JPH09265714A - Disk array device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To store long data and to improve reliability even for DRAW(direct read after write) type optical disk device or the like by using a non-rewritable storage device for data storage and a rewritable storage device for redundant data storage. SOLUTION: A disk array device 200 is composed of a CPU 201, DRAW type optical disk devices 211, 212, 213 and 214 for storing data, a DRAW type optical disk device 215 for storing parity data as redundant data, a magnetic disk device 217 for temporarily storing undetermined parity data. By storing redundant data (parity data) in the device 217, even if one of the five DRAW type optical disk devices and the magnetic disk device goes wrong, a function of generating data is provided. That is, by using the non-rewritable storage device such as the DRAW type optical disk device or the like, a large capacity and highly reliable device capable of reproducing data even if the storage device goes wrong is provided.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a disk array device including a plurality of storage devices including a storage device for storing at least one redundant data.

[0002]

2. Description of the Related Art A disk array device having a large storage capacity is the most needed device in the multimedia age.

Such a disk array device is rewritable as a storage device and has the same capacity when forming a disk array device comprising a plurality of storage devices including a storage device for storing at least one redundant data. It is composed of a plurality of magnetic disk devices and the like.

First, a conventional disk array device will be described with reference to FIGS. FIG. 20 shows a disk array device 100 including five magnetic disk devices. A SCSI controller (106, 107, 108, 109) that controls data transfer between the CPU 101 that controls the entire disk array device 100, the memory 102 that temporarily stores programs and data of the CPU 101, and the five magnetic disk devices. , 110)
And a magnetic disk device for storing data (111, 11
2, 113, 114, 115), the SCSI controller (106, 107, 108, 109, 110) and the memory 102, and data transfer means 105 for controlling data transfer of data stored in the magnetic disk device and redundant data. Of parity data and a parity generation unit 104 for generating data when a disk fails, and an interface 103 for connecting to a host computer. CPU 101, memory 102, interface 1
03, data transfer means 105, parity generation unit 104,
SCSI controller (106, 107, 108, 10
9, 110) are respectively connected to the bus 119.

In the present disk array device 100,
Storing redundant data (parity data) in one magnetic disk device 115 has a function of generating data even if one magnetic disk device of the magnetic disk devices storing data has a failure and is reliable. We provide disk devices with excellent performance.

That is, when a data write request is issued from the host computer via the interface 103, the CP
The U 101 allocates the magnetic disk device from the address and at the same time receives data from the host computer and temporarily stores it in the memory 102. On the other hand, under the control of the CPU 101, the data transfer means 105 is used to read the data of the sector of the magnetic disk device to be written and the data of the same sector of the parity data magnetic disk device 115 and store them in the memory 102. The CPU 101 writes write data (ND), old data (OD) written in the current data magnetic disk device, and old parity data (O) written in the parity data magnetic disk device 115.
PD), new parity data (NPD) is calculated, and then write data (ND) and new parity data (N) are calculated.
PD) is written in each magnetic disk device. The calculation of the new parity data at this time is performed by the parity generation unit 104 according to Equation 1 (+ is an exclusive OR).

[0007]

[Equation 1]

This will be described with reference to FIG. First, FIG. 21A shows the magnetic disk devices A111 and B1.
12, C113, D114, and P115 initial data are shown. Here, it is assumed that all are cleared to "00". Here, in order to simplify the explanation, 1 sector is 1
It is assumed to be byte and striped in byte units. The parity is even parity. Next, when a data write request is issued from the host computer, the data “E6” sent from the host computer is written to the memory 102 via the interface 103. At the same time, the magnetic disk device A111 is controlled by the CPU 101.
The older data “00” is the SCSI controller 106,
It is read out to the memory 102 via the data transfer means 105. Similarly, the old parity data “00” is read from the magnetic disk device P115, and the memory 102
Is stored. The parity generation unit 104 calculates parity data newer than these data. That is, N
PD = OD + ND + OPD = 00 + E6 + 00 = E6, and the write data “E6” is transferred to the data transfer means 105.
Then, it is read from the memory 102 and written to the magnetic disk device A111 via the SCSI controller 106. Similarly, the new parity data “E6” is written in the magnetic disk device P115. As a result, the data of the magnetic disk device (111, 112, 113, 114, 115) becomes as shown in FIG. Further, FIG.
As shown in (c), the data “2
Also when writing 1 ", the old data" 00 "is read from the magnetic disk device C113, the old parity data" E6 "is read from the magnetic disk P115, and the new parity data" C7 "is read by the parity generation unit 104. Calculate and write data “21” to the magnetic disk device C11.
The new parity data "C7" is written to the magnetic disk device P1.
Write to 15.

As described above, in response to a data write request from the host computer, it is necessary to calculate new parity data from old data and rewrite the old parity data in the storage device storing the parity data. . Next, regarding the reading of data from the host computer, as long as there is no failure in the magnetic disk device, it can be read simultaneously from each magnetic disk device and sent to the host computer via the interface 3. Here, the magnetic disk device A
If 111 fails and cannot be read, the other four magnetic disk devices (112, 113, 114,
115) Read out the data of the same sector and store it in the memory 1
02, and the parity generation unit 104 sets A = B + C.
By calculating the data as + D + P = 00 + 21 + 00 + C7 = E6, the data of the magnetic disk device A111 can be generated and sent to the host computer.

The above is the outline of the operation when a disk array device is constructed with a magnetic disk device as a conventional example.

[0011]

As described above, since the contents of the parity disk need to be changed when the contents are changed by writing data, it is impossible to rewrite the storage device data such as the write-once type optical disk and the data. It has been difficult to construct a disk array device by using it for storing parity data.

Conventionally, a rewritable magnetic disk device has been used as a storage device of a disk array device, but in a non-rewritable storage device such as a write-once optical disk,
Since it is necessary to rewrite redundant data, there is a problem that it is difficult to configure a highly reliable disk array device by RAID. Further, no measure for improving reliability has been taken for the ROM medium.

Further, in the drive device for handling portable media, different media can be handled by one drive, but media having different capacities could not be mixed in order to improve reliability due to arraying. .

Therefore, the present invention has been made in view of the above points, and its purpose is to use a non-rewritable storage device such as a write-once optical disc device, a ROM type medium, or a medium having a different capacity. Another object is to provide a highly reliable disk array device that stores redundant data.

[0015]

According to a first aspect of the present invention, there is provided a disk array device comprising a plurality of storage devices including a storage device for storing at least one redundant data. Among them, the storage device for storing the data is composed of a non-rewritable storage device, and is provided with a rewritable storage device as the storage device for storing the redundant data.

A disk array device according to claim 2 is
In a disk array device composed of a plurality of storage devices including a storage device for storing at least one piece of parity data, among the storage devices, the storage device for storing the data is composed of a write-once optical disc, The patite
A magnetic disk device is provided as a storage device for storing data.

A disk array device according to claim 3 is
In a disk array device including a plurality of storage devices including a storage device that stores at least one redundant data, the storage device is composed of a non-rewritable storage device, and further, redundant data is stored in the storage device. Temporary storage means for temporarily storing, and detection means for detecting that the redundant data has been fixed,
When the detecting means determines the redundant data, the redundant data
And a data transfer means for copying the data to the non-rewritable storage device.

A disk array device according to a fourth aspect is
In a disk array device including a plurality of storage devices including a storage device for storing at least one piece of parity data, the storage device is a write-once optical disk device, and a temporary storage for storing parity data. A storage means, a detection means for detecting that the parity data is fixed, and a copy of the parity data to the write-once optical disc when the parity data is fixed by the detection means.
And a data transfer means for performing the same.

A disk array device according to a fifth aspect is
In a disk array device including a plurality of storage devices including at least one storage device storing redundant data, a plurality of non-rewritable storage devices storing data and non-rewritable storage storing redundant data When data is written to a device and redundant data calculating means for calculating redundant data and a plurality of non-rewritable storage devices for storing data, the redundant data calculated by the redundant data calculating means is recorded as a history each time. Data transfer means for writing the non-rewritable storage device for storing the redundant data.

A disk array device according to claim 6 is
In a disk array device including a plurality of storage devices including a storage device that stores at least one piece of parity data, a plurality of write-once optical disks that store data, a write-once optical disk that stores parity data, and parity data are calculated. And a write-once optical disc that stores the parity using the parity data calculated by the parity data calculation unit as a history each time data is written to the plurality of write-once optical discs that store the data. A data transfer means for writing is provided.

A disk array device according to a seventh aspect is
In a disk array device composed of a plurality of storage means including a storage means for storing at least one redundant data, a detection means for detecting the type of the storage means and a table forming the disk array are created based on this information. When the table creation unit and the detection unit determine that the storage unit is a ROM type storage unit, a disk array is configured with a plurality of ROM type storage units, and the disk array is registered in the table creation unit. It is characterized by comprising redundant data calculating means for reading data and calculating redundant data, and storage means for storing the redundant data.

A disk array device according to claim 8 is
In a disk array device composed of a plurality of storage means including at least one storage means for storing parity data, an optical disk detection means for detecting whether the type of optical disk is a ROM type, a write-once type or a RAM type, and this information are used as a basis. A table creating unit for creating a table constituting the disk array and the R by the optical disk detecting means.
Multiple ROMs when judged to be OM type optical disks
Type optical discs to form a disk array, which is registered in the table creation unit, reads the data of the plurality of ROM type optical discs, calculates parity data, and a plurality of write-once optical discs for storing data. When data is written, the data transfer means writes the parity data calculated by the parity data calculation means as a history to the write-once type optical disk storing the parity each time.

A disk array device according to a ninth aspect is
In a disk array device composed of a plurality of storage means including a storage means for storing at least one redundant data, a detection means for detecting the capacity of the storage means and a table forming the disk array based on this information are created. Table creating section, redundant data calculating means for calculating redundant data, and data reproducing means and table creating section for reproducing broken data from data in other memory means constituting the disk array when the memory means fails. It is characterized by comprising a selecting means for recognizing the capacity of the storage means by the table created by and selecting the storage means for reproducing the data.

According to a tenth aspect of the present invention, there is provided a disk array device comprising a plurality of optical disks including a storage means for storing at least one piece of parity data, a detecting means for detecting the capacity of the optical disk, and this information. A table creating unit that creates a table that forms a disk array based on the above, a parity data calculation unit that calculates parity data, and, when the optical disk fails, reproduces broken data from the data of another optical disk that forms the disk array. It is characterized by further comprising: data reproducing means for selecting and an selecting means for recognizing the capacity of the optical disk by the table created by the table creating section and selecting the optical disk for playing the data.

[0025]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, a first embodiment of the present invention will be described with reference to the drawings. A disk array device including a non-rewritable storage device such as a write-once optical disc will be described below with reference to an embodiment of the present invention.

First, the first embodiment will be described with reference to FIG. FIG. 1 shows a disk array device 200 including a non-rewritable storage device. here,
A write-once optical disc device will be described as an example of a non-rewritable storage device.

The disk array device 200 controls the CPU 201 as a whole, a memory 202 for temporarily storing the programs and data of the CPU 201, and a memory 4 for storing data.
Recordable optical disk device (211, 212, 213,
214), a write-once optical disk device 215 that stores parity data that is redundant data, and a magnetic disk device 217 that temporarily stores undetermined parity data.
SCSI controllers (206, 207, 208, 2) that connect these disk devices and control the buses respectively.
09, 210, 216), data transfer means 205 for controlling data transfer of stored data, etc., parity generation section 204 for generating parity data which is redundant data and data at the time of disk failure, and host It is composed of an interface 203 for connecting to a computer.

The CPU 201, memory 202, interface 203, parity generation unit 204, data transfer means 205, SCSI controller (206, 207, 2)
08, 209, 210, 216) are respectively connected to the bus 219.

In this disk array device 200,
By storing redundant data (parity data) in one magnetic disk device 217, it is possible to generate data even if one of the five write-once optical disk devices and one magnetic disk device fails. It also provides a highly reliable disk device.

That is, when a data write request is issued from the host computer via the interface 203, the CP
The U 201 allocates the write-once optical disc device from the address, and at the same time, receives data from the host computer and temporarily stores it in the memory 202. On the other hand, under the control of the CPU 201, the data transfer means 205 is used to read the data of the same sector of the magnetic disk device for parity data 217 and store it in the memory 202. Since the data disk device is a write-once optical disk device, it cannot be rewritten.

Therefore, since writing to the same address can be performed only once, the old data is the initial data at the time of formatting and is in the clear state, so there is no need to read the old data as in the conventional example.

Therefore, the CPU 201 calculates new parity data (NPD) from the write data (ND) and the old parity data (OPD), and writes the write data (ND).
Is written to the write-once optical disc device, and new parity data (NPD) is written to the magnetic disc device 217. The calculation of the new parity data at this time is performed by Expression 2 (+ is an exclusive OR).

[0033]

[Equation 2]

This will be described with reference to FIG. 2 and the flow chart of FIG. First, FIG. 2A shows a write-once type optical disc 221, 222, 223, 2 loaded in a write-once type optical disc device 211, 212, 213, 214, 215.
24 and 225, and initial data of the magnetic disk device 217 are shown. Here, a disk array is configured with these five write-once optical disks as one unit, and data is stored in four write-once optical disk devices 221, 222, 223, and 224, and parity data is stored on the write-once optical disk 225. Remember. Further, the magnetic disk device 217 temporarily stores the parity data to be stored in the write-once optical disk 225. Here, all the data of the write-once type optical disk device is in the unrecorded state "-", and this state is "0".
It is assumed that the parity data is 0 ", the parity data is even parity, and is cleared to" 00 ".
When a data write request is issued from the host computer, the data “E6” is written in the memory 202 (step S
1). At the same time, the old parity data "00" is read from the magnetic disk device 217 storing the parity data under the control of the CPU 201 to the memory 202 via the SCSI controller 216 and the data transfer means 205 (step S2). The parity generation unit 204 calculates parity data newer than these data (step S).
3). That is, NPD = ND + OPD = E6 + 00 =
The write data “E6” is read from the memory 202 by the data transfer means 205, and the write-once type optical disk device 211 is read via the SCSI controller 206.
The write-once read-many optical disc 221 is loaded into the sector number 0 (step S4). Similarly, the new parity data “E6” is assigned to the sector number 0 of the magnetic disk device 217.
(Step S5). As a result, the data of each write-once type optical disc (221, 222, 223, 224, 225) and the magnetic disc device 217 becomes as shown in FIG. At this time, the striped parity data of sector number 0 is added to the write-once optical disks 222, 223, 2
Since the data of 24 is not determined, it is not written to the write-once type optical disc 225 for parity storage. Further, as shown in FIG. 2C, the data “0” is written on the write-once optical disc 222.
In the case of simultaneously writing "0", data "21" on the write-once optical disc 223 and data "00" on the write-once optical disc 224, the old parity data "E6" is read from the magnetic disk device 217, and the parity generation unit 204 reads it. Calculate new parity data “C7” and write data “0
Write data “2” to the write-once optical disc 222.
1 ”to the write-once optical disc 223 and write data“ 0 ”
0 "is written to the write-once optical disc 224, and new parity data" C7 "is written to the magnetic disc device 217. With respect to sector number 0, since the data of the four write-once optical discs for data has been determined (step S6), the magnetic disc device. Parity data “C temporarily stored in 217
7 "is read by the data transfer means 205, and SCSI
Parity data is written to the write-once optical disc 225 via the controller 210 and the write-once optical disc device 215 (step S7).

When the parity data is determined as described above, as shown in FIG.
Copy the parity data stored in. As described above, in response to a data write request from the host computer, new parity data is calculated from old parity data,
It is necessary to rewrite the old parity data in the storage device storing the parity data.

That is, in the invention according to the first embodiment, since the rewritable magnetic disk device 217 is used for parity data, the write-once optical disk 221,
Since the parity data can be rewritten even when 222, 223, and 224 are independently accessed, the disk array apparatus can be configured using a non-rewritable write-once optical disk apparatus.

As described above, since the optical disk can be used, a large-capacity disk array device can be constructed, and in the write-once optical disk device, it is not necessary to read old data, and the speed can be increased. In addition, regarding the reading of data from the host computer, the write-once optical disc device 2 is used.
As long as there is no failure in 11, 212, 213, 214 or the write-once optical disks 221, 222, 223, 224 become unreadable, they can be simultaneously read from each disk and sent to the host computer via the interface 203. . Here, when the write-once optical disc 221 becomes unreadable, the data of the same sector number of the other four write-once optical discs (222, 223, 224, 225) is read out and stored in the memory 202. By storing and calculating the data by A = B + C + D + P by the parity generation unit 204, the data of the write-once optical disc 221 can be reproduced and sent to the host computer. In the present embodiment, the parity data is copied from the magnetic disk device 217 to the write-once optical disc, but it may be written to both the write-once optical disc for parity 225 and the magnetic disk device 217 when the parity is determined. Also,
If there is a magnetic disk device with the same capacity as the optical disk, it is not necessary to provide a write-once optical disk for parity.
It suffices that only the magnetic disk device has the parity data.
In addition, five write-once type optical disk devices (211, 212,
213, 214, 225) are used, but reading and writing may be performed by using a library device and exchanging the write-once optical disc with one write-once optical disc device. In short, it suffices that the write-once type optical discs are arrayed. There is no problem whether the write-once optical disc is striped or what the RAID level is. Further, in the present embodiment, the magnetic disk device having the same capacity for temporarily storing the parity data is used, but when striping is performed and writing is performed in the order of sectors forming the parity of the plurality of write-once optical disks for data, for example, As shown in FIG. 3, the write-once optical disc (A,
(B, C, D) Writing data to each sector is A-0
Sector, B-0 sector, C-0 sector, D-0 sector,
In the order of A-1 sector ..., it is not necessary to have a magnetic disk device having the same capacity as the write-once optical disk, but a temporary storage buffer having a capacity of 1 sector is sufficient.

Next, a second embodiment will be described with reference to FIG. FIG. 5 shows a disk array device 300 composed of a non-rewritable storage device. Here, a write-once optical disc device will be described as an example of a non-rewritable storage device. A CPU 301 that controls the entire disk array device 300, a memory 302 that temporarily stores programs and data of the CPU 301, and four write-once optical disk devices (311, 312, 313, 3) that store data.
14) and a write-once optical disk device 315 that stores parity data that is redundant data, and a temporary storage unit 317 that temporarily stores data of a sector in which parity data is not fixed.
And a SCSI controller (306, 307, 308,) that connects these disk devices and controls the bus respectively.
309, 310) and the data transfer means 305 for controlling the data transfer of the stored data and the like, and the connection between the host computer and the parity generation unit 304 for generating the parity data which is the redundant data and the data at the time of the disk failure. And an interface 303 for performing. And C
PU 301, memory 302, interface 303, parity generation unit 304, data transfer means 305, SCSI
Controller (306, 307, 308, 309, 31
0) are connected to the buses 319, respectively.

In the present disk array device 300,
By storing data with undetermined parity in the temporary storage unit 317, the additional write-once type for parity after the data of the same sector of the four write-once data optical discs (321, 322, 323, 324) for generating parity is determined. 5 for writing and storing parity data on the optical disk 325
5 write-once optical disks (321, 322, 323, 3
24, 325) has a function of generating data even if reading becomes impossible, and provides a highly reliable disk device.

That is to say, referring to the flowchart of FIG. 19, when a data write request is issued from the host computer via the interface 303, the CPU 301 selects the write-once optical disk device and the sector number from the address. At the same time, the data is received from the host computer (step S11) and temporarily stored in the temporary storage unit 317 (step S13). At this time, the parity is four write-once data optical discs (321, 322, 323,
324), the same sector of the write-once type optical disc for parity 325 is allocated to the same sector, but four write-once type optical discs for data (321, 322, 323, 3) are assigned.
24) data is stored in the temporary storage unit 317 until the data of the same sector is determined, and when the data and the parity are determined, four write-once optical discs for data (321,
322, 323, 324) reads the data temporarily stored in the temporary storage unit 317 (step S14),
Further, the parity generation unit 304 generates parity data from these data (step S15), and writes the parity data to the write-once optical disc 325 for parity via the data transfer unit 305, the SCSI controller 310, and the write-once optical disc device 315. Is performed (step S16). Further, the data is written on each write-once optical disc (321 to 324). At this time, the parity data is calculated by Equation 3 (+ is an exclusive OR).

[0041]

(Equation 3)

This will be described with reference to FIGS. 6, 7 and 8 and the flow chart of FIG. First, FIG.
Is a write-once optical disc device 311, 312, 313, 3
Write-once optical disks 321 and 3 loaded in
Initial data of 22, 323, 324, and 325 are shown. Here, a disk array is configured with these five write-once optical disks as one unit, and the write-once optical disk 3
Data is stored on the four recording sheets 21, 322, 323, and 324, and parity data is stored on the write-once optical disc 325. Here, all of the data on the write-once optical disc is in the unrecorded state “−”, and this state is considered to be the state of “00”, and the parity data is assumed to be even parity and is cleared to “00”.

Next, FIG. 10 is a table showing the configuration of the disk array unit stored in the temporary storage unit 317. It can be seen that the unit 0 is composed of a write-once optical disc for data (A0, B0, C0, D0) and a write-once optical disc for parity data P0. Figure 7
Is a write-once optical disc for data (A0, B0, C0, D
It is a table showing whether the data of each sector of 0) has been confirmed. FIG. 8 is a table showing definite data before writing on the data write-once type optical discs (A0, B0, C0, D0) in the sector numbers shown in FIG. These are stored in the temporary storage unit 317.

Next, when a data write request is issued from the host computer, the data "E6" is temporarily written in the memory 202, and is stored in the temporary storage section 317 as the finalized data of the data write-once optical disc 321 (A0) as shown in FIG. At the same time, the value "1" is written in the data confirmation table as shown in FIG. 7 (b). At this time, only A0 of the data confirmation table is "1" and B0, C0, D0
Is “0” and undetermined, the write-once optical disks 321, 3
No data is written to 22, 323, 324, and 325. The same operation is repeated in response to the write request from the host computer, and the write-once optical disks 321, 322, 323, 32 forming the parity as shown in FIG.
When the data of sector 0 of 4 is confirmed and all the corresponding sectors of the data confirmation table become “1”, the data shown in FIG. 8C stored in the temporary storage unit 317 is read by the data transfer means 305, Write-once optical disc 32
Data is written to 1, 322, 323, and 324. In parallel with this, the parity generation unit 304 reads the confirmed data from the temporary storage unit 317, calculates the parity by the equation 4 (+ is an exclusive OR), and then writes the parity to the write-once type optical disc 325 for parity. Write data. As a result, as shown in FIG. 6B, the data of the sector 0 and the parity are write-once optical disks 321, 322,
It is written in 323, 324, and 325 and is determined.

[0045]

(Equation 4)

As described above, when the parity data is determined, the data and the parity are written to the write-once optical disc, so that the disk array (RAID) system can be configured by the write-once optical disc. Since the optical disc can be used in this manner, a large-capacity disc array device can be configured, and in the write-once optical disc device, there is no need to read old data, and high speed operation can be achieved. Regarding reading of data from the host computer, the write-once type optical disk devices 311, 312, 313,
314 has no failure, write-once optical disks 321, 32
Unless 2, 323, 324 become unreadable, they can be read simultaneously from each disk and sent to the host computer via the interface 303. If the write-once read-many optical disc 321 becomes unreadable, the unit configuration is recognized from the unit configuration table of FIG. 10, and the other four write-once optical discs (322, 32) are recognized.
3, 324, 325) of the same address are read and stored in the memory 302, and the parity generation unit 304 calculates the data by A0 = B0 + C0 + D0 + P0 to reproduce the data of the write-once optical disc 321, and the host computer Can be sent to. In the present embodiment, only the data of the same sector is stored in the temporary storage unit, and the parity data is calculated after the data is fixed. However, before the data is prepared, the calculation is sequentially performed and the temporary storage unit 317 stores the data. You may remember it. This also facilitates calculation of parity data.

Next, a third embodiment will be described with reference to FIG. FIG. 11 shows a disk array device 400 composed of a non-rewritable storage device. Here, a write-once optical disc device will be described as an example of a non-rewritable storage device. A CPU 401 that controls the entire disk array device 400, a memory 402 that temporarily stores programs and data of the CPU 401, and four write-once optical disk devices (411, 412, 41) that store data.
3, 414) and a write-once type optical disk device 415 for sequentially storing the history of parity data which is redundant data, and an SCS for connecting these disk devices and controlling the bus respectively.
I controller (406, 407, 408, 409, 4
10) and a data transfer unit 405 for controlling data transfer of stored data, a parity generation unit 404 for generating parity data which is redundant data and data for a disk failure, and an interface for connecting the host computer. And 403.

CPU 401, memory 402, interface 403, parity generator 404, data transfer means 4
05, SCSI controller (406, 407, 40
8, 409, 410) are respectively connected to the bus 419.

In this disk array device 400,
The parity history is stored in the parity-type write-once optical disc each time data is written. As a result, the writing history is saved, which helps prevent tampering and the like, and improves the evidence of the optical disc. That is, when writing is performed on four data write-once optical discs (421, 422, 423, 424) that generate parity, parity data is generated with the confirmed data even if the data in the same sector is not confirmed. Then, the parity data is written and stored in the write once optical disc for parity 425. Therefore, the five write-once type optical discs (421, 422, 423, 424, 42 that constitute the disc array are formed.
It provides a highly reliable disk device having the function of generating data even if the reading of 5) becomes impossible.

When a data write request is issued from the host computer via the interface 403, the CPU 401
Selects the write-once type optical disk device and the sector number from the address, receives the data from the host computer, and temporarily stores the data in the memory 402. At this time,
The parity is four write-once optical discs for data (42
1, 422, 423, 424), the same sector of the write-once type optical disc for parity 425 is allocated. For each writing, the parity data is generated by the data of the same sector of the four data write-once optical discs (421, 422, 423, 424), but if the data is undetermined, the data is set to "00". The parity generation unit 404 generates parity data, and the write-once type optical disc 42 for parity is generated via the data transfer unit 405, the SCSI controller 410, and the write-once type optical disc device 415.
The parity data is written in 5. At this time, the parity data is calculated by Equation 5 (+ is an exclusive OR).

[0051]

(Equation 5)

This will be described with reference to FIGS. 12 and 13 and the flow chart of FIG. First, FIG. 12A shows a write-once type optical disc 421, 422, 423, 42 loaded in the write-once type optical disc device 411, 412, 413, 414.
4 shows the initial data of 4. Here, a disk array is configured by using these five write-once optical disks as one unit, and the write-once optical disks 421, 422, 423, and 42.
Data is stored in 4 sheets of 4. In addition, FIG.
It is the contents of the write-once optical disc for parity, and since a history of parity is left every time data is written, the sector number indicating the address of the parity and the data indicating which disc of the four write-once optical discs for data has been determined. It is composed of fixed bits and actual parity data. Here, all of the data on the write-once optical disc is in the unrecorded state “−”, and this state is considered to be the state of “00”, and the parity data is assumed to be even parity and is cleared to “00”.

Next, when a data write request is issued from the host computer, the data "E6" and "00" are once written in the memory 402 (step S21), and the data write-once optical discs 421 (A0) and 422 (B0) are written. Memorized in. At the same time, the parity generation unit 404 generates a parity from the data of the write-once optical discs A0 and B0 (steps S22 and S23). At this time, since the data of the write-once optical discs C0 and D0 has not been decided, "00"
And Therefore, P0 = A0 + B0 = E6.
As shown in (b), the write-once type optical disc 425 for parity is used.
Information of sector number = 0, data confirmation bit = 0011, and parity data P0 = E6 is written in (step S25). Next, the data “21,0” is sent from the host computer.
When a write request of 0, 1C, 20, 81, 1F ″ occurs, the write request is once written in the memory 402, and the data transfer unit 4
By 05, data is written to the write-once optical disc through the SCSI controller and the write-once optical disc device.
This is shown in FIG. On the other hand, in parallel, the parity generation unit 404 generates parity data. By this writing, the data of the sector number 0 and the sector number 1 are fixed, and thus the parity is also fixed. Therefore, FIG.
As shown in (c), the history is added, and as a result, in sector number 0, the data confirmation bit is (1111),
The parity is fixed to "C7". In addition, the data confirmation bit of the sector number 1 is (1111), and the parity is confirmed to be "A2".

As described above, by leaving the history of the parity on the write-once type optical disc for parity 425, it becomes effective when the data has a failure or is tampered with. Further, the disk array can be constructed using only the write-once optical disk.

Next, a fourth embodiment will be described with reference to FIG. FIG. 15 shows a disk array device 500 including an optical disk device. Here, a CD drive will be described as an example of the optical disc device. Here, the CD drive can access a ROM type CD-ROM, a write-once type CD-R, and a rewritable type CD-E. A CPU 501 for controlling the entire disk array device 500, a memory 502 for temporarily storing programs and data of the CPU 501, four optical disk devices (511, 512, 513, 514) and an optical disk device 515 for storing parity data as redundant data. And a magnetic disk device 517 that stores a disk management table and the like, and SCSI controllers (506, 507, 5) that connect these disk devices and control the bus.
08, 509, 510, 516) and a data transfer means 505 for controlling data transfer of stored data and the like, a parity generation unit 504 for generating parity data as redundant data and data at the time of a disk failure, and a host computer. An interface 503 for connection with the optical disc device and an optical disc detection means 530 for detecting the type of the optical disc loaded in the optical disc device.

CPU 501, memory 502, interface 503, parity generator 504, data transfer means 5
05, optical disk detection means 530, SCSI controller (506, 507, 508, 509, 510, 51
6) are respectively connected to the bus 519. In the present disk array device 500, the optical disk detection means 53
0 to the optical disk device (511, 512, 513, 5
It is determined whether the optical disk loaded in 14) is a CD-ROM, a CD-R or a CD-E, and the data capacity (number of sectors) and the like. In the case of CD-R and CD-E, in order to write data, parity is created, an optical disk for parity is created, when disks are arrayed, the disks to be arrayed are determined, and parity is created at the time of writing, thereby making RAID. Is possible. However, since the CD-ROM is read-only, no writing occurs, and thus no parity is generated. However, even in the case of a CD-ROM, important information is stored when the optical disk device fails, is stored in a different autochanger, and the autochanger fails, or the CD-ROM itself becomes unreadable. You won't get it. Therefore, even in the case of a CD-ROM, by making a disk array and having a storage device for parity,
The reliability can be greatly improved. Therefore, in the present embodiment, the write-once optical disc 5 is used for storing the parity data.
25, the loaded optical disc is detected by the CD-ROM and the optical disc detection means 530, and when it is not RAID-ized, data is read from a plurality of CD-ROMs and the parity generation unit 504 determines the parity for each sector. Generate and write the parity data on the write-once optical disc. As a result, the five optical discs (521, 522, 523, 524, 52) forming the disc array are
It provides a highly reliable disk device having the function of generating data even if the reading of 5) becomes impossible. When a data read request is issued from the host computer via the interface 503, the CPU 501 reads data from the designated optical disk 521. At this time, the data transfer means 505 causes the SCSI controller 50 to
6. Read data via the optical disk device 511,
It is temporarily stored in the memory 502. After that, it is sent to the host computer via the interface 503. CD-RO
When M is loaded, the optical disc detection means 530 is a CD-R.
The header of the OM 521 is read (step S32), and the type of optical disc is determined. Since the optical disc is a CD-ROM, the CPU 501 reads the management information and determines whether the CD-ROM is RAID-configured (step S34). If not RAID-configured, the RAID is configured with other optical disks to generate parity. FIG. 16 is a disk management table stored in the magnetic disk device 517 (step S35). Each optical disc is composed of a unique disc number, a disc type, a data type, and a unit number constituting a RAID. For the disc type, "0" is CD-ROM, "1"
Is CD-R, "2" is CD-E (capacity 200B),
"3" represents a CD-E (capacity 100B), data type "0" represents a data optical disc, and "1" represents a parity optical disc. The unit number represents a unit that constitutes a disk array of a plurality of optical disks. "*"
Indicates that it is treated as a single unit and does not constitute a disk array. The CPU 501 reads this information from the magnetic disk device 517, and if the RAID is not configured, configures the RAID together with other disks. Here, when the optical discs with the disc numbers 0, 3, 4, and 5 are loaded, the optical disc detection means 530 reads the header of the optical disc, determines the type (step S36), and registers it in the disc management table (step S38). . At this time, the unit number is “*” because it has not yet been converted to RAID. Therefore, RAID is used to improve reliability. The RAID conversion may be automatically performed by the disk array device or may be instructed by the application of the host. The CPU 501 selects the optical disks constituting the unit, selects the disk numbers 0, 3, 4, and 5 as the data optical disks, selects the CD-R type optical disk of the disk number 1 as the parity optical disk, and selects the respective optical disks. Is set in the optical disk device. At this time, if an auto changer mechanism is used, it can be set automatically. Next, by the data transfer means, disk numbers 0, 3, which are data disks,
Data for one sector is read from the fourth and fifth CD-ROMs (step S39) and stored in the memory 502. The parity generation unit 504 generates parity data from the data of the four sectors (step S40), and the memory 5
02 is stored. This data is transferred by the data transfer means 505 to the SCSI controller 510 and the optical disk device 51.
5 via a write-once type optical disc (C
D-R) (step S41). By repeating this for all sectors, a parity optical disk can be created (FIG. 17). In the present embodiment, a write-once optical disc was used as the optical disc for parity.
Either an AM type disk or a magnetic disk may be used. In short,
Any writable storage device may be used.

Next, a fifth embodiment will be described.
The present invention realizes a disk array composed of a plurality of storage devices having different capacities. As optical discs, there are discs of various capacities, and the same drive device enables recording and reproduction of plural types of optical discs. Therefore, optical discs of different types and optical discs of different capacities are handled together by an autochanger or the like. Conventionally, it has been impossible to form a disk array using storage devices having different capacities, but the present invention makes this possible by having a management table. Hereinafter, an embodiment according to the present invention will be described with reference to FIGS. 15 and 16 and FIG. 19 which is a flowchart. Since FIG. 15 has been described in the fourth embodiment, description thereof will be omitted. Now, consider a case where a disk array is configured by disk numbers 2, 7, 8, 9, and 10 in the disk management table of FIG. The types of these optical disks are the CD-E, which is a rewritable RAM type optical disk of disk numbers 2, 7, 8, and 10.
Therefore, the capacity is 200B. Disc number 9 is CD
-E100B type. A disk array of unit number 1 is constructed by these five optical disks. At this time, the disc number loaded in the optical disc device 511 is 2, the disc number loaded in the optical disc device 512 is 7, the disc number loaded in the optical disc device 513 is 8, and the disc number loaded in the optical disc device 514 is 9. The disc number loaded in the optical disc device 515 is 10. Therefore, only the optical disk loaded in the optical disk device 514 of the data optical disk is 100B and the capacity is half that of the other optical disks. Generation of parity when data is written on these optical disks will be described. Interface 50 from host computer
When a data write request is issued via the CPU 3, the CPU 50
1 selects the optical disk 521 from the address and writes data. Now, sector number 0 of optical disk A0
When a write request of “44” comes to the interface 5,
Data is temporarily stored in the memory 502 from 03. Since the optical disk is a RAM type, it is necessary to read old data and old parity from the optical disk. So, CP
Using the data transfer means 505 under the control of U501,
The data “E6” of sector number 0 is read from the optical disc 521 of disc number 2 via the SCSI controller 506 and the optical disc device 511 and stored in the memory 502. Similarly, the data “C7” of sector number 0 is read from the optical disk 525 of disk number 10 via the SCSI controller 510 and the optical disk device 515 and stored in the memory 502. Next, the parity generation unit 50
4, the new parity is calculated by the following equation. That is,
The new parity data becomes NPD = 44 + E6 + C7 = 65 by the expression 6 (+ is the exclusive OR) and the temporary memory 50
Stored in 2. Next, the new data “44” is written to the sector number 0 of the optical disc 521 of the disc number 2 via the SCSI controller 506 and the optical disc device 511 by using the data transfer means 505, and similarly, the new parity data “65” is written. Is written to the sector number 0 of the optical disc 525 of the disc number 10 via the SCSI controller 510 and the optical disc device 515. Next, when a write request of "38" is sent to the sector number 100 of the optical disk B0, the interface 503 once outputs the memory 502.
Store the data in.

[0058]

(Equation 6)

Next, the old data and old parity data CPU
By using the data transfer means 505 under the control of 501, S
The data “00” of the sector number 100 is read from the optical disc 522 of the disc number 7 via the CSI controller 507 and the optical disc device 512 and stored in the memory 502. Similarly, the SCSI controller 510,
Data “00” of sector number 100 from the optical disc 525 of disc number 10 through the optical disc device 515.
Is stored in the memory 502. Next, the parity generation unit 504 calculates a new parity according to the following equation. That is, the new parity data becomes NPD = 38 + 00 + 00 = 38 by the expression 7 (+ is exclusive OR) and is stored in the temporary memory 502.

[0060]

(Equation 7)

Next, the new data "38" is transferred to the data transfer means 5
05 via the SCSI controller 507 and the optical disk device 512.
22 sector number 100, and similarly, the new parity data “38” is written to the SCSI controller 510,
Writing is performed on the sector number 100 of the optical disc 525 having the disc number 10 via the optical disc device 515. In this case, referring to the disk management table of FIG. 16, the parity data of the sector number 100 is the optical disk of the unit number 1, the disk type of the optical disk of the disk number 9 is "3", and the capacity is 100B. 1/2
It is. Therefore, in the present embodiment, for sector numbers 100 to 199, a disk array is configured by three data optical disks and one parity optical disk. This can be easily determined because the disk management table of FIG. 16 is created when configuring the unit. Therefore, considering the case where the optical disc of the disc number 2 cannot be read (step S51), first, when the read request of the sector number 0 occurs, the disc management table
Since the disk array is composed of five optical disks (step S53), A0 (0) = B0 (0) + C
(0) + D (0) + P (0) = 00 + 21 + 00 + C7
= Playback is possible with E6. That is, the data transfer control means 50
5, the data “00” of sector number 0 is read from the optical disc of disc number 7 loaded in the optical disc device 512 and stored in the memory 502. Similarly, the data “21” with the sector number 0 is read from the optical disc with the disc number 8 loaded in the optical disc device 513, and the data “00” with the sector number 0 is read from the optical disc with the disc number 9 loaded in the optical disc device 514. The data "C7" of sector number 0 is read from the optical disc of disc number 10 loaded in the optical disc device 515 and stored in the memory 502 (steps S54 and S55).
Next, the parity generation unit 504 reproduces the data by the above calculation (step S56), stores "E6" in the memory 502, and sends it to the host computer via the interface 503 (step S57).

Next, when a read request for the sector number 100 occurs, the disk array is composed of four optical disks excluding the optical disk D0 from the disk management table (step S55), so A0 (100). =
B0 (100) + C (100) + P (100) = 38 +
Playback is possible at 00 + 38 = 00. That is, the data transfer control means 505 reads the data “38” of sector number 0 from the optical disc of disc number 7 loaded in the optical disc device 512 and stores it in the memory 502. Similarly, the disc number 8 loaded in the optical disc device 513
The data "00" of sector number 0 is read from the optical disc of No. 10 and the data "38" of sector number 0 is read from the optical disc of No. 10 loaded in the optical disc device 515 and stored in the memory 502 (steps S54, S54).
55). Next, the parity generation unit 504 reproduces the data by the above calculation (step S56), and "0
"0" is stored in the memory 502 and sent to the host computer via the interface 503 (step S57).
As described above, by using the disc management table created by the optical disc detection means 503, a disc array can be realized even when discs having different capacities are mixed.

In the present embodiment, the optical disk has been described, but another storage device may be used. Although the capacities of the sectors are the same, if they are different, a sector management table may be created and the sectors of the respective storage devices may be associated with each other.

[0064]

As described above, according to the present invention, it is possible to use a non-rewritable storage device such as a write-once type optical disc device, which has a large capacity and is capable of reproducing data even if the storage device fails. A high disk array device can be constructed.

Since the optical disc is portable,
It is possible to further increase the capacity by using the library device (auto changer), which can reduce the bit unit price. In a disk array device composed of a plurality of non-rewritable storage devices, by using a rewritable storage device for at least one storage device that stores redundant data, the redundant data changes when data is written. Also, rewriting is possible and the disk array device can be configured.

Further, even if the storage device for storing the redundant data is a non-rewritable storage device, by providing the temporary storage means for temporarily storing the redundant data, the storage device for the redundant data after the redundant data is confirmed. The disk array device can be configured by writing to

Further, by sequentially storing the history of the parity, the evidence ability is improved and the reliability can be improved. Further, by providing the disc management table by the optical disc detection means, it is possible to add the parity optical disc to the ROM type optical disc and improve the reliability.

Furthermore, since disks having different data capacities can be mixed to form a disk array,
Even when a plurality of types of optical disks are mixed, such as when using an autochanger, the disk array can be flexibly configured.
As described above, a disk array device called RAID, which was conventionally only possible with a RAM type storage device, can be realized with a write-once type or ROM type storage device.

[Brief description of drawings]

FIG. 1 is a block diagram of a disk array device according to a first embodiment.

FIG. 2 is a diagram illustrating data on a disc according to the first embodiment.

FIG. 3 is a diagram illustrating a stripe structure according to the first embodiment.

FIG. 4 is a flowchart illustrating the first embodiment.

FIG. 5 is a block diagram of a disk array device according to a second embodiment.

FIG. 6 is a diagram illustrating data on a disc according to the second embodiment.

FIG. 7 is a diagram illustrating data management information in a temporary storage unit according to the second embodiment.

FIG. 8 is a diagram illustrating data in a temporary storage unit according to the second embodiment.

FIG. 9 is a flowchart illustrating a second embodiment.

FIG. 10 is a diagram illustrating a unit that constitutes the disk array according to the second embodiment.

FIG. 11 is a block diagram of a disk array device according to a third embodiment.

FIG. 12 is a diagram illustrating data on a data disk according to the third embodiment.

FIG. 13 is a diagram illustrating data on a parity data disk according to the third embodiment.

FIG. 14 is a flowchart illustrating a third embodiment.

FIG. 15 is a diagram illustrating a unit that constitutes a disk array according to fourth and fifth embodiments.

FIG. 16 is a disk management table according to the fourth and fifth embodiments.

FIG. 17 is a flowchart illustrating a fourth embodiment.

FIG. 18 is a diagram illustrating data on a data disc according to the fifth embodiment.

FIG. 19 is a flowchart illustrating a fifth embodiment.

FIG. 20 is a block diagram of a disk array device for explaining a conventional example.

FIG. 21 is a diagram illustrating data on a disc in a conventional example.

[Explanation of symbols]

201 ... CPU, 202 ... Memory, 203 ... Interface, 204 ... Parity generation unit, 205 ... Data transfer means, 211-215 ... Write-once optical disc device, 221-
225 ... Write-once optical disc.

Claims (10)

[Claims] 1. A disk array device comprising a plurality of storage devices including a storage device for storing at least one piece of redundant data, wherein the storage device for storing data is rewritten. A disk array device comprising a non-rewritable storage device and comprising a rewritable storage device as a storage device for storing the redundant data. 2. A disk array device comprising a plurality of storage devices including a storage device for storing at least one piece of parity data, wherein the storage device for storing the data among the storage devices is a write-once type. A disk array device comprising an optical disk and comprising a magnetic disk device as a storage device for storing the parity data. 3. A disk array device comprising a plurality of storage devices including a storage device for storing at least one redundant data, wherein the storage device is a non-rewritable storage device,
Further, a temporary storage means for temporarily storing redundant data, a detection means for detecting that the redundant data is fixed, and a redundant data when the redundant data is fixed by the detecting means.
A disk array device comprising: a data transfer means for copying data to the non-rewritable storage device. 4. A disk array device comprising a plurality of storage devices including a storage device for storing at least one piece of parity data, wherein the storage device comprises a write-once optical disc device, and further parity data. A temporary storage means for temporarily storing the parity data, a detection means for detecting that the parity data has been determined, and a data for copying the parity data to the write-once optical disc when the parity data is determined by the detection means. A disk array device comprising a transfer means. 5. A disk array device comprising a plurality of storage devices including at least one storage device for storing redundant data, and a plurality of non-rewritable storage devices for storing data, and for storing redundant data. When data is written to a non-rewritable storage device, redundant data calculation means for calculating redundant data, and a plurality of non-rewritable storage devices for storing data, it is calculated by the redundant data calculation means each time. A disk array device comprising: data transfer means for writing redundant data as a history in a non-rewritable storage device for storing the redundant data. 6. A disk array apparatus comprising a plurality of storage devices including at least one storage device for storing parity data, and a plurality of write-once optical discs for storing data and a write-once optical disc for storing parity data. And parity data calculating means for calculating parity data, and when data is written to a plurality of write-once optical discs for storing data, each time the parity is calculated using the parity data calculated by the parity data calculating means and the parity is stored. A disk array device comprising a data transfer means for writing to a write-once optical disc. 7. A disk array device comprising a plurality of storage means including at least one storage means for storing redundant data, wherein a disk array is formed based on this detection means for detecting the type of the storage means and this information. When the table creating unit for creating the table to be configured and the detecting unit determine that the ROM type storage unit is a ROM type storage unit, a disk array is configured with a plurality of ROM type storage units, and the disk array is registered in the table creating unit to store the plurality of ROMs. A disk array device comprising: a redundant data calculating unit for reading data from a storage unit of a type to calculate redundant data; and a storing unit for storing the redundant data. 8. A disk array device comprising a plurality of storage means including a storage means for storing at least one piece of parity data, wherein the type of optical disk is a ROM type, a write-once type or a RAM.
The optical disk detection means for detecting whether the type is a table, a table creating section for creating a table forming a disk array based on this information, and a disk array with a plurality of ROM type optical disks when the optical disk detection means determines that the optical disk is a ROM type optical disk. And writing to the plurality of write-once optical discs for storing the data, and the parity data calculating means for reading the data of the plurality of ROM type optical discs and calculating the parity data. At this time, each time, the disk array device is provided with a data transfer means for writing the parity data calculated by the parity data calculation means as a history into a write-once type optical disk storing the parity. 9. A disk array device comprising a plurality of storage means including a storage means for storing at least one redundant data, wherein a disk array is formed based on the detection means for detecting the capacity of the storage means and this information. A table creating unit that creates a table to be configured, a redundant data calculating unit that calculates redundant data, and a data reproducing unit that reproduces broken data from the data of another storage unit that constitutes the disk array when the storage unit fails. And a selection unit for recognizing the capacity of the storage unit by the table created by the table creation unit and selecting the storage unit for reproducing the data. 10. A disk array device comprising a plurality of optical disks including a storage means for storing at least one piece of parity data, wherein a disk array is constructed based on the detection means for detecting the capacity of the optical disk and this information. A table creating unit for creating a table, a parity data calculating unit for calculating parity data, a data reproducing unit for reproducing broken data from data of another optical disc forming the disc array when the optical disc fails, and the table producing unit A disk array device comprising a selection means for recognizing an optical disk capacity from a table created by the unit and selecting an optical disk for reproducing data.