JPH10254644A - High-reliability storage system using detachable storage medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To hold data with high reliability by using an irreducible necessary number of storage media by holding redundant data computed from data held in a data holding means in a redundant data storage means. SOLUTION: Two DVD storage drives 551 and 552 are connected to a RAID control means 600. Further, 12 DVD storage media 511 to 518 and 521 to 524 are provided. The DVD storage media 511 to 518 and 521 to 524 are held in corresponding storage medium holding means 558. Data holding means 551 constitute DVD storage device 551 and eight DVD storage media 551 to 518. Further, the DVD storage devices 552 and four DVD storage media 521 to 524 constitute a redundant data holding means 502. Thus, the DVD storage drives 551 and 552 are shared by the storage media 511 to 518 and 521 to 524. In the redundant data storage means, the redundant data computed from the data held in the data holding means are held.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention utilizes a removable storage medium and a storage device such as a magneto-optical disk device for reading and writing from and on the removable storage medium, an optical disk device, and a magnetic disk device. More particularly, the present invention relates to a control method for realizing a large-capacity and highly-reliable storage system, and more particularly to a control method for realizing a storage device with less than the total number of removable storage media.

[0002]

[Prior art]

● RAID is a technology that increases the read / write speed by moving multiple hard disk drives in parallel instead of the hard disk drive (HDD: Hard Disk Drive) that was conventionally used alone, and increases reliability by using a redundant configuration. Disk array device (RAID: Redundant A)
rrays of Inexpencive Disks) (Paper: "A Case for Redundant Arrays of Inexpencive Di
sks (RAID) ", David A. Patterson, Garth Gibson, and
Randy H. Katz, Computer Science Division Department
of Electrical Engineering and Computer Sciences,
University of Calfornia Berkeley). In the above paper, depending on the configuration of the disk array device,
Numbers from 5 to 5 are given. RAID level 3
Is expected to improve performance for sequential access that transfers a large amount of data, and it is well known that RAID level 5 is expected to improve performance for random access in which a large number of small-sized reads and writes occur (Nikkei Electronics 1993/4/26 No.579 P77-91 "Special Feature Disk Array Device").

[0003] Hereinafter, in this specification, a magnetic disk device using a non-removable storage medium is called a "hard disk device", and a storage device using a removable storage medium (a magneto-optical disk device, an optical disk device, Magnetic disk drive, etc.).

[0004] What is RAID control? Here, RAID control will be described with reference to FIGS. 9 and 10. FIG. As shown in FIG. 9, the disk array device 101 has a plurality of hard disk devices (111 to 115). RAID control refers to how data is distributed among a plurality of hard disk drives, how redundant data is created, and where redundant data is stored (RAID).
Control means 105).

The disk array device shown in FIG. 9 is composed of five hard disk devices, and has data storage areas from 0 to 15 as shown in FIG. This data storage area is called a "stripe" and is, for example, 32 KB.
Can be stored.

There are redundant data storage areas represented by P0-3, P4-7, P8-11, and P12-15. In P0-3, the exclusive OR (XOR) of the data stored in stripes 0 to 3 is calculated. (P0-3) = (data of stripe 0) XOR (data of stripe 1) XOR (data of stripe 2) XOR (data of stripe 3) is stored. Generally, in the case of Pm-n, the result of calculating the exclusive OR (XOR) of the data stored in the stripe m to the stripe n is stored. Such redundant data becomes effective when one of the hard disk devices constituting the disk array device is broken. This “one hard disk device is broken” state is referred to as a “degenerate state”. Read / write processing in this degenerate state will be described later.

[0007] Stripes n to m of the data storage area and the redundant data storage area Pm-n are put together.
Called "stripe group". When the disk array device is viewed from the host computer, stripes 0 to 1
It is recognized as a logical single storage device that continues up to the fifth. However, actually, a plurality of hard disk devices are used, and data is distributed and held.

First, a process in which the host computer 100 reads data from the disk array device 101 will be described. For example, to read data stored in stripes 4 and 5, the host computer 100 issues a request to read the data to the disk array device 101. The disk array device 101 reads the fourth stripe of the hard disk 111 and the fifth stripe of the hard disk 112 and transfers them to the host computer 100.

[Write Processing] Next, the host computer 100 operates the disk array device 1
The process of writing data to 01 will be described. The write processing of the disk array device 101 is different from that of a hard disk device used alone because there is redundant data. When data is written, redundant data also needs to be updated. There are two methods for updating this redundant data.
There are two ways.

[0010] Write processing (part 1) For example, to write data to stripe 6, first, data of stripe 4, stripe 5, and stripe 7 of the stripe group to which stripe 6 belongs are read, and (P4 -7) = (data of stripe 6) XOR (data of stripe 4) XOR (data of stripe 5) XOR (data of stripe 7) to calculate redundant data, and then write data to stripe 6 It is necessary to write redundant data to P4-7.

Write processing (No. 2) Similarly, to write data to stripe 6, first, write data of stripe 6 before writing (referred to as "data of old stripe 6") and P4 before writing. -7
(P4-7) = (data of stripe 6) XOR (data of old stripe 6) using data of stripe 6 to be read and written (called "old P4-7") (data of new stripe 6) Data) It is necessary to calculate XOR (former P4-7) to find redundant data, and then write data to stripe 6 and redundant data to P4-7.

In the disk array device, not only data but also redundant data needs to be updated and written, so the number of accesses to the hard disk device increases.

Read processing in a degenerate state Characteristic in a disk array apparatus using parity for redundant data,
The read processing in the degenerate state will be described. For example, in FIG. 10, it is assumed that the hard disk device 111 has failed and reading and writing cannot be performed. In this state, it is assumed that data stored in stripes 4 and 5 is read. First, the disk array device attempts to read the first half of data from stripe number 4, but cannot read due to a failure. In such a case, the disk array device 101 reads the redundant data P4-7, the data of the stripe No. 5, the data of the stripe No. 6, and the data of the stripe No. 7, and reads (data of the stripe 4) = (P4-7) XOR (the data of the stripe 5). Data) XOR (data of stripe 6) XOR (data of stripe 7) is calculated to restore the data of stripe 4 of the damaged hard disk device 111. The restored stripe No. 4 data and the read stripe No. 5 data are transferred to the host computer. Thus, data can be read even in a degraded state in which one hard disk drive is broken.

Of course, if the disk array device knows in advance that the hard disk device 111 is broken, the reading process for stripe 4 can be omitted.

Parity Adjustment In order to restore correct data in the degenerated state, correct redundant data (parity) must be generated. There are two methods for matching the consistency between the data and the redundant data (parity matching).

The first method is a method of calculating (Pmn) = (data of stripe m) to XOR (data of stripe n) for all stripe groups.

The second method is a system in which zero is written to all hard disk devices constituting a disk array. The second method simply writes fixed data to the hard disk device as compared with the first method requiring operation, and can simplify the processing. As long as the parity operation expression is satisfied, the data to be written need not be zero.

[0018]

[Problems to be solved by the invention]

● Removable storage media Conventional disk array devices are premised on using storage devices that use non-removable storage media such as hard disk devices. Therefore, removable storage media and storage devices are considered. Not.

The number of storage media = the number of storage devices For example, as shown in FIG.
215) and five storage devices (221 to 225),
If the five storage media are mounted on the five storage devices (the storage media are mounted on the storage devices in a state where they can be read and written), the conventional hard disk device shown in FIG. 9 can be used. Disk array device 10
1 can be configured in the same manner. However, when such a configuration as shown in FIG. 11 is used, the same number of storage devices as the number of storage media to be configured is required.

The number of storage media> the number of storage devices Therefore, in the present invention, a plurality of removable storage media is one.
By sharing one storage device, the advantage of a removable storage medium was taken advantage of to provide a storage system with reduced cost. In general, the cost of a removable storage medium is sufficiently smaller than the cost of its storage device (1).
/ 10 or less).

For example, as shown in FIG. 12, two storage devices 351 and 352 are prepared, the storage device 351 is shared by four removable storage media (311 to 314), and one storage device 352 is provided. Storage system occupied by the removable storage medium (315). Five storage media (311 to 311)
15) respectively correspond to the five hard disk devices (111 to 115) in FIG. By using this configuration, the number of storage devices is three compared to the storage system of FIG.
The number of units can be reduced, and the cost can be reduced.

Problem: medium exchange time However, in the storage system shown in FIG. 12, if the arrangement of the data and the redundant data shown in FIG. 10 is used, the exchange of the storage medium is required every time the read or write processing is performed. Therefore, a practical system cannot be constructed. Generally, the exchange of the storage medium takes several seconds to several tens of seconds, so that the processing time is much longer than the reading and writing processing time of tens of milliseconds.

For example, the stripe 6
In order to write data in the storage number, first, the storage device 351
The storage medium 313 storing the stripe No. 6 is mounted on the storage device 352, and the storage medium 315 storing the P4-7 is mounted on the storage device 352, and the stripe 6 before the writing is performed.
Number data (referred to as “old stripe 6 data”)
Read P4-7 (called "old P4-7") before writing, and calculate (P4-7) = (data of stripe 6) XOR (data of old stripe 6) XOR (old P4-7) It is necessary to find redundant data, and then write data to stripe 6 and write redundant data to P4-7.
Therefore, there is a problem that every time a read or write processing request is issued to a different storage medium, the storage medium needs to be replaced, and the processing time becomes longer.

Further, the stripe 6
In order to write data into the storage media, the storage media 311, 31
While exchanging 2, 314, the data of stripe 4, stripe 5, and stripe 7 are read to obtain redundant data, and further replaced with storage medium 313 to store data in stripe 6 and redundant data in P4-7. There is a need to. Therefore, the processing time is further increased.

Problem: Waste of capacity In the storage system shown in FIG. 11, it can be seen that the number of mounted storage media is determined by the number of storage devices. In the storage system of FIG. 11, one set includes four data storage media (211 to 214) and one redundant data storage medium (215). For example, the capacity per storage medium is 5GB (1GB = 1073741824Byte)
Then, 20 GB (= 5 GB × 4) of data can be held per set, and reliability is improved by redundant data of 5 GB (for one sheet). In this storage system, two sets (10) of storage media are required to store 25 GB of data, but the second set uses only 5 GB despite the capacity of 20 GB, There is a problem that capacity is wasted.

The object of the present invention is to take advantage of the characteristics of a removable storage medium,
An object of the present invention is to construct a large-capacity and highly reliable storage system with a smaller number of storage devices than the number of storage media. In the case of the present invention, the present invention can be realized if there are at least two storage devices: a plurality of storage media, a storage device that handles a removable storage medium for holding data, and a storage device that holds redundant data. .

Another object of the present invention is to hold data with high reliability by using a minimum number of storage media. In the case of the present invention, the number of storage media for storing data varies according to the amount of data to be stored, but the number of storage media for storing redundant data may be basically one.

[0028]

In order to achieve the above object, a data storage system comprising at least two or more removable storage media and at least one storage device for reading from and writing to the storage media. Means, at least one or more removable storage media, and redundant data storage means comprising at least one storage device for reading and writing to the storage medium, wherein the redundant data storage means comprises: The redundant data calculated from the data stored in the data storing means is stored.

As another means for achieving the first object, a storage device having a non-removable storage medium may be used as the redundant data holding means.

In order to share one or more storage devices with two or more removable storage media, a storage medium holding means for storing a removable storage medium, and a medium from the storage medium holding means to the storage device. A conveying means for conveying is provided.

[0031]

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

<< First Embodiment >> FIG. 1 is a diagram showing a schematic configuration of a storage system according to an embodiment of the present invention.

System Configuration Host Computer 100 and Disk Array Device 500
RAID control means 600 is a SCSI (Small Comp
(uter system interface) via a bus 109. In this embodiment, the SCSI bus is used. However, other connection means, for example, PCI (Peripheral Component Injection) is used.
terconnect), Fiber Channel, IEEE 1394
It can also be realized by using such a method. Also, the SCSI bus has
Several versions (SCSI, SCSI-2, SCSI-3, Wide-SCS
I, Ultra-SCSI, etc.), but the present invention can be realized using any version.

The RAID control means 600 includes two DVs.
D storage devices 551 and 552 are connected. The DVD storage device is a device that reads from and writes to a removable storage medium (DVD-RAM storage medium: hereinafter, referred to as “DVD storage medium”). It goes without saying that the present invention can be realized not only by the DVD storage device but also by using a magneto-optical disk device, an optical disk device, a magnetic disk device and the like, and a removable storage medium corresponding to the storage device.

In this embodiment, twelve DVD storage media 5
11 to 518 and 521 to 524 are provided. The DVD storage media are held in the corresponding storage medium holding means 558. The DVD storage device 551 and eight DVD storage media (511 to 518) constitute a data holding unit 501. Further, the DVD storage device 552 and the four DVD storage media (521 and 524) use the redundant data holding unit 502.
Is configured. Thus, the DVD storage device is shared by a plurality of DVD storage media.

The transport means 559 is a storage medium holding means 55
8 for transporting a DVD storage medium between the DVD storage device 8 and the DVD storage devices 551 and 552.

FIG. 2 shows the internal configuration of the RAID control means 600.
Shown in The RAID control unit 600 includes: a host connection unit 610 for connecting to the host computer 100;
Storage device connection means 601 for connecting a VD storage device
And the host connection unit 610 and the storage device connection unit 60
Storage means 650 for holding data to be transferred via
The data bus 620 and the control bus 6 for connecting the host connection means 610, the storage device connection means 601 and the storage means 650
30, a control unit 690 connected to the control bus 130 to control each unit, a data restoration unit 660 connected to the storage unit 650, a redundant data generation unit 670, and a transport control unit for controlling the transport unit 559. 640.

DVD Storage Medium In this embodiment, a DVD storage medium of 5 GB per sheet is used. In addition, in order to simplify parity adjustment, all DVD storage media have all data initialized to zero in advance and held in storage medium holding means 558.

Write Process First, a process in which the host computer 100 writes data to the disk array device 500 will be described. In this embodiment, the host computer 100 is connected to the storage device 190 connected to the host computer 100.
The process of saving data in the disk array device 500 will be described as an example. Such saving processing is generally called “backup”.

Start of mounting the storage medium The host computer 100 notifies the control unit 690 of the start of backup via the host connection unit 610. The control unit 690 controls the transport control unit 640
The DVD storage medium 511 is stored in the DVD storage
An instruction is given to mount each of the D storage media 521 on the DVD storage device 552. The transport control unit 640 controls the transport unit 559 to take out each DVD storage medium from the storage medium holding unit 558 and mount it on the DVD storage device.

Arrangement of Data and Redundant Data FIG. 3 shows an arrangement of data and redundant data when two DVD storage media are mounted. At this point, one storage medium (511) for holding data and one storage medium (521) for holding redundant data make up one set.

In this embodiment, the storage capacity per stripe is set to 32 KB and 163 per DVD storage medium.
It has 840 stripes (0 to 163839). D
The VD storage medium 511 holds data, and the DVD storage medium 521 holds redundant data. Pn holds the same data as the data of the stripe n.

When the mounting is completed, the control unit 690 notifies the host computer via the host connection unit 610 that the preparation has been completed. Upon receiving the preparation completion notification, the host computer 100 starts transferring data from the storage device 190 to the disk array device 500.

Data Writing The host computer 100 writes 32 KB of data according to the size of the stripe in the disk array device 5.
Transfer to 00. The disk array device 500 receives data via the host connection means 610 and
Hold at 50. The control unit 690 further writes the data held in the storage unit 650 to the stripe 0 and P0 via the storage device connection unit 601. Thereafter, the transmitted data is sequentially written to stripe No. 1 and P1, and stripe No. 2 and P2.

Then, the stripe 163839 and P1
When the data has been written up to 63839, all data has been written on the DVD storage medium 511, so that further writing is not possible. Therefore, the control means 690 controls the DVD
The transfer control unit 640 is instructed to exchange the storage medium 511. At this time, there is no need to exchange the DVD storage medium 521.

The exchange of the storage medium The control means 690 instructs the transport control means 640 to take out the DVD storage medium 511 from the DVD storage device 551 and return it to the predetermined storage medium holding means 558. Further, an instruction is given to mount the DVD storage medium 512 from the storage medium holding means 558 to the DVD storage device 551.

During the exchange of the DVD storage medium, it can be notified that the data transfer from the host computer 100 is temporarily interrupted. Alternatively, the storage unit 650
, And can be written after the replacement.

Change of arrangement of data and redundant data (second) The arrangement of data and redundant data at the end of mounting the DVD storage medium 512 is shown in FIG. DVD storage medium 512
At the end of the mounting of the storage medium 2 for holding the data
A total of three sheets (511 and 512) and one storage medium (521) holding redundant data constitute one set.

The newly mounted DVD storage medium 512
As a result, 163840 stripes are added, and stripes 163840 to 327679 are added. In Pn, the result of calculating (Pn) = (data of stripe m) XOR (data of stripe n) (where m = n + 163840) is held. By this, two DVDs
Even if one DVD storage medium among the storage media (511 and 512) becomes unreadable for some reason, R
Reading and writing can be continued by the AID mechanism.

Since the DVD storage medium is initialized to zero and stored in the storage medium holding means 558, there is no need to perform parity adjustment when newly mounted.

Using the above equation, for example, in order to write data to stripe 163840, data of stripe 0 is required. In this embodiment, the DVD
Since the storage medium 511 is already unmounted (the storage medium is taken out of the storage device and stored in the storage medium holding means) and stored in the storage medium holding means 558, in order to read the data of stripe 0, the DVD storage is required. The media must be changed. However, this requires a long replacement time and does not become a practical system. Therefore, in the present invention, Pn is calculated as follows.

First, the data of the m-th stripe before writing (referred to as “old stripe m data”) and the Pn before writing (referred to as “old Pn”) are read, and the data of the m-th stripe to be written (“m”). (Referred to as “data of new stripe m”), (Pn) = (data of new stripe m) XOR (data of old stripe m) XOR (old Pn) (where m = n + 163840). Thus, by obtaining Pn, the DVD
Reading from the storage medium 511 is no longer necessary, and
No exchange of D storage media is required. Also, since the DVD storage medium is initialized to zero and stored in the storage medium holding means 558, “data of the old stripe m” is zero, and (Pn) = (data of the new stripe m) XOR (zero) ) XOR (old Pn) (where m = n + 163840) Therefore, it can be simplified as (Pn) = (data of new stripe m) XOR (old Pn) (where m = n + 163840).

After the exchange is completed, when the exchange is completed, data is written in order from the stripe 163840, and the redundant data generating means 670 is written.
The result of calculating (P0) = (data of the new stripe 163840) XOR (old P0) is written in P0. After that, stripe 1638
Processing is performed in the order of No. 41 and P1.

Change of arrangement of data and redundant data (third sheet) Further, data writing continues, and stripe 3276
When data has been written up to number 79, the DVD storage medium 51
Replace 2. As in the previous case, there is no need to replace the DVD storage medium 521 at this time.

The control means 690 includes a transport control means 640
Is instructed to take out the DVD storage medium 512 from the DVD storage device 551 and return it to the storage medium holding means 558. Further, the DVD storage medium 513 is
Instruct 51 to mount.

FIG. 5 shows the arrangement of data and redundant data at the end of mounting the DVD storage medium 513. As shown in FIG. 5, a stripe 16 is newly added at the end of mounting.
3,840 stripes are added, and stripes 327,680 to 491,519 are added. And Pn
Holds the result of calculating (Pn) = (data of stripe k) XOR (data of stripe m) XOR (data of stripe n) (where m = n + 163840 k = n + 327680)

Since the DVD storage medium is initialized to zero and stored in the storage medium holding means 558, there is no need to perform parity adjustment when newly mounted.

When the replacement is completed, the stripe 32768
Data is written in order from No. 0, and the result of calculating (Pn) = (data of new stripe k) XOR (old Pn) (where k = n + 327680) by using redundant data generation means 670 Write. After that, stripe 32768
Processing is performed in the order of 1 and P1.

Thereafter, similarly, every time the DVD storage medium becomes full, the exchange of the DVD storage medium is repeated, and writing can be continued. In this embodiment, it is assumed that the backup of the storage device 190 of the host computer 100 has been completed in the middle of the DVD storage medium 513. Therefore, four DVD storage media (511, 512, 513, and 521)
Becomes one set.

[Effect] As described above, by performing the processing, even if a plurality of DVD storage media share two DVD storage devices, the exchange of the storage media is minimized and the data is protected by redundant data. A highly reliable storage system can be provided.

Further, since the storage media are exchanged according to the capacity of the data to be written, only the required number of storage media are required. For example, in the case of the present embodiment, 4
DVD storage media were used. When the backup processing of the present embodiment is performed by the system shown in FIG. 11, five DVD storage media are used. This is because the number of storage media is determined by the number of storage devices. In the present invention,
Even if the amount of data to be written fluctuates, only one storage medium holds redundant data, and data can be held efficiently and with high reliability.

● Batch reading Also, at the time of creating redundant data, Pn before writing (old Pn)
Although n) is required, the control unit 690 collectively reads out the old Pn to the storage unit 650 via the storage device connection unit 601. Thereby, the speed of the Pn generation process can be increased.

[Read Processing] The host computer 100 is connected to the disk array device 500
A read process for reading data from the device will be described. D
The VD storage medium is the four D media used in the description of the write process.
A VD storage medium (511, 512, 513, 521) is used.

When the host computer 100 has stripe 0
A description will be given by taking as an example the case of reading the number data. The read request from the host computer 100 is notified to the control unit 690 via the host connection unit 610. When the control unit 690 determines that the read operation is stripe 0, the control unit 690 instructs the transport control unit 640 to mount the DVD storage medium 511 on the DVD storage device 551. The transport control unit 640 controls the transport unit 559 to transfer the DVD storage medium 51 from the storage medium holding unit 558.
1 is taken out and mounted on the DVD storage device 551.

When the control means 690 receives a report from the transport control means 640 that the mounting has been completed, the storage means control means 601 sends the stripe number 0 from the DVD storage medium 511 mounted on the DVD storage device 551. The data is read and stored in the storage unit 650. The control unit 690 transfers the data of stripe 0 stored in the storage unit 650 to the host computer 100 via the host connection unit 610.

Multi-processing If the host computer 100 issues a read request for stripe number 163840, the DV
The D storage medium 512 can be replaced with the DVD storage medium 511 and mounted on the DVD storage device 551,
Further, the DVD storage medium 512 can be mounted on the DVD storage device 552. By mounting two storage media (511 and 512 in this case) holding data on two DVD storage devices at the same time, data can be transferred to the host computer at higher speed.

Next, read processing when a failure occurs during reading will be described. The DVD storage media are the four DVD storage media (511, 512, and 5) used in the description of the write process.
13 and 521) are used. At the time of write processing, DVD
Since redundant data is created in the storage medium 521, of the three DVD storage media (511, 512, and 513),
Even if one DVD storage medium becomes unreadable for some reason, reading can be performed by the RAID mechanism.

When the host computer 100 has stripe 0
A description will be given by taking as an example the case of reading the number data. The read request from the host computer 100 is notified to the control unit 690 via the host connection unit 610. When the control unit 690 determines that the read operation is stripe 0, the control unit 690 instructs the transport control unit 640 to mount the DVD storage medium 511 on the DVD storage device 551. The transport control unit 640 controls the transport unit 559 to transfer the DVD storage medium 51 from the storage medium holding unit 558.
1 is taken out and mounted on the DVD storage device 551.

When the control unit 690 receives a report from the transport control unit 640 that the mounting has been completed, the control unit 690 sends the data of the stripe 0 from the DVD storage medium 511 mounted on the DVD storage device 551 via the storage device control unit 601. When trying to read data, an error occurs and the data cannot be read. The control unit 690 attempts to read some stripes in order to check whether only the stripe 0 of the DVD storage medium 511 cannot be read or whether the DVD storage medium 511 itself cannot be read. At the time of partial failure where only stripe 0 can not be read,
At the time of a storage medium failure that the DVD storage medium itself cannot be read,
It is divided into two processing methods.

In the event of a partial failure When it is determined that only a stripe 0 is a partial failure to be read, the control means 690 determines that (data of the stripe 0) = (P0) XOR (data of the stripe 163840) XOR (data of the stripe 327680) To calculate the data of stripe 0,
The DVD storage device 552 has three DVD storage media (512
, 513, and 521) are sequentially mounted, and the respective data are read out to the storage unit 650. The control means 650 includes:
The data of stripe 0 is restored using the data restoration unit 660 and transferred to the host computer 100 via the host connection unit 610. Then, the control means 690
Attempts to write the restored data of stripe 0 in the DVD storage medium 511 by allocating an alternate area. In general, a storage device has a function of allocating another area (alternate area) on a storage medium for a partial failure of the storage medium and retaining data. After the actual writing, if the verify is performed to confirm that the data of the stripe 0 has been correctly written, the reliability is further improved.

In the event of a failure in the storage medium If it is determined that the failure in the storage medium is such that the DVD storage medium itself cannot be read, the control means 690 newly sets the DVD storage medium 5
11 is executed. First, the control unit 690 sends the DVD storage medium 511 to the transport control unit 640.
Is instructed to mount the unused storage medium 514 on the DVD storage device 551. As shown in FIG. 6, R0 to R1638 are stored on the DVD storage medium.
There are up to 39 stripes.

The DVD storage device 552 has a DV
The D storage medium 521 is mounted, and the DVD storage medium 521 is mounted.
Is copied to the DVD storage medium 514 (R0 to R163839).

When the copy is completed, the control means 690
The DVD storage medium 521 of the DVD storage device 552 is unmounted, and the DVD storage medium 512 is mounted. The control means 690 calculates Rn as follows.

First, Rn before writing (referred to as “old Rn”) is read, and data of the m-th stripe (m = n + 1) is read.
(Rn) = (old Rn) XOR (data of stripe m) (where m = n + 163840). The control means 690 calculates using the storage means 650 and the data restoration means 660,
Update up to 39.

Further, the control means 690 unmounts the DVD storage medium 512 of the DVD storage device 552,
The VD storage medium 513 is mounted. Control means 690
Calculates Rn as follows.

First, Rn (called “old Rn”) before writing is read, and data of the k-th stripe (k = n + 3) is read.
Using (27680), (Rn) = (old Rn) XOR (data of stripe k) (where k = n + 327680) is calculated. The control means 690 calculates using the storage means 650 and the data restoration means 660,
Update up to 39.

As described above, the same data (stripe 0 to stripe 163839) as the lost DVD storage medium 511 can be restored to the stripes R0 to R163839 of the DVD storage medium 514. Hereafter, DV
By using the DVD storage medium 514 instead of the D storage medium 511, it is possible to completely recover from a failure.
Similarly, if the DVD storage medium 511 itself is lost for some reason, it can be restored by the same procedure.

<< Second Embodiment >> FIG. 7 is a diagram showing a schematic configuration of a storage system according to an embodiment of the present invention.

System Configuration The system configuration of FIG. 7 is different from the system configuration of FIG.
Hard disk drive 5 instead of DVD storage 552
53 is used. The hard disk device 553 uses a non-detachable storage medium. Also, 5
Although a single DVD storage medium is used, different numbers 531 to 535 are given to distinguish them from the first embodiment.

A DVD storage medium and a hard disk device In this embodiment, a 5 GB DVD storage medium is used for each DVD. The hard disk drive has 10 GB (= 5 GB
× 2) is used. Instead of a 10 GB hard disk, two 5 GB hard disk devices may be used.

FIG. 8 shows the arrangement of data and redundant data. In the present embodiment, a storage medium (531 to 53) holding four pieces of data is used.
4) and the hard disk drive 553 form a set.

In this embodiment, the storage capacity per stripe is set to 32 KB, and 1638 per DVD storage medium is set.
It has 40 stripes (0 to 163839). Since the hard disk device 553 has a capacity of 10 GB,
It can have 27680 stripes. This 3
27680 stripes are A0 to A163839,
It is divided into two of B0 to B163839 and used.

In implementing this system, the data in the DVD storage medium and the hard disk are all initialized to zero.

[Write Process] The case where the host computer 100 writes data in stripe 0 will be described as an example. The write request from the host computer 100 is sent to the host connection unit 610.
Is notified to the control means 690 via the. Control means 69
When 0 determines that the writing is stripe 0, it instructs the transport control means 640 to mount the DVD storage medium 531 on the DVD storage device 551. The transport control unit 640 controls the transport unit 559 to take out the DVD storage medium 531 from the storage medium holding unit 558 and mount it on the DVD storage device 551. Control means 6
Numeral 90 receives data to be written to stripe 0 from the host computer 100 and stores the data in the storage unit 650.

When the control unit 690 receives a report from the transport control unit 640 that mounting has been completed, the control unit 690 writes data from the storage unit 650 to the stripe 0 through the storage device control unit 601. Further, A0 is calculated as follows.

First, data of stripe 0 before writing (called “old stripe 0 data”) and A0 before writing (called “old A0”) are read, and data of stripe 0 to be written (“old A0”) is read. (Referred to as “new stripe 0 data”), (A0) = (data of new stripe 0) XOR (data of old stripe 0) XOR (old A0) is calculated. The result is written into A0.

When a write request is received for a stripe stored on a different DVD storage medium, the control means 6
90 performs processing after mounting the corresponding DVD storage medium using the transport control unit 640 and the transport unit 559. Generally, for writing in which stripes are continuous (writing continuously with stripes n, n + 1, n + 2,...), The exchange frequency of the DVD storage medium is small and there is no practical problem.

[Read Process] The host computer 100 operates in the disk array device 700
Since the read processing for reading data from the first embodiment may be performed in the same manner as in the first embodiment, it is omitted here.

Read processing at the time of failure At the time of a partial failure, the same processing as in the first embodiment may be performed, so only the read processing at the time of a failure of the recording medium will be described. In the present embodiment, the D of the DVD storage medium 531
A description will be given as a storage medium failure in which a failure has occurred in the VD storage medium and reading is impossible over the entire DVD storage medium.

In the event of a failure in the storage medium, a case where the host computer 100 reads the data of stripe 0 will be described as an example. The read request from the host computer 100 is sent to the host connection unit 610.
Is notified to the control means 690 via the. Control means 690
Determines that the data is read from stripe 0, mounts the DVD storage medium 531 on the DVD storage device 551, and executes reading of data of stripe 0. However, an error occurs and data cannot be read. The control unit 690 attempts to read some stripes of the DVD storage medium 531 and determines whether the failure is a partial failure or a storage medium failure. In this embodiment, the control unit 690 determines that the DVD storage medium 531 itself cannot be read, and executes a process of newly creating a copy of the DVD storage medium 531.

The control means 690 unmounts the DVD storage medium 531 to the transport control means 640,
Instruct to mount 2. When the mounting is completed, the control means 690 calculates Bn as follows.

(Bn) = (An) XOR (data of stripe m) (where m = n + 163840) is calculated. The control means 690 calculates using the storage means 650 and the data restoration means 660,
Create up to 39.

Further, the control means 690 unmounts the DVD storage medium 532 of the DVD storage device 551,
The VD storage medium 533 is mounted. Control means 690
Calculates Bn as follows.

First, Bn before writing (called “old Bn”) is read, and data of the k-th stripe (k = n + 3) is read.
Using (27680), (Bn) = (old Bn) XOR (data of stripe k) (where k = n + 327680) is calculated. The control means 690 calculates using the storage means 650 and the data restoration means 660,
Update up to 39.

Further, the control means 690 unmounts the DVD storage medium 533 of the DVD storage device 551,
The VD storage medium 534 is mounted. Control means 690
Calculates Bn as follows.

First, Bn before writing (referred to as “old Bn”) is read, and data of the j-th stripe (j = n + 4) is read.
Using (91520), (Bn) = (old Bn) XOR (data of stripe j) (j = n + 491520) is calculated. The control means 690 calculates using the storage means 650 and the data restoration means 660,
Update up to 39. From the above, B0 to B1638
Thus, all the data in the DVD storage medium 531 that was lost in 39 could be restored.

Then, the control means 690 unmounts the DVD storage medium 534 of the DVD storage device 551 and mounts an unused DVD storage medium 535. When the mounting is completed, the control means 690 transfers the data restored from B0 to B163839 of the hard disk device to a DVD.
All are copied to the storage medium 535. Thereafter, by using the DVD storage medium 535 instead of the DVD storage medium 531, it is possible to completely recover from the failure.

If the unused DVD storage medium 535 is
When not prepared in the disk array device 700,
The data from B0 to B163839 is stored in stripe 0
It can also be handled as data from the data number to the stripe number 163839.

After that, when a new unused DVD storage medium is inserted, it can be copied from the hard disk device 553.

[0100]

As described in the present specification, the system is configured to hold data and redundant data, thereby taking advantage of the characteristics of the removable storage medium, and achieving a large-capacity and highly-reliable storage with a small number of storage devices. A system can be built. In the case of the present invention, the present invention can be realized if there are at least two storage devices: a plurality of storage media, a storage device that handles a removable storage medium for holding data, and a storage device that holds redundant data. .

Although the number of storage media for storing data changes according to the amount of data to be stored, the number of storage media for storing redundant data may be basically one, and the number of storage media for a small number of storage media is small. Data can be held with high reliability.

[Brief description of the drawings]

FIG. 1 is a diagram illustrating a schematic configuration of a storage system according to a first embodiment (a configuration of two DVD storage devices).

FIG. 2 is a diagram showing an internal schematic configuration of a RAID control unit.

FIG. 3 is a diagram showing an arrangement of data and redundant data in an initial state.

FIG. 4 is a diagram illustrating an arrangement of data and redundant data when one DVD storage medium is added from the state of FIG. 3;

FIG. 5 is a diagram showing an arrangement of data and redundant data when two DVD storage media are added from the state of FIG. 3;

FIG. 6 is a diagram showing an arrangement of data and redundant data when an unused storage medium is mounted.

FIG. 7 is a diagram illustrating a schematic configuration of a storage system according to a second embodiment (one DVD storage device and one hard disk device).

FIG. 8 is a diagram illustrating an arrangement of data and redundant data in the storage system according to the second embodiment;

FIG. 9 is a diagram showing a conventional disk array device using a hard disk device.

FIG. 10 is a diagram showing the arrangement of conventional data and redundant data.

FIG. 11 is a diagram showing a disk array device configured using a removable storage medium in the related art.

FIG. 12 is a diagram showing a disk array device configured using the present invention.

[Explanation of symbols]

100 host computer, 101 disk array device, 109 SCSI bus, 105 RAID control means, 190 storage device, 201 disk array device
205 RAID control means, 301 disk array device, 305 RAID control means, 351 storage device, 3
52: storage device, 500: disk array device, 501
... data holding means, 502 ... redundant data holding means, 55
1: DVD storage device, 552: DVD storage device, 553
... Hard disk drive, 558 ... Storage medium holding means, 5
59 transport means, 610 host connection means, 601 storage device connection means, 620 data bus, 630 control bus, 640 transport control means, 650 storage means, 660
... data restoration means, 670 ... redundant data generation means, 69
0: control means, 700: disk array device, 701:
Data holding means, 702... Redundant data holding means

──────────────────────────────────────────────────の Continued on the front page (51) Int.Cl. ⁶ Identification code FI G11B 20/18 570 G11B 20/18 570Z 572 572F

Claims (7)

[Claims] 1. A storage device comprising at least two removable storage media, at least one storage device for reading from and writing to the storage medium, and at least one removable storage device. A redundant data storage means comprising at least one storage device for reading and writing data from and to the storage medium, wherein the redundant data storage means has a redundancy calculated from data held in the data holding means; A storage system for storing data. 2. A storage medium comprising at least two or more removable storage media, at least one storage device for reading and writing to and from the storage medium, and a non-detachable storage medium. A redundant data storage means comprising at least one storage device for reading and writing is provided, and the redundant data storage means holds redundant data calculated from data held in the data holding means. Storage system. 3. The storage system according to claim 1, wherein said storage medium holding means stores said plurality of removable storage media, and said storage medium is transported from said storage medium holding means to said storage device. A storage system comprising a transport unit. 4. The storage system according to claim 1, wherein the number of redundant data storage media is a, the number of data storage media is b, and the number of storage devices that handle removable storage media is c. Then, a storage system characterized by c <a + b. 5. A storage device comprising at least three or more removable storage media and at least two or more storage devices, wherein said storage device is a storage device for reading and writing to said removable storage medium. Wherein at least one of the storage media is a redundant data storage medium for holding redundant data, and at least two or more storage media are data storage media for holding data; Storing redundant data calculated from the data stored in the data storage medium. 6. A storage medium having at least two or more removable storage media and at least two or more storage devices, wherein at least one of the storage devices is connected to the removable storage medium. At least one storage device is a storage device that reads and writes data on a non-removable storage medium and holds data on the removable storage medium. A storage system having a non-removable storage medium holds redundant data calculated from data stored in the removable storage medium. 7. A plurality of (two or more) removable storage media, and a device for reading and writing to and from the storage media.
Data storage means comprising one storage device; one removable storage medium; and redundant data storage means comprising one storage device for reading and writing to the storage medium. The storage system stores redundant data calculated from the data stored in the data storage unit.