JPH07271647A - Duplex file system - Google Patents

Abstract

PURPOSE:To efficiently perform the processes for writing data to duplex files by asynchronously performing the process for writing data to the main file and the process for writing data to the sub-file. CONSTITUTION:A 1st data writing means 15 writes data to the master file 12 at a data writing process request made by an AP 11 and stores information on the data written to the sub-file 13 in a 1st repeating area 14. A 2nd data writing means 16 operates asynchronously with the 1st data writing means 15 to copy the data stored in the 1st repeating area 14 to a 2nd repeating area 17 and the initialize the 1st repeating area 14, and further read the data out of the 2nd repeating area 17 and writes them to the sub-file 13.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION The present invention uses the same information for different 2
The present invention relates to a dual file system in which data is written to a file in one external storage device and information is read from a file in one external storage device.

[0002]

2. Description of the Related Art In a conventional dual file system, a writing process is performed on both a primary file and a secondary file, and a reading process is performed only on the primary file, which is adopted to improve file reliability. System.

Japanese Unexamined Patent Publication No. 56-129964 discloses a dual file system in which two files, a primary file and a secondary file, are written in order to improve the efficiency of the data writing process. There is described a technique of sharing a direct memory access channel with a buffer register.

[0004]

In either of the above-mentioned conventional dual file system or the technique described in Japanese Patent Laid-Open No. 56-129964, the writing process of data to the primary file and the secondary file is performed on the primary file. Since the writing process to the secondary file is performed after the writing process is completed, the output completion wait time is longer than that of the writing process when the duplication is not performed. .

An object of the present invention is to provide a duplicated file system in which the data writing process to a duplicated file realizes the same performance as the data writing process to a non-duplicated file.

Another object of the present invention is to provide a dual file system for efficiently performing dual file data write processing by asynchronously performing data write processing to a primary file and data write processing to a secondary file. To provide.

Still another object of the present invention is to continue all the input / output processing without delay by executing all the input / output processing for the secondary file when a failure occurs in the primary file. The purpose is to provide a dual file system that can do this.

Still another object of the present invention is to save the contents of the relay area to an external storage device when a failure occurs in a computer system including a duplicated file system so that the data to be written is not lost. An object of the present invention is to provide a redundant file system that can restart the operation of the redundant file system after restarting the computer system.

[0009]

According to a first duplicated file system of the present invention, a file for storing data of an application program is constituted by a pair of two files, one of which is a primary file and the other of which is a file. Is a secondary file, the writing process is performed for both the primary file and the secondary file, and the reading process is performed for the primary file only. In a dual file system, on the main storage device that temporarily stores update information for the secondary file Of the first relay area and the second relay area, and the update information of the data to be written to the primary file in response to the data write processing request from the application program and to be written to the secondary file. A first data writing means for storing in the area;
Of the first relay area, the content of the first relay area is copied to the second relay area, the first relay area is initialized, and the second relay area is stored in the second relay area. And second data writing means for reading the written data and writing it in the sub file.

A second duplicated file system of the present invention is the same as the first duplicated file system according to the first duplicated file system.
And the second relay area, the buffer unit storing the write destination block number of the data in the sub-file and the write data, and the first data writing unit adds data to the first relay area. The end record pointer indicating the additional position when the data is read, the first record pointer indicating the position of the first record when the second data writing unit acquires data from the second relay area, and the number of unused buffers are stored. And a control unit composed of a free buffer number counter used for this purpose.

A third duplicated file system of the present invention is the same as the second duplicated file system according to the second duplicated file system.
Writing means for copying the contents of the first relay area to the second relay area and setting the first relay area as an exclusive control target during the process of initializing the first relay area. The writing of information from the first writing means to the first relay area is put on hold.

According to a fourth duplex file system of the present invention, in the first, second, and third duplex file systems, the first data writing means performs input / output processing when a failure occurs in the primary file. When it is determined that the second data writing means is not executed, all the input / output processing is directly performed on the sub file, and the primary file and the first relay area are processed. All input / output processing is not performed.

According to a fifth duplex file system of the present invention, in the first, second, third, and fourth duplex file systems, when a failure occurs in a computer system including the duplex file system, A relay area saving means for saving the contents of the first relay area and the second relay area to an external storage device, and the first relay area and the first backup area saved to this external storage device after restarting the computer system. The relay area loading means is configured to load the contents of the second relay area into the first relay area and the second relay area.

A sixth duplex file system of the present invention is the first, second, third and fourth duplex file system, wherein the first relay area and the second relay area are provided on an external storage device. First relay file and second
Is configured as a relay file.

According to a seventh duplex file system of the present invention, in the first, second, third, and fourth duplex file systems, data is written to a primary file and the primary file data is written to store the write destination block number. Step, a normal file data write status determination step to determine whether the data write to the primary file has completed normally, and a case where the data write to the primary file has been normally completed in this primary file data write status determination step , A free buffer existence determining step of determining whether or not a free buffer exists in the first relay area, and if a free buffer does not exist, again determining whether or not a free buffer exists after a lapse of a predetermined time. In the free buffer existence determination step, it is determined that there is a free buffer. In this case, the first relay that writes the block number in which the data was written to the primary file in the primary file data writing step to the write destination block number of the area pointed to by the end record pointer of the first relay area and the data body is written to the write data Area data writing step, first relay area end record pointer updating step for positioning the end record pointer of the first relay area to the next record, first relay area free space for subtracting 1 from the free buffer number counter of the first relay area When it is determined in the buffer number counter updating step and the primary file data writing status determining step that the data writing to the primary file has terminated abnormally, both the end record pointer and the head record pointer of the first relay area are the head of the buffer section. Determine whether or not However, if it does not point to the beginning of the buffer section, a first relay area initialization completion determination step of determining whether or not it points to the beginning of the buffer section again after a certain period of time, and this first relay area initialization A sub-file data writing step of writing data to a sub-file when it is determined that both the end record pointer and the top record pointer of the first relay area point to the head of the buffer section in the conversion completion determination step, Data writing means, and a first relay area exclusion step for performing exclusive control on the first relay area,
Relay area data copy step of copying the contents of the first relay area to the second relay area, first relay area initialization step of initializing the first relay area, and release of exclusive control of the first relay area First relay area exclusion releasing step,
In the second relay area data read completion determination step of determining whether or not the first record pointer and the end record pointer of the second relay area are equal, in the second relay area data read completion determination step, The second relay area data reading step of reading the write destination block number and the write data from the buffer pointed to by the first record pointer in the second relay area when it is determined that the first record pointer and the end record pointer are not equal, Second relay area end record pointer updating step of locating the end record pointer of the relay area of the next relay area to the next record, second relay area free buffer number counter updating step of adding 1 to the free buffer number counter of the second relay area, and Indicates the write block number in the subfile And a second data writing means including a sub-file data writing step to write the write data in the area.

An eighth redundant file system of the present invention is the seventh redundant file system, wherein when a failure occurs in a computer system including the redundant file system, the first relay area and the second redundant file system are used. A relay area saving means for saving the contents of the relay area in an external storage device, and a content of the first relay area and the second relay area saved in the external storage device after the computer system is restarted. One relay area and relay area loading means for loading the second relay area are included.

According to a ninth duplex file system of the present invention, in the first, second, third, and fourth duplex file systems, data is written to the primary file and the primary file data is written to store the write destination block number. Step, a normal file data write status determination step to determine whether the data write to the primary file has completed normally, and a case where the data write to the primary file has been normally completed in this primary file data write status determination step , A free buffer existence determining step of determining whether or not a free buffer exists in the first relay file, and if a free buffer does not exist, again determining whether or not a free buffer exists after a certain period of time, It is determined that there is a free buffer in the free buffer existence determination step. In this case, the block number at which the data is written to the primary file in the primary file data writing step is written to the writing destination block number of the area pointed to by the end record pointer of the first relay file, and the data body is written to the writing data. Relay file data writing step, first relay file end record pointer updating step for locating end record pointer of first relay file to next record, first relay file subtracting 1 from free buffer number counter of first relay file When it is determined in the free buffer number counter updating step and the normal file data writing status determining step that the data writing to the primary file has terminated abnormally, the end record pointer and the first record pointer of the first relay file are both buffered. The first relay file initialization complete. It is determined whether it is pointing to the beginning of the copy section, and if it is not pointing to the beginning of the buffer section When it is determined in the determination step and the first relay file initialization completion determination step that both the end record pointer and the start record pointer of the first relay file point to the beginning of the buffer section, the data is written to the sub file. First data writing means including a writing step for writing sub-file data, a first relay file exclusion step for performing exclusive control on the first relay file, and copying the contents of the first relay file to the second relay file Relay file data copy step, and a first relay file initialization step for initializing the first relay file. First release the exclusive control of the first relay file
Relay file exclusion release step, a second relay file data read completion determination step that determines whether the first record pointer and the end record pointer of the second relay file are equal, and this second relay file data read completion determination step If the first record pointer and the end record pointer of the second relay file are not equal to each other, the second relay that reads the write destination block number and the write data from the buffer pointed to by the first record pointer in the second relay file File data reading step, second relay file end record pointer updating step for locating end record pointer of second relay file to next record, second relay file free space for adding 1 to free buffer number counter of second relay file Bag And a second data writing means including § number counter update step, and the area indicated by the write block number of the secondary file and a sub file data writing step to write the write data.

[0018]

An embodiment of the present invention will be described in detail below with reference to the drawings.

Referring to FIG. 1, a duplicated file system according to a first embodiment of the present invention is an application program (hereinafter referred to as AP) 11 and a duplicated file, which is an external storage for storing user data. A regular file 12 in the device and a duplicated file, a regular file 1
When an error occurs in the second storage device, a sub-file 13 in the external storage device for backup to be used, a first relay area 14 in the main storage device that temporarily stores information of data to be written to the sub-file 13, and 2 relay area 17 and AP 11
In accordance with the data writing processing request of the first data writing program to the primary file 12 and the information of the data to be written to the sub file 13 is stored in the first relay area 14. Data writing means 15) and the first data writing means 15 operate asynchronously, copy the data stored in the first relay area 14 to the second relay area 17, and then A second data writing program (hereinafter referred to as a second data writing means) 16 which is a program for initializing the relay area 14, reading the data in the second relay area 17, and writing the data in the secondary file 13 ing.

Referring to FIG. 2, in the first relay area 14 and the second relay area 17, the data write destination block number 31 and the write data 3 in the sub file 13 are written.
2 shows a buffer unit 24 for storing 2, a terminating record pointer 21 indicating an addition position when data is added to the first relay area 14, and a position of a leading record when data is acquired from the second relay area 17. First record pointer 2
2, and a control unit 25 including a free buffer number counter 23 used to store the number of unused buffers.

Next, a duplicated file system according to the first embodiment of the present invention will be described with reference to FIGS.

The first data writing means 15 is activated in response to the data writing processing request from the AP 11.

The first data writing means 15 first writes the data to the primary file 12 and stores the block number of the writing destination of the data (step 31).

Next, the first data writing means 15 refers to the free buffer number counter 23 of the first relay area 14 and determines whether or not the free buffer number counter 23 is 0 (step 32). When the free buffer number counter 23 is 0 in step 32, the process returns to step 32 after a lapse of a certain time in order to wait until a free area is formed in the first relay area 14 (step 33).

When the free buffer number counter 23 is not 0 in step 32, that is, when there is a free area in the first relay area 14, the first data writing means 15 first writes in the first relay area 14. The data information to be written is written in the area indicated by the end record pointer 21 in the format of the write destination block number 31 and the write data 32 (step 34). At this time, the block number indicating the position where the data was written to the primary file in step 31 is written to the write destination block number 31, and the data body is written to the write data 32. Further, the end record pointer 21 in the first relay area 14 is positioned at the next record (step 35), and 1 is subtracted from the free buffer number counter 23 (step 36).

The second data writing means 16 first performs an exclusive control on the first relay area 14 (step 41),
Next, the contents of the first relay area 14 are copied to the second relay area 17 (step 42), and the first relay area 14 is further copied.
After initializing (step 43), the first relay area 14
The exclusive control for is canceled (step 44).

First relay area 14 in step 43
In the initialization of, the end record pointer 21 and the start record pointer 22 both point to the start of the buffer unit 24, and the free buffer number counter 23 indicates the number of buffers in the buffer unit.

Next, the second data writing means 16 judges whether or not the start record pointer 22 and the end record pointer 21 in the second relay area 17 are equal (step 45), and if they are equal, the processing is executed. finish.

If the start record pointer 22 and the end record pointer 21 in the second relay area 17 are different in step 45, the second data writing means 16 starts the top record pointer 22 in the second relay area 17.
The write destination block number 31 and the write data 32 are read from the buffer indicated by (step 46), the head record pointer 22 in the second relay area 17 is positioned at the next record (step 47), and the free buffer number counter 23 is incremented by 1. (Step 48).

Further, the second data writing means 16 is
The write data 32 is written to the area indicated by the write destination block number 31 in the sub-file 13 (step 49), and the process returns to step 45.

With the above, the processing of the duplicated file system according to the first embodiment of the present invention is completed.

The dual file system according to the first embodiment of the present invention includes a data writing process to a primary file,
By asynchronously performing the data writing process to the sub file, the data writing process of the duplicated file can be efficiently performed.

The duplicated file system according to the second embodiment of the present invention is the first data writing means 15 in the first, second and third duplicated file systems.
When it is determined that a failure occurs in the primary file 12 and the input / output processing cannot be executed, the second data writing means 1
After waiting for the completion of the processing of No. 6, all the input / output processing is directly performed on the sub-file 13, and the primary file 12 and the first
The program is configured so that input / output processing with respect to the relay area 14 is not performed.

Next, a duplicated file system according to a second embodiment of the present invention will be described with reference to FIGS.

After writing the data to the primary file 12 in step 31, the first data writing means 15 determines whether or not the writing process is normally completed (step 37). Step 32 if completed normally
Leading to.

When the first data writing means 15 determines in step 37 that the data writing process for the primary file 12 has terminated abnormally, the end record pointer 21 and the top record pointer in the first relay area 14 It is determined whether or not both 22 point to the head of the buffer section 24 (step 38), and if they do not point, step 38 is executed again for a certain period of time (step 39).

When the end record pointer 21 and the head record pointer 22 in the first relay area 14 both point to the head of the buffer section 24 in step 38, the first data writing means 15 sub-data. Write to file 13 (step 3A).

With the above, the processing of the duplicated file system according to the second embodiment of the present invention is completed.

In the duplicated file system according to the second embodiment of the present invention, when a failure occurs in the primary file, all input / output processing is performed on the secondary file, thereby delaying the input / output processing. It has the effect that it can be continued without being allowed to.

The duplex file system according to the third embodiment of the present invention is used in the first, second, third and fourth duplex file systems when a failure occurs in a computer system including the duplex file system. Is a program for saving the contents of the first relay area 14 and the second relay area 17 to the external storage device 1C, a relay area save program (hereinafter, referred to as a relay area save means) 18, and a restart of the computer system. Later, the first relay area 14 and the second relay area 17 saved in the external storage device 1C
Contents of the first relay area 14 and the second relay area 17
And a relay area loading program (hereinafter, referred to as a relay area loading means) 19 which is a program to be loaded into.

Next, a duplicated file system which is a third embodiment of the present invention will be described with reference to FIGS. 1, 2 and 5.

When a failure occurs in the computer system including the duplicated file system, the relay area saving means 18 is activated from the exception processing subsystem of the operating system.

The relay area saving means 18 saves the contents of the first relay area 14 and the second relay area 17 as they are in the same format to the external storage device 1C (step 51).

When the computer system including the duplicated file system is restarted, the relay area loading means 1 is loaded from the initialization processing subsystem of the operating system.
9 is activated.

The relay area loading means 19 stores the contents of the first relay area 14 and the second relay area 17 saved in the external storage device 1C by the relay area saving means 18 in the first relay area 14 and the second relay area. The area 17 is loaded (step 52).

With the above, the processing of the duplicated file system according to the third embodiment of the present invention is completed.

The redundant file system according to the third embodiment of the present invention saves the contents of the relay area in the external storage device when a failure occurs in the computer system including the redundant file system, thereby making it a target of writing. It has an effect that the operation of the redundant file system can be restarted after the computer system is restarted without losing the data.

The dual file system according to the fourth embodiment of the present invention temporarily stores the information of the data to be written in the secondary file 13 in the first, second, third and fourth dual file systems. First on external storage
Relay file 1A and second relay file 1B.

Next, a duplicated file system according to a fourth embodiment of the present invention will be described with reference to FIGS.

The first data writing means 15 is activated in response to the data writing processing request from the AP 11.

The first data writing means 15 first writes the data in the primary file 12 and stores the block number of the writing destination of the data (step 81).

Next, the first data writing means 15 refers to the free buffer number counter 23 of the first relay file 1A and determines whether or not the free buffer number counter 23 is 0 (step 82). If the free buffer number counter 23 is 0 in step 82, the process returns to step 82 after a lapse of a fixed time in order to wait until a free area is formed in the first relay file 1A (step 83).

If the free buffer number counter 23 is not 0 in step 82, that is, if there is a free area in the first relay file 1A, the first data writing means 15 first writes in the first relay file 1A. The data information to be written is written in the area indicated by the end record pointer 21 in the format of the write destination block number 31 and the write data 32 (step 84). At this time, the block number indicating the position where the data is written to the primary file in step 81 is written to the write destination block number 31, and the data body is written to the write data 32. Further, the end record pointer 21 in the first relay file 1A is positioned at the next record (step 85), and 1 is subtracted from the buffer number counter 23 (step 86).

After writing the data to the primary file 12 in step 81, the first data writing means 15 determines whether or not the writing process is normally completed (step 87). Step 82 if completed normally
Leading to.

When the first data writing means 15 determines in step 87 that the data writing process for the primary file 12 has terminated abnormally, the end record pointer 21 and the top record pointer in the first relay area 14 It is determined whether or not both 22 point to the head of the buffer section 24 (step 88), and if not pointed, step 88 is executed again for a certain period of time (step 89).

When the end record pointer 21 and the top record pointer 22 in the first relay area 14 both point to the top of the buffer section 24 in step 88, the first data writing means 15 sub-data. Write to file 13 (step 8A).

The second data writing means 16 first applies the exclusive control to the first relay file 1A (step 9).
1) Next, the contents of the first relay file 1A are copied to the second relay file 1B (step 92), and the first relay file 1B is further copied.
After initializing the relay file 1A of (step 93),
The exclusive control for the first relay file 1A is released (step 94).

In the initialization of the first relay file 1A in step 93, the end record pointer 21 and the start record pointer 22 both point to the start of the buffer section 24, and the free buffer number counter 23
Indicates the number of buffers in the buffer section.

Next, the second data writing means 16 determines whether or not the start record pointer 22 and the end record pointer 21 in the second relay file 1B are equal (step 95). finish.

If the start record pointer 22 and the end record pointer 21 in the second relay file 1B are different in step 95, the second data writing means 16 starts the start record pointer 22 in the second relay file 1B. From the buffer indicated by
1 and write data 32 are read (step 96),
First record pointer 22 in the second relay file 1B
Is positioned as the next record (step 97), and the free buffer number counter 23 is incremented by 1 (step 98).

Further, the second data writing means 16 is
The write data 32 is written in the area indicated by the write destination block number 31 in the sub-file 13 (step 99), and the process returns to step 95.

With the above, the processing of the duplicated file system according to the fourth embodiment of the present invention is completed.

The dual file system according to the fourth embodiment of the present invention includes a data writing process to a primary file,
By performing the data write process to the secondary file asynchronously, the data write process of the duplicated file can be performed efficiently, and when a failure occurs in the primary file, all input / output processing is performed on the secondary file. By carrying out the above, there is an effect that the input / output processing can be continued without being delayed. Further, by setting the first relay file and the second relay file in a high-speed external storage device such as an electronic disk device, there is an effect that the data writing process of the duplicated file can be performed more efficiently. .

[0064]

As described above, the present invention has an effect that the data writing process to the duplicated file can be realized with a minimum delay as compared with the data writing process to the non-duplicated file. There is.

[Brief description of drawings]

FIG. 1 is a block diagram showing first to third embodiments of the present invention.

FIG. 2 is a diagram showing a data storage format of a first relay area 14 and a second relay area 17 in the first to third embodiments of the present invention.

FIG. 3 is a first view of the first and second embodiments of the present invention.
6 is a flowchart showing the processing of the data writing means 15 of FIG.

FIG. 4 is a flowchart showing the processing of the second data writing means 16 in the first exemplary embodiment of the present invention.

FIG. 5 is a flowchart showing the processing of a relay area saving means 18 and a relay area loading means 19 in the third embodiment of the present invention.

FIG. 6 is a block diagram showing a fourth embodiment of the present invention.

FIG. 7 is a diagram showing a data storage format of a first relay file 1A and a second relay file 1B according to a fourth embodiment of the present invention.

FIG. 8 is a flowchart showing the processing of the first data writing means 15 in the fourth embodiment of the present invention.

FIG. 9 is a flowchart showing the processing of the second data writing means 16 in the fourth embodiment of the present invention.

[Explanation of symbols]

 11 AP 12 Primary File 13 Secondary File 14 First Relay Area 15 First Writing Means 16 Second Writing Means 17 Second Relay Area 18 Relay Area Saving Means 19 Relay Area Loading Means 1A First Relay File 1B First 2 relay file 1C external storage device 21 end record pointer 22 start record pointer 23 free buffer number counter 24 buffer unit 25 control unit 31 write destination block number 32 write data

Claims (6)

[Claims] 1. A file for storing data of an application program is configured as a pair of two files, one of which is a primary file and the other of which is a secondary file, and the writing process is performed for both the primary file and the secondary file. In a duplex file system in which a file is read and a read process is performed only from a primary file, a first relay area and a second relay area on a main storage device for temporarily storing update information for the sub-file; First data writing means for writing data to the primary file in response to a data writing request from an application program and storing update information of the data to be written to the sub file in the first relay area; Of the first relay area and the second data writing section of the second relay area. A second data writing means for initializing the first relay area after copying to the second relay area, reading the data stored in the second relay area, and writing the data to the sub file. Characteristic duplex file system. 2. The first relay area and the second relay area include a buffer unit for storing a write destination block number of data in the sub-file and write data, and the first data writing unit for the first data writing unit. An end record pointer indicating an additional position when data is added to one relay area, a head record pointer indicating a position of a head record when the second data writing unit acquires data from the second relay area, 2. The duplicated file system according to claim 1, further comprising a control unit including a free buffer number counter used for storing the number of unused buffers. 3. The second writing means copies the contents of the first relay area to the second relay area,
During the process of initializing the relay area, the first relay area is subject to exclusive control, and writing of information from the first writing means to the first relay area is kept waiting. The duplicated file system according to claim 2, wherein 4. When the first data writing means waits for the processing completion of the second data writing means when it is determined that a failure has occurred in the primary file and the input / output processing cannot be executed, 3. The input / output processing of No. 1 is directly performed on the sub file, and the input / output processing for the primary file and the first relay area is not performed. Duplicated file system described. 5. A relay area saving means for saving the contents of the first relay area and the second relay area to an external storage device when a failure occurs in a computer system including the duplicated file system, After restarting the computer system, a relay area load for loading the contents of the first relay area and the second relay area saved in the external storage device into the first relay area and the second relay area 5. The duplex file system according to claim 1, claim 2, claim 3, and claim 4, further comprising: 6. The first relay area and the second relay area are provided on an external storage device as a first relay file and a second relay file, respectively.
The duplicated file system according to claim 2, claim 3, or claim 4.