S ₁ = 00h (10) S ₀ = 00h (11)

Next, when there is an error in the read data of the HDD (i) but no error information is notified from the HDD (i), V (X ) =
W (X) + E _i × X ⁱ , and the syndrome is expressed by the following equations (12) and (13).

S ₁ = E _i × 02h ⁱ (12) S ₀ = E _i × 01h ⁱ = E _i (13)

[0069] Therefore, the equation (13), Motomari magnitude E _i of the error, subscript indicated by the size E _i and wherein the error (12), the read data for any HDD (i) is whether the error i is obtained, which enables error correction of read data. This is called one error correction.

Next, two HDD (i) and HDD (j)
If there is an error in the read data (j is an integer from 0 to 15) and error information is notified from the HDD (i) and the HDD (j), based on the equations (6) and (7), Te, since the V (X) = W (X ) + E i × X i + E j × X j, syndromes following equation (14) and (15).

S ₁ = E _i × 02h ⁱ + E _j × 02h ^j (14) S ₀ = E _i × 01h ⁱ + E _j × 01h ^j = E _i + E _j (15)

Here, since it is assumed that error information has been notified from HDD (i) and HDD (j), the values of subscripts i and j are known. Therefore, the values of 02h ⁱ and 02h ^j in equation (14) can be calculated, and E _i
And the values of E _j are obtained by a binary simultaneous equation, which enables error correction of read data. This is 2
This is called error erasure correction.

In decoding in the Reed-Solomon code,
In this way, the value of the syndrome is obtained, and the position and magnitude of the error are obtained using the value of the syndrome.

The data bytes W ₁₅ , W ₁₄ ,..., W ₂ in the above description of the Reed-Solomon code correspond to the divided data in the present embodiment, and the parity bytes W ₁ , W ₀ correspond to the present embodiment. Code data P in the form
1, P2.

Next, the operation of the data recording / reproducing apparatus 11 according to the present embodiment will be described. The following description also serves as a description of the data recording / reproducing method according to the present embodiment.

First, the write operation of the data recording / reproducing device 11 will be described. Input data D _I is input to the data distributor 22 a plurality of divided data is generated, the divided data is distributed in the input memory 23 ₁ ~ 23 _K with once recorded, the redundant code generator 26 Is entered. At this time, as a data distribution method, for example, data strings D1, D2, D
, D4, D5,..., The data D1 is distributed to the _first HDD 211, and the data D2 is
The data is sequentially distributed, for example, distributed to the _second HDD 212, and the data DK is distributed to the last K-th HDD 212.
Once it has distributed to the D21 _K, again, first of HDD21 ₁
There is a method of distributing data in order from the beginning.

The redundant code generator 26 is provided with the data distributor 22
The redundant code data P1 and P2 in the Reed-Solomon code are generated and output based on the divided data output from. The input memories 27 ₁ and 27 ₂ temporarily record the redundant code data P1 and P2. After that, each HDD 21 ₁
Controllers 24 _{1 to} 24 of up to 21 _K , 28 ₁ and 28 _2
K, 29 _1, 29 _2, based on the control of the CPU 33,
Reading the input memory _{23 1 ~23 K, 27 1,} 27 2 divided data and the redundant code data P1, P2 from and writes to the _{HDD21 1 ~21 K, 28 1,} 28 2.

Next, the data reading operation of the data recording / reproducing apparatus 11 will be described. Under the control of the CPU 33, each of the controllers 24 _{1 to} 24 _K , 29 ₁ , 29
₂ reads the divided data and the redundant code data P1 and P2 from each of the HDDs 21 _{1 to} 21 _K , 28 ₁ and 28 ₂ ,
The data is written in the output memories 25 _{1 to} 25 _K , 30 ₁ and 30 ₂ respectively. At this time, the HDDs 21 _{1 to} 21 _K , 28 ₁ ,
28 when a read error occurs in _2, error information indicating to that effect, HDD 21 ₁ through 21 _K, 28
From the control unit _1, 28 _2, sent as status data to the controller 24 ₁ to 24 _K, 29 _1, 29 _2,
Furthermore, the controller 24 to _{1 ~24 K, 29 1, 29} 2 from the CPU 33, the error information ER ₁ ~ER _K, ER _P1,
Sent as ER _P2 .

Output memories 25 _{1 to} 25 _K , 30 ₁ , 30
Each data recorded in ₂ is output to the error corrector 31 in synchronization. At this time, if a read error has occurred, error information indicating that
The signal is sent from 33 to the error corrector 31. This error information includes information for specifying the HDD in which the read error has occurred. The error corrector 31 restores the divided data based on the error information and the redundant code data P1 and P2 as necessary, and outputs the data to the data multiplexer 32. The data multiplexer 32 rearranges the divided data output from the error corrector 31 into the original data sequence, and outputs it as output data _Do to the outside.

In this embodiment, the error correction code is
Since a Reed-Solomon code capable of correcting an error in a plurality of divided data is used, the divided data can be repaired in the following cases.

(1) When there is an error in the read data of one HDD In this case, not only the error information indicating that the read error has occurred in the HDD, but also the error information is not notified. Even in such a case, it is possible to repair the divided data. (2) When there is an error in the read data of the two HDDs and there is notification of error information indicating that a read error has occurred in those HDDs

When an error (hereinafter referred to as a write error) occurs in a write operation in any of the HDDs, it is necessary to repair the data in which the write error has occurred. The process of restoring data for a part of the recording area on such a recording medium (hard disk) is referred to as a partial rebuild process (partial Rebuild) in the present embodiment. Further, when a defective sector that always causes an error occurs in the write operation or the read operation, as described above, the reassignment process (Reassign) that changes the correspondence between the LBA and the physical sector number. Need to do. In addition, any HD
When D is exchanged, it is necessary to reconstruct the original data in a new HDD. The process of restoring data for the entire recording area on such a recording medium (hard disk) is referred to as an overall rebuild process (entire Rebuild) in the present embodiment.

Hereinafter, the data recording / reproducing apparatus 1 including the above-described partial reconstruction processing, reallocation processing, and overall reconstruction processing will be described.
An example of the overall operation of No. 1 will be described. In the following example, the operation modes of the data recording / reproducing apparatus 11 include a normal mode, a reallocation processing mode, and a whole reconstruction processing mode. The transition from the normal mode to the reallocation processing mode is performed when a predetermined condition is satisfied in the normal mode or when an instruction from a higher-level device, that is, the input / output processor unit 12 in FIG. Occurs when there is. When the processing in the reallocation processing mode is completed, the mode shifts to the normal mode. The transition from the normal mode to the entire reconstruction processing mode is performed when the data recording / reproducing device 11 receives an instruction from a higher-level device, that is, the input / output processor unit 12 in FIG. Own HDD
Occurs when it is detected that has been replaced. When the processing in the whole reconstruction processing mode is completed, the mode is shifted to the normal mode.

FIGS. 3 and 4 are flowcharts showing the operation of the data recording / reproducing apparatus 11 in one time slot in the normal mode. In this operation, first, the CPU 33 determines whether or not a command to instruct writing has been received (step S101). If a command to instruct writing has been received (Y), the above-described writing operation is performed (step S101). S102). CPU 33 whether the determination has been received write instruction command, for example, the higher the input-output processor unit 13 ₁ to 13 ₆ CPU 33 commands transferred via the bus 16 from the received, this command is a write instruction It can be performed by determining whether the command is a command or not. After the end of the write operation, the CPU 3
3, based on the _{HDD21 1 ~21 K, 28 1,} 28 information from _2, to determine whether the writing of data is successful (step S103). If the data writing is successful (Y), the operation in one time slot ends. If the data writing has failed (step S103; N), the CPU 33 stores the HDD and the address where the writing has failed (step S10).
4) The operation in one time slot ends.

When the CPU 33 has not received the command for writing (step S101; N),
It is determined whether a command instructing reading has been received (step S105). If a command instructing reading has been received (Y), the above-described reading operation is performed (step S106). Determination also whether it has received this read command, the higher the input-output processor unit _{131-134 6}
The command transferred from the CPU 16 via the bus 16
Can be performed by receiving. After completion of the read operation, CPU 33, each HDD 21 ₁ through 21 _K, 2
8 _1, 28 error information ER ₁ to Er _K from _2, E
Based on the presence or absence of R _P1, ER _P2, it is determined whether the data read is successful (step S107). If the data reading is successful (Y), the operation in one time slot is ended. If the data reading has failed (step S107; N), the CPU 33
HDD and address where reading failed, and the HD
The number of read failures at D and the address is stored (step S108). Next, the CPU 33 determines that the number of read failures stored in step S108 is equal to the predetermined value N1.
(N1 is an arbitrary integer value of 1 or more) is determined (step S109), and the number of read failures is reduced to a predetermined value N.
If it is not one or more (N), the operation in one time slot ends. When the number of read failures is equal to or greater than the predetermined value N1 (Y), the CPU 33 shifts to a reallocation processing mode described later (step S115).

When neither a command instructing writing nor a command instructing reading is received (step S105; N), the CPU 33 determines whether or not there is an HDD and an address to be subjected to the partial rebuilding process. (Step S110). Here, there are the following three types of HDDs and addresses to be subjected to the partial reconstruction processing. That is, the first is the HDD and the address where the writing failed in step S104. The second is the HDDs and addresses that have been reallocated. Third,
As will be described in detail later, the HDD and the address stored during the stop of the write operation in the reallocation processing. The CPU 33 stores the HDD and the address at which the partial rebuilding process is to be performed. It should be noted that the HDD and the address at which the partial rebuilding process is to be performed may be stored in a higher-level device than the data recording / reproducing device 11, that is, the input / output processor unit 12 in FIG. Incidentally, there is no write instruction (step S101; N),
Further, when there is no read instruction (step S105; N), there is no operation instruction from the host processor 12 side. In other words, the upper processor 12 operates in the assigned time slot when there is a data write instruction or a data read instruction from the upper level, and when there is no such instruction, the operation in the assigned time slot is performed. It's already finished.

If there is no HDD and address to perform the partial rebuilding process (step S110; N), the process is terminated. If there is an HDD and an address to be subjected to the partial reconstruction process (step S110; Y), the partial reconstruction process is performed (step S111). This partial reconstruction processing will be described later in detail.

After the completion of the partial reconstruction process, the CPU 33
It is determined whether or not the partial reconstruction process has succeeded (step S112). If the partial reconstruction process is successful (Y),
The processing operation ends. If the partial rebuilding process has failed (N), the CPU 33 stores the HDD where the partial rebuilding process failed, the address, and the number of times the partial rebuilding process failed (step S113). Next, the CPU 33
Determines whether the number of failures in the partial reconstruction process stored in step S113 is equal to or greater than a predetermined value N2 (N2 is an arbitrary integer of 1 or more) (step S114), and determines the number of failures in the partial reconstruction process. Is not greater than or equal to a predetermined value N2 (N), 1
The operation in the time slot ends. If the number of failures in the partial reconstruction process is equal to or more than the predetermined value N2 (Y), the CPU
33 shifts to the reallocation processing mode described later (step S115).

As described above, in the present embodiment, the partial rebuilding process is performed without interrupting the writing operation for the data transferred from the upper side or the reading operation of the recorded data to be output to the upper side. It is executed in a time slot without a write instruction or a read instruction from, that is, in an idle operation time.

Next, the partial reconstruction process (step S111) will be described in detail with reference to the flowchart of FIG. This partial reconstruction process is performed in sector units. In this part reconstruction process, first, CPU 33 specifies the address (sector) to perform the partial reconstruction processing to perform the read operation for each controller _{24 1 ~24 K, 29 1,} 29 2. In response, the controller _{24 1 ~24 K, 29 1,} 29 2 , each HDD 21 ₁ ~
21 _K, 28 _1, from 28 ₂ to read the data at the specified address (step S201). The read data are stored in output memories 25 _{1 to} 25 _K , 30 ₁ and 30 _2.
Are input to the error corrector 31. At this time, C
The PU 33 instructs the error corrector 31 not to use the data read from the HDD to be subjected to the partial reconstruction processing. Error corrector 31 outputs the memory _{25 1 ~25 K, 30 1,} 30 2 of the data outputted from, using data other than the data read from the HDD to perform the partial reconstruction process, the partial data Is restored (step S202), and the restored partial data is output to the data multiplexer 32. The data multiplexer 32
The divided data output from the error corrector 31 is rearranged into the original data sequence and output as output data _Do.

[0091] Next, the control of the CPU 33, receives the output data D _o from the data multiplexer 32 to the data distributor 22, by performing the same writing operation and writing the input data D _I, partial reconstruction process The repaired partial data is written to the HDD to be subjected to the
03), end the partial restructuring process. In the write operation, partial reconstruction processing may be performed to write data in the HDD only be performed, but in a simplified manner, all the _{HDD21 1 ~21 K, 28 1,} 28 2 may be carried out writing in .

In this embodiment, the error correction code is
Since a Reed-Solomon code capable of correcting an error of a plurality of divided data is used, even if a partial reconstruction process is being performed,
The same data error detection and correction capabilities as those of the data recording / reproducing apparatus using AID-3 can be maintained. That is, even if a read error occurs in another HDD other than the HDD to be subjected to the partial rebuilding process during the partial rebuilding process, the data can be correctly restored.

[0093] In the above description, in the partial reconstruction process, each was to perform a write operation by sending the output data D _o of the data multiplexer 32 to the data distributor 22, and is outputted from the error corrector 31 the divided data may be performed write processing is sent to each of the input memory 23 ₁ ~ 23 _K and the redundancy code generator 26.

In the above description, the CPU 33 detects the HDD and the address where the partial rebuilding process is to be performed, and
The input / output processor unit 12 shown in FIG. 1 stores the HDD and the address at which the partial rebuilding process is to be performed with respect to the data recording / reproducing apparatus 11. For 11,
An instruction for the partial rebuilding process may be sent by designating the HDD and the address where the partial rebuilding process is to be performed, and the data recording / reproducing apparatus 11 may perform the partial rebuilding process in response to the instruction. Also in this case, the partial rebuilding process is executed in a time slot without a write instruction or a read instruction.

Next, the operation in the reallocation processing mode will be described with reference to the flowchart of FIG. The transition from the normal mode to the reallocation processing mode is performed when a predetermined condition is satisfied (step S115) in the normal mode shown in FIGS.
This occurs when there is an instruction from a higher-level device, namely, the input / output processor unit 12 in FIG.

In the operation of the reallocation processing mode, first,
The CPU 33 stops the write operation and the read operation in the HDD to be reallocated (step S30).
1) Excluding HDDs to be reallocated from the recording / reproduction targets of the divided data. Next, the CPU 33 specifies a sector to be re-allocated to the HDD to be re-allocated and starts the re-allocation process (step S30).
2). Note that the sectors to be reallocated are the sectors in which the number of read failures is N1 or more and the sectors in which the number of failures in the partial reconstruction process is N2 or more. As described above, the reallocation processing is to change the correspondence between the LBA and the physical sector number. When the CPU 33 receives a command instructing writing during the reallocation processing, the CPU 33 causes the HDD other than the HDD to be reallocated to perform the write operation, and the HDD to be reallocated and the LBA to which the write has been instructed. Is stored. When the CPU 33 receives a command instructing reading during the re-allocation processing, the CPU 33 causes the HDD other than the HDD to be re-allocated to perform a read operation, and performs H re-allocation to the error corrector 31.
An instruction is issued to invalidate data from the DD and perform error correction.

Next, the CPU 33 determines whether or not the reallocation processing has been completed (step S303), and if not completed (N), repeats this determination. When the reallocation processing is completed (Y), the suspension of the write operation and the read operation in the HDD to be reallocated is released (step S304), and the mode shifts to the normal mode (step S304).
305).

As described above, in the present embodiment, even during the reallocation process, the write operation and the read operation are executed without interruption. Moreover, in the present embodiment, a Reed-Solomon code capable of correcting an error in a plurality of divided data is used as the error correction code. The same data error detection and correction capability as the device can be maintained. That is, in the read operation during the reallocation process, the HD performing the reallocation process
Even if a read error occurs in another HDD other than D, the data can be correctly restored.

Next, the whole reconstruction process will be described.
The whole rebuilding process is executed when any one of the HDDs is replaced. As described above, the transition from the normal mode to the whole reconstruction processing mode is performed when the data recording / reproducing apparatus 11 receives an instruction from a higher-level device, that is, the input / output processor unit 12 in FIG. This occurs when the playback device 11 itself detects that any HDD has been replaced. The instruction from the input / output processor unit 12 includes information for specifying the HDD on which the entire rebuilding process is to be performed. The following method is used as a method in which the data recording / reproducing device 11 itself detects that any HDD has been replaced. That is, first, the detection operation is performed when the data recording / reproducing device 11
This is performed by the CPU 33 when inserting and removing the DD. The fact that the HDD has been removed and inserted indicates that the controller 2
_{4 1 ~24 K, 29 1,} 29 2 is detected, transmitted to CPU 33. It determines whether HDD is replaced, using the serial number of the HDD to be detected by the controller _{24 1 ~24 K, 29 1,} 29 2, or a specific information of the HDD recorded in a predetermined area of the HDD In advance, using that information.

Next, the operation in one time slot in the whole reconstruction processing mode will be described with reference to the flowchart of FIG. In this operation, first, the CPU 33 determines whether or not a command for instructing writing has been received (step S401). If a command for instructing writing has been received (Y), a writing operation is performed (step S4).
02). The write operation at this time is the same as in the normal mode. After the end of the write operation, the CPU 33
DD21 based on _{1 ~21 K,} 28 _1, 28 information from _2, to determine whether the writing of data is successful (step S403). If the data writing is successful (Y), the operation in one time slot ends. When data writing has failed (step S40)
3; N) indicates that the CPU 33
D and the address are stored (step S404),
The operation in the time slot ends.

When the CPU 33 has not received the command for writing (step S401; N),
It is determined whether or not a command to instruct reading is received (step S405). If a command to instruct reading is received (Y), a reading operation is performed (step S406). However, in the read operation at this time, C
The PU 33 instructs the error corrector 31 not to use the data read from the HDD for which the entire rebuilding process is to be performed. Error corrector 31 outputs the memory _{25 1 ~25 K, 30 1,} 30 2 of the data outputted from, using data other than the data read from the HDD to perform the entire rebuild process, the partial data Is restored, and the restored partial data is output to the data multiplexer 32. After the end of the read operation, the CPU 33
_{DD21 1 ~21 K, 28 1,} 28 error information ER ₁ to Er _K from _2, based on the presence or absence of ER _P1, ER _P2, it is determined whether the data read is successful (step S407). If the data reading is successful (Y), the operation in one time slot is ended. When the data reading has failed (step S407; N), the CPU 33 stores the HDD and the address from which the reading failed and the number of times of the reading failure in the HDD and the address (step S408). End the operation.

When neither a command instructing writing nor a command instructing reading is received (step S405; N), the CPU 33 performs an overall rebuilding process (step S409). The whole reconstruction process performs the same process as the partial reconstruction process for the entire recording area of the hard disk. Next, the CPU 33 determines whether or not all the entire rebuilding processes have been completed (step S41).
0), if all the rebuilding processes have not been completed (N), the operation in one time slot is completed. When the entire rebuilding process has been completed (Y), the mode is shifted to the normal mode (step S411).

As described above, in the present embodiment, the entire rebuilding process is executed in a time slot without a write instruction or a read instruction without interrupting a write operation or a read operation.

In this embodiment, a Reed-Solomon code capable of correcting an error in a plurality of divided data is used as an error correction code. The same data error detection and correction capabilities as those of the data recording / reproducing apparatus used can be maintained. That is, even if a read error occurs in another HDD other than the HDD for which the entire rebuilding process is to be performed during the entire rebuilding process, the data can be correctly restored. Further, in the read operation during the whole reconstruction processing mode, H to perform the whole reconstruction processing
Even when a read error occurs in another HDD other than the DD, the data can be correctly restored.

In the whole rebuilding process, similarly to the partial rebuilding process, in the data writing operation after the restoration, the data may be written only in the HDD for which the whole rebuilding process is to be performed. All HD
_{D21 1 ~21 K, 28 1,} 28 may be made to write in _two.

[0106] In the above description, each of the rebuilding process a whole, but to perform a write operation by sending the output data D _o of the data multiplexer 32 to the data distributor 22, and is outputted from the error corrector 31 the divided data may be performed write processing is sent to each of the input memory 23 ₁ ~ 23 _K and the redundancy code generator 26.

As described above, in the data recording / reproducing apparatus 11 according to the present embodiment, when recording data, the input data is divided into predetermined units to generate a plurality of divided data, and based on the input data. To generate error correction code data (redundant code data P1 and P2) based on an error correction code (Reed-Solomon code) capable of correcting a plurality of divided data with respect to input data. Code data is stored in separate HD
_{D21 1 ~21 K, 28 1,} 28 2 by recording, during data reproduction, each HDD21 ₁ ~21 _K, 28 _1,
From 28 _2, reproduces a plurality of divided data and the error correction code data using the error correction code data reproduced, performs error correction processing for the reproduced plurality of divided data, each after the error correction processing The divided data is multiplexed and output. Therefore, the data recording / reproducing apparatus 11 according to the present embodiment can be used for RAID-3 or RAID-5.
Is superior in data error detection and correction capability as compared with a data recording / reproducing apparatus using a RAID, and can improve the reliability of the system. Can be written in a single access, error correction after data reading can be performed immediately, and it is suitable for processing that requires real-time characteristics, and real-time characteristics such as a multi-channel video server are required. Suitable for equipment.

More specifically, the data recording / reproducing apparatus 11 according to the present embodiment generates redundant code data P1 and P2 capable of correcting the error of two divided data. Even if a failure occurs, the same data error detection and correction capability as that of a data recording / reproducing apparatus using RAID-3 can be maintained. In addition, even when one HDD reads incorrect data without a read error, it can be detected and the data can be restored. Further, even when two HDDs read illegal data, if a read error in each HDD is detected, the data can be restored.

Further, according to the data recording / reproducing apparatus 11 according to the present embodiment, even when the partial reconstructing process is being performed, the RA
It is possible to maintain the same data error detection and correction capabilities as those of the data recording / reproducing apparatus using ID-3, and improve the reliability and maintainability of the system.

Further, according to the data recording / reproducing apparatus 11 according to the present embodiment, the partial reconstruction processing is executed in a time slot without a write instruction or a read instruction. The partial rebuilding process can be performed while operating a device that requires reliability.

Further, according to the data recording / reproducing apparatus 11 according to the present embodiment, even when the entire reconstruction process is being performed, the RA
It is possible to maintain the same data error detection and correction capabilities as those of the data recording / reproducing apparatus using ID-3, and improve the reliability and maintainability of the system.

Further, according to the data recording / reproducing apparatus 11 according to the present embodiment, the whole reconstruction processing is executed in a time slot without a write instruction or a read instruction. It is possible to perform the entire rebuilding process while operating a device that requires reliability.

Further, according to the data recording / reproducing apparatus 11 according to the present embodiment, the write operation and the read operation are executed without interruption even during the reallocation processing. The reallocation process can be performed while operating a device that requires real-time processing such as the above.

Further, according to the data recording / reproducing apparatus 11 according to the present embodiment, even if the reallocation process is being performed, the RA
It is possible to maintain the same data error detection and correction capabilities as those of the data recording / reproducing apparatus using ID-3, and improve the reliability and maintainability of the system.

Finally, with reference to FIG. 10, an example of a result of comparing reliability between the data recording / reproducing apparatus according to the present embodiment and a data recording / reproducing apparatus using RAID-3 will be described. In this example, the data recording / reproducing apparatus according to the present embodiment includes 14 divided data recording HDDs and 2 HDDs.
And an HDD for recording redundant code data using one Reed-Solomon code. In contrast, RAI
A data recording / reproducing apparatus using D-3 includes seven HDDs for recording divided data and one HDD for recording parity data.
D and one set, and two sets of this combination are provided. In this example, the data recording / reproducing apparatus according to the present embodiment and the data recording / reproducing apparatus using RAID-3 both use 16 HDDs, have the same hardware scale, and have redundancy. Are equal in one eighth.

In FIG. 10, the horizontal axis represents the error occurrence probability P _U of the HDD alone, and the vertical axis represents the error occurrence probability P _S of the entire data recording / reproducing apparatus. FIG.
At 0, the line indicated by reference numeral R6 represents a relationship between the error occurrence probability P _S of the entire error occurrence probability P _U and apparatus in HDD itself in the data recording and reproducing apparatus according to this embodiment, by the symbol R3 The line shown represents the relationship between the error occurrence probability P _U of the HDD alone in the data recording / reproducing device using RAID-3 and the error occurrence probability P _S of the entire device.

The line indicated by reference numeral R6 is represented by the following equation. _{P S = 16 C 3 P U} 3 (1-P U) 13

Similarly, the line indicated by reference numeral R3 is represented by the following equation. P _S = 2 × ₈ C ₂ P _U ² (1-P _U ) ⁶

[0119] In FIG. 10, reference numeral D is between the data recording and reproducing apparatus using the data recording and reproducing apparatus and RAID-3 according to this embodiment, apparatus for error occurrence probability P _U in the same HDD alone Error probability P _{S as a} whole
Represents the difference between As can be seen from such a difference D,
The data recording / reproducing apparatus according to the present embodiment has the same HDD as that of the data recording / reproducing apparatus using RAID-3.
Error probability P _S of the entire device with respect to error occurrence probability P _U by itself has become smaller, it can be said to be reliable.

In FIG. 10, the line indicated by reference symbol C represents the relationship where P _U = P _S when no RAID is used. In the area below this line C, RAID
As compared with the case of not using the error occurrence probability P _S of the entire data recording and reproducing apparatus for error occurrence probability P _U in the same HDD itself is decreased, the reliability is improved. Line R
6, the point where R3 intersects with the line C is a branch point where the reliability starts to improve as compared with the case where RAID is not used. Line R
6, R3 that intersects the line C, that a branch point each A, when is B, the data recording and reproducing apparatus according to this embodiment, when a small probability P _U of the branch point A, RA
Reliability is improved compared to the case where no ID is used.
-3 data recording and reproducing apparatus using, when the probability P _U is smaller than the branch point B, reliability can be improved as compared with the case of not using the RAID.

[0121] Because of the greater probability P _U at the branch point A as compared with the probability P _U at the branch point B, RAID-
It can be seen that the data recording / reproducing apparatus according to the present embodiment has higher resistance to non-stationary errors (random errors) and higher reliability than the data recording / reproducing apparatus using No. 3.

According to the above comparison results, the data recording / reproducing apparatus according to the present embodiment has the same hardware scale and redundancy as the data recording / reproducing apparatus using RAID-3, while maintaining the reliability. Performance can be improved.

Note that the present invention is not limited to the above embodiment. For example, in the above embodiment, a hard disk is used as a recording medium for recording the divided data and the error correction code data. The invention is not limited to this, and may be a magneto-optical disk, a semiconductor memory, or the like.

The error correction code data is not limited to data based on the Reed-Solomon code, but may be data based on other error correction codes as long as the error correction of a plurality of divided data can be performed.

Further, in the above embodiment, the error correction code data (redundant code data) is set to 2 bytes and two H
An example in which the data is recorded on the DD has been described.
Bytes or more may be recorded on three or more HDDs. As the number of bytes of the error correction code data increases, the redundancy increases, but the number of error-correctable divided data increases, and the reliability improves.

Although the above embodiment has been described on the assumption that a multi-channel video server is used for recording / reproducing video / audio data in a broadcasting station or the like, the present invention is not limited to this. Instead, the present invention can be applied to an apparatus used for recording and reproducing other types of data.

[0127]

As described above, claims 1 to 8 are described.
According to the data recording / reproducing apparatus described in any one of the above, while recording a plurality of divided data obtained by dividing the input data in a predetermined unit on a plurality of first non-linearly accessible recording media, A divided data recording / reproducing means for reproducing the divided data recorded on the recording medium, generating error correction code data of the divided data, and recording the plurality of error correction code data on a second non-linearly accessible recording medium; Error correction code data recording / reproducing means for reproducing error correction code data recorded on the second recording medium, and reallocation processing for changing the correspondence between logical addresses and physical addresses in the recording area of the first recording medium Of the first recording medium on which the recording and reproduction are to be performed by the divided data recording / reproducing means, except for the first recording medium on which the recording and reproduction are performed. Control means for controlling the divided data recording / reproducing means so that the data recording / reproducing means can be performed, so that data errors can be detected and corrected even during the reallocation processing. This makes it possible to perform the reallocation processing without performing the processing, and to realize a data recording / reproducing apparatus suitable for processing that requires real-time processing.

Further, according to the data recording / reproducing apparatus of the seventh or eighth aspect, the data is reproduced from the divided data recording / reproducing means by using the error correction code data reproduced by the error correction code data recording / reproducing means. Error correction processing means for performing error correction processing on the divided data; and a repair target among a plurality of mutually corresponding divided data and error correction code data recorded on the first recording medium and the second recording medium, respectively. Data other than the divided data is reproduced by controlling the divided data recording / reproducing means and the error correction code data recording / reproducing means, and an error correction process is performed on the divided data based on the reproduced divided data and the error correction code data. The error correction processing means is controlled so that the divided data to be repaired, which has been corrected, is recorded on the first recording medium. By controlling, since a data recovery processing control means for performing recovery processing of the divided data,
Further, there is an effect that error detection and correction of data can be performed even in data restoration processing after the reallocation processing.

Further, according to the data recording / reproducing apparatus of the eighth aspect, the data restoration processing is performed when the divided data is not recorded on the recording medium or reproduced from the recording medium. Further, there is an effect that data restoration processing can be performed without interrupting the operation of the data recording / reproducing apparatus.

According to the AV server of the ninth aspect, data including video and audio data input from the outside is converted into data recordable on a non-linearly accessible recording medium, and output. A plurality of input / output processing means for converting data output from the medium into a format that can be output to the outside and outputting the converted data; and a plurality of data obtained by dividing data output from each of the plurality of input / output processing means into predetermined units. The divided data recorded on the first non-linearly accessible recording medium, reproduces the divided data recorded on the first recording medium, and generates error correction code data for the divided data. Recording on a second non-linearly accessible recording medium and reproduction of error correction code data recorded on the second recording medium The divisional data recording / reproducing means records the correction code data recording / reproducing means and the first recording medium on which the reallocation processing for changing the correspondence between the logical address and the physical address in the recording area of the first recording medium is to be performed. And a re-allocation processing control means for controlling the divided data recording / reproducing means so as to perform the re-allocation processing by excluding it from the other first recording medium for reproduction.
Data errors can be detected and corrected even during the re-allocation process. As a result, the re-allocation process can be performed without interrupting the operation of data recording and reproduction, and is suitable for a process that requires real-time processing. There is an effect that an AV server can be realized.

According to the data recording / reproducing method according to any one of claims 10 to 17, a plurality of divided data obtained by dividing input data into predetermined units is recorded on a first non-linearly accessible recording medium. A first step of recording and generating a plurality of error correction code data for the divided data and recording the generated data on a second non-linearly accessible recording medium; and a division step recorded on the first recording medium in the first step. A second step of reproducing data and reproducing the error correction code data recorded on the second recording medium, and a step of changing the correspondence between the logical address and the physical address in the recording area of the first recording medium. When performing the allocation processing, the allocation processing is excluded from the other first recording media on which recording and reproduction are performed in the first step and the second step, and the reallocation processing is performed. Cormorants since a third step, enables error detection and correction data even during re-allocation processing, as a result, it is possible to perform re-allocation processing without interrupting the operation of the data recording and reproduction,
In addition, there is an effect that it is possible to realize data recording / reproducing processing suitable for processing requiring real-time properties.

Further, according to the data recording / reproducing method of the present invention, further, of the divided data and the error correction code data which are recorded on the first and second recording media, respectively, correspond to each other. The divided data other than the target divided data and the error-correcting code data are reproduced from the first recording medium and the second recording medium, respectively, and at least the data to be restored is made from the reproduced divided data and the error-correcting code data. An error correction process is performed on the divided data, and a restoration process step of restoring the divided data by recording the error-corrected divided data to be repaired on the first recording medium is provided. This has the effect that data error detection and correction can be performed even in data restoration processing after processing.

According to the data recording / reproducing method of the present invention, when the divided data is not recorded on the recording medium or reproduced from the recording medium, the data restoration processing is performed. Further, there is an effect that it is possible to perform data restoration processing without interrupting the operation of data recording and reproduction.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating an example of a configuration of a video server including a data recording / reproducing device according to an embodiment of the present invention.

FIG. 2 is a block diagram showing a configuration of a data recording / reproducing device according to one embodiment of the present invention.

FIG. 3 is a flowchart showing an operation in one time slot in a normal mode of the data recording / reproducing apparatus according to one embodiment of the present invention.

FIG. 4 is a flowchart following FIG. 3;

FIG. 5 is a flowchart showing an operation of a partial reconstruction process in FIG. 4;

FIG. 6 is a flowchart showing an operation in a reallocation processing mode of the data recording / reproducing apparatus according to one embodiment of the present invention.

FIG. 7 is a flowchart showing an operation in one time slot in the whole reconstruction processing mode of the data recording / reproducing apparatus according to one embodiment of the present invention.

FIG. 8 is an explanatory diagram for describing a format on a hard disk according to an embodiment of the present invention.

FIG. 9 is an explanatory diagram for explaining a correspondence relationship between a logical address and a physical sector number.

FIG. 10 is a characteristic diagram showing an example of a result of comparing reliability between the data recording / reproducing apparatus according to the embodiment of the present invention and a data recording / reproducing apparatus using RAID-3.

FIG. 11 is a block diagram illustrating an example of a configuration of a data recording / reproducing device using RAID-3.

FIG. 12 is a block diagram illustrating an example of a configuration of a data recording / reproducing apparatus using RAID-5.

[Explanation of symbols]

10 video server, 11 data recording / reproducing device, 12
... I / O processor unit, 21 _{1 to} 21 _k , 28 ₁ , 28
₂ ... HDD, 22 ... data distributor, 23 ₁ ~23 _k, 2
7 ₁ , 27 ₂ ... input memory, 24 _{1 to} 24 _k , 29 ₁ ,
29 ₂ ... _{controller, 25 1 ~25 k, 30 1} , 30
₂ ... output memory, 26 ... redundant code generator, 31 ... error corrector, 32 ... data multiplexer, 33 ... CPU.

──────────────────────────────────────────────────の Continued on the front page (51) Int.Cl. ⁶ Identification code FI H04N 5/92 H04N 5/92 H (72) Inventor Satoshi Katsura 6-7-35 Kita Shinagawa, Shinagawa-ku, Tokyo Sony Corporation Within (72) Inventor Jun Yoshikawa 6-7-35 Kita-Shinagawa, Shinagawa-ku, Tokyo Sony Corporation (72) Inventor Satoshi Aburaya 6-35, Kita-Shinagawa, Shinagawa-ku, Tokyo Inside Sony Corporation ( 72) Inventor Koichi Sato 6-7-35 Kita-Shinagawa, Shinagawa-ku, Tokyo Inside Sony Corporation (72) Inventor Tomohisa Shiga 6-35-35 Kita-Shinagawa, Shinagawa-ku, Tokyo Inside Sony Corporation (72) Inventor Hiroyuki Fujita 6-35, Kita-Shinagawa, Shinagawa-ku, Tokyo, Japan Sony Corporation (72) Inventor Masaki Hirose 6-35, Kita-Shinagawa, Shinagawa-ku, Tokyo Sony Corporation

Claims (17)

[Claims] 1. A plurality of divided data obtained by dividing input data by a predetermined unit are respectively recorded on a plurality of first non-linearly accessible recording media, and the divided data recorded on the first recording medium are recorded. Dividing data recording / reproducing means for reproducing data; generating error correction code data of the divided data;
An error correction code data recording / reproducing means for recording a plurality of the error correction code data on a non-linearly accessible recording medium and reproducing the error correction code data recorded on the second recording medium; For the first recording medium on which the reallocation process for changing the correspondence between the logical address and the physical address in the recording area of the first recording medium is to be performed, the other first recording medium on which the divided data recording / reproducing means performs recording / reproduction. And a reallocation processing control means for controlling the divided data recording / reproducing means so as to perform the reallocation processing. 2. The divided data generating means for dividing the input data into predetermined units to generate a plurality of the divided data; and the error correction code based on the divided data generated by the divided data generating means. Error correction code data generation means for generating data; and error correction processing for the divided data reproduced from the divided data recording / reproduction means using the error correction code data reproduced by the error correction code data recording / reproduction means. 2. The data recording / reproducing apparatus according to claim 1, further comprising: an error correction processing unit that performs the following, and a multiplexing unit that multiplexes and outputs each of the divided data output from the error correction processing unit. 3. The error correction processing means, when the divided data to be reproduced from the divided data recording / reproducing means cannot be obtained, the divided data reproduced from another divided data recording / reproducing means, and 3. The data recording / reproducing apparatus according to claim 2, wherein the divided data not obtained from the error-corrected code data reproduced from the error-corrected code data recording / reproducing means is error-corrected and restored. 4. The data recording / reproducing apparatus according to claim 1, wherein said error correction code data recording / reproducing means records said plurality of error correction codes on a plurality of said second recording media, respectively. 5. The data recording / reproducing apparatus according to claim 1, wherein the error correction code data is Reed-Solomon code data. 6. The data recording / reproducing apparatus according to claim 1, wherein the first and second non-linear accessible recording media are hard disks. 7. An error correction process for performing an error correction process on the divided data reproduced from the divided data recording / reproducing device using the error correction code data reproduced by the error correction code data recording / reproducing device. Means, among the plurality of divided data and the error correction code data, which are recorded on the first recording medium and the second recording medium, respectively, other than the divided data to be repaired, Controlling and reproducing the divided data recording / reproducing means and the error correction code data recording / reproducing means,
Controlling the error correction processing means to perform the error correction process on the divided data based on the reproduced divided data and the error correction code data, the error-corrected divided data of the repair target 2. A data restoration processing control means for performing a restoration processing of the divided data by controlling the divided data recording / reproducing means so as to record the divided data on a first recording medium.
A data recording / reproducing apparatus according to claim 1. 8. The data restoration processing control means records the divided data on the first recording medium or stores the divided data on the first recording medium.
8. The data recording / reproducing apparatus according to claim 7, wherein the restoration processing of the divided data is performed when the reproduction from the recording medium is not performed. 9. Converting data including video and audio data input from the outside into data recordable on a non-linearly accessible recording medium and outputting the data, and outputting the data output from the recording medium to the outside. A plurality of input / output processing means for converting and outputting the data output from each of the plurality of input / output processing means into a predetermined unit; A divided data recording / reproducing means for recording on the non-linearly accessible recording medium and reproducing the divided data recorded on the first recording medium; and generating a second non-linear data by generating error correction code data of the divided data. The error correction code data recorded on the accessible recording medium and the error correction code data recorded on the second recording medium are reproduced. Correction code data recording / reproducing means; and a first recording medium to be reallocated to change the correspondence between logical addresses and physical addresses in a recording area of the first recording medium. An AV server comprising: reallocation processing control means for controlling the divided data recording / reproducing means so as to perform the reallocation processing while excluding it from other first recording media for performing recording / reproduction. 10. A plurality of divided data obtained by dividing input data into predetermined units is recorded on a first non-linearly accessible recording medium, and a plurality of error correction code data for the divided data is generated. A first step of recording on the second non-linearly accessible recording medium; and a step of reproducing the divided data recorded on the first recording medium in the first step and recording the divided data on the second recording medium. Second reproducing the error correction code data
And performing the reallocation processing for changing the correspondence between the logical address and the physical address in the recording area of the first recording medium, wherein the recording and reproduction are performed in the first and second steps. Excluded from the first recording medium,
A third step of performing the reallocation processing. 11. A fourth step of performing an error correction process on the plurality of divided data reproduced in the second step from the error correction code data reproduced in the second step. 11. The data recording / reproducing method according to claim 10, wherein: 12. The method according to claim 5, further comprising: receiving the divided data output from the fourth step, multiplexing the divided data, and outputting the multiplexed data.
12. The data recording / reproducing method according to claim 11, further comprising the steps of: 13. The first step is to record the plurality of divided data on a plurality of first recording media, respectively, and to record the plurality of error correction code data on a plurality of second recording media. 11. The data recording / reproducing method according to claim 10, wherein: 14. The data recording / reproducing method according to claim 10, wherein said error correction code data is Reed-Solomon code data. 15. The data recording / reproducing method according to claim 10, wherein said first and second non-linearly accessible recording media are hard disks. 16. The method according to claim 16, further comprising the steps of:
The divided data and the error correction code data other than the divided data to be repaired among the divided data and the error correction code data corresponding to each other are referred to as the first recording medium and the second recording medium, respectively. From the reproduced divided data and the error correction code data to perform error correction processing on at least the divided data to be repaired, the error-corrected divided data of the repair target is the 11. The data recording / reproducing method according to claim 10, further comprising a restoration processing step of performing restoration processing of the divided data by recording the divided data on one recording medium. 17. The restoration process is performed when the division data is not recorded on the first recording medium or the division data recorded on the first recording medium is not reproduced. 17. The data recording / reproducing method according to claim 16, wherein: