JP2004046352A - Data storage device and method, and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a data storage device that can flexibly vary data multiplicity in compliance with data volume and disk capacity and can maximize disk reliability and a data transfer rate. <P>SOLUTION: With a storing block number set at a final block number plus one, whether there are free areas in a number of volumes equal to current multiplicity 33 or not is determined, and if there are such free areas, duplicates are stored in file blocks of multiplicity equal to the current multiplicity. If there are not such free areas, the current multiplicity is decremented by one, and areas already storing data are overwritten for storage. About the written volumes, a current block count 25 is updated, file block information 35 is added and the final block number 34 is updated. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a data storage device and a data storage method for copying and writing data to a plurality of secondary storage devices.
[0002]
[Prior art]
In recent years, the capacity of secondary storage devices (hereinafter, referred to as disks) has been increased, and accordingly, high-speed data transfer from disks and high reliability of disks have been required.
There is a disk array device as a means for dealing with these, and several data storage methods as described below have been proposed for the disk array device.
[0003]
One of the data storage methods of a disk array device is mirroring. Mirroring is a method for improving the reliability of disks by writing the same data to two disks. The mirror disk is also called RAID1 (RAID: Redundant Arrays of Inexpensive Disks).
[0004]
FIG. 15 is a diagram illustrating an example of data storage by mirroring. In the figure, # D1 to # D4 indicate physical disks, and B1 to B24 indicate file block numbers. The physical disk # D1 and the physical disk # D2 form a pair, and the physical disk # D3 and the physical disk # D4 form a pair. Each pair forms one logical disk. The same data is written in the paired physical disks, and the operating system operates to access # D2 when a failure occurs in # D1, for example, and when the failure occurs in # D2, the operating system operates on # D1. Operate to access. Therefore, data can be accessed as long as two paired physical disks do not fail simultaneously.
[0005]
Some mirroring has a load distribution function that selects a disk to which data is transferred so that the load on each physical disk is equalized during data transfer, and improves the overall data transfer efficiency of the disk. This is because when a data transfer request is issued to a logical disk that has a mirroring configuration, the next data transfer request is sent to the corresponding physical disk during the waiting time for the requested data to be transferred from the physical disk. It is issued for a disc. Thereby, the data transfer is performed in parallel and the data transfer efficiency is improved.
[0006]
As described above, by adopting a mirroring configuration in which the same data is stored in two disks, high reliability and high-speed transfer of the disks can be realized.
[0007]
Japanese Patent Application Laid-Open No. 6-18722 discloses a method of storing a copy of data on a plurality of disks. FIG. 16 and FIG. 17 are diagrams showing storage examples. In each figure, the same reference numerals are given to the same elements as those in FIG. 15, and the description thereof will be omitted. The storage example of FIG. 16 is based on a method of storing a copy of data on each disk in a distributed manner on all other disks. The data storage example of FIG. 17 is based on a method of storing a copy by dividing it into two or more other duplication disks.
[0008]
In the data storage method shown in FIGS. 16 and 17, when the stored data is transferred, the data is transferred in parallel from each disk, so that high-speed transfer can be realized. When data transfer requests are concentrated on a specific disk, the load can be distributed by transferring the same data from another disk. Furthermore, when recovering from a failure after a disk failure, data can be transferred from multiple disks in parallel and written to a spare disk, making recovery more difficult than a mirroring system that stores copies on only one disk. It also has the advantage of being fast.
[0009]
[Problems to be solved by the invention]
Since the conventional data storage device is configured as described above, the multiplexing degree of data at the time of disk connection, that is, the number of data copies is determined, so that it is possible to increase or decrease the amount of stored data or increase or decrease the number of connected disks. There has been a problem that the multiplicity cannot be flexibly changed accordingly.
[0010]
The present invention has been made in order to solve the above-described problems, and can flexibly change the degree of multiplexing of data according to the amount of data and the capacity of a disk, and can improve the reliability and data transfer efficiency of the disk. It is an object to obtain a data storage device and a data storage method that can be maximized.
[0011]
[Means for Solving the Problems]
A data storage device according to the present invention includes a volume management table storing a first identifier for identifying each volume constituted by a storage device and capacity information of each volume, and a copy number of a file block stored in the volume. A file block management table for storing the current multiplicity, which is the minimum value, a second identifier for identifying the volume in which each file block is to be stored, and the read priority of each file block, a first identifier, and a capacity. File block management means for controlling the writing of file blocks to each volume based on the information, the current multiplicity and the second identifier, and controlling the reading of file blocks from each volume in accordance with the priority. is there.
[0012]
In the data storage device according to the present invention, the file block management means determines whether or not free space exists in a number of volumes equal to the current multiplicity, and if so, the multiplicity of the file block is equal to the current multiplicity. In this case, the file block is duplicated and stored.
[0013]
In the data storage device according to the present invention, the file block management means overwrites a file block to be stored with an area storing a file block having a higher multiplicity than the current multiplicity when no free space exists in all volumes. Is what you do.
[0014]
In the data storage device according to the present invention, when reading data from each volume, the file block management means refers to the priority of the file block management table and reads out the file block having the highest priority.
[0015]
In the data storage device according to the present invention, when a certain volume becomes inaccessible, the file block management means reads a copy of the file block stored in the accessible volume.
[0016]
In the data storage device according to the present invention, the file block management means stores a file block having a lower multiplicity than the current multiplicity in the newly added volume.
[0017]
In the data storage device according to the present invention, the file block management means issues a warning to the user when the current multiplicity is changed.
[0018]
In the data storage device according to the present invention, the file block management means causes an error when the current multiplicity is changed.
[0019]
According to the data storage method of the present invention, a volume management table storing a first identifier for identifying each volume constituted by a storage device and capacity information of each volume, and a copy number of a file block stored in the volume And a file block management table for storing a second identifier for identifying a volume in which each file block is to be stored, and a read priority of each file block. A data storage method, which controls writing of a file block to each volume based on a first identifier, capacity information, current multiplicity, and a second identifier, and reads a file block from each volume according to a priority. The control is performed.
[0020]
A data storage program according to the present invention includes a volume management table for storing a first identifier for identifying each volume composed of a storage device and capacity information of each volume, and a copy number of a file block stored in the volume. A file block management table for storing the current multiplicity which is the minimum value, a second identifier for identifying a volume in which each file block is to be stored, and a read priority of each file block, a first identifier, capacity information, The computer functions as file block management means for controlling the writing of file blocks to each volume based on the current multiplicity and the second identifier, and controlling the reading of file blocks from each volume according to the priority. .
[0021]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, an embodiment of the present invention will be described.
Embodiment 1 FIG.
FIG. 1 is a diagram showing a configuration of a system including a data storage device according to Embodiment 1 of the present invention. In the figure, reference numeral 1 denotes a network (transfer path) connected to the buses (transfer paths) 2 and 2a by the network coupling devices 3a and 3b. Reference numeral 2 denotes a bus (transfer path) connecting the central processing unit 15 and the main storage device 10, which is connected to a bus (transfer path) 2b via a bus coupling device 3c for connecting a plurality of buses. . The bus coupling device 3c is, for example, a bridge that relays between different buses and networks. Reference numerals 6a and 7a denote magnetic disk devices (storage devices), which are connected to the bus 2a by a disk controller 4a and constitute one volume 8a. Similarly, magnetic disks 6b and 7b are also connected to the bus 2b by the disk controller 4b and constitute one volume 8b. Reference numeral 5a denotes an I / O processor connected to the bus 2a. The volume 8a corresponds to a network drive and is controlled by the I / O processor 5a. The volume 8b corresponds to a local drive. Here, the volume indicates a storage unit composed of one or more magnetic disk devices managed through an operating system.
[0022]
Reference numeral 10 denotes a main storage device, which stores a volume management unit 11, a file control unit 12, an operating system 13 for controlling the entire system, a file block management unit (file block management unit) 14, and the like. Reference numeral 15 denotes a central processing unit (CPU) which executes an operating system 13 or the like to control each device of the entire system.
[0023]
In this embodiment, the data storage processing according to the present invention is realized by the file block management unit 14 stored in the main storage device 10. The file block management unit 14 may be mounted on the main storage device 10 as software programmed separately from the operating system 13 to make a file control request to the operating system 13, File control may be performed integrally with the system 13.
[0024]
FIG. 2 is a diagram showing the interrelation of the components of the data storage device according to the first embodiment of the present invention. In the figure, reference numeral 8 denotes a plurality of volumes composed of magnetic disk devices, 12 denotes a file control unit for controlling the writing and reading of data to and from the volume 8 via an operating system, and 16 denotes data between the volume 8 and the operating system 13. The device driver 20 that relays transmission and reception issues a request from the volume management unit 11 to the operating system to obtain information such as the volume name (first identifier) of each volume and the volume type indicating the distinction between network drive / local drive. And a volume management table 30 generated based on these pieces of information. Reference numeral 30 denotes a volume name (second identifier) in which each file block is stored by the file block management unit 14 with respect to the operating system 13, a multiplicity of each file block, etc. of It issues a resulting request, indicating the file block management table generated based on this information.
[0025]
FIG. 3 is a diagram showing a configuration of the volume management table 20 in FIG. In the figure, reference numeral 21 denotes volume information, which includes a volume name (first identifier) 22, a type 23, a maximum block number (capacity information) 24, a current block number 25, and an active flag 26. The volume information 21 is stored in the volume management table 20 by the number of volumes. The volume name 22 is an identifier of a volume identified by the operating system 13. The type 23 is a type of a volume, and indicates a distinction between a network drive, a local drive, and the like. The maximum block number 24 represents the maximum number of file blocks that can be stored in the volume, and is obtained by dividing the volume capacity by one file block size. The current block number 25 is the number of file blocks stored in each volume, and is 0 in the initial state. The active flag 26 indicates whether or not the drive is used as part of a redundant configuration, or whether the drive is offline due to an error or the like.
[0026]
FIG. 4 is a diagram showing a configuration of the file block management table 30 in FIG.
In the figure, reference numeral 31 denotes file block management information, which comprises a maximum multiplicity 32, a current multiplicity 33, and a last block number 34. Reference numeral 35 denotes file block information, which includes a volume name (second identifier) 36, a block number 37, a time stamp 38, and a rank (priority) 39. The file block management table 30 is composed of file block management information 31 and file block information 35. One file block management information 31 is stored in the file block management table 30, and the file block information 35 is stored by the number of file blocks. You. Here, the file block is a numerical sequence obtained by dividing data to be stored in the storage device by the same capacity unit, for example, a numerical sequence that is the minimum unit of data to be stored in the storage device.
[0027]
In FIG. 4, the maximum multiplicity 32 is the maximum value of the number of copies of the stored file block, the current multiplicity 33 is the minimum value of the number of copies of the stored file block, and the last block number 34 is the block number among the block numbers in all volumes. Indicates the last number. The volume name 36 indicates the identifier of the volume in which the file block is stored. The block number 37 indicates a serial number assigned to a file block stored in each volume, and means that file blocks having the same block number 37 have the same data content. The time stamp 38 indicates the date and time when the file block was stored in the volume, and the rank 39 indicates the priority when reading (reading) the file block having the same block number.
[0028]
Next, an operation of writing data to the volume 8 will be described.
FIG. 5 is a flowchart showing a file block storage procedure in this embodiment. FIG. 6 is an example of file block storage according to the procedure of FIG. 5. In FIG. 6, # V1 to # V4 are volume names, and B1 to B8 are Block numbers are shown and common elements are shown in the following figures, and the description thereof is omitted. FIG. 6 shows a state where data is stored in all areas of four volumes having a maximum volume number of 24 = 8.
The procedure for storing data from the initial state in which all the volumes in FIG. 6 are free areas will be described with reference to FIG.
[0029]
In the initial state of FIG. 6, the block management unit 14 sets the last block number 34 to 0, and sets the maximum multiplicity 32 and the current multiplicity 33 to 4 equal to the number of volumes. First, the file block management unit 14 acquires the final block number 34 = 0 from the file block management table 30 (step ST1), and sets the acquired (final block number 34) + 1 = B1 as the block number to be stored (step ST2). ). Subsequently, the file block management unit 14 determines whether a free area exists in a number of volumes equal to the current multiplicity of 33 (step ST3). For example, the determination can be made by the file block management unit 14 acquiring the maximum number of blocks 24 from the volume management table 20 via the volume management unit 11 and comparing this with the block number to be stored obtained in step ST2. If the block number to be stored is smaller than the maximum number of all blocks 24, a free area exists in a number of volumes equal to the current multiplicity of 33. In this case, since the maximum number of all blocks 24 is 8 and the block number to be stored is 1, it can be determined that there is a free area.
[0030]
If there is no free area, the process proceeds to block overwrite update processing (step ST4). If there is a free area, the file block management unit 14 acquires the name of the volume in which the free area exists via the volume management unit 11 and sets the multiplicity of the write data to these volumes to be equal to the current multiplicity 33. A request to copy and write data is made to the operating system 13 (step ST5). The volume management unit 11 may acquire a volume in which the current block number 25 is smaller than the maximum block number 24 as a volume having a free area. The operating system 13 stores data in a free area of the volume according to a request from the file block management unit 14. In FIG. 6, one data is stored in each of the volumes # V1 to # V4 such that the multiplicity is 4.
[0031]
Subsequently, the volume management unit 11 updates the current block number 25 of each of the volume names # V1 to # V4 of the volume information 21 on the volume management table 20 to 1 (step ST6). The file block management unit 14 adds the file block information 35 of the block number B1 (step ST7), and updates the last block number 34 of the file block management information 31 to B1 (step ST8). Subsequently, when storing the next data, the process starts from step ST1 of the flowchart again and stores the next data. FIG. 6 shows a state where data is stored by repeating the above procedure, and the multiplicity of all file blocks = 4 and the final block number = B8.
[0032]
Next, a file block read operation will be described.
First, the file control unit 12 passes the read target block number to the file block management unit 14. The file block management unit 14 refers to the rank 39 of the number obtained from the file block management table 30, obtains the volume name of the file block information 35 having the highest priority, and returns it to the file control unit 12. Subsequently, the file control unit 12 requests the operating system 13 to read the read target block stored in the volume having the acquired volume name.
The operating system 13 sends a file read request to the device driver according to the request from the file control unit 12. The device driver acquires the read data from the volume and returns it to the operating system 13.
[0033]
Here, an operation of updating the rank 39 will be described.
FIG. 7 is a flowchart showing a procedure for updating the rank 39 of the file block information 35 on the file block management table 30. FIG. 8A is a diagram showing a read file block selected according to the rank 39 as an ellipse. Description will be made with reference to these figures.
[0034]
The file block management unit 14 initializes all ranks 39 of the file block information 35 on the file block management table 30 (step ST11), and first obtains the file block information 35 of the block number B1 as a rank update target file block. (Step ST12). Subsequently, the file block management unit 14 refers to the volume name of the acquired file block information 35 and calculates the total value of the rank 39 of the file block information 35 having the same volume name (step ST13). In FIG. 8A, the volume names of the file block information 35 of the block number B1 are # V1 to # V4, and the ranks 39 corresponding to all the volume names are all initialized. The total value is all 0.
[0035]
Subsequently, the file block management unit 14 compares the total value of the ranks 39 of the respective volumes, and updates the rank 39 of the rank update target block number 37 corresponding to the volume having the smallest total value to, for example, 1 (step ST14). ). When there are a plurality of ranks 39 having the smallest total value of the ranks 39, the ranks 39 may be updated according to the order of the volume information 21 on the volume management table 20. In FIG. 8A, the volume # V1 is selected according to the order of the volume information 21 out of # V1 to # V4 having the smallest total value of the rank 39, and the file block information 35 of the block number B1 stored in the volume # V1 is selected. Is updated to 1.
[0036]
Subsequently, the file block management unit 14 similarly updates the rank 39 for the file block information 35 of the block number B2. Since the rank 39 of the block number B1 of the volume # V1 has been updated to 1, the total rank of the volume # V1 is 1, and the total rank of the volumes # V2 to # V4 is 0. Therefore, the file block management unit 14 selects the volume # V2 according to the order of the volume information 21 among the volumes # V2 to # V4 having the smallest total rank value, and updates the rank 39 of the block number B2 of the volume # V2 to 1. I do. Steps ST12 to ST14 are similarly repeated for the file blocks of block numbers B3 to B8 (step ST15).
[0037]
By determining the read priority according to the rank 39 as described above, file blocks can be read from different volumes in parallel, and the overall data transfer speed and reliability are improved.
[0038]
FIG. 8B is a diagram showing, by an ellipse, the read file blocks that are read when four read requests for the block number B3 are made from the file control unit 12 to the file block management unit 14. As in the case of FIG. 8A, the file block management unit 14 may update the rank 39 of each file block and determine the read priority of the file block according to the rank 39. When the ranks 39 are the same, the priority may be determined according to the order of the volume information 21 on the volume management table 20. When the file block management unit 14 selects a read file block according to the priority, when a read request for the same file block occurs, the file blocks with the same block number in different volumes can be read simultaneously. As a result, the data transfer waiting time can be eliminated and the throughput can be improved.
[0039]
The file block management unit 14 may perform the rank updating process immediately before reading the file block, or may perform it after the writing is completed. Further, after setting the rank 39 for the block numbers 37 of B1 to B8 according to the above procedure, the file block management unit 14 again targets the block numbers 37 corresponding to the unset ranks 39 in step ST12 of the flowchart of FIG. And step ST14 may be repeated to set the rank 39. In this case, the initial value is set to 4, and the value is decremented each time the setting is repeated thereafter. Thus, different ranks 39 are set for different file block information 35 having the same block number.
[0040]
Next, a data write operation including an overwrite process will be described.
FIG. 9 is a flowchart showing a data writing procedure including an overwriting process. FIG. 10A is a diagram in which a file block stored as shown in FIG. 6 is overwritten and updated, and the configuration of FIG. 6 is overwritten according to FIG. The procedure to be performed will be described.
[0041]
The file block management unit 14 acquires the last block number 34 = B8 from the file block management information 31 stored in the file block management table 30 (step ST16), and stores the acquired (last block number 34) + 1 = B9. (Step ST17). Subsequently, the volume management unit 11 acquires a volume name in which an empty area exists, and passes the acquired name to the file block management unit 14. In the case of FIG. 6, since the current block number is 25 = 8 and the maximum block number is 24 = 8 in all the volumes, it can be seen that there is no free area.
[0042]
If there is a free area, the file block management unit 14 makes a request to the operating system 13 to write data to the free area, and the operating system 13 stores the data in accordance with the request of the file block management unit 14. The data is duplicated for 33 times and stored (step ST18). The volume management unit 11 updates the current block number 25 of the volume information 21 of the volume in which the data is written by adding 1 (step ST19).
[0043]
Subsequently, the file block management unit 14 determines whether or not there is an unwritten block, that is, whether or not the number obtained by subtracting the number of volumes written from the current multiplicity 33 is 0 (step ST20). Does not exist, the number of remaining writings of B9 is four. Subsequently, the file block management unit 14 acquires the volume in which the stored file block = B9 is not written, that is, the block number 37 of the file block information 35 of the volume names # V1 to # V4 (step ST21). The number of the file block information 35 of the number is counted for each block number (step ST22). The block numbers of the file block information 35 of the volume names # V1 to # V4 are B1 to B8, and the number of the file block information 35 of all the numbers is four. Here, the block number 37 of the volume to which the file block is not written is obtained so that the same block is distributed and written to a plurality of volumes.
[0044]
The file block management unit 14 compares the number of file block information 35 with the current multiplicity 33 and selects a block having a larger number of file block information 35 as a block to be overwritten (step ST23). Since the weight is 33 = 4, there is no block to be overwritten. In this case, the current multiplicity is reduced by 1 (step ST23). At this time, the file block management unit 14 may request the operating system 13 to issue a warning to the user or request an error.
[0045]
As a result, the file blocks with the numbers B1 to B8 become candidates for blocks to be overwritten. As described above, when there are a plurality of blocks in which the number of the file block information 35 is larger than the current multiplicity 33, the block having the smallest block number 37 may be selected as the overwritten block. The volume to be stored may be determined according to the order of the volume information 21 on the volume management table 20. Therefore, the file block management unit 14 selects the block with the block number B1 of the volume # V1 as the block to be overwritten. The process of step ST23 is repeated until the number of blocks to be overwritten reaches the current multiplicity 33, that is, the number obtained by subtracting the number of volumes already written in step ST18 by the current multiplicity 33 (step ST24). Since the block B1 is selected as the block to be overwritten, the blocks having the file block information 35 larger than the current multiplicity are B2 to B8, and the block B2 of the volume # V2 is selected as the next block to be overwritten. . Similarly, block B3 of volume # V3 is selected as a block to be overwritten.
[0046]
Subsequently, the file block management unit 14 determines whether a block to be overwritten has been selected (step ST25), and issues a request to the operating system 13 to write to each volume (step ST26). The operating system 13 overwrites the storage target block. Subsequently, the file block management unit 14 updates the file block management table with the block numbers 37 = B1, B2, and B3 of the overwritten block in the file block information 35 and the overwritten and stored block number = B9 (step ST27). , The last block number 34 is updated to B9 (step ST28). When there is no remaining block to be written in step ST20, and when a block to be overwritten is selected in step ST25, steps ST27 and ST28 are performed.
[0047]
In FIG. 10A, blocks overwritten by the above procedure are indicated by circles. According to the above procedure, the file blocks are overwritten so that the multiplicity of each file block is maximum and the same blocks are evenly distributed to each volume. FIG. 10B is a diagram showing a case where a file block is overwritten by the above procedure and the multiplicity of all the file blocks becomes 1. In this example, no more file blocks can be stored.
[0048]
FIG. 11A is a diagram showing an ellipse of a file block read from the file block shown in FIG. The read priority is determined according to the rank 39 of the file block information 35 of the file block management table 30, and the rank 39 is set according to the procedure shown in FIG. The procedure for setting the rank 39 will be described with reference to FIGS. 7 and 11A.
[0049]
The file block management unit 14 initializes all the ranks 39 of the file block information 35 of the file block management table 30 (step ST11), and acquires the file block information 35 of the rank update target block number B1 (step ST12). Subsequently, the total of the ranks 39 of the file block information 35 having the same volume name is calculated with reference to the volume name of the obtained file block information 35 (step ST13). In this case, the volume names of the acquired file block information 35 are # V2 to # V4, and the total value of the rank 39 corresponding to each volume name is all 0.
[0050]
Subsequently, the file block management unit 14 compares the total value of the ranks 39 of the respective volumes, and updates the rank 39 of the rank update target block number 37 corresponding to the volume name having the smallest total value to, for example, 1 (step). ST14). When there are a plurality of ranks 39 having the smallest total value of the ranks 39, the ranks 39 may be updated according to the order of the volume information 21 on the volume management table 20. In this case, the volume # V2 is selected in accordance with the order of the volume information 21 from # V2 to # V4 having the smallest total value of the rank 39, and the rank 39 of the file block information 35 of the block number B1 stored in the volume # V2 is changed. Update to 1.
[0051]
Subsequently, the file block management unit 14 similarly updates the rank 39 for the file block information 35 of the block number B2. Since the rank 39 of the block number B1 of the volume # V2 has been updated to 1, the total rank value of the volume # V2 is 1, and the total rank value of the volumes # V1, # V3, and # V4 is 0.
Accordingly, the volume # V3 is selected from the volumes # V1, # V3, and # V4 having the smallest rank total value in accordance with the order of the volume information 21, and the rank 39 of the block number B2 of the volume # V3 is updated to 1. Steps ST12 to ST14 are similarly repeated for the file blocks of block numbers B3 to B8 (step ST15).
[0052]
By determining the read priority according to the rank 39 updated by the above procedure, file blocks can be read from different volumes in parallel, and the overall data transfer speed and reliability are improved.
[0053]
FIG. 11B is a diagram showing, by ellipses, file blocks that are read when three read requests for the block number B3 are made. As in the case shown in FIG. 11A, the file block management unit 14 may update the rank 39 of each file block and determine the read priority of the file block according to the rank 39.
When the ranks 39 are the same, the priority may be determined according to the order of the volume information 21 on the volume management table 20. By selecting the read file block according to the priority, when a read request for the same file block is issued from the file control unit 12, the file blocks of the same block number in different volumes can be read simultaneously. As a result, the data transfer waiting time can be eliminated and the throughput can be improved.
[0054]
FIG. 11C is a diagram showing an ellipse of a read file block in a state where the file block is overwritten and the current multiplicity is 2, and FIG. 11D is a diagram in which the file block is overwritten and all the file blocks are overwritten. It is a figure which shows the read file block in the state in which the multiplicity was set to 2 by an ellipse. 11A, the file block management unit 14 may update the rank 39 of each file block and determine the read priority of the file block according to the rank 39, as in the case shown in FIG. When the ranks 39 are the same, the priority may be determined according to the order of the volume information 21 on the volume management table 20.
[0055]
As described above, according to the first embodiment, the file block management unit 14 determines whether or not a volume equal to the current multiplicity 33 has an empty area. Since the file blocks are duplicated and stored so that the weight becomes equal to the current multiplicity of 33, the file blocks can be stored so that the multiplicity of all the file blocks is maximized according to the number of volumes. The effect of being able to flexibly determine the number of disks and maximizing the reliability of the volume is obtained.
[0056]
When a file block is stored in a free area and there is a block to be written when stored in the entire area, the file block management unit 14 reduces the current multiplicity 33 to the area where data is already stored. Since the data is overwritten, it is possible to avoid that data cannot be stored due to lack of free space in the volume, and the effect of changing the capacity of the volume according to the data size can be obtained.
[0057]
In addition, the file block management means 14 issues a warning to the user when the current multiplicity 33 is changed. Therefore, if necessary, the volume is increased by the user, so that the reliability of the required multiplicity can be improved. The effect that the property can be maintained is obtained.
[0058]
Further, the file block management means 14 sets an error when the current multiplicity 33 is changed, so that it is possible to prevent data from being stored at a multiplicity lower than the minimum required multiplicity as a system. Thus, an effect that the reliability of the minimum multiplicity can be guaranteed can be obtained.
[0059]
In reading data from each volume, the file block management unit 14 refers to the rank 39 of the file block management table 30 so that the file block having the highest priority is read, so that the file blocks are read in parallel from different volumes. And the effect of maximizing the overall data transfer efficiency can be obtained.
[0060]
Embodiment 2 FIG.
In the second embodiment of the present invention, a file block read operation when a certain volume cannot be accessed due to a disk failure or a network failure will be described. In addition, a description is given of file block duplication for recovery of an inaccessible volume or addition of a new volume.
[0061]
FIG. 12A shows this embodiment. In a case where the volume # V2 fails and becomes inaccessible in a state where the file blocks of the block numbers B1 to B10 are written, data is read from volumes other than the volume # V2. It is a figure which shows a file block with an ellipse. When the operation system 13 detects an error of the device driver or the designation of the user, the volume management unit 11 sets the active flag 26 of the volume information 21 and can determine whether or not the volume is inaccessible by referring to this. Since the file blocks stored in the inaccessible volume cannot be accessed, the file block management unit 14 selects the read file block again.
[0062]
The rank updating process for selecting a read file block is performed according to the procedure shown in the flowchart of FIG. 7, and the procedure for determining the rank of FIG. 12A according to FIG. 7 will be described. The file block management unit 14 initializes all ranks 39 of the file block information 35 and acquires the file block information 35 of the block number B1. Subsequently, the file block management unit 14 calculates the total value of the rank 39 of the file block information 35 having the same volume name as the volume names # V3 and # V4 of the obtained file block information 35, and both of them become 0. Therefore, the file block management unit 14 updates the rank 39 of the file block information 35 of the block number B1 stored in the volume # V3 to 1.
[0063]
Subsequently, the file block management unit 14 acquires the file block information 35 of the block number B2. When calculating the total value of the rank 39 of the file block information 35 having the same volume name as the volume names # V1, # V3, and # V4 of the acquired file block information 35, the file block management unit 14 calculates 0, 1, 0, respectively. It becomes. Therefore, the file block management unit 14 selects the volume # V1 according to the volume information 21 on the volume management table 20 with the smallest total rank value, and selects the block number B2 of the volume # V1 as a read block. Similarly, the rank 39 is updated for the file blocks of the block numbers B3 to B10.
[0064]
FIG. 12B is a diagram showing a read file block by an ellipse when the volume # V2 fails while the file block numbers B1 to B16 are written. In the figure, the read file block is selected according to the rank 39 on the file block management table 30, and the rank 39 is updated by the file block management unit 14 according to the procedure shown in FIG. 7, as in the case of FIG.
[0065]
Next, an operation of storing data in the added volume will be described.
FIG. 13 is a flowchart showing a procedure for storing data in the added volume. FIG. 14 is a diagram showing a case where the volume # V5 is stored in the volumes # V1 to # V4 in which data is stored in a state where the last block number is B12 and the current multiplicity is 2. FIG. 14 is a diagram illustrating an example in which a file block is added and stored in volume # V5 according to the procedure of FIG. 13; A procedure for storing a file block from a state where all the volumes # V5 are free areas will be described with reference to FIGS.
[0066]
The volume management unit 11 adds the volume information 21 of the added volume # V5 to the volume management table 20 (Step ST27). Subsequently, the file block management unit 14 acquires the file block information 35 from the file block management table 30 (step ST28), counts the number of the file block information 35 having the same block number (step ST29), and obtains the block numbers B1 and B2. , B11, and B12, the number of file block information 35 is 2, and the number of block numbers B3 to B10 of the file block information 35 is 3.
[0067]
Subsequently, the file block management unit 14 determines whether or not the number of the file block information 35 is larger than the current multiplicity 33 in descending order from the block number B12 (step ST30). Blocks in which the number of file block information 35 is equal to or less than the current multiplicity of 33 are stored in # V5. Since the number of the file block information 35 of the block number B12 is 2 and equal to or less than the current multiplicity, the file block management unit 14 requests the operating system 13 to copy and write the data of the block number B12. The operating system 13 stores the data of the block number B12 in the volume # V5 according to the request of the file block management unit 14 (Step ST31).
[0068]
Subsequently, the file block management unit 14 adds the file block information 35 of the block number B12 to the file block management table 30 (Step ST32). The file block management unit 14 performs the processing of steps ST28 to ST32 for all block numbers 37 (step ST33). If the number of the file block information 35 counted in step ST30 is larger than the current multiplicity 33, the processes of steps ST28 to ST32 are performed for the next block number. The file block management unit 14 may perform the processing of steps ST28 to ST32 in ascending order of the block numbers or in descending order. If performed in descending order, a copy of the new file block will be stored preferentially.
[0069]
If a copy of the file block is stored in the added volume and the free space remains after all the file blocks have the same multiplicity, the file block management unit 14 increases the current multiplicity by one and copies the file block. May be created and stored in a copy of a file block that does not store the. Alternatively, a new file block to be added in the future may be stored in the empty area. According to the above procedure, the multiplicity of all the file blocks can be made the same and maximum.
[0070]
FIG. 14 shows that, after the file blocks of the block numbers B12, B11, B2, and B1 are sequentially stored in the volume # V5 and the multiplicity of all the file blocks becomes 3, the block having the larger block number, that is, the block number B10, This shows a state in which blocks B9, B8, and B7 are stored. FIG. 14 shows, by an ellipse, a read file block that is read when a read request for a file block of block numbers B1 to B12 occurs. The read file block may be selected according to the rank 39 of the volume information 21 on the volume management table 20. When the ranks 39 are the same, the priority may be determined according to the order of the volume information 21 on the volume management table 20. The rank update processing may be performed at the time of reading a file block, or may be performed after file block replication storage at the time of adding a volume.
[0071]
When the failed volume is recovered, the file block is duplicated and stored in the same procedure as the volume addition procedure.
[0072]
As described above, according to the second embodiment, when a certain volume becomes inaccessible due to a physical disk failure, a network failure, or the like, the file block management unit 14 deletes the file stored in the accessible volume. Since the copy of the block is read, an effect of preventing a read error can be obtained.
[0073]
In addition, the file block management unit 14 refers to the rank 39 of the file block management table 30 to read the file block stored in the volume with the highest priority. , Data can be read out in parallel, and the effect of maximizing the overall data transfer efficiency can be obtained.
[0074]
In addition, the file block management unit stores a copy of a file block having a multiplicity smaller than the current multiplicity in the newly added volume, so that it can flexibly cope with a case where a new volume is added. The effect is obtained.
[0075]
【The invention's effect】
As described above, according to the present invention, the volume management table storing the first identifier for identifying each volume constituted by the storage device and the capacity information of each volume, and the volume management table stored in all the volumes A file block management table storing a current multiplicity, which is the minimum number of file block copies, a second identifier for identifying a volume in which each file block is stored, and a reading priority of each file block; File block management means for controlling writing of file blocks to each volume based on the first identifier, capacity information, current multiplicity, and second identifier, and controlling reading of file blocks from each volume in accordance with priority; Configuration, so that all files can be Since the multiplicity of yl block can store so as to maximize the number of disks connected with flexibility in determining, also it has the effect that can maximize the reliability of the volume.
[0076]
According to the present invention, with the above configuration, data can be read out in parallel from different volumes, and the effect of maximizing the overall data transfer efficiency can be obtained.
[Brief description of the drawings]
FIG. 1 is a diagram showing a configuration of a system including a data storage device according to a first embodiment of the present invention.
FIG. 2 is a diagram showing a mutual relationship between components of the data storage device according to the first embodiment.
FIG. 3 is a diagram showing an internal configuration of a volume management table in FIG. 2;
FIG. 4 is a diagram showing an internal configuration of a file block management table in FIG. 2;
FIG. 5 is a flowchart showing a file block storage procedure according to the first embodiment.
FIG. 6 is a storage diagram of file blocks according to the first embodiment.
FIG. 7 is a flowchart showing a rank update procedure for selecting a read file block according to the first embodiment.
FIG. 8 is a diagram showing reading of a plurality of file blocks according to the first embodiment.
FIG. 9 is a flowchart showing a file block storage procedure including an overwriting process according to the first embodiment.
FIG. 10 is an overwrite storage diagram of a file block according to the first embodiment.
FIG. 11 is a diagram showing reading of a plurality of blocks in a state where a file block according to the first embodiment is overwritten.
FIG. 12 is a diagram showing a plurality of blocks read when a certain volume becomes inaccessible according to the second embodiment of the present invention;
FIG. 13 is a flowchart showing a copy block storage procedure of a file block when a volume is added in the second embodiment.
FIG. 14 is a diagram showing a copy of a file block when a volume is added in the second embodiment.
FIG. 15 is a storage diagram of a file block applied in the mirroring system.
FIG. 16 is a storage diagram of a file block applied in the mirroring system.
FIG. 17 is a storage diagram of a file block applied in the mirroring system.
[Explanation of symbols]
1 network, 2, 2a, 2b bus, 3a, 3b network coupling device, 3c bus coupling device, 4a, 4b disk controller, 5a I / O processor, 6a, 7a, 6b, 7b magnetic disk device, 8, 8a, 8b Volume, 10 main storage device, 11 volume management unit, 12 file control unit, 13 operation system, 14 file block management unit, 15 central processing unit, 16 device driver, 20 volume management table, 21 volume information, 22 volume name, 23 Type, 24 maximum block number, 25 current block number, 26 active flag, 30 file block management table, 31 file block management information, 32 maximum multiplicity, 33 current multiplicity, 34 last block number, 35 file block information, 36 volume Name, 37 block number, 38 time stamp, 39 rank, # D1, # D2, # D3, # D4 physical disk, # V1, # V2, # V3, # V4, # V5 volume, B1-B12 file block .

Claims (10)

In a data storage device that divides data into file blocks and stores the data in a storage device connected via a transfer path,
A volume management table for storing a first identifier for identifying each volume composed of the storage device and capacity information of each volume;
The current multiplicity, which is the minimum value of the number of copies of the file blocks stored in the volume, the second identifier for identifying the volume in which each of the file blocks is to be stored, and the read priority of each of the file blocks are stored. A file block management table,
Based on the first identifier, the capacity information, the current multiplicity, and the second identifier, write control of file blocks to each volume is performed, and read control of file blocks from each volume is controlled according to the priority. A data storage device comprising: The file block management means determines whether or not free space exists in the number of volumes equal to the current multiplicity, and if so, copies the file block so that the multiplicity of the file block becomes equal to the current multiplicity. The data storage device according to claim 1, wherein the data is stored. The file block management means overwrites a file block to be stored in an area storing a file block having a higher multiplicity than the current multiplicity when all volumes have no free area. Data storage device as described. 2. The data storage device according to claim 1, wherein the file block management means refers to the priority of the file block management table when reading data from each volume, and reads the data from the file block having the highest priority. When one or more volumes become inaccessible, the file block management means deletes the file from the accessible volume in which the file block having the same number as the file block stored in the inaccessible volume is stored. 2. The data storage device according to claim 1, wherein a copy of the file block is read. 2. The data storage device according to claim 1, wherein the file block management means stores a file block having a lower multiplicity than the current multiplicity in the newly added volume. 3. The data storage device according to claim 2, wherein the file block management unit issues a warning to the user when the current multiplicity is changed. 3. The data storage device according to claim 2, wherein the file block management unit generates an error when the current multiplicity is changed. An apparatus for dividing data into file blocks and storing the data in a storage device connected via a transfer path,
A volume management table for storing a first identifier for identifying each volume composed of the storage device and capacity information of each volume;
The current multiplicity, which is the minimum value of the number of copies of the file blocks stored in the volume, the second identifier for identifying the volume in which each of the file blocks is to be stored, and the read priority of each of the file blocks are stored. In a data storage method of a data storage device having a file block management table,
Based on the first identifier, the capacity information, the current multiplicity, and the second identifier, write control of a file block to each volume is performed, and read control of a file block from each volume is performed according to the priority. A data storage method characterized by performing. When data is divided into file blocks and stored in a storage device connected via a transfer path,
A volume management table for storing a first identifier for identifying each volume composed of a storage device and capacity information of each volume;
The current multiplicity, which is the minimum value of the number of copies of the file blocks stored in the volume, the second identifier for identifying the volume in which each of the file blocks is to be stored, and the read priority of each of the file blocks are stored. File block management table,
Based on the first identifier, the capacity information, the current multiplicity, and the second identifier, write control of file blocks to each volume is performed, and read control of file blocks from each volume is controlled according to the priority. A program that causes a computer to function as a file block management unit.

Images