JPH0883150A - Magnetic disk device - Google Patents

Abstract

PURPOSE: To improve the access processing efficiency and to improve reliability of a magnetic disk as the whole by lowering destruction probability for significant data relatively in magnetic disk device in which plural disk driving devices are assembled. CONSTITUTION: This device is equipped with plural cache memorys 9, 10 provided corresponding to the plural disk driving devices 5. 6. respectively, and an access control part 11 which stores the data of an inputted write request in the cache memory 9, 10 designated in advance corresponding to attribute information included in the write request. Moreover, it is equipped with a cache memory managing part 12 which receives a write command from the access control part 11 and writes the data written on the cache memory 9, 10 on the disk driving devices 5, 6 in accordance with the cache memory.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a magnetic disk device used as an external storage device in an information processing system, and more particularly to a magnetic disk device incorporating a plurality of desk drive devices.

[0002]

2. Description of the Related Art A computer in an information processing system has a main memory such as a semiconductor memory such as a RAM and an auxiliary memory such as a magnetic disk device as a storage element for storing various information. For example, a central processing unit such as a CPU executes a data access by directly designating an address to a semiconductor memory such as a RAM directly connected to a data bus or an address bus. On the other hand, when the central processing unit accesses data to the magnetic disk unit, it is executed via the basic software (OS) operating in this central processing unit.

When this basic software (OS) manages writing and reading of data to and from the magnetic disk device,
A series of data related to access to the magnetic disk is managed as a file. Then, it is necessary to secure the area necessary for storing the data of this file on the magnetic disk. Methods for securing a necessary area for a file on this magnetic disk are roughly classified into the following two types.

(1) Areas are set in advance on a magnetic disk, and the start position (start address) and end position (end address) of each area, the effective range of actual data, etc. are stored in a RAM or the like on the central processing unit side. It is registered in the management table formed in the storage element. Then, when an access request is issued, the actual access to the magnetic disk is executed based on the registration information of this management table.

However, in this method, since the size of the area is fixed in advance, there arises a problem that each data having various data lengths cannot be efficiently stored and held. (2) Every time a write access request is generated, an area of the required size is secured on the magnetic disk at that time,
Write processing is executed. In this case, information such as in which basin each data was written, and the start position and end position of the corresponding region is registered in the management table formed in the magnetic disk device.

In this method, since the area having the size corresponding to the data length is secured, the storage capacity of the magnetic disk can be effectively used. In addition, since there is a certain limit in the storage capacity of one magnetic disk, when storing and holding a large amount of various kinds of data, a plurality of disk drive devices each having a magnetic disk are incorporated in the magnetic disk. Use a disk device.

[0007]

[Problems to be Solved by the Invention] However, (2)
Even in the magnetic disk drive, which has a structure in which an area having a size corresponding to the required data length is formed on the magnetic disk every time a write request is generated, the following problems to be improved was there.

That is, the data to be stored in the magnetic disk device as the auxiliary storage device is of various types as described above. For example, once written data, such as data in a database or dictionary, is rarely rewritten, and data that is almost read only, or data that has been externally fetched and subjected to predetermined data processing is executed. There is data that is read relatively frequently, such as data that is output later.

Further, the length of the data forming one file also changes depending on the type of data. As described above, in the magnetic disk device, the write request data are all the same type of data in response to the access request from the basic program (OS) of the central processing unit, although the data attributes are different from each other. The same treatment was carried out.

In general, the life of a magnetic disk depends on the number of accesses, but when the number of accesses is the same, the read access only tends to have a longer life than the write access. In general, data in a database or dictionary is used for a long period of time and is often more important than general data that is frequently rewritten.

Therefore, when such important data and relatively unimportant data are stored and held on the same magnetic disk, the life of this highly important data depends on the life of the lower magnetic disk. As a result, there is a concern that reliability will decrease.

Generally, the same kind of data tends to be continuously written or read, but when various data are randomly stored in a plurality of magnetic disks, it is extremely difficult. In this case, when data is continuously accessed, the disk drive device to be accessed is changed for each access, and there is a problem that the access processing efficiency as a whole is lowered.

Furthermore, when writing long data and short data to the same magnetic disk or the same area within the same magnetic disk, if writing, reading and erasing are repeated on the magnetic disk, the data on the magnetic disk is rewritten. In this case, a large number of small writable areas are scattered, and when a long data write request occurs, it is difficult to secure a large continuous write area, and it is necessary to set the write areas in a distributed manner. Therefore, the write processing efficiency decreases. Also, when reading the data written in the dispersed areas, the read processing efficiency similarly decreases.

The present invention has been made in view of such circumstances, and the access processing efficiency of the entire apparatus is determined by distinguishing the write destination magnetic disk or area of the corresponding data by the attribute of the write request data. It is an object of the present invention to provide a magnetic disk device that can improve the reliability of the entire device by improving the reliability of the device and the data destruction probability of important data.

[0015]

In order to solve the above-mentioned problems, the invention of claim 1 incorporates a plurality of disk drive devices each having a magnetic disk, and a plurality of disk drive devices are provided according to a write request inputted from the outside. Select one of the disk drive devices to write data, store and hold the write position in the management table, and from the position stored in the management table in response to a read request input from the outside. In a magnetic disk device for reading and sending out data, a plurality of cache memories provided respectively corresponding to a plurality of disk drive devices and data of an input write request correspond to attribute information included in this write request Stored in a pre-specified cache memory and outputting a write command, and a write from the access control unit. In response to a command, in which a cache memory management unit for writing to the disk drive apparatus corresponding to data written to the cache memory to the cache memory.

According to the second aspect of the invention, the attribute information included in the write request in the above-mentioned invention is such that the data of this write request is the data to be written and read, or is mainly read only. It is access type information that distinguishes whether the data is to be performed.

Further, in another invention of claim 3,
The attribute information included in the write request in the above invention is
This is data length information that distinguishes whether the data of the write request is long data over a plurality of loading requests or short data completed by one write request.

Further, in the magnetic disk device according to the invention of claim 4, the data of the write request formed in the write request is data which is frequently written and read, or is mainly read only. Access type information for distinguishing whether the data is to be performed, and data length information for distinguishing whether the data of the write request is long data over a plurality of loading requests or short data completed by one write request; A plurality of cache memories provided respectively corresponding to the disk drive devices, and a cache in which the data of the input write request is designated in advance corresponding to the access type information and the data length information included in the write request. An access control unit that stores the data in the memory and outputs a write command, and a write command from the access control unit receives the data written in the cache memory. Together written to the disk drive apparatus corresponding to Shumemori, and a cache memory management unit to write to different areas in accordance with the data length information in the disk drive apparatus corresponding data.

[0019]

In the magnetic disk device thus constructed, a plurality of disk drive devices each having a magnetic disk and a plurality of cache memories corresponding to the respective disk drive devices are provided. Also,
The write request input from the outside includes attribute information indicating the nature of this data in addition to the data. Then, the plurality of disk drive devices and the cache memory are allocated corresponding to the attribute information.

Then, the input write request data is written in the disk drive device corresponding to the attribute information by the access control unit and the cache memory control unit. Therefore, since similar data are written on the same magnetic disk, the data access processing efficiency is improved.

According to the second aspect of the present invention, in the case of data that is mainly read, such as data in a database or a dictionary, the data is written and read on a separate magnetic disk. Written. Therefore, as described above, it is possible to improve the data access processing efficiency and extend the life of the magnetic disk of the disk drive device in which the data that is mainly read only is written.

Further, according to the third aspect of the invention, the long data is written on another magnetic disk separately from the short data. Therefore, as described above, in the case of long data, a large area can be easily secured, and the data access processing efficiency can be further improved.

In the fourth aspect of the invention, the write request includes the access type information and the data length information as attribute information. Then, the write request data is written in different magnetic disks according to the access type information, and is written in different areas on the same magnetic disk according to the data length. Therefore, the effect of extending the life of the magnetic disk and the effect of improving access processing efficiency can be obtained at the same time.

[0024]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing a schematic configuration of the magnetic disk device of the embodiment. This magnetic disk device 1 is a central processing unit (C
PU) 2 via a bus 3a. A main memory 4 formed of a RAM element or the like is connected to the central processing unit (CPU) 2 via a bus 3b.

In the magnetic disk device 1, a first disk drive device 5 for storing data, such as general data, which can be read and written substantially uniformly.
And a second disk drive device 6 for storing mainly read-only data such as data in a data database or a dictionary. Further, one magnetic disk 7 or 8 is provided in each disk drive device 5 or 6, respectively.
Is incorporated.

The magnetic disk 7 of the first disk drive device 5 has a long data area 7a exclusively storing and holding long data and a short data area 7b exclusively storing and holding short data. This long data area 7
The boundary between a and the short data area 7b is not fixed in advance and moves according to the amount of each data.

Further, in the magnetic disk device 1, each data to be written to each disk drive device 5, 6 is temporarily stored, or each data read from each disk drive device 5, 6 is temporarily stored. Dedicated cache memories 9 and 10 are provided.

Further, in the magnetic disk device 1, an access control unit 1 composed of, for example, a microcomputer is provided.
1 and a cache memory management unit 12 are provided.
A management table 13 shown in FIG. 3 is formed in the cache memory management unit 12.

The central processing unit 2 sends a write request and a read request to the magnetic disk unit 1 as needed. FIG. 2A is a diagram showing the data format of the write request 14. The access type information 14a is set at the head position of the write request 14, and the data length information 1 is secondly set.
4b is set. Next, the identifier 14c for specifying the data 14d is set, and finally the data 14d to be written is set.

The access type information 14a and the data length information 14b are attribute information indicating the characteristics of the data 14d. The access type information 14a is, for example, reading information indicating that the data 14d is data in which reading and writing are executed almost equally like normal data, and
For example, read information indicating that the data is mainly read only, such as data in a data database or a dictionary, is prepared.

Further, the data length information 14b is long data information indicating that the data 14d is a part of data constituting not only this one write request 14 but also a plurality of loading requests 14. And this one write request 1
The short data information indicating the short data completed in 4 is prepared.

The identifier 14c is information such as a file name for specifying the data 14d consisting of, for example, one file.
It becomes search data when reading from No. 8.

FIG. 2B shows the data format of the read request 15. An identifier 15a is set at the head position of the read request 15, and a request size 15b that specifies the size of the data to be read second is set.

In the management table 13 shown in FIG. 3, for each identifier 13a, an HD for identifying the disk drive device 5, 6 in which the data identified by the corresponding identifier 13a is stored.
Number 13b, start and end address 13c of corresponding data
Is remembered.

In the magnetic disk device 1 having such a configuration, when the access request is input from the central processing unit 2, the access control unit 11 performs access processing to the disk drive devices 5 and 6 according to the flowchart shown in FIG. Is configured to run.

When the flow chart is started and an access request is received at P (program step) 1, it is judged whether the access request is a write request 14 or a read request 15 (P2). In the case of the write request 14, the leading access type information 14a is checked. For reading information, go to P4,
The next data length information 14b is checked.

If the data length information 14b is long data information, the process proceeds to P5, and the data 14d of the write request 14 is written in the empty area of the long data area 9a of the cache memory 9 corresponding to the reading disk drive device 5. Put in. afterwards,
A write command designating the write destination disk drive device 5 is sent to the cache memory management unit 12 (P6).

If the data length short information 14b is short data information at P4, the process proceeds to P7, and the data of the write request 14 is written to the short data area 9b of the cache memory 9 corresponding to the reading disk drive device 5. Write 14d. After that, a write command is sent to the cache memory management unit 12 (P6).

Further, in P3, when the leading access type information 14a is read information, the process proceeds to P8, and the data 14d of the write request 14 is written to the data area 10a of the cache memory 10 corresponding to the read disk drive device 6. Write in. After that, a write command is sent to the cache memory management unit 12 (P6).

Further, when the input access request is the read request 15 at P2, the process proceeds to P9 and the read command is sent to the cache memory management unit 12. After that, the areas 9a, 8b, 10 of one of the cache memories 9, 10
When it is confirmed that the data has been read from the disk drive units 5 and 6 in a, the read data is read again and sent to the central processing unit 2 via the bus 3a (P1).
0).

When the write command or the read command is input from the accessor control unit 11, the cache memory management unit 12 follows the flow chart shown in FIG.
The program is configured to execute the actual access processing for the data.

The flow chart of FIG. 5 is started, and it is judged at Q1 whether the input command is a write command or a read command. In the case of the write command, the process proceeds to Q2, and the corresponding cache memories 9 and 10 are judged from the write destination disk drive devices 5 and 6 (Q2). In the case of the cache memory 9 for reading, Q
Then, the process proceeds to step 3 to check whether data is stored in the long or short data areas 9a and 9b.

In the case of the long data area 7a, the data 14d stored in the long data area 7a is read and written in the empty area of the long data area 7a of the magnetic disk 7 in the disk drive device 5 (Q4).

After that, the process proceeds to Q5, and the identifier 14a of the corresponding data 14d is stored in the empty area of the management table 13 shown in FIG.
(13a), the HD number 13b that specifies the written disk drive device, and the write start address and write end address 13c on the magnetic disk 7 are set (Q5).

When the data 14d is stored in the short data area 9b of the cache memory 9 in Q3, the data 14d is written in an empty area of the short data area 7b of the magnetic disk 7 in the disk drive device 5. (Q6). Then, the process proceeds to Q5, and the identifier, HD number, and address are registered in the management table 13.

Further, in Q2, in the case of the cache memory determined from the write destination disk drive device 5, 6 or the read cache memory 10, the process proceeds to Q7 and is stored in the data area 10a of this cache memory 10. The read data 14a is read and written in the free area of the data area 8a of the magnetic disk 8 in the disk drive device 6. Then, the process proceeds to Q5, and the identifier, HD number, and address are registered in the management table 13.

If the command input from the access control unit 11 is a read command in Q1, the process proceeds to Q8 and the HD number 13b and start address corresponding to the identifier 13a designated by the read command from the management table 13 are entered. And the end address 13c are read (Q8).

Then, the maximum readable data amount within the required size 15b is read from the start address on the magnetic disk 7, 8 of the designated disk drive device 5, 6 and the corresponding cache memory 9, 10 is read. Data is written in the data areas 9a and 10a (Q9).

If the reading of all the data designated by the requested size 15b has not been completed, the process returns to Q9,
Data is read from the next address and written into the corresponding data areas 9a, 10a of the cache memories 9, 10.

When the reading of all the data designated by the requested size 15b is completed, the current reading process is completed. In the magnetic disk device 1 configured as described above, the first
Write to the magnetic disk 7 of the disk drive device 5 of
The general data to be read is mainly written. On the other hand, on the magnetic disk 8 of the second disk drive device 6, relatively important data such as data of a database or a dictionary exclusively for reading is written.

Therefore, such important data is stored and held in the read-only disk drive device 6. That is, by discriminatingly storing and storing this important data in the disk drive device 5 in which reading and writing are frequently performed, compared with other general data,
Data reliability can be maintained higher.

Further, even in the data which is frequently written and read, the long data and the short data are written in different data areas 7a and 7b in the magnetic disk 7 of the same disk drive device 5, respectively.

Therefore, in the case of long data, it is possible to easily secure a large area in the long data area 7a, and a series of data will be written to consecutive addresses, and a large number of discrete small areas will be written. The data access processing efficiency can be improved as compared with the case where data is written over the area.

Also, in the short data area 7b, since only relatively small data is written, it is possible to reduce the occurrence probability of the worm eating state in the data area and to effectively use the data area.

Data write / read processing for the respective cache memories 9 and 10 is executed by the access processing section 11, and actual access processing for the disk drive devices 5 and 6 is executed by the independent cache memory management section 12. ing.

As a result, when data of a large requested size is read, writing to the cache memory and reading from the cache memory can be performed independently, so that the access processing efficiency seen from the side of the central processing unit 2 is high. Can be improved.

FIG. 6 is a block diagram showing a schematic structure of a magnetic disk device according to another embodiment of the present invention. The same parts as those in the embodiment shown in FIG. 1 are designated by the same reference numerals. Therefore, detailed description of the overlapping portions is omitted.

The apparatus of this embodiment is provided with a first disk drive device 21 which mainly stores and holds long data and a second disk drive device 22 which mainly stores and holds short data.

Then, in response to the write request 14 input from the central processing unit 2, the access control unit 23 stores the data 1 in the cache memory 9 for long data in the case of long data.
4d is written, and in the case of short data, it is written in the cache memory 10 for short data.

Even with such a configuration, depending on the type of data, the write destination disk drive device 21,
Since 22 are distinguished, it is possible to obtain an effect similar to that of the embodiment shown in FIG.

[0061]

As described above, according to the magnetic disk device of the present invention, the write destination magnetic disk or area of the corresponding data is distinguished by the attribute of the data of the write request. Therefore, the access processing efficiency of the entire apparatus can be improved, the data destruction probability for important data can be relatively reduced, and the reliability of the entire apparatus can be improved.

[Brief description of drawings]

FIG. 1 is a block diagram showing a schematic configuration of a magnetic disk device according to an embodiment of the present invention.

FIG. 2 is a diagram showing a data format showing a write request and a read request input to the apparatus of the embodiment.

FIG. 3 is a diagram showing a management table formed in the apparatus of the embodiment.

FIG. 4 is a flowchart showing the operation of the access control score in the apparatus of the embodiment.

FIG. 5 is a flowchart showing an operation of a cache memory management unit in the apparatus of the embodiment.

FIG. 6 is a block diagram showing a schematic configuration of a magnetic disk device according to another embodiment of the present invention.

[Explanation of symbols]

1 ... Magnetic disk device, 2 ... Central processing unit, 5, 21 ...
First disk drive device, 6, 22 ... Second disk drive device, 7, 8 ... Magnetic disk, 9, 10 ... Cache memory 11, 23 ... Access control unit, 12 ...
Cache memory management unit, 13 ... Management table, 14 ...
Write request, 15 ... Read request

Claims (4)

[Claims] 1. A plurality of disk drive devices each having a magnetic disk are built-in, and one of the plurality of disk drive devices is selected in accordance with a write request input from the outside to write data. writing,
In a magnetic disk device for storing and holding a writing position in a management table and reading and sending data from a position stored in the management table in response to a read request inputted from the outside, the magnetic disk device comprises: A plurality of cache memories provided corresponding to each other, and the data of the input write request is stored in a cache memory designated in advance corresponding to the attribute information included in the write request, and a write command is issued. And a cache memory management unit for receiving the write command from the access control unit and writing the data written in the cache memory into the disk drive device corresponding to the cache memory. Magnetic disk device. 2. The attribute information included in the write request is access type information for distinguishing whether the data of the write request is data to be written and read or data mainly to be read. The magnetic disk drive according to claim 1, wherein the magnetic disk drive is provided. 3. The attribute information included in the write request distinguishes whether the data of the write request is long data over a plurality of loading requests or short data completed by one write request. The magnetic disk drive according to claim 1, wherein the magnetic disk drive is information. 4. A plurality of disk drive devices each having a magnetic disk as a recording medium are built-in, and one of the plurality of disk drive devices is selected in response to a write request inputted from the outside. In a magnetic disk device for writing data, storing and holding the writing position in the management table, and reading and sending the data from the position stored in the management table in response to a read request input from the outside. Is included in the write request, and the data of this write request is
Access type information for distinguishing between data to be written and read or data to be mainly read, and whether the data of the write request is long data over a plurality of loading requests or one write request Data length information for distinguishing whether the data is short data, a plurality of cache memories provided respectively corresponding to the plurality of disk drive devices, and the data of the input write request is included in this write request. Corresponding to the access type information and the data length information to be stored, the access control section storing the pre-designated cache memory and outputting the write command, and the cache memory receiving the write command from the access control section. While writing the data written in to the disk drive device corresponding to this cache memory,
A magnetic disk device comprising: a cache memory management unit that writes the data to different areas according to data length information in the corresponding disk drive device.