JP3480282B2 - Duplicated file I/O device, method thereof, and information processing device using the same - Google Patents

Description

Detailed Description of the Invention

[0001]

The present invention relates to a duplex file I/O.
The present invention relates to an O device, a method thereof, and an information processing device using the same.

[0002]

2. Description of the Related Art Conventionally, in this type of duplicated file I/O device, write operations are performed on both the primary file and the secondary file, and read operations are performed only from the primary file; this is a system adopted to improve the reliability of files.

Japanese Patent Publication No. 56-129964 describes a technique for sharing the direct memory access channel and buffer register corresponding to each file in order to improve the efficiency of the process of writing data to two files, a main file and a sub file, in such a dual file system.

[0004]

The first problem with this prior art is that when reading a file, generally the primary file is always read, so if the I/O load on the device where the primary file is located is high, the processing efficiency drops.
This is because the /O load is not taken into consideration.

The second problem is that when writing a file, the processing time of the program increases, because the execution of the program is stopped until writing of both the primary and secondary files is completed.

An object of the present invention is to provide a duplicated file I/O device, method and information processing device which provide fault tolerance by duplicating files while suppressing a decrease in the processing efficiency of user programs.

[0007]

The duplicated file of the present invention
The I/O device has a dual file system consisting of a primary file and a secondary file.
The original file and device information indicating the device on which the original file is located
information, device information indicating the device on which said subfile resides, and finally
The location information indicating the location of the written block and the last
The license to either the main file or the sub file
A file that remembers file information indicating which
and load information indicating the I/O load of the device.
The I/O load table to be stored and the I/O request if it is a read
In this case, the position information in the file status table is referenced.
The I/O request is a request to a block indicated by the location information.
At some time, the file information in the file status table is
Refer to the original file indicated by the file information or
Issue a read to the subfile and the I/O request
If the request is not for the block indicated by the location information,
The load information and the file status of the I/O load table
The main file device information and the sub file device information in the state table
Refer to the information and check that the positive file is in a device with low I/O load.
issue a read to the file or to the subfile,
If the request is a write,
Load information and the original file device of the file status table
The I/O load is calculated by referring to the file device information and the secondary file device information.
The primary file or the secondary file in the device with the lowest
After issuing a write to the other subfile or
issues a write to the original file and
/O is completed for the main file or the sub file
Store information in the file status table
The position of the written block is stored in the file status
and means for storing the position information in a table.
Let us assume that.

The dual file I/O method of the present invention comprises:
A computer receives an I/O request, and the I/O request is
If the file is a duplicated file, the primary file is
Device information showing the device where the duplicated file is located,
Device information indicating the device on which the subfile resides,
The location information and the last block written to the
Either the main file or the sub file is
A file that stores file information indicating which file was completed first
The I/O request is
is a request to a block indicated by the location information,
Refer to the file information in the file status table,
The primary file or secondary file indicated by the file information
and the I/O request issues a read to the
When there is no request to the block shown, the I
The I/O load table stores load information indicating an I/O load.
Load information and the original file device of the file status table
The I/O load is calculated by referring to the file device information and the secondary file device information.
The primary file or the secondary file in the device with the lowest
If a read is issued to the I/O request and the I/O request is a write
The load information of the I/O load table and the file
The main file device information and the sub file device information in the file status table
Refer to the device information and select the positive one in the device with low I/O load.
After issuing a write to a file or its subfile,
The other secondary file or the primary file is
At the same time, the I/O for the primary frame that was completed first is issued.
The information of the file or the subfile is stored in the file status.
The block that was written and stored in the file information of the table
The location of the file is stored in the location information of the file status table.
The present invention is characterized by including a step of:

[0009]

[0010]

[0011]

[0012]

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

Referring to FIG. 1, which shows a block diagram of one embodiment of the present invention, the duplex file I/O of this embodiment is
The O device includes an I/O destination selection means 1 and a file status table 2.
, an I/O load table 3 , an I/O means 4 , and a duplicated file 5 .

The file status table 2 stores the devices on which the primary and secondary files of the duplicated file 5 are located, the location of the last written block, and whether writing to the primary or secondary file was completed first.

The I/O load table 3 stores information on the I/O load of each disk device configured in the computer system, such as the I/O busy rate. This information is updated as needed by the I/O means 4 and the OS components that process general disk I/O.

In response to a read request from a user program 6, the I/O destination selection means 1 checks whether the position of the block to be read is the same as the position stored in the file status table 2. If they are the same, the file to which data was previously written is selected as the first file. If the position of the block to be read is different from the position stored in the file status table 2, the I/O destination selection means 1 checks the I/O load table 3 and
A file in the device with the lightest load is selected. In response to a write request from the user program 6, the I/O destination selection means 1 first confirms that the previous write has been completed, then looks at the I/O load table 3 and selects a file in the device with the lightest load as the first file, and the other device as the second file.

If the request is a read request, the I/O means 4 issues a read command to the first file selected as described above and waits for completion. If the request is a write request, the I/O means 4 issues a read command to the first and second files selected as described above and waits for completion.
Then, the write is issued to the file, and the necessary information is registered again in the file status table 2 after waiting for the write of either file to be completed.

Next, the operation of this embodiment will be described with reference to FIG. 2, which shows the flow of operation of this embodiment, in addition to FIG.

From the user program 6 to the duplicated file 5
When an I/O is requested, the I/O destination selection means 1 judges the type of the requested I/O (step A1). In the case of a read, the position of the block to be read is checked in the file status table 2.
If the position is different from the previous write position stored in the I/O load table 3, the primary or secondary file in the device with the lower load is selected as the first file, and the other is selected as the second file (steps A2, A5 to
A7). In the case of a read, the second file is not used even if it is selected. If the position of the block to be read is the same as the position where it was written last time, the primary or secondary file in which the previous write was completed first is selected as the first file from the file status table 2 (steps A2 and A3).
If the type of I/O is a write, the I/O destination selection means 1 first waits until the previous write is completed (step A4), and then selects the first and second files from the I/O load table 3 according to the load, as described above (step A5).
5-A7).

In response to the request, the I/O means 4 issues a read or write command to the first file (step A8). In the case of a write command, it then issues a read command to the second file (steps A9 and A10).
The process waits for one I/O to be completed (step A11). In the case of a write, the file for which the I/O has been completed first is stored in the file status table 2 together with the write position (steps A12 and A13). Note that normally, the I/O load is low,
The write to the first file selected should be completed first.

Next, the operation of this embodiment will be described using a concrete example with reference to FIGS. 3 to 5 in addition to FIGS.

As shown in Fig. 3, for example, the file status table 2 stores the device where the primary file is located, the device where the secondary file is located, the last write position, and the file for which the last write was completed first. Also, as shown in Fig. 4, the I/O load table 3 stores the load of the device at any given time.

When the user program 6 is at position P2 (P2≠P
When an attempt is made to read the file from the file status table 2 (FIG. 3), the I/O destination selection means 1 learns that the file is in a different location from the location where it was last written (step A1,
Next, from I/O Load Table 3 (FIG. 4), it is found that the load of device D2 is lower than that of device D1, and the secondary file is selected as the first file and the primary file as the second file (steps A5 and A7).
O means 4 therefore reads from the subfile and waits for its completion (steps A8, A9, A11).

When the user program 6 attempts to write to the position P3, the I/O destination selection means 1 first waits for the previous write to be completed (step A4), and then checks the I/O load table 3.
From FIG. 4, select the secondary file as the first file and the primary file as the second file (step A
5, A7).

In the I/O means 4, a write is first issued to the subfile, and then a write is issued to the main file (steps A8 to A11). If the write to the subfile is completed first, the contents of the file status table 2 will be as shown in FIG.

[0026]

As described above, the effects of the present invention are as follows:
The present invention makes it possible to utilize fault tolerance by duplicating files while suppressing a decrease in the processing efficiency of user programs.

The reason for this is that read/write requests are issued preferentially to files in devices with low I/O loads, and
In the case of write, I/O for either the primary or secondary file
This is to enable execution of the user program when the process is completed.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of an embodiment of the present invention.

FIG. 2 is a flow chart showing an example of the operation in this embodiment.

FIG. 3 is a diagram showing a specific example of a file status table according to the embodiment.

FIG. 4 is a diagram showing a specific example of an I/O load table according to this embodiment.

5 is a diagram showing a specific example of a state changed from that shown in FIG. 3 in the file status table of this embodiment.

[Explanation of symbols]

1 I/O destination selection means 2 File status table 3 I/O load table 4 I/O means 5 Duplicated file 6 User program

─────────────────────────────────────────────────────────── Continued from the front page (58) Field surveyed (Int.Cl. ⁷ , DB name) G06F 12/00 G06F 12/16 G06F 3/06

Claims (2)

(57) [Claims] Claim 1: A duplicate file consisting of a main file and a sub file.
The file, Device information indicating the device where the primary file is located,
Device information showing the device on which the file is located, the block that was last written,
Location information indicating the location of the check and finally the original file
Or, which of the subfiles is written first?
a file status table for storing file information indicating whether An I/O load information indicating an I/O load of the device is stored.
Load table and If the I/O request is a read, the file status table
and the I/O request is
If the request is for the block indicated by the information,
The file information in the state table is referenced.
A read is made to the main file or the sub file indicated by
The I/O request is issued to the block indicated by the location information.
When the request is not to
The file device of the shipment information and the file status table
The I/O load is calculated by referring to the file device information and the file device information.
The primary file or the secondary file in the lower device
If a read is issued and the I/O request is a write,
The load information in the I/O load table and the file
The primary file device information and the secondary file device information in the file status table
The primary folder in the device with the lowest I/O load is selected by referring to the location information.
After issuing a write to the file or its subfile,
Write to the other secondary file or primary file.
and issues the positive file where the I/O was completed first.
The information of the file or the sub-file is stored in the file status table.
The file information is stored in the block
The position of the file is stored in the position information of the file status table.
and a duplex file I/O means.
Device. Claim 2: A computer receives an I/O request and
If the I/O request is a read request,
Device information indicating the device where the primary file is located,
The device on which the subfile, the other half of the file, is located.
The location of the last written block.
Location information and finally the main file or the sub file
The file that indicates which write to which file was completed first.
Refer to the location information in the file status table that stores the information
The I/O request is directed to a block indicated by the location information.
If it is a request, the file in the file status table
Emotion The file information indicates the correct file.
or issue a read to the subfile and
When the request is not a request for a block indicated by the location information
load information indicating the I/O load of the device is stored in
The load information in the I/O load table and the file status table
The primary file device information and the secondary file device information
Refer to the above primary file on the device with low I/O load.
or issue a read to the subfile, and
If it is a write, the load in the I/O load table
information and the primary file device information of the file status table
The I/O load is reduced by referring to the information on the secondary file device and the information on the secondary file device.
The original file or the secondary file in the device that is not connected to the
After issuing the file, the other subfile or
Write to the file and perform I/O first.
Information on the main file or the sub file for which
The file information is stored in the file status table.
The position of the block that was inserted is added to the front of the file status table.
The method further comprises the step of storing the position information in the
Dual file I/O method.