JPH07302217A - Fault recovery system for file in distributed processing system - Google Patents

Abstract

PURPOSE:To quickly recover a file by providing a means for updating information relating to the state of the file managed by a file state management means when a fault is generated in the file and when the file provided with the fault is recovered, retrieving the file provided with the fault and recovering it. CONSTITUTION:A file state updating means 14 updates the information relating to the state of the file 11 managed by a file state file 12 and a file management table 21 when the fault is generated in the file 11 and when the file 11 provided with the fault is recovered by a file recovery means 15. Thus, latest information relating to the state of the entire file 11 is stored in the file state file 12 and the file management table 21 at all times except the time when the fault is generated in the file state file 12 itself. Then, the file recovery means 15 recovers the file 11 provided with the fault in work stations 10 and 30 at a prescribed timing set beforehand.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a file failure recovery system for recovering a failed file in a duplex type distributed processing system environment in which files are duplicated among a plurality of workstations.

[0002]

2. Description of the Related Art In a distributed processing system in which a plurality of workstations perform distributed processing, as a method of distributing files, there is a duplex type distributed processing environment which allows the same file to be duplicated in a plurality of workstations. In this distributed processing environment, the same file is duplicated on multiple workstations,
Even if a file fails in one place, there is an advantage that it can be easily restored by copying the same file in another workstation.

Conventionally, in order to recover a file failure in a distributed processing environment as described above, a recovery process is manually performed for each workstation having the file in which the failure has occurred, such as when the system is terminated. Was being done.

[0004]

However, in the above-described conventional file failure recovery method, it is troublesome to manually recover the file individually for each workstation having the file in which the failure has occurred. There was a drawback.

SUMMARY OF THE INVENTION It is an object of the present invention to solve the above-mentioned conventional drawbacks and to provide a failure recovery system capable of promptly recovering a file in which a failure has occurred, targeting a duplex type distributed processing system. .

[0006]

In order to achieve the above object, the present invention provides a file for storing and managing information regarding the state of a file in a distributed processing system having a plurality of workstations with overlapping files. Status management means and information about the status of the files managed by the file status management means for the files managed by the file status management means when a failure occurs in the file and when the failed file is restored And a file restoration means for searching for a faulty file and restoring the file by referring to information relating to the state of the file managed by the file status management means.

In another aspect, the file status management means comprises a file management table created by an operating application program and a file status file permanently installed in each workstation.

In another aspect, when the file recovery means receives a file recovery request from an operating application program, it refers to the file management table and stores the same normal file stored in another workstation. When data is copied from a file and the faulty file is restored, and when a file restoration request is received from a waiting application program, at least two file status files are compared and the latest file status file is selected and referenced. Then, data is copied from the same normal file stored in another workstation to recover the faulty file.

In another file failure recovery method for achieving the above object, in a distributed processing system in which files are duplicated among a plurality of workstations, a file status management means for storing and managing information on the status of files is provided. , When an access request is received from an application program, the file status management means refers to the information about the status of the managed file, and accesses any file when the status of all duplicated files is normal. However, when there is a faulty file among the duplicated files, a fault occurs in the file access means for selecting and accessing a normal file and the file managed by the file status management means. In case and when the file with the failure is restored, the above file A file status updating means for updating information about the status of the file managed by the state management means,
Search for a faulty file by referring to the information on the state of the file managed by the file status management means,
A file recovery means for recovering the file is provided.

In another aspect, the file status management means comprises a file management table created by an operating application program, and a file status file permanently installed in each workstation, and the file access means comprises: The file management table is referenced to search for a normal file.

According to another aspect, when the file recovery means receives a file recovery request from an operating application program, it refers to the file management table and stores the same normal file stored in another workstation. When data is copied from a file and the faulty file is restored, and when a file restoration request is received from a waiting application program, at least two file status files are compared and the latest file status file is selected and referenced. Then, data is copied from the same normal file stored in another workstation to recover the faulty file.

[0012]

[Operation] According to the file failure recovery method in the distributed processing system of the present invention, it is possible to automatically find a failed file and restore it to a normal state in accordance with a request from an application program.

[0013]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing the configuration of a distributed processing system that realizes a file failure recovery system according to an embodiment of the present invention.

In the present embodiment, the workstations constituting the distributed processing system each perform a predetermined process according to an application program, but the workstation in a state in which the application program is operating and executing the process is a work system. An application program that is actually running on the station 10 and the working workstation 10 is called a working application 20. A workstation in a state where the application program is not running and processing is not executed is called a standby system workstation 30, and an application program in standby of the standby system workstation 30 is called a standby system application 40. Working workstation 10 and standby workstation 3
0, and the operating system application 20 and the standby system application 40 are structurally the same, and when focusing on an arbitrary workstation and application, there is a difference in whether or not the application is in the operating state. Only. Therefore, if it is not necessary to distinguish between the active system and the standby system, the workstation 1
It is expressed as 0, 30, and applications 20, 40.

The workstations 10 and 30 constituting the distributed processing system of this embodiment access the file 11 storing predetermined data, the file status file 12 storing information on the status of the file 11, and the file 11. It comprises a file access means 13, a file status updating means 14 for updating the information of the file status file 12, a file recovery means 15 for recovering the faulty file 11, and applications 20, 40 for performing data processing. In addition, the operating system application 20 of the present embodiment creates a file management table 21 that stores information regarding the state of the file 11 and retains it in the memory of the operating system workstation 10.

The applications 20 and 40 perform arithmetic processing of data according to various purposes. Then, if necessary, the file 11 is accessed through the file access unit 13, necessary data is read, and the processed data is stored. In this embodiment, the operating system application 20 creates the file management table 21. Figure 2
An example of the structure of the file management table 21 is shown in FIG. As shown in the figure, the file management table 21 includes workstations 10 and 30 for storing primary system files 11 and workstations 10 and 30 for storing secondary system files 11.
It stores a workstation name for identifying each of the files 11 and information indicating whether each file 11 is a primary system or a secondary system and the state of the file 11 is normal or faulty. The primary system and the secondary system of the file will be described later.

The file 11 stores various data to be processed by the distributed processing system of this embodiment. The distributed processing system of this embodiment allows files to be duplicated among a plurality of workstations.
Two identical files 11 storing the same data are created (duplication), two workstations 10, 3
It is managed at 0.

The file status file 12 is all the files 11 managed by the workstations 10 and 30.
Stores information about the state of. FIG. 3 shows an example of the data structure of the file information file 12. As shown in the figure, the file status file 12 includes the recording date and time of the information for indicating the version, the workstations 10 and 30 for storing the primary file 11, and the secondary file 11 as with the file management table 21. It stores the workstation names of the workstations 10 and 30 that store the information, and the information indicating whether each file 11 is a primary system or a secondary system and the state of the file 11 is normal or faulty.

Here, the above-mentioned file management table 2
1 and the file 11 in the file status file 12
The primary system and the secondary system of the file 11 are the file 11 and the file 11
Shows the relationship with the workstations 10 and 30 in which is stored. That is, the file 11 is the same workstation 10 or 30 as the file status file 12.
The file 11 is referred to as the primary system when it is stored in the sub-system, and the file 11 is referred to as the secondary system when stored in different workstations 10 and 30. In addition, the file management table 21 and the file status file 12
Stores information about the status of the file status file 12 itself. Therefore, the file status file 12 is included in the file 11 unless otherwise specified. The recording date and time stored in the file 11 indicates the version of the file 11, and may not necessarily be the recording date and time, but may be a version number or the like. Also, the workstation name can be substituted by an appropriate identifier that can identify the workstation.

When the file access means 13 receives an access request to the file 11 from the operating system application 20, the file access means 13 searches the file management table 21 on the memory and accesses the corresponding file 11. The files 11 accessed at this time are all the files 11 corresponding to the access request. That is, if it corresponds to the access request, the working workstation 1
Both the primary file 11 stored in 0 and the secondary file 11 stored in the standby workstation 30 are accessed. However, the file management table 21
When a file 11 having a failure is searched as a result of referring to, only the normal file 11 is accessed. Further, the file access unit 13 outputs an update request of information regarding the state of the file 11 to the file state update unit 14 when a failure occurs in the file 11 while accessing the predetermined file 11.

The file status updating means 14 uses the file 1
When a failure occurs in 1 and the failed file 1
When 1 is recovered by the file recovery means 15, the information on the status of the file status file 12 and the corresponding file 11 managed in the file management table 21 is updated. In particular, when a failure occurs in the file 11 while the file access unit 13 is accessing the file 11, a file update request is received from the file access unit 13, and the file status file 12 and the file management table 21 are received in accordance with this request. Update information on. This allows the file status file 12
The file management table 21 always stores the latest information regarding the status of all the files 11 except when a failure has occurred in the file status file 12 itself.

The file restoration means 15 is adapted to the workstations 10 and 30 at a predetermined timing set in advance.
Then, the file 11 having the fault in the above is restored. The recovery of the file 11 is performed by the other workstations 10 and 30.
In step 1), the duplicated normal file 11 is copied. Here, in order for the file restoration means 15 to search for the file 11 having a failure, it is necessary to refer to the file management table 21 or the file status file 12. Therefore, the method of searching for a faulty file 11 differs depending on whether the workstation 10, 30 including the file restoration means 15 is the active workstation 10 or the standby workstation 30.

First, in the case of the file recovery means 15 of the active workstation 10, since the file management table 21 created by the active application 20 is on the memory of the active workstation 10, the file recovery means 15 uses this. The file management table 21 is referred to search for the file 11 having a failure. On the other hand, in the case of the file recovery means 15 of the standby workstation 30, since the file management table 21 is not created, the file recovery means 15 uses the file status file 12
The file 11 having the fault is searched by referring to. However, when the file status file 12 itself has a failure, normal information cannot be obtained by referring to only the file status file 12 of the standby workstation 30. Therefore, the file restoring means 15 compares the file status file 12 of the standby system workstation 30 with the file status file 12 duplicated in the other workstations 10, 30 and refers to the newest file status file 12. Faulty file 1
Search for 1.

The file recovery means 15 can recover the faulty file 11 at the time of starting up or ending the operation of the workstations 10, 30 or periodically during the operation of the workstations 10, 30. However, the file management table 21 is created by the operating system application 20, and the file 1 is processed according to the processing while the workstations 10 and 30 are operating.
Considering that the data of No. 1 is updated, etc., it is preferable to do it at the end of the operation of the workstations 10, 30.

Next, the operation of updating the file status when a failure occurs in the file 11 while accessing the file 11 will be described with reference to FIG. As an initial state, it is assumed that the working application 20 is running in the working workstation 10 (step 401).

When the operating system application 20 issues a file access request to the file access means 13 (step 402), the file access means 13
Refer to the file management table 21 and duplicate
Of the two identical files 11, the file 11 whose file status is normal is accessed. (Step 403). When the file states of the duplicated same file 11 are all normal, the file 11 to be accessed is determined by an appropriate means such as giving priority to the normal system file.

If a failure occurs in the file 11 being accessed during the file access, the file access means 13 issues a file update request to the file status update means 14 (step 404). Upon receiving the file update request, the file status updating means 14 accesses the file management table 21 on the memory and updates the information on the file status of the file 11 in which the failure has occurred (step 405). At the same time, each of the duplicated file status files 12 is accessed to write the contents and recording date and time of the file management table 21.

Next, the recovery operation of the faulty file 11 at the end of the operation of the workstations 10 and 30 will be described with reference to FIGS. First, the recovery operation of the file 11 in the working workstation 10 will be described with reference to FIG. As an initial state,
It is assumed that the operating system application 20 is running on the operating system workstation 10 (step 50).
1).

The operating system application 20 issues a file recovery request to the file recovering means 15 when the operation of the operating system workstation 10 is completed (step 50).
2). Upon receiving the file restoration request, the file restoration means 15 refers to the contents of the file management table 21 in the memory to check the state of the file 11, and if there is a faulty file in the working workstation 10, The normal duplicated identical file 11 is copied from the other standby workstation 30 (step 50).
3, 504). This completes the restoration of the file 11.

Next, the recovery operation of the file 11 in the standby workstation 20 will be described with reference to FIG. In the initial state, the standby workstation 3
At 0, the standby application 40 is in a state of waiting for the input of an operation command (step 601).

The standby system application 40 issues a file recovery request to the file recovery means 15 at the end of the operation of the standby system workstation 30 (step 60).
2). Upon receiving the file restoration request, the file restoration means 15 compares the recording date and time of the file status file 12 of the standby workstation 30 with the recording date and time of the duplicated file status file 12 of the other workstations 10 and 30. The file having the newest recording date and time is regarded as the normal file status file 12, and the status of the file 11 is checked by referring to the file status file 12 (step 603). If the standby workstation 30 has a failed file, the duplicate normal file 11 is copied from the other workstations 10 and 30 (steps 604 and 60).
5). This completes the restoration of the file 11.

Although the present invention has been described above with reference to the preferred embodiments, the present invention is not necessarily limited to the above embodiments. For example, in this embodiment, the files are duplicated,
Although it has been decided to create two identical files, in order to recover the files according to the present invention, it is sufficient that the same files are duplicated in the distributed processing system, and even if two or more files are created and managed. Good. In this case, it is necessary to identify the files so that not only the primary file and the secondary file but also the third and subsequent files can be identified.

Further, in the present embodiment, the operating system application creates the file management table and uses it when recovering the faulty file in the operating system workstation, but it is not always necessary to use the file management table. As with the standby workstation, the file status file may be used to restore the file.

Further, in this embodiment, in response to the access request from the operating system application, the file access means targets all the duplicated corresponding files, but accesses only the files stored in the operating system workstation. You may make it a target.

[0035]

As described above, according to the file failure recovery method in the distributed processing system of the present invention, the file in which the failure has occurred is automatically detected in accordance with the request from the application program, and the normal state is obtained. Since there is no need to restore the file, it is possible to reduce the time and effort required to restore the file and to quickly restore the file.

Further, the file recovery table can be further simplified by creating a file management table while the application program is operating and referring to the file management table when the file recovery means recovers the file. The effect is that you can do it.

Further, according to the present invention, the file access means selects a normal file having no trouble from the duplicated files and accesses the file, thereby processing even if the file to be accessed has a trouble. There is an effect that can be continued.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a distributed processing system for realizing a file failure recovery system according to an embodiment of the present invention.

FIG. 2 is a diagram showing a configuration example of a file management table in FIG.

FIG. 3 is a diagram showing a configuration example of a file status file in FIG.

FIG. 4 is a flowchart showing an updating operation by the file access means and the file status updating means of FIG.

5 is a flowchart showing an operation of a file recovery means in the working workstation of FIG.

FIG. 6 is a flowchart showing the operation of a file recovery means in the standby workstation of FIG.

[Explanation of symbols]

 10 Working Workstations 11 Files 12 File Status Files 13 File Access Means 14 File Status Updating Means 15 File Restoring Means 20 Working Applications 21 File Management Tables 30 Standby Workstations 40 Standby Applications

Claims (6)

[Claims] 1. In a distributed processing system in which files are duplicated among a plurality of workstations, a file status management unit that stores and manages information related to file status, and a file managed by the file status management unit are targets. As a file status update means for updating information on the status of a file managed by the file status management means when a failure occurs in the file and when the failed file is restored, and the file status management means manages the file status management means. Find the faulty file by looking at the information about the status of the deleted file,
A file failure recovery method in a distributed processing system, comprising: a file recovery means for recovering the file. 2. The file status management means comprises a file management table created by an operating application program, and a file status file permanently installed in each workstation. A failure recovery method for files in a distributed processing system. 3. When the file recovery means receives a file recovery request from an operating application program, it refers to the file management table and retrieves data from a normal identical file stored in another workstation. When a file with a failure is copied and restored, and a file recovery request is received from a waiting application program, at least two file status files are compared, the latest file status file is selected and referenced, and another file status file is selected. 3. The file failure recovery method in the distributed processing system according to claim 2, wherein data is copied from a normal identical file stored in the workstation to recover the failed file. 4. A distributed processing system in which files are duplicated among a plurality of workstations, and a file status management means for storing and managing information on the status of files and an access request from an application program are received. , The file status management means, referring to the information about the status of the managed files, when the status of all the duplicated files is normal, an arbitrary file is accessed, and the file having the failure in the duplicated files is accessed. If there is a file, the file access means for selecting and accessing a normal file, and the file managed by the file status management means are targeted when a failure occurs in the file and when the failed file is restored. Information on the status of files managed by the file status management means File status updating means for updating information, and searching for a file with a failure by referring to information relating to the status of files managed by the file status managing means,
A file failure recovery method in a distributed processing system, comprising: a file recovery means for recovering the file. 5. The file status management means comprises a file management table created by an operating application program, and a file status file permanently installed in each workstation, and the file access means comprises the file management table. 5. The file failure recovery method in the distributed processing system according to claim 4, wherein a normal file is searched by referring to. 6. When the file recovery means receives a file recovery request from an operating application program, it refers to the file management table and retrieves data from the same normal file stored in another workstation. When a file with a failure is copied and restored, and a file recovery request is received from a waiting application program, at least two file status files are compared, the latest file status file is selected and referenced, and another file status file is selected. 6. The file failure recovery method for a distributed processing system according to claim 5, wherein data is copied from the same normal file stored in the workstation to recover the failed file.