JPH1125574A - Recording / reproduction control device capable of parallel writing or parallel reading - Google Patents

Abstract

(57) [PROBLEMS] To provide a library device that realizes continuous writing and reading without requiring a waiting time due to disk replacement in a RAID system of a portable medium such as an optical disk. SOLUTION: A plurality of recording / reproducing apparatuses 2a to 2e for reading / writing from / to readable / writable portable media 1a to 1e,
When writing data in a library apparatus having a mechanism 3 for transporting portable media and a shelf 4 capable of storing a plurality of portable media 5a to 5f, a physical recording start position (sector or the like) of each medium. ) Is shifted intentionally to change the medium replacement time, and the medium to be replaced at one time is always limited to one. The spare recording / reproducing device 6 and the spare portable medium 1f are used together so that data can be written and read by combining a plurality of portable media without causing a waiting time due to medium exchange.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to a control system for realizing writing and reading in a RAID system using a portable medium (such as an optical disk) without causing a waiting time due to medium exchange.

[0002]

2. Description of the Related Art Conventionally, parallel processing, that is, parallel writing /
Distributed management of data by arranging magnetic disks in an array as a system that achieves high reliability by realizing high-speed data access by reading and adding redundancy by adding parity to data at the same time So-called RAID (Redundant arrays of inexpensive di
sk) A system is known, but recently, a RAID is used to disperse and manage data by filling a recording / reproducing device with a portable medium such as an optical disk instead of a fixed medium such as a magnetic disk and arranging the data in an array. The system has been invented.

[0003] An outline of a RAID system using a portable medium will be described with reference to FIG. 9. The conventional example of FIG.
A plurality of drives 2a to 2e for recording and reproducing data to and from a group of optical disks having a plurality of readable and writable optical disks 1a to 1e, and a controller 7 for these drives
a to 7e, an accessor device 3 for transporting these optical discs and its controller 9, a shelf 4 for storing a plurality of new optical discs or optical discs permitted to be overwritten, and enabling distributed writing and reading by RAID. And an array controller 23.

When receiving a write request from the host computer 11, the array controller 23 temporarily stores the content of the write request,
Each of the drives 2a to 2 which constitutes a RAID is a portable medium writable from the shelf 4 using the accessor 3 via the
e. When the drives 2a to 2e are all filled with the portable media 1a to 1e, the array controller adds parity to the temporarily stored write data from the host computer, and the drive controllers 7a to 7e. 7e each drive 2a-2
The distributed writing is performed on each of the portable media 1a to 1e in e. When the drives 2a to 2e are already filled with the writable portable medium, the filling operation by the accessor 3 does not occur.

Similarly, upon receiving a read request from the host computer 11, the array controller 23 uses the accessor 3 via the accessor controller 9 to transfer the portable medium containing the read data from the shelf 4 to the R.
The drives 2a to 2e constituting the AID are all filled. Each of the portable media 1a is attached to each of the drives 2a to 2e.
When all of .about.1e are filled, the array controller performs decentralized reading from the respective portable media 1a.about.1e in the respective drives 2a.about.2e via the respective drive controllers 7a.about.7e. If the drives 2a to 2e are already filled with a portable medium having read data, the accessor 3 does not perform a filling operation.

[0006] RAI by portable medium using FIG. 9 described above
As can be seen from the brief description of the D system, the biggest difference between a RAID system using a portable medium such as an optical disk and a RAID system using a magnetic disk is that a new write is performed or the remaining capacity of the medium during writing is reduced. For example, when performing a new read when the data is completely consumed, or when the data to be read next during the read is not present in the medium in the current recording / reproducing apparatus, the RAI at that time is used.
The same group of physical media constituting D (hereinafter referred to as a RAID physical group. On the other hand, in a single or a plurality of RAID physical groups, a group constituting a RAID with logically continuous data is referred to as a RAID physical group. RA
It will be called an ID logical group. ) Must be replaced all at once. That is, when a single device 3 transports the portable medium, the RA
The filling operation for the number of drives 2a to 2e constituting the ID occurs, and this takes a very long time.

In order to solve a long medium exchange time in a portable medium RAID system, Japanese Patent Laid-Open Publication No. Hei 7-36634 sets a certain threshold value in writing to a disk array device, and sets a data size relative to the threshold value. The distributed writing using RAID and the centralized writing are used together.

Here, the data arrangement in the portable medium in this system will be described with reference to FIG. JP-A-7-366
According to Japanese Patent No. 34, data smaller than a certain threshold value is written on a single medium, and when the remaining capacity of the medium is consumed, the free space of the medium filled in another recording / reproducing apparatus. The speed is increased by limiting the medium exchange to only one medium. This has the same data arrangement as the centralized data 1 in FIG. 10 in the portable medium constituting the RAID.

For data larger than a certain threshold value, the time required for writing itself is increased by writing data in parallel on a plurality of media to compensate for the time required for medium exchange. This is because the distributed data 1 or 2 shown in FIG.
The same data arrangement is used.

On the other hand, a certain threshold value is set for reading, and for data accessed more frequently than this threshold value, the data is stored in an empty area on a different medium currently filled in the recording apparatus. To read from the medium where the copy data exists even if the medium where the original data exists is not in the recording / reproducing apparatus,
Higher speed is achieved by reducing the opportunity for media exchange. This has the same data arrangement as the centralized data 1 in FIG. 10 in the portable medium constituting the RAID.

Further, the time required for reading itself can be increased by dispersing and re-recording data written on a single medium on a plurality of media currently filled in the recording / reproducing apparatus, A method for compensating for the time required for medium exchange has been adopted. This has the same data arrangement as the distributed data 1 or 2 in FIG. 10 in the portable medium configuring RAID. However, it is very difficult to virtually eliminate the medium exchange time.

Further, in addition to the above, Japanese Patent Application Laid-Open No. 4-3106767
In Japanese Unexamined Patent Application Publication No. Hei 8-249794,
Prior to media exchange, a method of holding a readable / writable medium in the vicinity of the recording / reproducing device in advance, or a method of once transporting a plurality of media by a medium transport device having a medium arrangement shape suitable for a plurality of recording / reproducing devices. It has been realized that the transfer time and distance of the transfer device can be shortened and the medium exchange speed can be increased by the transfer method. However, it is very difficult to substantially eliminate the medium exchange time itself.

[0013]

The above prior arts in which a portable medium is used in a RAID system are described as follows.
Although the medium exchange time at the time of reading is shortened, the waiting time for medium exchange cannot be completely eliminated, so that continuous writing and reading cannot be performed continuously.

[0014] As a solution for shortening the medium exchange time and increasing the speed, the opportunity for medium exchange is reduced. In order to reduce the chances of medium exchange, the number of media to be recorded depends on the length of data to be recorded and the frequency of access thereto. In the method of determining and recording whether to perform distributed recording or centralized recording on a single medium, in addition to the problem that complete continuous writing is not possible, as shown in the data arrangement in the portable medium in FIG. There is a problem that the control method and the medium management become complicated such that original data and copy data (centralized data 1 in FIG. 10) and data written in a distributed manner (distributed data 1 or 2 in FIG. 10) are mixed. However, there is a problem in that the reliability is reduced as compared with the distributed writing RAID in which the centralized recording is performed on a single medium and the redundancy is provided.

Furthermore, in order to reduce the medium exchange time and increase the speed, a method of holding a readable / writable medium on a shelf near or near a recording / reproducing apparatus, a method of arranging a plurality of recording / reproducing apparatuses, and the like. In the method of exchanging a plurality of media at a time by using a medium transport device having a medium arrangement shape, in addition to the problem that complete continuous writing cannot be performed, the configuration of the media transport device or the control method thereof is complicated, Accordingly, there is a problem that the entire library apparatus becomes expensive.

Therefore, in the present invention, by shifting the individual physical recording start positions (sectors and the like) of a plurality of readable and writable portable media constituting the RAID, a difference is intentionally generated in the medium replacement time. The medium to be exchanged in one exchange operation is limited to one, and the recording / reproducing apparatus in which the medium exchange does not occur and the spare recording / reproducing apparatus are used in combination to continuously write and read during the medium exchange time. It is an object of the present invention to provide a library device which enables a user to be unaware of a waiting time due to medium exchange.

[0017]

In order to solve the above problems, the present invention mainly employs the following configuration.

A plurality of recording / reproducing mechanisms for writing / reading to / from a portable medium, a transport mechanism for transporting the portable medium to the recording / reproducing mechanism, and a shelf capable of storing a plurality of the portable media; A recording / reproduction control device capable of performing parallel writing or parallel reading on the portable medium using a plurality of recording / reproduction mechanisms for input data, wherein the recording / reproduction mechanism is configured to perform parallel recording on a plurality of portable media by a plurality of recording / reproduction mechanisms. When writing or reading data, writing or reading is performed on each portable medium from a position where the recording capacity of each portable medium is shifted by a size equally divided by an integer larger than the number of recording and reproducing mechanisms. A recording / reproducing control device for starting and replacing each portable medium sequentially at different times.

Also, a plurality of recording / reproducing mechanisms for writing / reading to / from a portable medium, a transport mechanism for transporting the portable medium to the recording / reproducing mechanism, and a shelf capable of storing a plurality of the portable media are provided. A recording / reproducing control device capable of performing parallel writing or parallel reading on said portable medium using a plurality of recording / reproducing mechanisms for input data, wherein said plurality of portable media are provided by a plurality of recording / reproducing mechanisms. When writing or reading data in parallel, the writing or reading is performed on each portable medium from the position where the recording capacity of each portable medium is shifted by a size equally divided by an integer larger than the number of recording and reproducing mechanisms. A spare portable medium and a spare recording / reproducing mechanism for starting reading and writing or reading data during replacement of any one portable medium are provided, and the medium exchange is performed. Without causing the waiting time of the data writing or reading by recording and reproducing control apparatus to perform a continuous writing or reading data.

[0020]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A first embodiment according to the present invention will be described below. 1 and 2, reference numerals 1a to 1f denote optical disks filled in an optical disk drive, 2a to 2e denote regular optical disk drives, 3 denotes an optical disk transport accessor, 4 denotes a shelf for accommodating a plurality of optical disks, 5a to 5a
f is an optical disk placed on a shelf for storing a plurality of optical disks, 6 is a spare optical disk drive, 7a to 7e are regular optical disk drive controllers, 8 is a spare optical disk drive controller, 9 is an optical disk transport accessor controller, 10 is a sequential access controller, 11 is a host computer, 12 is a CPU of a sequential access controller, 13 is a memory of a sequential access controller, 14 is an I / O controller, 15 is a sector controller, 16 is a drive switching controller, 17 is a hard disk of a sequential access controller, and 17a is a sequential Management file in the hard disk of the access controller.

FIG. 1 is a schematic diagram of an optical disk library apparatus utilizing the present invention. This library apparatus includes five drives 2a to 2e for recording and reproducing data to and from a group of optical disks, each of which includes five readable and writable optical disks 1a to 1e, and controllers 7a to 7e for these drives.
e, an accessor device 3 for transporting these optical disks and its controller 9, a new optical disk or a shelf 4 for storing a plurality of optical disks 5a to 5e permitted to be overwritten, and enabling continuous writing and high-speed reading. Access controller 10, a spare drive 6 whose use is defined by the controller 10, and its controller 8.

Next, a sequential access controller 10 of a library device utilizing the present invention will be described. The sequential access controller 10 of FIG.
Is a characteristic controller for enabling continuous writing and high-speed reading in an optical disk library device using the present invention. Hereinafter, this characteristic controller is referred to as a sequential access controller.

Sequential access controller 10
Is a CPU 12 for controlling the entire library apparatus in accordance with a command from the host computer 11, and a memory 13 for temporarily storing commands from the host computer 11 and temporarily storing data coming from the library apparatus. And an I / O controller 14 for controlling I / O to each drive of the library device, a sector controller 15 for directly controlling a head position of each drive of the library device, and a drive for writing or reading data. And a management file 17a composed of a medium management table created by the sector controller 15 and the drive switching controller 16 via the CPU 12, and a hard disk 17 for storing and storing the same.

Next, a writing method of the optical disk library apparatus using the present invention will be described with reference to the flowcharts of FIGS. 2, 3 and 4. Host computer 11
When the CPU 12 issues a write request, the CPU 12 first fills each of the drives 2a to 2e and 6 with the optical disks 1a to 1f, and temporarily stores the write data contents in the memory 1
3 is stored.

Next, the CPU 12 reads each of the drives 2a to 2e from the specific management file 17a owned by the hard disk 17 of the sequential access controller 10.
Then, as shown in the flowchart of FIG. 3, the write start sector position of No. 6 is read and set in each of the drives 2a to 2e through the sector controller 15. Hereinafter, the five drives 2a to 2e that do not include the spare drive 6 are collectively referred to as regular drives. When the current data writing is the first time in this optical disk library apparatus, similarly, from the specific management file 17a owned by the hard disk 17 of the sequential access controller 10, the writing start sector position of each drive 2a to 2e, 6 Is read.

The following description will be continued on the assumption that the write start sector position for each of the drives 2a to 2e, 6 is at T = t1 in FIG. Regular drive 2 next to Fig. 5
The deviation width of the recording start sectors a to 2e is determined by each drive 2a.
２2e, the single capacity of the disks 1a〜1f filled in 6 is equal to the number n of regular drives 2a〜2e (in the example of FIG. 5, the number of regular drives is 5, so n = 5)
The size must be set equal to the size obtained by equally dividing the integer. This is the regular drive 2 described later
This is a method for keeping the replacement time of the disk in the spare drive 6 from ae to 2e.

When the write start sector position for each of the drives 2a to 2e, 6 is set, the CPU 12 calls the I / O controller 14 and writes the write data from the host computer stored in the memory 13 to the five regular drives. Redundancy, that is, parity is added to 2a to 2e and written.

Eventually, the disk at which the writing is started at the sector position closest to the end, the optical disk 1e in FIG.
When the replacement time arrives, the CPU 12 follows the processing flow of the flowchart of FIG.
Drive 2 requiring disk replacement through 6
The spare drive 6 is selected as the write destination instead of e, and then the write is continued through the I / O controller 14. The switching of the writing destination drive at this time is instantaneous only of the switching of the electric signal, and is not conscious of the user.

While writing to the disk 1f in the spare drive 6 is continued, the drive switching controller 16 determines whether the data currently being written is
A table indicating when the drive switching has occurred is created, and the sequential access controller 1
0 to the hard disk 17. The contents of this table are written in the management file 17a owned by the hard disk 17 of the sequential access controller 10, and serve as a pointer for switching drives when reading data. Writing to the disk 1f in the spare drive 6 is performed while the disk replacement time of the drive 2e elapses.

When the disk exchange of the drive 2e is completed, the CPU 12 again selects the drive 2e as a write destination through the drive switching controller 16, and
The writing is continued through the O controller 14. Switching of the drive at the writing destination at this time is also instantaneous switching on the electric signal, and is not conscious of the user. When writing to the newly filled disc in the drive 5e is resumed, the drive switching controller 1
Reference numeral 6 updates a table indicating the point in time at which drive switching has occurred for the data currently being written, and transmits the updated data to the hard disk 17 of the sequential access controller 10.

As described above, the numerical value obtained by dividing the total capacity of one optical disk by the number of regular drives is basically the black band shown in FIG. 5, that is, the size. However,
In FIG. 5, the size obtained by equally dividing by 5, which is the number of regular drives, is not the size, but as described above, it is sufficient to divide equally by an integer of 5 or more. Further, as another parameter for determining the size, the drive replacement period must be considered. For example, T = t3 in FIG.
The disk replacement period in the drive 2e in the above must be ended within the period of writing the size to the spare drive 6. Otherwise, continuous data cannot be written continuously.

According to a series of flows shown in FIGS.
As shown in FIG. 5, the replacement time of the disks 1a to 1e is as follows.
Although a fifth drive 2e, a fourth drive 2d, a third drive 2c, a second drive 2b, and a first drive 2a occur sequentially, a spare drive 6
Is used in combination with the other regular drive groups 2a to 2e that do not require disk exchange, thereby enabling continuous and continuous writing.

In the library apparatus utilizing the present invention, the above-mentioned contrivance, that is, the single capacity of the disks 1a to 1f filled in the respective drives 2a to 2e and 6 is determined by the number n of the regular drives 2a to 2e (see FIG. In the example of No. 5, the number of regular drives is 5, so n = 5) By setting the size equal to a size equally divided by an integer larger than an integer, the time when the disk 1f filled in the spare drive 6 is to be replaced (see FIG. 5). T = t8) corresponds to the other regular drives 2a to 2e.
The replacement is performed so as not to coincide with the replacement time of the discs 1a to 1e filled in the. That is, the user is not aware of the replacement of the disk in the spare drive 6, and the writing at this time is performed by the regular drives 2a to 2e.
It is performed continuously by using.

As shown in FIG. 5, in the library apparatus utilizing the present invention, the physical data recording start positions of the disks 1a to 1e filled in the regular drives 2a to 2e constituting the RAID are shifted from each other. In addition, it is possible to deliberately shift the replacement time of each disk, and it is possible to continuously write continuously without requiring a waiting time for disk replacement.

Since the replacement timing of the disks 1a to 1e is intentionally shifted as described above, it is not necessary to replace the disks 1a to 1e belonging to the same RAID physical group at once, and the replacement is performed in one replacement operation. Since a single disk is used, this series of movements can be realized by a single accessor 3.

Therefore, even in a conventional single accessor 3 or an optical disk library device having only a smaller number of accessors (n = 5 in FIG. 5) than the regular drives constituting a RAID, the disk replacement time is not easily considered. Continuous writing can be realized.

Further, by utilizing the present invention, it is possible to adopt a single accessor 3 or an accessor system smaller than the number n of regular drives constituting a RAID, thereby simplifying the structure of the accessor 3 and its control system. It is possible to provide an inexpensive library device in addition to the fact that the number of accessors 3 is physically small. That is, according to the recording method of the present invention, RAID
In addition to the high speed due to the parallel processing and the high reliability due to the redundant data, continuous writing in the true sense can be realized by a library device having a single accessor 3 of an inexpensive and simple mechanism.

Next, a reading method of the optical disk library apparatus utilizing the present invention will be described with reference to the flowcharts of FIGS. 2, 3 and 4. Host computer 11
When a read request is issued from the CPU 12, the CPU 12 first finds out the RAID logical group to which the data to be read belongs from the specific management file 17a owned by the hard disk 17 of the sequential access controller 10, the RAID physical group and the disk to which the data belongs, and the disks. Location (regular drive 2a ~ 2e
Or on the shelf 4), the recording start sector position at the time of writing the data to be read, and the like. If necessary, the disk groups 1a to 1f in which the data to be read exist are spared with the regular drives 2a to 2e. The drives 6 are filled respectively. Thereafter, as shown in the flowchart of FIG. 3, the read start sector position is set in each of the drives 2a to 2e through the sector controller 15.

When the reading start sector position for each of the drives 2a to 2e, 6 is set, the CPU 12 starts reading through the I / O controller 14. Eventually, when the replacement time comes to the disk at which the sector position from which the reading has been started is closest to the end, for example, 1e, the CPU 12 proceeds with the management file 17a of the hard disk 17 of the sequential access controller 10 according to FIG.
As shown in the processing flow of the flowchart of FIG. 1, the spare drive 6 is selected as a read source instead of the drive 2e which needs the disk exchange through the drive switching controller 16, and then the I / O controller 14
To continue reading. The switching of the read-out drive at this time is instantaneous only of the switching of the electric signal, and is not conscious of the user. Reading from the disk 1f in the spare drive 6 is performed while the disk replacement time of the drive 2e has elapsed.

When the disk exchange of the drive 2e is completed, the CPU 12 again selects the drive 2e as a read source through the drive switching controller 16, and
Reading is continued through the I / O controller 14. Switching of the read source drive at this time is also instantaneous switching on the electric signal and is not conscious of the user. Thereafter, the reading is similarly continued.

According to a series of flows shown in FIGS. 2 to 4,
As shown in FIG. 5, the replacement time of the disks 1a to 1e is as follows.
A fifth drive 2e, a fourth drive 2d, a third drive 2c, a second drive 2b, and a regular drive which are generated sequentially for the first drive 2a but do not require the spare drive 6 to be replaced with another disk. When used in combination with the groups 2a to 2e, reading can be performed continuously and continuously.

As shown in FIG. 5, in the library apparatus utilizing the present invention, the physical data recording start positions of the disks 1a to 1e filled in the regular drives 2a to 2e constituting the RAID are shifted from each other. Since it is possible to intentionally shift the replacement time of each disk, when reading data belonging to the same RAID logical group continuously with respect to the elapse of writing time,
As in the case of writing, the disks 1a to 1e belonging to the same RAID physical group do not need to be replaced at once, and only one disk is replaced in one replacement operation, so that there is no waiting time due to disk replacement.

Therefore, in the case where data belonging to the same RAID logical group is continuously read out with respect to the elapse of writing time, the conventional single accessor 3 or RA
Even in an optical disk library apparatus having fewer accessors than the number n of regular drives constituting the ID, it is possible to easily realize continuous reading without regard to the disk exchange time.

As a case where data can be read continuously with respect to the lapse of time of writing belonging to the same RAID logical group, there are restoration of long full backup data, continuous reproduction of advertising video, and the like. By using the present invention for these data, a single accessor 3 or an accessor method smaller than the number n of regular drives constituting a RAID is adopted, and the structure of the accessor 3 and its control method are simplified. In addition to the fact that the number of the accessors 3 is physically small, it is possible to provide an inexpensive library device.

That is, according to the recording method of the present invention, the RAI
In addition to the high speed due to the parallel processing possessed by D and the high reliability due to the redundant data, continuous writing in the true sense can be realized by a library device having a single accessor 3 with an inexpensive and simple mechanism.

Next, a second embodiment according to the present invention will be described below. First, FIG. 6 shows an assumed data arrangement state in each of the disks 1a to 1f when reading is performed after data is recorded using the present invention.

In the case of the pattern 1 in FIG. 6, the data to be read are data 1 and data 2 in the figure, and both belong to the same RAID physical group of the same RAID logical group. On the other hand, in the case of the pattern 2 in FIG. 6, the data to be read are the data 1 and the data 2 in the figure. These data belong to the same RAID logical group, but have different RAID physical groups. .

In the case of the pattern 1, the reading is realized only by the sector movement of the head of each of the drives 2a to 2e and 6, but in the case of the pattern 2, it is necessary to exchange the RAID physical groups for the reading. That is, pattern 1
In the case of (1), the user can read data without being aware of the disk exchange. In the case of the pattern 2, the user is aware of the waiting time for the disk exchange. FIG. 7 shows an outline of a library apparatus using the present invention so as to correspond to an example of use such as pattern 2.

The feature of this library device is that it has a tray-shaped accessor 19 that can transfer a plurality of disks belonging to the same RAID physical group at a time while having a much simpler mechanism than the ordinary accessor 3. And a shelf 18 for arranging the disks 5a to 5f on standby in a horizontal plane so as to easily carry out the transfer. The shelf 18 on which the waiting disks 5a to 5f are arranged in a horizontal plane will be described later in detail.

On the other hand, each row of the shelf 18 in FIG.
FIG. 8 illustrates a state in which the disks 5a to 5e to be filled in the regular drives 2a to 2e are arranged in order from the top, which are defined as storage columns dedicated to the disks to be filled in a to 2e and 6. In FIG. 8, when the arrangement status of the disk where the distributedly written data exists is shown according to time, it is from T = t1 to T = t6.

That is, at a certain time T = t1, data exists only in Column 1 of the shelf 18, and at T = t2, data exists in Column 1 for Rows 1 to 4, and exists in Column 2 only for Row 5. When the time further advances and at time T = t5, the data is C only for Row1.
It exists in column 1 and Rows 2 to 5 exist in column 2. At T = t6, all data exists only in column 2.

As can be seen, in the library apparatus utilizing the present invention, the adjacent regular drives 2
The disc (1a) filled in each of the regular drives (2a to 2e) has the deviation width of the recording start sector of a to 2e.
~ 1e) with a single drive of regular drive (2a ~ 2e)
e), the number of regular drives (in the example of FIG. 8, the number of regular drives is 5, so n = 5) is set to a size equally divided by an integer equal to or larger than an integer, and disks (2a to 2e) to be filled in each regular drive (2a to 2e) When the shelves 18 shown in FIG. 8 are arranged so that 5a to 5e) are sequentially arranged vertically downward, the data recorded in the same RAID physical group at a certain time spans each RAID configuration disk at most.
It can be seen that the column is within the range of two columns.

Noting this, the library apparatus according to the second embodiment of the present invention differs from the library apparatus according to the first embodiment in that, as shown in FIG. / A configuration in which the discharge port 20 is added. The tray-like accessor 19 is provided with individual disks 19a to 19 for arranging five disks 5a to 5e to be filled in a regular drive in a horizontal plane in a partitioned state.
e, the entire tray-shaped accessor 19 moves vertically in the form of an elevator, and the accessors 19a to 19e, on which individual disks are mounted, simply reciprocate the disk back and forth in the horizontal direction orthogonal to the direction of movement of the entire tray 19. With the simple function of.

Also, the disc loading / unloading port 2 shown in FIG.
0 denotes a drive (2a to 2a) included in the library device utilizing the present invention in combination with the tray accessor 19.
e, 6), the same number of disks (n = 6 in FIG. 7) can be simultaneously loaded / ejected.

As described above, in a library apparatus utilizing the present invention, the same RAID
Since the data recorded in the physical group spans each RAID configuration disk at most within the range of two columns in the column of FIG. 8, for example, when reading data written at the time T = t3 in FIG. First, the tray accessor 19 first moves to Column 1 of the shelf 18 in FIG. 8, and the individual accessors 19a to 19c constituting the tray accessor 19 simultaneously pull out three disks from Row1 to Row3. Subsequently, the tray-like accessor 19 moves to Column 2 of the shelf 18 in FIG. 8, where the individual accessors 19d to 19e constituting the tray-like accessor 19 simultaneously pull out two disks from Row4 to Row5.

The five disks constituting the RAID physical group at a certain time are stored in the tray accessor 19.
, The tray-shaped accessor 19 moves to the positions of the regular drives 2a to 2e, and installs these five disks at a time in the regular drives 2a to 2e. On the other hand, the disk 5f filled in the spare drive 6
Is transported by a single accessor 3. Here, the reason why the spare drive accessor 3 is provided independently is that the disc replacement position in the spare drive is often located in a column separated from the replacement column of the regular drive.

As described above, by providing a simple mechanism such as the shelf 18 and the tray-like accessor 19 of the second embodiment of the library apparatus utilizing the present invention, the data arrangement as shown in the pattern 2 in FIG. A library device capable of high-speed reading can be provided.

The two embodiments of the invention described above are not limited to the present invention. The portable medium described in the embodiment of the present invention is not limited to an optical disk, but may be a magnetic tape, a floppy disk, or another portable medium. The present invention is also applicable to a case where a fixed medium such as a hard disk can be handled as a portable medium. The function of the sequential access controller 10 does not need to be limited to a hardware-based realization method. For example, application software may be realized on the host computer 11 by owning the management file 17a in software. In the present invention, RAID levels and regular drives (2a to
The number of 2e) and the number of spare drives 6 are not limited.

Further, in the second embodiment of the present invention, only the reading operation has been described.
9 is used when writing is performed from a state where no disk is present in the regular drives 2a to 2e or when the user
It can also be used when loading / ejecting the disk groups 1a to 1e belonging to the AID logical group or RAID physical group, or simply when loading / ejecting disks collectively on the shelf 18. In addition, tray-shaped accessor 19
Is not limited to being able to simultaneously transport disks equal to the number of regular drives (2a to 2e), but is capable of simultaneously transporting disks equal to the number of all drives (2a to 2e, 6) including the spare drive 6. It may be.

As described above, the present invention includes the following configuration examples.

A library device using the present invention transports a plurality of recording / reproducing devices for reading / writing from / to a readable / writable portable medium (such as an optical disk), and transports a portable medium readable / writable to / from these recording / reproducing devices. Transport mechanism, shelves capable of storing a plurality of these portable media, and a plurality of recording / reproducing devices on a portable medium capable of reading and writing input data. Writing, that is, RAID, and a first unit that enables these portable media to be handled as a specific combination of logical and physical.

Further, the library apparatus using the present invention shifts the physical recording start position (sector, etc.) of each of a plurality of readable and writable portable media constituting the RAID, thereby making it possible to change Differences,
The second means capable of limiting the medium to be exchanged in one exchange operation to one and the recording / reproducing apparatus in which the medium exchange does not occur even when the medium exchange occurs. When the medium is exchanged with a spare recording / reproducing device for enabling writing or reading, the write destination or read source is switched to the spare recording / reproducing device, and after the medium exchange is completed, writing is performed on the original recording / reproducing device. A third means capable of switching the destination or the reading source and making the user unaware of the waiting time due to medium exchange.

[0063]

As described above, according to the present invention, when writing distributed data on a plurality of media by a plurality of recording / reproducing devices, the physical recording start position of each medium filled in each recording / reproducing device. (Sectors and the like) are shifted to intentionally cause a change in the medium replacement time, and the medium to be replaced at one time is always limited to one. A high-reliability, high-speed RAID system that uses a spare recording / reproducing device and improves the processing performance by simultaneously satisfying the writing and reading of data by combining multiple portable media without generating a waiting time due to medium exchange. Can be realized.

Further, according to the present invention, since the medium to be exchanged at a time is always limited to one by intentionally causing a difference in the exchange time of the medium, a single medium transport apparatus (or a recording / reproducing apparatus) is used. The effect described in the preceding paragraph can be exerted by a small number of medium transport devices. Therefore, the structure of the medium transport device of the present invention and the control method thereof are simpler than a library having the same number of media transport devices as the number of recording / reproducing devices, and are inexpensive because of its simplicity. Therefore, as a result, R by a portable medium having an inexpensive and simple mechanism is obtained.
An AID system can be realized.

Further, in the present invention, different RAID physical media can be provided simply by providing a simple tray-shaped media transport device capable of simultaneously transporting a plurality of media constituting the RAID described in the second embodiment and a shelf supporting the same. A RAID system that can access data belonging to a group at high speed can be realized.

[Brief description of the drawings]

FIG. 1 is a schematic diagram for explaining a library device adopting a RAID system using a portable medium according to the present invention.

FIG. 2 is a block diagram for explaining a sequential access controller in a library device using the present invention.

FIG. 3 is a flowchart for explaining processing of a sector controller in a library device using the present invention.

FIG. 4 is a flowchart for explaining processing of a drive switching controller in a library device utilizing the present invention.

FIG. 5 is a diagram illustrating an RA in a library device using the present invention;
FIG. 4 is a diagram for explaining a relationship between a change in the usage of a disk constituting an ID logical group and time.

FIG. 6 shows an RA in a library device utilizing the present invention.
FIG. 3 is a diagram for explaining data arrangement in disks constituting an ID logical group;

FIG. 7 is a schematic diagram illustrating a library device according to a second embodiment of the present invention.

FIG. 8 shows an RA in a library device using the present invention.
FIG. 3 is a diagram for explaining a relationship between a data arrangement in a disk constituting an ID physical group and time.

FIG. 9 is a schematic diagram for explaining a conventional library device adopting a RAID system using a portable medium.

FIG. 10 is a diagram for explaining a data arrangement in a medium of a conventional library device adopting a RAID system using a portable medium.

[Description of sign]

1a to 1f Optical disk filled in optical disk drive 2a to 2e Regular optical disk drive 3 Optical disk transport accessor 4 Shelf for storing a plurality of optical disks 5a to 5f Optical disk placed on a shelf for storing a plurality of optical disks 6 Spare optical disk drives 7a to 7e Regular optical disk drive controller 8 Spare optical disk drive controller 9 Optical disk transport accessor controller 10 Sequential access controller 11 Host computer 12 Sequential access controller CPU 13 Sequential access controller memory 14 I / O controller 15 Sector controller 16 Drive switching controller 17 Sequential access controller Hard disk 17a sequence Management file in the hard disk of the local access controller 18 Shelf for storing the same number of optical disks as the optical disk drive in the same horizontal plane 19 Tray-shaped optical disk accessors 19a to 19e Individual accessors constituting the tray-shaped optical disk accessor 20 Optical disk loading / unloading port 21 Optical Disk Library Device (Side) 22 Optical Disk Library Device (Front) 23 Array Controller

Claims (4)

[Claims] A plurality of recording / reproducing mechanisms for writing / reading to / from a portable medium, a transport mechanism for transporting the portable medium to the recording / reproducing mechanism, and a shelf capable of accommodating a plurality of the portable media. A recording / reproduction control device capable of performing parallel writing or parallel reading on said portable medium using a plurality of recording / reproduction mechanisms for input data, comprising: When writing or reading data in parallel to each other, writing or reading is started from physically different positions on each portable medium, and each portable medium is sequentially replaced at different times. Recording and reproduction control device. 2. A plurality of recording / reproducing mechanisms for writing or reading to / from a portable medium, a transport mechanism for transporting the portable medium to the recording / reproducing mechanism, and a shelf capable of storing a plurality of the portable media. A recording / reproduction control device capable of performing parallel writing or parallel reading on said portable medium using a plurality of recording / reproduction mechanisms for input data, comprising: When writing or reading data in parallel, the data is written or read on each portable medium from the position shifted by the recording capacity of each portable medium by a size equally divided by an integer larger than the number of recording / reproducing mechanisms. A recording / reproducing control device for starting reading and replacing each portable medium sequentially at different times. 3. A plurality of recording / reproducing mechanisms for writing or reading to / from a portable medium, a transport mechanism for transporting the portable medium to the recording / reproducing mechanism, and a shelf capable of storing a plurality of the portable media. A recording / reproduction control device capable of performing parallel writing or parallel reading on said portable medium using a plurality of recording / reproduction mechanisms for input data, comprising: When writing or reading data in parallel, the data is written or read on each portable medium from the position shifted by the recording capacity of each portable medium by a size equally divided by an integer larger than the number of recording / reproducing mechanisms. A read-out is started, and a spare portable medium and a spare recording / reproducing mechanism for writing or reading data during the exchange period of any one portable medium are provided. That data write or without generating read latency, the recording and reproducing controller, characterized in that to perform a continuous writing or reading data. 4. The recording / reproducing control device according to claim 1, wherein the transport mechanism has a plurality of accessors corresponding to the number of mediums of the series of portable media that has been written or read in parallel. A tray-shaped transport mechanism, when writing or reading data in parallel on a plurality of portable media using a plurality of recording and playback mechanisms, by the tray-shaped transport mechanism, from the shelf to the recording and playback mechanism, A recording / reproducing control device for taking out, transporting, and storing the series of portable media once or twice.