JP2000076022A - File storage device and recording medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a file storage area for managing empty areas with plural disk devices as one logical disk device. SOLUTION: This device is provided with an entry managing table 3 for managing continuous empty areas with one logical disk device as a prescribed managing unit, a bit managing table 2 for managing empty areas with one logical disk device as a prescribed managing unit and a means for finding one or plural empty areas capable of storing a designated size by retrieving the entry managing table 3 when the size designated at the time of write request is larger than plural managing units or for finding any empty area capable of storing the designated size by retrieving the bit managing table 2 when the designated size is smaller than prescribed one or plural managing units.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a file storage device and a recording medium for managing an empty area by using a plurality of disk devices as one logical disk device.

[0002]

2. Description of the Related Art Conventionally, there are a disk array device and a RAID device as a technique for physically treating a plurality of disk devices as one logically and as a disk device having a larger capacity. In these devices, when managing the empty area, generally, there are many parts depending on the operating system on which the device operates,
In many cases, a management unit (for example, a block or a cluster) is used / empty state is managed using a table.
When newly arranging data, the table is searched to find an empty management unit, and the data is stored in that area.

[0003]

However, for efficient file access in a disk device, it is necessary to store data in a continuous area with respect to the disk device. This is because when data is distributed to disk devices,
It is necessary for the head of the disk device to seek a position of a different track for each block, and there is a problem that a time dependent on the seek position is required and quick access is not possible. For this reason, even when a plurality of disk devices are logically treated as one disk device, it is an important issue how to find a continuous empty area at high speed.

[0004] The present invention solves these problems,
When treating a plurality of disk devices as one logical disk device, an entry management table for managing continuous free space in a predetermined management unit and a bit management table for managing a small area smaller than the predetermined management unit are provided. When the write size is larger than one or a plurality of predetermined management units, a continuous empty area is searched for and found from the entry management table, and when the write size is smaller, an empty area is searched for and found from the bit management table, and is optimal for the access size. It is an object of the present invention to realize highly efficient disk access and to store information in a parent disk device in units of blocks and store 0 or 1 in a child disk device to improve the reliability in a mirror configuration.

[0005]

Means for solving the problem will be described with reference to FIG. In FIG. 1, a bit management table 2 manages an empty area in one or a plurality of predetermined management units, and manages, for example, a bit unit as illustrated.

The entry management table 3 manages continuous empty areas in predetermined management units, for example, manages continuous areas in block units. The data section 4 is an area for storing data in one logical disk device.

The control means 5 searches the entry management table 3 to find an empty area when the write request is equal to or more than a predetermined one or a plurality of management units. The management table 2 is searched to find an empty area.

Next, the operation will be described. When the size specified at the time of the write request is equal to or larger than one or more predetermined management units, the control means 5 searches the entry management table 3 to find one or a plurality of empty areas where the specified size can be stored, or The bit management table 2 is searched when the number is equal to or smaller than a predetermined one or a plurality of management units to find an empty area in which the designated size can be stored, and the found empty area is allocated.

At this time, the same area of a plurality of disk devices is collectively used as a management unit. In addition, a plurality of collected management units are collectively managed as a continuous empty area.

When allocating empty areas from the entry management table 3, consecutive empty areas are allocated in order from the largest area. Further, a parity disk device is provided for each of the collected management units, and when a failure occurs, data in the failed disk device is restored.

Also, a parent disk device, one or more child disk devices, and a parity disk device are combined into one logical disk device as a plurality of disk devices, and 0 or 0 is set as an arbitrary management unit of the child disk device. 1 is written so that the parent disk device and the parity disk device are mirrored for an arbitrary management unit.

Therefore, when a plurality of disk devices are treated as one logical disk device, an entry management table 3 for managing a continuous empty area in a predetermined management unit, and an entry management table 3 for one or more predetermined management units or less. A bit management table 2 for managing a small empty area is provided. When the light size is larger than one or a plurality of predetermined management units, a continuous empty area is searched for and found from the entry management table. By searching for and finding an empty area, it is possible to realize efficient disk access that is optimal for the access size, store information in the parent disk device in a predetermined management unit, and write 0 or 1 in the child disk device. Is stored in a mirror configuration to improve reliability.

[0013]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, an embodiment and operation of the present invention will be sequentially described in detail with reference to FIGS.

FIG. 1 shows a system configuration diagram of the present invention.
In FIG. 1, an area management index 1 is an index for managing an area. As shown in the figure, the data section 4, the bit management table 2, and the head addresses of the plurality of entry management tables 2 and information on their sizes are respectively stored. They are set up and accessed quickly.

The bit management table 2 manages an empty area in a predetermined management unit. Here, the bit management table 2 manages an empty area in bit units as shown in the figure. The entry management table 3 manages continuous empty areas in predetermined management units, and manages empty areas in continuous block units. A plurality of entry management tables are provided as shown in the figure.

The data section 4 is an area for storing data, and stores an area for storing data (contents) as shown in the figure and an empty area for the present invention as shown in the figure.
It is managed by the bit management table 2 and the plurality of entry management tables 3 according to the head address and the size.

The control means 5 searches the entry management table 3 to find an empty area when the write request is equal to or more than a predetermined one or a plurality of management units. The overall control, such as finding an empty area by searching the management table 2, is performed (to be described later with reference to FIGS. 2, 3, and 5).

Next, the operation of acquiring a continuous empty area at the time of a write request under the configuration of FIG. 1 will be described in detail according to the sequence shown in the flowchart of FIG. FIG. 2 is a flowchart illustrating the operation of the present invention (empty area acquisition).

In FIG. 2, S1 issues a region acquisition command by designating a necessary region (the number of bytes) and an address for storing allocated block information.
This is an area in which the number of bytes required for writing by the access request source (here, the write request source) and the address of the area for storing the block information (the start address of the acquired block and information on its size) are specified by parameters. An acquisition command is issued to the control means 5 of FIG.

In step S2, the number of blocks is obtained from the number of bytes (rounded up). In this case, the number of blocks to be continuously obtained is obtained from the number of byte sizes of the area to which the data notified in S1 is written. For example, in a general disk device, the size of a logical sector is 512 bytes, and a plurality of (for example, 1024) logical sectors are combined into a record, and records existing at the same physical position of each disk device are combined into a stripe. Further, a plurality of stripes are collectively formed as a block. This block (for example, 512 bytes × 102
One or more of 4 × 5 = 2.62 Mbytes) is a predetermined management unit of the empty area.

In step S3, it is determined whether the number of blocks is 33 or more. This is because the number of blocks required for writing obtained in S2 is 33
Determine if it is more than a block. In the case of YES, the processes from S4 to S8 are performed. On the other hand, in the case of NO, it is determined that the number of blocks is equal to or less than 33 blocks. Therefore, it is determined in S9 whether the number is equal to or less than the total number of blocks of the area managed by bits, and in the case of NO, the process proceeds to S4. The head LBA (head logical block address) of the number of blocks obtained is obtained, and the obtained area is allocated.

In S4, it is determined that the number of blocks is 33 or more or less than 33 blocks in YES in S3, but is equal to or more than the total number of blocks in the area managed by bits in YES in S9. The head LBA of the largest continuous free area is allocated.

In step S5, the latest maximum continuous free area is searched (the entry management table 3 is searched). In step S6, the entry management area is updated. This updates the information of the head LBA of the largest continuous empty area found by searching the entry management table 3 in S5 and its size.

A step S7 decides whether or not the capacity is satisfied. This determines whether the number of blocks for storing the number of bytes requested to be written is satisfied. In the case of YES, the first LBA (there may be a plurality) allocated in S8 is returned (S
The information of the obtained head LBA and size is set in the area for storing the block information notified in step 1 and the fact is returned to the write request source by an interrupt), and the processing is terminated. On the other hand, in the case of NO in S7, since the capacity is not satisfied, the flow returns to S4 and the allocation of the next largest empty area is repeated.

As described above, the number of blocks is obtained based on the number of bytes required at the time of a write request. When the obtained number of blocks is equal to or larger than one or a plurality of predetermined management units (here, 33 blocks), an entry management table is obtained. 3, the required number of blocks is allocated in order from the largest continuous empty area. On the other hand, when the obtained number of blocks is equal to or smaller than one or a plurality of management units (here, 33 blocks), the capacity of the bit management table 2 is determined. In the following cases (in the above case, the entry management table 3 is searched for and allocated), the bit management table 2 is searched and an empty area is allocated.
It is possible to allocate a continuous empty area that is optimal for the number of bytes at the time of a write request, and realize efficient high-speed access by making access to the disk device a continuous area.

Next, the operation of releasing the acquired area under the configuration of FIG. 1 according to the order shown in the flowchart of FIG. 3 will be described in detail. FIG. 3 is a flow chart for explaining the operation of the present invention (area release).

In FIG. 3, S11 is a content ID
Issuing the delete command by specifying. In this case, the access request source issues a delete command specifying the content ID whose area is to be released to the control means 5 of FIG.

In step S12, a list of blocks to be deleted is created. In S13, the bit management table 2 is updated when the number of blocks is 32 or less, and otherwise, the entry management table 3 is updated. This means that the bit management table 2 is updated (the block to be deleted is added to the bit management table 2) if the number of blocks to be deleted created in S12 is 32 or less, and the entry management table 3 is updated otherwise. Update (add the block to be deleted to the entry management table 3).

At S14, in the bit management table 2,
If there are 33 or more continuous blocks, the continuous blocks are moved to the entry management table.
S15 searches for continuity of the entry management table 3,
If there are consecutive entries, they are combined into one block.

At S16, the entry management table 3
An LBA (head logical block address) indicating the largest empty block is searched for and stored. As described above, the content I
In response to the issuance of the delete command designating D, when the area of the content ID to be deleted is 32 blocks or less, it is added to the bit management table 2. Moves it to the entry management table 3, on the other hand, adds it to the entry management table 3 when the area of the content ID to be deleted is 33 blocks or more, re-searches for the largest empty block after addition, and stores it. As a result, preparation for allocating a continuous maximum empty area from one logical disk apparatus in which a plurality of disk apparatuses are integrated when the area is released is completed.

FIG. 4 shows an example of a directory table according to the present invention. This directory table, as shown,
This is a table for registering and managing a set necessary for storing content as a set of an LBA number (logical block number, block number) of an empty area acquired in association with a content ID and its size. Thus, when the content to be written does not fit in one continuous empty area, the content is repeatedly allocated to the next largest continuous empty area by repeatedly storing the content in the next largest continuous empty area. It is possible to realize continuous access as much as possible.

FIG. 5 is a flowchart illustrating another operation of the present invention. As shown in FIG. 6, which will be described later, this is one logical disk device including a physical parent disk device, one or more child disk devices, and a parity disk device. When the information of the directory table of FIG. 6 is stored, all 0s or 1s are written to one or more child disk devices and the parity data is written to the parity disk device, when a read error occurs in the parent disk device,
This is an explanation of the operation when the contents of the parity disk device can be restored simply by copying the contents of the parity disk device to a newly replaced parent disk device as it is, so to speak, the mirror configuration.

In FIG. 5, in S21, a read error of the parent disk device occurs. In step S22, a restoration process is performed. In this restoration process, the contents of the parity disk device are restored by copying the contents of the parity disk device directly to the replaced parent disk device. This is because the directory management information is stored only in the parent disk device, 0 or 1 is stored in one or more other child disk devices, and the parity bit is stored in the parity disk device. If the parity is set to an odd number (when 0 is written to the child disk device) or the parity is set to be an odd number (when 1 is written to the child disk device), the data of the parent disk device and the data of the parity disk device become the same. It becomes the same as a mirror, and when a read error occurs in the parent disk device, it is possible to automatically repair the data only by replacing the parent disk device and copying the data of the parity disk device.

FIG. 6 shows another explanatory diagram of the present invention. When the parent disk device, one or more child disk devices, and the parity disk device in the figure are one logical disk device (a disk array device or a RAID device), information is stored in the parent disk device. For the blocks in the figure in which 0 (NULL) is written to the child disk device and parity data is written to the parity disk device, the same data is stored in the parent disk device and the parity disk device and becomes a mirror. As described above, when a read error occurs in the parent disk device, data is automatically read only by reading data from the parity disk device, or by replacing the parent disk device and copying the data of the parity disk device to a new parent disk device. It becomes possible.

As described above, when the parent disk device, one or more child disk devices, and the parity disk device are one logical disk device, all 0 (or 1) is written to the child disk device. In the case of a block, the data in the parent disk device and the data in the parity disk device are the same and become a mirror, so that even if a read error occurs in one of them, data can be read from another disk device and the read error After replacing the disk device in which the error occurred, the data can be automatically restored only by copying data from another parent disk device or parity disk device, and the reliability can be improved in block units.

[0036]

As described above, according to the present invention,
When treating a plurality of disk devices as one logical disk device, an entry management table 3 for managing an area in a predetermined management unit and a bit management table 2 for managing a small area smaller than the predetermined management unit are provided. When the write size is larger than one or a plurality of predetermined management units, a continuous empty area is searched and found and assigned from the entry management table 3, and when it is smaller, an empty area is searched and found and assigned from the bit management table 2. Since the configuration employs a plurality of disk devices as one logical disk device, it is possible to realize efficient disk access that is optimal in access size and to store information in the parent disk device in block units. Since a configuration in which 0 or 1 is stored in the child disk device is adopted, a mirror configuration is used for each block. To thereby improve the reliability.

[Brief description of the drawings]

FIG. 1 is a system configuration diagram of the present invention.

FIG. 2 is a flowchart for explaining the operation of the present invention (empty area acquisition).

FIG. 3 is a flowchart for explaining the operation of the present invention (area opening).
It is.

FIG. 4 is an example of a directory table according to the present invention.

FIG. 5 is a flowchart illustrating another operation of the present invention.

FIG. 6 is another explanatory diagram of the present invention.

[Explanation of symbols]

 1: area management index 2: bit management table 3: entry management table 4: data section 5: control means

 ──────────────────────────────────────────────────の Continued on the front page F term (reference) 5B065 BA01 CA30 CC03 CH18 5B082 CA02 CA08

Claims (7)

[Claims] 1. A file storage device for managing an empty area by using a plurality of disk devices as one logical disk device, wherein an entry for managing an empty region in which said one logical disk device is continuous in a predetermined management unit. A management table; a bit management table for managing the free area of the one logical disk device within a predetermined management unit; and a bit management table for controlling the size specified at the time of a write request to one or more predetermined management units. The entry management table is searched to find one or more empty areas where the specified size can be stored, or the bit management table is searched when the size is equal to or smaller than the predetermined one or more management units, and the specified size is searched. Means for finding an empty area in which a file can be stored. 2. The file storage device according to claim 1, wherein the same area of the plurality of disk devices is collectively used as the management unit. 3. A file storage device, wherein a plurality of management units according to claim 2 are further combined to manage a continuous empty area. 4. The file storage device according to claim 1, wherein, when allocating empty areas from the entry management table, consecutive empty areas are allocated in order from the largest area. . 5. The file storage device according to claim 1, wherein a disk device for parity is provided to restore data in the failed disk device when a failure occurs. 6. A plurality of disk devices, a parent disk device, one or a plurality of child disk devices, and a parity disk device are combined into one logical disk device. 6. The file storage device according to claim 1, wherein all 0s or 1s are written and the parent disk device and the parity disk device are mirrored for the arbitrary management unit. 7. An entry management table for managing a continuous empty area of one logical disk device in a predetermined management unit, and a bit for managing an empty area of one logical disk device in a predetermined management unit. When the size specified at the time of the write request is one or more predetermined management units or more, the entry management table is searched to find one or a plurality of empty areas in which the specified size can be stored, or A computer-readable recording medium in which a program for functioning as a means for searching the bit management table and finding an empty area where a designated size can be stored when the number is equal to or smaller than the predetermined one or a plurality of management units is recorded.