JP2000010868A - Decentralized system and its backup method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a decentralized system which places a small load on computers and a communication line and is reliable, and its backup method. SOLUTION: Master data 3 are held on an in-operation system computer 1A and backup data 4 corresponding to the contents of the master data 3 at a specific point of time and update history files 5 corresponding to an updating process on the in-operation system computer 1A are held on computers 1B and 1C. The in-operation system computer 1A holds an update history file 5 corresponding to an updating process for the master data 3 and also manages a position and order information file 6 for specifying the positions of update history files 5 divided and held between the backup computers 1B and 1C. Further, the backup computers 1B and 1C manage position and order information files 6 for specifying the positions of the update history files 5 divided and held between the backup computers 1B and 1C.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a distributed system in which a plurality of computers are connected via a communication path, and more particularly to a backup method for a database or the like in the distributed system.

[0002]

2. Description of the Related Art In general, backup of a database or the like in a computer is performed based on backup data corresponding to the contents of master data at a predetermined point in time and an update history corresponding to an update process performed after recording the backup data. Have been done. In such a computer, when the master data is destroyed, the backup data is updated based on the update history, and the latest data is restored.

FIG. 16 is a diagram for explaining a backup method of a database or the like in a conventional distributed system. As shown in FIG. 16, the distributed system is composed of three computers 1A, 1B, and 1C connected via a communication path 2, and includes a computer (operating computer) 1A including a database update process. Business such as online processing is performed, and the computer (backup destination computer) 1B
The process A and data are backed up. Here, the master data 3 is stored in the active computer 1A.
And the backup destination computer 1B holds the backup data 4. The operating system computer 1A
The update history corresponding to the database update process performed in step 1 is transferred from the active computer 1A to the backup destination computer 1B. In such a distributed system, when a failure occurs in the active computer 1A for some reason, or when the master data 3 held by the active computer 1A is lost.
Becomes unavailable, the backup destination computer 1
B updates the backup data 4 based on the update history file 5, restores the latest data, and updates the active computer 1A.
Business is being taken over.

[0004] In such a distributed system, when a failure occurs in the backup destination computer 1B, backup is continued mainly by the following two methods.

That is, in the first conventional method, recording of the update history is started at the active computer 1A as the backup source, and when the backup destination computer 1B is restored, the update history is transferred from the active computer 1A to the backup destination computer 1B. This allows the backup destination computer 1B to update the backup data 4 based on the update history transferred from the active computer 1A and keep the backup data 4 in the latest state.

[0006] The second conventional method uses the active computer 1.
In this method, the backup destination of A is changed from the backup destination computer 1B to another computer (for example, the computer 1C in FIG. 16).
The backup of A can be continued.

[0007]

However, in the first and second conventional methods described above, when the backup destination computer is restored or the backup destination computer is changed, the active computer 1A and the backup destination computer 1B are used.
Alternatively, since the update history has to be transferred to the backup destination computer 1C, a heavy load is temporarily applied to the computer and the communication path, which may affect other operations other than the database update processing.

In the first and second conventional methods described above, when the computers 1B and 1C fail, only the active computer 1A holds the latest data. If the data held by the computer 1A becomes unusable, there is a possibility that serious damage to the user may be caused not only by stopping the business but also by losing the latest data.

SUMMARY OF THE INVENTION The present invention has been made in consideration of the above points, and an object of the present invention is to provide a highly reliable distributed system having a low load on a computer and a communication path, and a backup method thereof.

[0010]

A first feature of the present invention is as follows.
An active computer that updates the master data,
Backup data corresponding to the content of the master data at a predetermined point in time, and at least two backup computers that hold an update history corresponding to the update process, wherein each of the backup computers is A distributed system for managing position information for specifying a position of each update unit of the update history which is divided and held.

In a first aspect of the present invention, the active computer holds an update history corresponding to the update processing, and stores the update history of the update history divided and held between the backup computers. It is preferable to manage position information for specifying a position for each division unit, and
It is preferable that each of the backup computers collects an update history from another backup computer or the active computer based on the location information.

A second feature of the present invention is that an active computer for updating master data and at least two backup computers for holding backup data corresponding to the contents of the master data at a predetermined time. Transferring the update history corresponding to the update process from the active computer to the backup computer serving as a backup destination, and dividing the backup computer between the backup computers. Managing position information for specifying a position of each of the update histories to be held for each division unit.

In the second aspect of the present invention, it is preferable that the method further includes a step of collecting an update history from another backup computer or the active computer based on the location information in any one of the backup computers.

According to a third feature of the present invention, in the above-mentioned second feature, when it becomes necessary to restore the latest data in any one of the backup computers, another backup computer is restored based on the position information. Compare the data amount of the update history to be collected from the computer or the active computer, and collect the update history from another computer based on the location information in the backup computer having the smallest data amount, from the backup data. A backup method for a distributed system, further comprising the step of restoring the latest data.

According to a fourth feature of the present invention, in the above-described second feature, when it is necessary to select a backup destination, the data is collected from another backup computer or the active computer based on the location information. 5. The distributed system backup method according to claim 4, further comprising the step of comparing the data amounts of update history to be performed and selecting a backup computer having the smallest data amount as a new backup destination.

According to the first and second aspects of the present invention,
A series of update histories corresponding to the update processing performed by the active computer is divided and held between multiple backup computers, and the update histories held by the active computer and each backup computer are exchanged based on the location information. It is not necessary to transfer a large amount of past update history to the backup destination computer when the backup destination computer is changed or when the backup destination computer is restored, and the load at night is reduced. The update history may be transferred during a light time period or a time period during which no work is performed, so that the load on the computer and the communication path can be reduced and the influence on other work can be minimized. Also, at the time of failure of the active computer or change of the backup destination computer, at least two computers are in the state of holding the latest data or in the state of being able to restore the latest data, Even if the data held by the active computer becomes unusable in this state, the latest data can be restored based on other backup computers, etc., thereby preventing the latest data from being lost. Reliability can be improved. Further, since the same update history is maintained between the active computer and each backup computer, necessary update histories can be collected from a plurality of computers. By collecting the history, the load on the computer can be reduced. Further, by collecting update histories so that a load is not concentrated on a specific computer or a communication path, the transfer time of the update histories can be reduced.

According to the third feature of the present invention, when it becomes necessary to restore the latest data in one of the backup computers due to a failure of the active computer or the like, the latest data is collected from another backup computer. Since the backup computer having the smallest data volume of the update history to be selected is selected as the restoration destination, the load on the computer or the communication path can be reduced, and the latest data can be restored in a short time.

According to the fourth feature of the present invention, when it becomes necessary to select a new backup destination due to restoration of a backup computer or the like, the amount of update history data to be collected from another computer is the smallest. Since the backup computer is selected as the backup destination, it is possible to reduce the load on the computer or the communication path, restore the latest data in a short time, and transfer the update history file during business operation. However, it is possible to reduce the load on the computer or the communication path and reduce the influence on the business.

[0019]

DESCRIPTION OF THE PREFERRED EMBODIMENTS First Embodiment An embodiment of the present invention will be described below with reference to the drawings. FIGS. 1 to 13 are diagrams for explaining a first embodiment of a distributed system and its backup method according to the present invention.

First, a system configuration of a distributed system will be described with reference to FIG. As shown in FIG. 1, the distributed system includes three computers 1 connected via a communication path 2.
A, 1B, and 1C, the computer 1A performs operations such as online processing including database update processing, and the computers 1B and 1C back up the processing and data of the computer 1A. I have. That is, the computer 1A is an active computer that performs an update process on the master data 3, and the computers 1B and 1C are backup data 4 corresponding to the contents of the master data 3 at a predetermined time, and the update process in the active computer 1A. Is a backup computer that holds the update history file 5 corresponding to. Note that master data 3 and backup data 4
And the update history file 5 are stored in each computer 1A, 1
B and 1C are stored in a predetermined storage device.

Here, the active computer 1A holds the update history file 5 corresponding to the update processing for the master data 3, and stores the update history file 5 divided and held between the backup computers 1B and 1C. A position / order information file (position information) 6 for specifying the position is managed. Each backup computer 1B, 1C
Manages a position / order information file 6 for specifying the position of the update history file 5 that is divided and held between the backup computers 1B and 1C. The position / order information files 6 held by each of the computers 1A, 1B, 1C have the same contents, and the update history file name divided and held, the computer name holding the update history file, and the order information of the update history file It is included.

Next, various processes in the distributed system having such a configuration will be described.

(At the Start of Business and Backup) First, referring to FIG. 2, a description will be given of a process for starting a business and backup in the distributed system shown in FIG.

As shown in FIGS. 1 and 2, when starting the operation of online processing and backup of the active computer 1A, first, the backup computer (in FIG. 1, the backup computer) is used as the backup destination. Calculator 1
B) is selected (step 101), and then the position / order information file 6 is updated by the active computer 1A (step 102). Here, in the active computer 1A, a position / order information file 6 as shown in FIG. 3 is created. As shown in FIG. 3, the position / order information file 6 includes an update history file sequence number (“1”), an update history file name (“log1”), and a computer name (“computer A”) holding the update history file. , "Computer B"), and the update state of the update history file ("Updating").

Thereafter, from the active computer 1A, the position / sequence information file 6 is transferred to all computers (in FIG. 1, backup computers 1B and 1 in FIG. 1) incorporated in the system.
C) via the communication path 2 (step 103), and the work and backup of the active computer 1A are started (step 104). In step 104, when the active computer 1A updates the master data 3 as a database, the update history corresponding to the update process is stored in the update history file (log1) 5, and the active computer 1A is updated. Computer 1A to backup computer 1B
Transfer to In the backup computer 1B, the update history transferred from the active computer 1A is stored in the update history file (log1) 5.

(When a backup destination computer fails)
With reference to FIGS. 4 to 6, a description will be given of a process in a case where a failure occurs in the backup destination computer of the active computer 1A and the backup destination is switched in such a distributed system.

As shown in FIGS. 4 and 5, when a failure occurs in the backup computer 1B, which is the backup destination, and the backup of the active computer 1A cannot be performed, the active computer 1A is disabled. A failure of the backup computer 1B is detected (Step 201), and the operation and backup of the active computer 1A are temporarily suspended (Step 202), and the failed backup computer 1B is disconnected from the system (Step 20).
3).

Thereafter, the position / sequence information file 6 is updated by the active computer 1A (step 204). Here, in the active computer 1A, a position / order information file 6 as shown in FIG. 6A is created. Then, the position / order information file 6 is transferred from the active computer 1A to all the computers (the backup computer 1C in FIG. 4) incorporated in the system via the communication path 2 (step 205). The operation and backup of the computer 1A are restarted (step 206). When the operation and backup of the active computer 1A are restarted, the processing shown in FIG. 2 is performed again. That is, the backup computer (the backup computer 1C in FIG. 4) is selected in the active computer 1A, the position / order information file 6 as shown in FIG. 6B is created, and the backup computer 1C is sent to the backup computer 1C. After the transfer, the operation and backup of the active computer 1A are started.

(When the backup computer is restored)
With reference to FIGS. 7 to 10, a description will be given of a process in a case where a backup computer in which a failure has occurred in such a distributed system is restored and a backup destination is switched.

As shown in FIGS. 7 and 8, when the backup computer 1B disconnected due to a failure recovers and is re-integrated into the system, the backup computer 1B to be incorporated is incorporated into the system. The other computers (in FIG. 7, the active computer 1A and the backup computer 1C in FIG. 7) are requested to be incorporated (step 301). Here, the active computer 1A or the backup computer 1C, which has received the request, sends the position / order information file 6 to the backup computer 1B to be incorporated in order to match the contents of the position / order information files 6 in all the computers. Is transferred (step 302). As a result, the backup computer 1B is incorporated into the system, and shifts to a state where work and backup can be performed (step 303).

When the backup computer 1B is disconnected from the system or restored as described above, as shown in FIGS. 7 and 9, the backup computer 1B is installed (step 401). ), The operation and backup of the active computer 1A are temporarily suspended (step 402), and a computer (backup computer 1B in FIG. 7) to be a backup destination is selected by the active computer 1A (step 403).

After that, the operating computer 1A updates the position / order information file 6 (step 404). Here, in the active computer 1A, a position / order information file 6 as shown in FIG. 10A is created. Then, the position / order information file 6 is transferred from the active computer 1A to all the computers (backup computers 1B and 1C in FIG. 7) incorporated in the system via the communication path 2 (step 405). The operation and backup of the active computer 1A are resumed (step 406). In addition,
When the operation and backup of the active computer 1A are restarted, the processing shown in FIG. 2 is performed again. That is,
The active computer 1A selects a computer to be a backup destination (the backup computer 1B in FIG. 7), and FIG.
After creating the position / sequence information file 6 as shown in (b) and transferring it to the backup computers 1B and 1C, the work and backup of the active computer 1A are started.

When the backup computer 1B is restored in this way, as shown in FIG. 7, the update history transferred from the active computer 1A is divided and held in the backup computers 1B and 1C. Will be done. Here, the backup computers 1B and 1C specify the position of the update history file 5 held by the active computer 1A or the other backup computers 1C and 1B based on the position / order information file 6 and update the necessary update history. Can be collected. For example, in the case shown in FIG. 7, the backup computer 1B collects an update history file (log2) from the backup computer 1C, and updates the collected update history file and the update originally held by the backup computer 1B. By updating the backup data 4 based on the history files (log1, log3), the latest data can be restored from the backup data 4. Such transfer of the update history file and restoration of the latest data need not be performed when the backup destination computer is changed or when the backup destination computer is restored. It can be performed at a time when it is not performed.

(At the time of failure of the active computer) Next, referring to FIG. 11, the processing when a failure occurs in the active computer in such a distributed system will be described.

For example, in the distributed system shown in FIG.
When a failure occurs in the active computer 1A as the backup source, as shown in FIG. 11, a failure in the active computer 1A is detected (step 501), and the work and backup of the active computer 1A are temporarily suspended. At the same time (step 502), the active computer 1A is disconnected from the system (step 503).

Then, a new active computer (for example, the backup computer 1) taking over the work of the active computer 1A
After selecting B) (step 504), the new active computer 1B updates the position / sequence information file 6 (step 505) and the new active computer 1B.
From this, the position / sequence information file 6 is used for all computers (backup computer 1) incorporated in the system.
C) via the communication path 2 (step 506).

Thereafter, the new active computer 1B specifies the position of the update history file 5 held by the other backup computer 1C based on the position / order information file 6, and collects necessary update histories. After updating the backup data 4 and restoring the latest data based on the collected update history file and the update history file originally held by the new operating computer 1B (step 507), the new operation The operation and backup are restarted in the system computer 1B (step 508).

(At the time of restoring the latest data) Next, referring to FIGS. 12 and 13, in order to restore the latest data in the event of a failure of the active computer, a specific computer (transfer destination computer) is transferred to another computer (transfer destination computer). A process when the update history is transferred from the transfer source computer) will be described. Here, FIG.
As shown in FIG. 2, it is assumed that the backup computer 1C restores the latest data in a state where both the active computer 1A and the backup computers 1B and 1C are incorporated in the system.

As shown in FIGS. 12 and 13, first,
The backup / computer 1C, which is the transfer destination computer, updates the position / order information file 6 (step 601), and then, from the backup computer 1C, all the computers incorporated in the system (the active computers 1A and 1A in FIG. 12). The position / order information file 6 is transferred to the backup computer 1B) (step 602).

Thereafter, the update history file 5 necessary for restoring the latest data from the active computer 1A and the backup computer 1B, which are transfer source computers, to the backup computer 1C is transferred (step 603). After updating the position / order information file 6 at 1C (step 604), the backup computer 1
All computers incorporated into the system from C (the active computer 1A and the backup computer 1B in FIG. 12)
To the position / order information file 6 (step 60).
5).

In the case shown in FIG. 12, the update history files (L1, L2) 5 are transferred from the active computer 1A and the backup computer 1B to the backup computer 1C. 12) is held in both the active computer 1A and the backup computer 1B, so that the update log file (L2) is transmitted from the active computer 1A as shown in FIG.
5, and the update history file (L1) 5 is transferred in parallel from the backup computer 1B.

As described above, according to the first embodiment of the present invention, a series of update histories corresponding to the update processing performed by the active computer 1A is stored in the backup computers 1B and 1B.
The update history files 5 held by the active computer 1A and each of the backup computers 1B and 1C can be mutually collected based on the position / order information file 6 while being divided and held. Therefore, in a distributed system whose main purpose is to maintain the latest data (for example, a distributed system that only backs up the latest data without taking over processing), when the backup destination computer is changed or when the backup destination computer is restored. For example, there is no need to transfer a large amount of past update history to the backup destination computer, etc., and it is sufficient to transfer the update history during a time when the load is light, such as at night, or during a time when business is not performed. The load on computers and communication paths can be reduced to minimize the impact on other tasks.

According to the first embodiment of the present invention,
Even when the active computer 1A fails or when the backup destination computer is changed, at least two computers are in a state where the latest data is held or the latest data can be restored. Even if the data held by the active computer 1A becomes unavailable in this state,
The latest data can be restored based on another backup computer or the like, so that the latest data can be prevented from being lost and reliability can be improved.

Further, according to the first embodiment of the present invention, the active computer 1A and each backup computer 1B,
Since the update history file 5 having the same contents is held between the computer 1C and the computer 1C, necessary update histories can be collected from a plurality of computers. Load can be reduced. Further, by collecting update histories so that a load is not concentrated on a specific computer or a communication path, the transfer time of the update histories can be reduced.

Second Embodiment Next, a second embodiment of the present invention will be described with reference to FIG. The second embodiment of the present invention relates to a method of selecting an optimal restoration destination when restoring the latest data on a backup computer due to a failure of an active computer or the like. The basic configuration, processing, and the like of the distributed system according to the second embodiment of the present invention are substantially the same as those of the first embodiment shown in FIGS. The same parts as those of the first embodiment are denoted by the same reference numerals, and detailed description is omitted.

As shown in FIGS. 14A and 14B, for example, in the distributed system shown in FIG. 7, when a failure occurs in the active computer 1A as the backup source, the backup computer 1C switches from the backup computer 1C to the backup computer. Transfer the update history file (L2) 5 to the backup computer 1B by transferring the update history file (L2) 5 to the backup computer 1B, or transfer the update history file (L1) 5 from the backup computer 1B to the backup computer 1C. 1C
Need to restore the latest data.

In this case, each backup computer 1
B, 1C, the data amount of the update history to be collected from the other backup computers 1C, 1B based on the position / order information file 6 (update history file (L1, L2)
Are compared (see FIG. 14A). And
The backup computer having the smallest data amount is selected as a computer for restoring the latest data, and the selected backup computer collects update history from another backup computer based on the position / order information file 6. To restore the latest data from the backup data (see FIG. 14B). In the case shown in FIG. 14, since the size of the update history file (L2) 5 is smaller than the size of the update history file (L1) 5, the update history file (L2) 5 is transmitted from the backup computer 1C to the backup computer 1C. 1B, and the backup computer 1B restores the latest data and takes over the processing.

As described above, according to the second embodiment of the present invention, when it is necessary to restore the latest data in one of the backup computers 1B and 1C due to a failure of the active computer 1A or the like. , Another backup computer 1
Since the backup computer that has the least amount of update history data to be collected from C and 1B is selected as the restoration destination,
The load on the computer or the communication path can be reduced, so that the latest data can be restored in a short time.

Third Embodiment Next, a third embodiment of the present invention will be described with reference to FIG. The third embodiment of the present invention relates to a method of selecting an optimal backup destination when changing a backup destination computer due to restoration of a backup computer or the like. The basic configuration, processing, and the like of the distributed system according to the third embodiment of the present invention are substantially the same as those of the first embodiment shown in FIGS. The same parts as those of the first embodiment are denoted by the same reference numerals, and detailed description is omitted.

As shown in FIG. 15, for example, in the distributed system shown in FIG. 5, when the backup computer 1B is restored, the backup computer 1C which is the current backup destination is continuously set as the backup destination or restored. It is necessary to select whether the backup computer 1B is a new backup destination. Here, FIG.
As shown in FIG. 5, the active computer 1A holds an update history file (L1, L2, L3, L4, L5) 5, and the backup computer 1B holds the update history file (L1, L1).
L2, L3) 5 are held and the backup computer 1C
Holds an update history file (L4, L5) 5, and in the backup computers 1B, 1C, the update history file (L1, L2, L4) 5 has already been reflected in the backup data. Each of the backup computers 1B and 1C holds, in addition to the position / order information file 6, information indicating which update history file 5 has been reflected in the backup data of each backup computer 1B and 1C. Shall be

In this case, the backup computers 1B and 1C, which may be selected as backup destinations, collect from the other backup computers 1C and 1B or the active computer 1A based on the position / order information file 6. The data amount of the update history to be compared is compared. At this time, the entire size S1 of the update history file 5 held by the backup computer 1C as the current backup destination and the restored backup computer 1B is calculated based on the position / order information file 6, and this size is calculated. From S1, the entire size S2 of the update history file 5 already reflected in the backup data is calculated, and the size S
The backup computer having the smallest value S3 obtained by subtracting the size S2 from 1 is selected as the next new backup destination.

That is, if the size S3 of the update history file 5 of the restored backup computer 1B is smaller than the size S3 of the update history file 5 of the backup computer 1C that is currently the backup destination, the restored backup The computer 1B is selected as a new backup destination. On the other hand, when the size S3 of the update history file 5 of the restored backup computer 1B is larger than the size S3 of the update history file 5 of the backup computer 1C which is the current backup destination, the backup currently being the backup destination is performed. Computer 1C is continuously selected as the backup destination.

In the case shown in FIG. 15, the size S3 (the size of the update history file (L3) 5) of the update history file 5 of the restored backup computer 1B is the backup computer 1C which is the current backup destination. Is larger than the size S3 of the update history file 5 (the size of the update history file (L5) 5), the backup computer 1C which is currently the backup destination is continuously selected as the backup destination.

As described above, according to the third embodiment of the present invention, when it becomes necessary to select a new backup destination due to the recovery of the backup computer 1B, the update history to be collected from another computer. The backup computer with the least amount of data is selected as the backup destination, so the load on the computer or communication path can be reduced and the latest data can be restored in a short time. Even if the data is transferred, the load on the computer or the communication path can be reduced, and the effect on the business can be reduced.

[0055]

As described above, according to the present invention, it is possible to provide a distributed system with a low load on a computer and a communication path and a high reliability, and a backup method thereof.

[Brief description of the drawings]

FIG. 1 is a diagram showing a system configuration of a first embodiment of a distributed system according to the present invention.

FIG. 2 is a flowchart for explaining operations at the start of business and backup in the first embodiment of the distributed system according to the present invention.

FIG. 3 is a view showing an example of a position / order information file held by each computer of the distributed system shown in FIG. 1;

FIG. 4 is a view for explaining processing when a backup destination is switched in the first embodiment of the distributed system according to the present invention;

FIG. 5 is a flowchart for explaining an operation when a backup destination computer fails in the first embodiment of the distributed system according to the present invention;

FIG. 6 is a view showing an example of a position / order information file held by each computer of the distributed system shown in FIG. 4;

FIG. 7 is a diagram for explaining processing when a failed computer is restored and a backup destination is switched in the first embodiment of the distributed system according to the present invention;

FIG. 8 is a flowchart for explaining the operation of the distributed system according to the first embodiment of the present invention when a computer is incorporated.

FIG. 9 is a flowchart for explaining an operation at the time of restoration of the backup computer in the first embodiment of the distributed system according to the present invention.

FIG. 10 is a view showing an example of a position / order information file held by each computer of the distributed system shown in FIG. 7;

FIG. 11 is a flowchart for explaining an operation at the time of failure of the active computer in the first embodiment of the distributed system according to the present invention;

FIG. 12 is a diagram for explaining an operation at the time of collecting update history files in the distributed system according to the first embodiment of the present invention.

FIG. 13 is a flowchart for explaining an operation at the time of transferring an update history file in the distributed system according to the first embodiment of the present invention.

FIG. 14 is a diagram for explaining a method of selecting a latest data restoration destination in the distributed system according to the second embodiment of the present invention.

FIG. 15 is a view for explaining a method of selecting a backup destination in the third embodiment of the distributed system according to the present invention.

FIG. 16 is a view for explaining a backup method of a database or the like in a conventional distributed system.

[Explanation of symbols]

 1A Active computer 1B, 1C Backup computer 2 Communication path 3 Master data 4 Backup data 5 Update history file 6 Position / sequence information file (position information)

Claims (7)

[Claims] 1. An active computer for performing an update process on master data; backup data corresponding to the content of the master data at a predetermined time; and at least two backups holding an update history corresponding to the update process. And a backup computer, wherein each backup computer manages location information for specifying a location for each division unit of the update history that is divided and held between the respective backup computers. Distributed system. 2. The operating system computer holds an update history corresponding to the update processing and specifies a position of each update unit divided and held among the backup computers for each division unit. 2. The distributed system according to claim 1, wherein location information for managing the information is managed. 3. The distributed system according to claim 1, wherein each of the backup computers collects an update history from another backup computer or the active computer based on the location information. 4. A backup system for a distributed system, comprising: an active computer for updating master data; and at least two backup computers for holding backup data corresponding to the contents of the master data at a predetermined time. In the method, a step of transferring an update history corresponding to the update processing from the active computer to a backup computer serving as a backup destination; and Managing position information for specifying a position for each division unit. 5. The distributed system according to claim 4, further comprising a step of collecting an update history from another backup computer or the active computer based on the location information in any one of the backup computers. Backup method. 6. A data amount of an update history to be collected from another backup computer or the active computer based on the location information when it is necessary to restore the latest data in one of the backup computers. And recovering the latest data from the backup data by collecting update history from another computer based on the location information in the backup computer having the smallest data amount. The method for backing up a distributed system according to claim 4. 7. When it becomes necessary to select a backup destination, the data amount of update history to be collected from another backup computer or the active computer is compared based on the location information. 5. The distributed system backup method according to claim 4, further comprising the step of selecting the smallest backup computer as a new backup destination.