JP2865028B2 - Distributed duplex type fault prevention device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a distributed duplex failure countermeasure device for a processing information system used in a general business entity and a public organization and performing online transaction processing on the premise that service can be performed for 24 hours. The present invention relates to a distributed duplex failure countermeasure device capable of horizontal distribution, a database recovery method thereof, and a data transfer method.

[0002]

2. Description of the Related Art Conventionally, there are mainly two types of failure countermeasures for a distributed processing information system. One is the mainstream method,
Specifically, this is a method of multiplexing system elements having a high frequency of failure occurrence. The other is a method of distributing replicas of center data in distributed devices. Unlike the replica distribution by multiplexing data storage devices, this is a method of explicitly performing replica distribution using communication means.

For example, “Nikkei Electronics No. 6
09 101-110, June 6, 1966, Masahiro Nakamura.

FIG. 9 illustrating this technique is a schematic diagram of an apparatus which realizes a method of multiplexing system elements having a high frequency of occurrence of a failure in the mainstream method. There are two ways to take measures against failures by multiplexing.

[0005] One is a dual type in which the processing contents of the normal processing system and the standby system are simultaneously performed. On the other hand, the standby system performs processing contents different from those of the normal processing system. This is a duplex type in which the standby system performs the processing contents of the normal processing system on behalf of the normal processing system.

FIG. 9 illustrates an apparatus which realizes a duplex system which is conceptually the same as the present invention. This method is an application of the method described in the above document.

[0007] The server function means 151 is a service processing means 1
55, a database realizing means 156, and a replica creating function means 157. Further, in order to multiplex the server function means 151, another server function means 15 is required.
1A, the separate server function means 151A includes the same business processing means 155A as the business processing means 155, the same database realizing means 156A as the database realizing means 156, and the replica creation function means 157. It is configured to include the same replica creation function unit 157A and another business processing unit 158.

In order for a user to make a business request to the system,
When an input request primitive 159 is issued from the client function means 153, the client function means 153
The business processing means 154 existing in the server is activated. The business processing unit 154 issues a processing commission request primitive 160 to the business processing unit 155 in the server function unit 151 via the communication unit 152. When the business processing means 155 receives the processing commission request primitive 160, it issues a transaction request primitive 161 to the database realization means 156 to perform business processing.

The server function means 151 and 151A have the replica creation function means 157 and 157A, respectively.
The data contents of 6A are matched.

When the user issues the input request primitive 159 from the client function means 153 in order to make a business request to the system again, the business processing means 154 existing in the client function means 153 is restarted. The business processing unit 154 is provided with the communication unit 1.
52, the processing commission request primitive 160 is issued to the business processing means 155 in the server function means 151. If a failure occurs in the server function means 151 and an error occurs, the business processing means 154
Re-issues the processing commission request primitive 160 to the business processing means 155A in the separate server function means 151A via the communication means 152, and continues the processing. As a result of the processing, data inconsistency occurring between the database realizing means 156 and the different database realizing means 156A will be described later.
It will be resolved at 57A.

FIG. 10 is a schematic diagram of an apparatus adopting a method of distributing a replica of center data to distributed apparatuses. This method is disclosed in Japanese Patent Application Laid-Open No. 04-274532.
It is described in.

What is defined as a distributed system in the above publication is redefined as client function means 173.
What is defined as the distributed system database in the above publication is redefined as the standby database realizing means 174. What is defined as a transmission line in the above publication is redefined as communication means 172. What is defined as the maintenance system in the above publication is redefined as the server function means 171. What is defined as the maintenance system database realizing means in the above publication is redefined as the database realizing means 175 in the server function means.

When working with the apparatus shown in FIG. 10,
The business processing means 1 via the client function means 173
At 78, an input request primitive 176 is issued. Then, the business processing unit 178 generally issues a request primitive 177 for inquiring of the standby database realizing unit 174, and obtains a result response 179.

The standby database realizing means 174
And the database realizing means 175 in the server function means.
Are required to match, but the database realizing means 175 in the server function means sends the communication means 1
72, the standby database realizing means 174
There is no clear rule on how to send the contents of the database realizing means 175 to the server. Therefore, it is assumed that an implementation procedure according to the following general method is adopted.

A transfer processing unit 180 is provided in the server function unit 171, and a database is realized from the database realizing unit 175 in the server function unit to the standby database realizing unit 174 via the communication unit 172. Sending the contents 186 of the means 175 is implemented.

If the server function means 171 fails, the standby database realizing means replica 174A is transferred from the standby database realizing means 174 in the client function means 173 to the server 174 using the communication means 172. It is created in the function means 171.

When the server function means 171 recovers,
Periodic database realizing means Saved data reading means 18
1, the database realization means data 182 recorded
Is read by the database realizing means creating means 183, and the database realizing means 175 in the server function means is read.
Create A in the state immediately before saving.

The database realizing means comparing means 184 in the server functional means 171 compares the contents of the standby database realizing means replica 174A and the contents of the server functional means database realizing means 175A in the state immediately before the storage, and generates the data immediately after the storage. The database realization means update 185 that has been performed is taken out and a notification of the correction is issued.

[0019]

In the conventional fault elimination device employing the duplex system shown in FIG.
There is a point problem.

The first problem is that the system configuration becomes expensive. The reason is that the server function means 15
1 needs to be multiplexed, so that another server function means 15
It is necessary to install two or more 1A units, and the server function units 151 and 151A are provided with client function units 153.
This is because, in general, the system configuration becomes expensive when the duplex system is adopted, because the system is expensive.

The second problem is that the operation becomes uneconomical. The reason is that in the conventional duplex type, there is no specific index for the multiplexing function, and the server function means 151 and the separate server function means 1
Immediately after the new installation of 51A, the replica creation function 15
7, 157A and the communication means 152, it is necessary to transfer the data content of the database realization means 156 to 156A. In this method, the same operation is necessary even when the reliability of the entire system is high or when the reliability of the entire system is low, and it becomes uneconomical in operation at the initial stage of the system maintenance. .

The apparatus employing the method shown in FIG. 10 has the following three problems.

The third problem is that data consistency cannot be guaranteed in some cases. The reason for this is that in the apparatus shown in FIG. 10, the server unit 171 is regarded as a standby system for maintenance, and a recovery unit when a failure occurs in the server function unit 171 is defined. However, for the client function means 173 which is the main body of the business,
There is no clear method rule and the client function means 17
This is because there is no means for guaranteeing the data consistency of the standby database realizing means 174 inside the 3.

The fourth problem is that, in a distributed processing information system used in a general business entity and a public organization, which performs online transaction processing on the premise that service can be performed for 24 hours, an expensive system is required if the method shown in FIG. 10 is adopted. Configuration. The reason is that the client function unit 173 becomes the main system in the business in the method of FIG. 10, and therefore, a large-scale one is required for the standby database realizing unit 174. is there. This is because each client function means has a large-scale configuration, so that the entire system has an expensive configuration.

The fifth problem is that the operation becomes uneconomical. The reason for this is that, in the method of FIG. 10, there is no clear index for the backup timing and the number of servers in the server function unit 171, and the creation of the replica 174 A of the standby database realizing unit 174 requires the reliability of the entire system. This is because the same operation is required even when the reliability is high or when the reliability of the entire system is low, and it becomes uneconomical in operation at the initial stage of the system maintenance.

The object of the present invention is twofold.

A first object of the present invention is to provide a distributed processing information system for performing online transaction processing on the premise that service can be performed for 24 hours, without providing server function means in duplicate, and without reducing the overall reliability. Is to provide a configuration.

The second purpose is a low priority purpose.
That is, the present invention provides a means for eliminating operation uneconomics by performing operation while measuring the reliability of the entire configuration according to the present invention.

In order to achieve this object, it is necessary that the reliability of the standby client function means is relatively high so that the values derived from the following equations (5) and ( 6) become small. If this condition is not satisfied, this is not the purpose.

[0030]

Distributed duplex type failure countermeasure device of the present invention, in order to solve the problems] includes (a) a database implementation means, the server function unit for managing holding the (b) the database realizing means, (c) the Said database realization means in server function means
Standby client that functions as a terminal when accessing the stage
And Ant functional means, the data contents included in; (d) a database implementing means one
The same data content as that of the standby
Standby database realizing means located in the remote function means
If, (e) the server function unit and communication means for providing access path between the standby client unit, (f) the arranged server function unit the standby database data content of the database implementation means A failure avoidance function processing means for automatically transferring data to the realization means via the communication means periodically; and (g) matching between an update request to the database realization means and an update request to the standby database realization means. Update log means arranged in each of the standby client function means and the server function means, and (h) system time managed in the standby client function means and in the server function means. In order to synchronize with the system time to be managed, in the standby client function means and in the server function means (I) grasping the operating status information of the standby client function unit disposed in the standby client function unit and reporting the operation status information to the server function unit via the communication unit; Operation management processing means; (j) calculating the cumulative operation time of all of the standby client function means after receiving the operation status information from the operation management processing means via the communication means and arranged in the server function means; And (k) a cumulative operation status data management means arranged in the server function means for centrally managing the cumulative operation time information and the reliability information of the standby client function means and the server function means. (L) log information managed by the update log means is distributed and managed, and the database realizing means and the standby system data are managed. And a distributed log management means arranged in each of the database realization means.

The database recovery method of the present invention comprises the steps of (a)
A "log save transfer" step of copying the log content of the server function means to a standby client function means; and (b)
And synchronize "System Time Synchronization" step of the system time in the standby client function unit and the system time in the server function unit, (c) the server function hand
The log information data up to immediately before the stage failure
Obtained from client function means or the server function means
"Automatic log acquisition" step and "Automatic log
If the acquisition fails in the `` obtain
Manually perform the same function as the "Automatic log acquisition" step
And (d) overwriting the log information data obtained in the “automatic log acquisition” step or the “manual log acquisition” step with the data of the standby database realizing means. A "mode pre-processing 1"step; (e) a "failure mode pre-processing 2" step in which a job is processed by the standby client function unit when a failure occurs in the server function unit; and (f). A "server recovery process 1" step of making the data contents of the server function unit sound upon recovery of the failure of the server function unit; and (g) a request for collecting the log information data copied to the standby client function unit. "Temporary update log collection request 1" step to create;
(H) a “temporary update log collection request 2” step of performing simultaneous broadcast in response to the collection request of the “temporary update log collection request 1”; and (i) creating a collection request response to the collection request. A "temporary update log collection response 1"step; (j) a "temporary update log collection response 2" step of transferring the collection request response; and (k) the standby system client function means during recovery of the server function means. and "temporary update limit" step in which the limits database access means for implementing, in the standby system in the client function unit when it becomes unnecessary temporary update log information data contained in the collection request response (l) A "temporary update log deletion" step to be executed; and (m) arranging the temporary update log information data in chronological order within the server function means when the temporary update log information data becomes unnecessary. (Temporary update log editing) step executed to obtain the information, and (n) a "system administrator judgment 1" step of displaying a message for asking the user in the editing executed in the "temporary update log editing" step. And (o) the "system administrator judgment 1"
A "system administrator decision 2" step in which the system administrator inputs decision data after the step is completed; and (p) a completion message is issued in the server function means when the "system manager decision 2" step is completed. "Failure handling mode end processing 1" step and (q) "failure handling mode end processing 2" step of simultaneously broadcasting the completion message of the "failure handling mode end processing 1" step to the standby client function means. And (r) a "failure-handling mode end process 3" step of resuming connection between the standby client function unit and the database realizing unit in the server function unit.

[0032] In the data transfer method of the present invention, in the distributed duplex failure countermeasure apparatus , ( b) a "transfer data generation" step of extracting the data contents from the database realizing means so as to be distributed; A "transfer data transfer" step of transferring the distributed data contents from the server function means including the realizing means to the client function means including the standby database realizing means; and (d) "transferring the data contents". "Data holding" step, (e) "transfer data recovery" step which is executed when an error occurs in the "data transfer" step, and (f) creating contents of the standby database realizing means from the data contents. (Standby system construction) step and (g) the "standby system construction" step And (h) from the client function means including the standby database realizing means to the server function means including the database realizing means, and pull back the same data content and "response data creation" step, (i) the "response data operation
Is executed when transfer is not possible in the
"Response data recovery" switch to failure mode
Step (j) in the above-mentioned "response data creation" step
And when the transfer is completed, the "Response data creation"
"Response" to delete the data contents created in step
A data post-processing "step, is described (k) and" commit test "step of confirming the reciprocating said data content, (l) enforced correctly processing request when the abnormality is not detected in the" commit test "step A "commit authorization transfer" step of transferring the confirmation file to the standby client function means, and (m) a database implementation means deletion request which is executed when an abnormality is detected in the "commit check" step. (A) a “commit-impossible transfer” step of transferring a confirmation file to the standby client function means; and (n) a “commit” operation that is performed after the “commit authorization transfer” step is completed and the corresponding “commit” is received in response to the confirmation file (Measures) step, (o) the "transferred data recovery" step, and the "standby" It includes a “post-processing” step which is performed after the “system construction recovery” step and the “response data recovery” step and arranges the environment in order to start over from the “transfer data creation” step.

[0033]

Next, the present invention will be described with reference to mathematical expressions.

The distributed processing information system to which the present invention is applied must satisfy the following conditions.

The first condition is that the following expressions (1), (2) and (3) are satisfied for the client function means.

[0036]

The second condition is that the following expressions (4), (5) and (6) are satisfied with respect to the elaps time of the transaction processing service started from the client function means.

[0038]

Next, the present invention will be described in detail with reference to the drawings.

In the present invention, the client function means in which the standby system database realizing means is arranged does not need to have the same reliability as the server function means, so it is not necessary to multiplex and introduce the expensive server function means. Absent. Since the same function can be provided by disposing the standby database realizing means which is disassembled on a small scale in the inexpensive client functional means, the first object can be easily realized.

FIG. 1 is a block diagram of a distributed duplex failure countermeasure device according to a first embodiment of the present invention.

FIGS. 2, 3, 4 and 5 are block diagrams showing a transfer method according to the present invention in a normal processing mode. FIG. 7 shows a data transfer according to a second embodiment of the present invention. 5 is a flowchart of a method.

FIG. 6 is a block diagram of a processing procedure in the failure handling mode when a failure occurs.

FIG. 8 is a flowchart of a database recovery method according to the third embodiment of the present invention.

The distributed processing system shown in FIG. 1 comprises a client function unit 3, a server function unit 1, and a communication unit 2.
It is composed of

In this apparatus, the client function means includes:
There are two types of standby client function means 3 and general client function means 4.

The standby client function means 3 includes, in addition to the system input / result display function means 5 as a system terminal function, a standby database realizing means 6,
There are distributed log management means 7 and various recovery function processing means. General client function means 4 described below includes:
Only the system input / result display function 5 and the business processing means 9 as the system terminal function means exist.

When the input request primitive 44 is issued from the standby client function unit 3 and the general client function unit 4 using the system input / result display function unit 5, the business processing unit 9 in the client function unit 4 is executed. It activates and issues a processing commission request primitive 45 to the business processing means 19 in the server function means 1 via the communication means 2 as necessary.

The job processing means 19 in the server function means
Multiplies the database realizing means 20 in the server function means 1 by a transaction request primitive 47 for inquiring or updating the corresponding data, and thereafter receives a transaction result primitive 48. After that, the business function processing means 19 in the server function means sends the result reply primitive 46 to the business processing means 9 in the client function means via the communication means 2.

When the job processing means 19 in the server function means accesses the database realizing means 20 and processes a transaction request primitive 47 that changes the state of the database realizing means 20, the business processing means The means 19 is provided in the server function means 1 and notifies the update log means 22 of the update content primitive 49 in order to record the access contents separately from the administrator processing means 21 for managing the database realizing means 20.

When an error occurs when the update log unit 22 records the access content, the update log unit 22 notifies the business processing unit 19 of a result response primitive 50 indicating the error. As a result, the business processing means 19 rolls back the transaction request primitive 47 and sends a result reply primitive 4 indicating an error to the business processing means 9 in the client function means.
Send 6. If the update log means 22 has successfully recorded the access content, the update log means 22
Notifies the business processing means 19 of the normal result response primitive 50, and in response, the business processing means 19 sends a normal result response primitive 46 to the business function processing means 9 in the client function means.

After recording the access contents, the update log means 22 issues a distributed storage request primitive 51 to the distributed log management means 23 so as to store the log contents in the standby client function means 3 in a distributed manner.

The distributed log management means 23 manages the update log means 22 and selects the additional log information data 52 passed to the distributed log management means 23 from the standby client function means 3 in advance. Transfer to several machines. Thereafter, the distributed log management unit 23 issues a revision request primitive 53 to the distributed log management unit 7 in the standby client function unit 3. Upon receiving the revision request primitive 53, the standby client distributed log management means 7 issues a storage log revision request primitive 53 to the update log means 8 in the standby client function means 3.

Distributed log management means 2 in server function means 1
3 is the minimum number n that satisfies the following expression (7), the number of the standby client function means 3 transferring the additional log information data 52. Since the amount of data recorded in the log is not large, it is not necessary to consider the influence on communication.

[0055]

The time synchronization processing means 24 resides in the server function means 1 and periodically adds a value obtained by adding 5.0 seconds to the system time of the server function means 1 to the standby client function means 3 in the synchronization message. 54. The time synchronization processing means 10 is also provided in the standby client function means 3.
When the synchronization message 54 is received from the server function unit 1, the system clock function unit 18 adjusts the time of the internal system clock function unit 18 to the time described in the synchronization message 54. Then, a synchronization request primitive 55 is sent to 18, and then the apparatus enters a sleep state for a time represented by the following equation (8). Thereby, synchronization with the system time in the server function means 1 is obtained.

[0057]

When the power switch is turned on / off, the standby client function means 3 has the following definition (9).
The operation management processing means 11 for sending the expression operation report message 56 to the server function means 1 is mounted.

[0059]

In the server function means 1, terminal management processing means 25 for managing the accumulated operation status information of all the standby client function means 3 is resident. The terminal management processing means 25 of the server function means 1 transmits the operation report message 5
Upon receipt of No. 6, the cumulative operating time of each standby-system client function unit 3 is immediately calculated, and the result data 57 is updated in the cumulative operating status data file 31.

The server function means 1 has a standby system function existing in the standby system client function means 3 so that even if a failure occurs in the main body of the server function means 1, the work handled by the entire means does not stagnate. A failure avoidance function processing unit 28 that distributes and reflects the contents of the database implementation unit 20 in the server function unit 1 with reference to the cumulative operation status data file 31 is resident in the database implementation unit 6. Works regularly.

When the fault avoidance function processing means 28 starts its operation, it stores the server function means cumulative operation time data 58, the server function means MTBF data 59, the standby client function means cumulative operation time from the cumulative operation state data file 31. With reference to the data 60 and the standby-system client function means MTBF data 61, the standby-system client function means 3 to be distributed and reflected and the number n of the standby-system client function means 3 are as follows:
0) and (11).

[0063]

The reliability of the entire system is as follows (1)
It is set so as to satisfy the expressions 2) and (13).

[0065]

When the standby client function means 3 to be distributed and reflected is determined, the fault avoidance function processing means 28
An export request primitive 62 is issued to the database realizing means 20 in the server function means, the database realizing means content data 63 is taken out, and written to the transfer file 32 in the sequence format.

Thereafter, the transfer file 32 is transferred to the plurality of standby client function means 3 in the step of the "data transfer method" of the present invention described below. FIG. 7 is a flow chart for explaining the steps.

Next, the operation of the embodiment of the "data transfer method" of the present invention will be described with reference to FIG.

Step S1 is a pre-process, in which the number of the standby database realizing means 6 to be created is obtained. As a result, the processes from step S2 to step S14 are continuously performed as many times as the number of creations. This step S1 is referred to as "preprocessing".

In step S1A, the following processing is performed. The server function unit 1 manages the created number of the standby database realizing unit 6, and each time the step S14 described later is performed, the created number is added and updated. At the end of step S1, the created number is set to “0”. In this step, the standby database realizing means 6 obtained in step S1 and the number of already prepared
Are compared with each other, and if they do not match, the process from step S2 described below is performed. If they match, the process ends.

FIG. 2 is an explanatory diagram of step S2. In step S2, the server function means internal fault avoidance function processing means 28 issues an export request primitive 62 to the database realizing means 20 to select the corresponding data from the horizontally distributable table 33, In addition to the data of the file creation semaphore 35,
After the ON message 64 is issued, the transfer file 32 in the sequence format is created. If the transfer file 32 has been successfully created, an off message 65 is issued to turn off the file creation semaphore 35 in order to turn it off. In the server function means 1, there is a file transfer resident processing means 26 for transferring the transfer file 32 to the standby client function means 3. When the transfer file 32 is generated and the file creation semaphore 35 is off, the transfer file 32 is transferred by the file transfer resident processing means 26. This step S2 is referred to as “transfer data generation”.

In step S2A, step S2
It is determined whether or not the transfer has been successful.

An explanatory diagram of step S3 is also included in FIG.
In step S3, if a failure has occurred in the standby client function unit 3 during the transfer of the transfer file 32, an error notification primitive 66 is received from the communication unit 2, so that the file transfer resident processing unit 2
6 transfers a file to another standby client function unit 3. Again, the file transfer resident processing means 2
6 is notified of the error from the communication means 2,
An error response primitive 67 is notified to the fault avoidance function processing means 28. The failure avoidance function processing means 28
Stores the transfer file 32 under the error directory 36 in the server function means 1 and prepares for retransmission later.

If a failure occurs in the communication means 2 and the file transfer resident processing means 26 detects an error,
A reset response primitive 68 is sent to the fault avoidance function processing means 28. Upon receiving the reset response primitive 68, the server function means internal failure avoidance function processing means 28 returns to the state before it was started. This step S3 is called "transfer data transfer".

An explanatory diagram of step S4 is also included in FIG.
In step S4, if the transfer is successfully completed, the normal response primitive 69 is returned from the file transfer resident processing unit 26 to the failure avoidance function processing unit 28 later.
7, the transfer file 32 is stored. (The transfer file 32 has a uniquely defined naming rule with the intention of clarifying the correspondence of the response file 39 described below.
Named. This step S4 is referred to as "transfer data holding"
It is named.

An explanatory diagram of step S5 is also included in FIG.
In step S5, if there is a failure in the standby client function unit 3 and the transfer file 32 cannot be transferred, the failure avoidance function processing unit 28
It will be resent to the standby client function means 3 again. Here, the process skips once to step S15. When resending later, all the transfer files 38 in the error directory 36 are transferred. If the transfer is successful,
The corresponding transfer file 38 is deleted from the error directory 36 and moved to the transfer directory 37. This step S5 is named "transfer data recovery".

FIG. 3 is an explanatory diagram of step S6. In step S6, the transfer monitoring processing unit 12 is resident in the standby client function unit 3, and monitors whether or not the transfer file 32 exists.
If the file transfer processing means 17 does not exist in the standby client function means 3 and the transfer file 32 exists, the transfer monitoring processing means 12 drops the standby database realization means 6 once. Therefore, a drop command 70 is issued to the administrator processing unit 13 existing in the standby client function unit 3.
Then, the administrator processing means 13 issues a delete command 71 to the standby system database 6 and deletes it. Thereafter, in order to re-create the standby database realizing means 6 based on the transfer file 32, the transfer monitoring processing means 12 issues an import request primitive 72 to the administrator processing means 13.
Thereafter, the administrator processing unit 13 issues a creation command 73 for the standby database realizing unit 6. This step S6 is called "standby system construction".

In step S6A, it is determined whether or not the standby system construction in step 6 has been completed.

An explanatory diagram of step S7 is included in FIG.
In step S7, if an error occurs during creation in the standby database realizing means 6, the administrator processing means 13 issues an error notification primitive 74 to the transfer monitoring processing means 12. The transfer monitoring processing means 12 obtains the name 119 of the transfer file 32 to notify the server function means 1 of the error, and then corresponds to the transfer file 32 one-to-one and writes the error content primitive 75. The response file 39 is created, and the process ends. The file transfer processing means 1
7 is resident, the response file 39 is transferred to the server function means 1, and a series of processing is completed. afterwards,
Skip to step S15. The transfer file 32
Has been left as it is, the transfer monitoring processing means 12
Thus, the update is performed again. This step S7 is named "standby system construction recovery".

An explanatory diagram of step S8 is also included in FIG.
In step S8, if the reflection on the standby system database realizing means 6 is successful, the transfer monitoring processing means 12 obtains the binary contents 77A from the transfer file 32 and writes the same contents as the transfer file 32. A response file 39A having a sequence configuration is created. Thereafter, the transfer monitoring processing means 12 transmits the transfer file 3
A delete command 76 is issued in order to delete 2. Thereafter, the resident file transfer processing means 17 obtains the binary data 77 of the response file 39A and transfers it to the server function means 1 via the communication means 2. Thereafter, the transfer monitoring processing unit 12 is in a waiting state until the file transfer processing unit 17 issues a normal response primitive 78 indicating that the response of the confirmation file 40 has been received from the communication unit 2. This step S8 is named "response data creation".

In step S8A, step S8
It is determined whether or not the response transfer has been successful.

An explanatory diagram of step S9 is also included in FIG.
In step S9, when the transfer is not possible due to a failure in the communication means 2 or the server function means 1 at the time of transfer, the file transfer processing means 17 sends an error response primitive 79 to the transfer monitoring processing means 12. Send it. Thereafter, the transfer monitoring processing means 12 immediately stops the process and skips to step S15. As a result, the mode is switched to the failure handling mode. This step S9 is named "response data recovery".

An explanatory diagram of step S10 is also included in FIG. In step S10, the server function means 1
If the transfer is successful, the transfer monitoring processing unit 12 issues the delete command 81 because the response files 39 and 39A of the standby client function unit 3 become unnecessary. This step S10 is referred to as "response data post-processing".

FIG. 4 is an explanatory diagram of this step S11. In step S11, a transfer monitoring processing unit 29 that monitors the existence of the response file 39A and the operation status of the file transfer resident processing unit 26 is resident in the server function unit 1.

The transfer monitoring processing means 29 reads all binary data 80, from the transfer file 32 and the response file 39A in the transfer directory 37 corresponding to the transferred response file 39A.
80A is removed and then binary compared. If the binary data 80 and 80A match, a delete command 81 for the transfer file 32 and a delete command 81A for the response file 39A are issued to delete both the transfer file 32 and the response file 39A.
This step S11 is named "commit check".

In step S11A, step S11
In step 11, it is determined whether the binary data matches, and the process branches to step S12 and step S13.

An explanatory diagram of step S12 is also included in FIG. In step S12, after step S11,
A confirmation file 40 is created, and a normal value response primitive 82 is written therein. Thereafter, the confirmation file 4 is transferred to the standby client function unit 3 for file transfer.
0 name data 83 in the file transfer resident processing means 2
Hand over to 6. This step S12 is named “commit authorization transfer”.

An explanatory diagram of step S13 is also included in FIG. In step S13, the transfer file 32
The transfer file-side all binary data 80 and the response file-side all binary data are respectively extracted from the response file 39A, and as a result of a binary comparison, if there is a discrepancy between the two binary data, the confirmation file 40 And a confirmation file name data 83.
Is transferred to the file transfer resident processing means 26. Therefore, it is considered that there is no series of processes performed from step S2 to step S11, and the process is performed again from step S2 for the other standby client function unit 3 partner. This step S13 is named "uncommittable transfer".

FIG. 5 is an explanatory diagram of step S14. In step S14, the transfer monitoring processing unit 12 in the standby client function unit 3 monitors the existence of the confirmation file 40 and the operation status of the file transfer processing unit 17, and is in a waiting state. The file transfer processing means 17 obtains the confirmation file 40 including the normal value primitive 85, and
Is the normal value primitive 85 from the confirmation file 40
Is read out, the series of processes is terminated, and the process waits for another transfer file 32 again.

The file transfer processing means 17 obtains the confirmation file 40 in which the database realization means deletion request primitive 86 is described, and the transfer monitoring processing means 12 obtains the database realization means deletion request primitive from the confirmation file 40. If you read 86,
The transfer monitoring processing means 12 issues the drop command 87 to the administrator processing means 13 in order to delete the standby database realizing means 6 which has been created, terminates a series of processing, and again executes the transfer file 3.
Wait two. Thereafter, the administrator processing means 13 issues a delete command 88 to the standby system database realizing means 6. This step S14 is named “commit handling”.

Step S15 is a post-processing of the forced termination performed in step S5, step S7, and step S9. As a result, the process returns to the determination processing of step S1A. In step 15, the environment is organized to start over from step S2. This step S15 is referred to as "post-processing".

As a result of the transfer method in the above steps, the standby database realizing means 6, which is a subset of the database realizing means 20 in the server functional means 1, is always created in the standby client functional means 3.

FIG. 1 shows the subsequent procedure again.
For all standby client function means 3 to be sent,
In the above steps, the transfer file 32 is transferred. After that, the failure avoidance function processing means 28 in the server function means 1 sends a clear request primitive 89 to the distributed log management means 23 in order to clear the additional log information data 52 previously taken. The distributed log management means 23
Receives the clear request primitive 89, issues a log clear request primitive 90 to the update log means 8 via the distributed log management means 7 in the standby client function means 3, and resets the log.

FIG. 6 shows a failure handling mode when a failure occurs. When the general client function unit 4 issues an input request primitive 44 using the system input / result display function unit 5, the job processing unit 9 in the client function unit is activated. Then, the business processing means 9
Attempts to issue a processing commission request primitive 45 to the business processing unit 19 of the server function unit 1 via the communication unit 2, but the server function unit 1 has failed or the communication unit 2 has failed. In this case, the task processing means 9 in the general client function means 4 reports the error information primitive 91 to the system input / result display function means 5 and urges the standby client function means 3 to work.

When the standby client function means 3 issues an input request primitive 44 via the system input / result display function means 5, the business processing means 9 in the standby client function means 3 is activated. An attempt is made to issue a processing commission request primitive 45 to the business processing unit 19 of the server function unit 1 via the communication unit 2, but a failure of the server function unit 1 or a failure of the communication unit 2 The task processing means 9 in the standby client function means 3 detects an error and issues a start command 92 for starting the recovery processing means 14.

The failure of the server function means 1 is dealt with by the "database recovery method" of the present invention.

FIG. 8 is a flowchart showing the processing steps of the database restoration method according to the present invention. Hereinafter, each processing step of the database recovery method will be described. Step A1 and step A2 will be described with reference to FIG.
Step A3 and subsequent steps will be described with reference to FIG.

Step A1 will be described with reference to FIG. In step A1, the distributed log management unit 23 manages the update log unit 22 and selects the additional log information data 52 passed to the distributed log management unit 23 from among the standby client function units 3 in advance. Transfer to several units. Thereafter, the distributed log management unit 23 requests the update log unit 8 in the standby client function unit 3 to update the storage log via the distributed log management unit 7 in the standby client function unit 3. This step A1 is named "log save transfer".

Step A2 will be described with reference to FIG. In step A2, the time synchronization processing means 24 is resident in the server function means 1, and periodically transfers the synchronization message 54 to the standby client function means 3. The time synchronization processing means 10 also resides in the standby client function means 3 and synchronizes with the system time in the server function means 1. This step A2 is named “system time synchronization”.

Step A3 will be described with reference to FIG. In step A3, the recovery processing unit 14 issues a request command 93 to the update log unit 8 and a request command 93A to the distributed log management unit 7 in the standby client function unit 3, and the server function unit 1 fails. The log information data 94 up to just before the time is obtained from the other standby client function unit 3 or the server function unit 1.
This step A3 is named "automatic log acquisition".

In step A3A, it is determined whether or not automatic log acquisition has succeeded, and the flow branches to step A4 and step A5.

Step A4 will be described with reference to FIG. In step A4, when a failure occurs in the communication unit 2 in step A3, the recovery processing unit 14 issues a communication unit failure notification message 95 to the system input / result display function unit 5, and provides the user with: The magnetic medium 43 prompts the user to obtain the log information data 94A. When the log information data 94A is obtained from the magnetic medium 43, the loader processing unit 15 in the standby client function unit 3
The log information data 94A is sucked up from the magnetic medium 43, read into the standby client function unit 3, and the log information data 94A is delivered to the recovery processing unit 14. Meanwhile, the job processing means 9 in the standby client function means 3 enters a waiting state. This step A4 is named "manual log acquisition".

In step A5, upon obtaining the log information data 94 or 94A, the recovery processing means 14 in the standby-system client function means 3 starts importing the log information data 94 or 94A from the standby-system database realizing means 6 and reflects the data. The transaction request all primitives 97 to the database realizing means 20 in the server functional means 1 added later are issued to the standby database realizing means 6, and the contents of the database realizing means 20 in the server functional means 1 are The state is shifted to a state in which the standby system database realizing means 6 is horizontally distributed, and the state shifts to the failure handling mode. At this time, the recovery processing unit 14 issues a creation command 99 for creating the temporary update log 98 in the standby client function unit 3. Step A5 is referred to as “failure-handling mode preprocessing 1”.

In step A6, in the failure handling mode, the business processing means 9 in the standby client function means 3 is present in the standby client function means 3 and the business processing means 9 in the server function means 1 In order to reconnect to the standby database realizing means 6 via the standby client function means business processing means 118 which shows the same behavior as the means 19, the processing commission request primitive 10
0 is reissued to the job processing means 118. Each time the task processing means 118 issues the transaction request primitive 47 for changing the state to the standby system database realizing means 6 and completes it, the temporary update log 98 records a record message 101 of the content and time. . Step A6 is referred to as “failure-handling mode preprocessing 2”.

In step A7, when the server function means 1 in the faulty state is restored, the administrator processing means 21 for managing the database realizing means 20 in the server function means 1 immediately starts up and has the inside of the administrator processing means 21. An update request primitive 103 is issued from the log 102 to the database realizing means 20 to reproduce the state immediately before the failure of the server function means 1. This step A7 is referred to as “server recovery processing 1”.
It is named.

In step A8, after step A7, the update log means 22 starts up continuously, and the distributed log management means 23 sends the temporary update log collection message 104
Send. This step A8 is named “temporary update log collection request 1”.

In step A9, after step A8, the distributed log management means 23 transmits the temporary update log collection message 105 by simultaneous broadcasting of the distributed log management means 7 in the standby client function means 3. This step A9 is named “temporary update log collection request 2”.

In step A 10, upon receiving the temporary update log collection message 105, the distributed log management unit 7 in the standby client function unit 3 sends the temporary update log information to the resident failure recovery function processing unit 16. The message 106 is sent. This step A10 is named “temporary update log collection response 1”.

In step A11, when the failure recovery function processing means 16 receives the temporary update log information message 106, it extracts the content data 107 of the temporary update log 98, and then sends it to the update log processing means 22 of the server function means 1. Then, the content data 107 is transferred to a file. This step A11 is referred to as "temporary update log collection response 2".
It is named.

In step A12, the failure recovery function processing means 16 in the standby client function means 3 sets the service processing means 9 in the standby client function means 3 to
Is a newly updated transaction request primitive 4
7 is issued to the administrator processing means 13 so as to temporarily strengthen the database access authority in the standby database realizing means 6 so as not to issue the restriction request primitive 7. As a result, only the transaction request primitive 47 for search can be issued to the standby database realizing means 6. At the same time, the failure recovery function processing means 16
A warning message 109 is issued to the system input / result display function means 5 of the standby client function means 3, issued from the update log means 22 in the server function means 1, and the distributed log management in the standby client function means 3 is issued. This state is maintained until the recovery completion message 110 received from the means 7 is received. This step A12 is named “temporary update restriction”.

In step A13, upon receiving the recovery completion message 110, the failure recovery function processing means 16 issues a command 111 for erasing the temporary update log 98 in the standby client function means 3, and issues a command to the server function means. 1 and reconnect to the database realizing means 20. This step A13 is named "temporary update log deletion".

At step A14, the system time of all the standby client function means 3 is set at step A1.
Is synchronized. The update log means 22 in the server function means 1 issues a start command 112 to the recovery processing means 27. When the recovery processing unit 27 obtains the content data 107 of the temporary update log 98 in all the standby client function units 3 via the update log unit 22, the recovery processing unit 27 The data 107 is rearranged in time order. This step A14
Is named "temporary log editing".

In Step A15, Step A1
After 4, a modified transaction request primitive 114 is issued so as to be reflected in the database implementing means 20 in the server function means 1. During the reflection, if there is any of the following rules 1) and 2), the recovery processing unit 27 issues a message 115 to the console 30 of the server function unit 1 to determine the system administrator. obtain. The default countermeasures are described below. This step A15 is referred to as “system administrator judgment 1”.
It is named.

Rules 1) In the data contents of the database realizing means 20 in the server function means 1 distributed to different standby client function means 3, a transaction request for prompting a change in the value of the same data in the same table separately. When the primitive 47 is issued or the transaction request primitive 47 indicating deletion is multiplied by the transaction request primitive 47 for prompting a change of a value.

Countermeasure 1) If the transaction request primitive 47 for prompting the change of the value is independently requested twice or more, the last transaction request primitive 47 in time is prioritized. When two of the transaction request primitive 47 indicating deletion and the transaction request primitive 47 prompting a change of a value are requested, the transaction request primitive 47 indicating deletion is given priority.

Rule 2) When data of the same primary key is to be inserted into the same table in the data contents of the database realizing means 20 in the server function means 1 distributed to different standby client function means 3.

Countermeasure 2) The last transaction request primitive 47 requested in terms of time is selected.

In step A16, the system administrator inputs a screen selection message 116 to the console 30 of the server function unit 1 as a result of the above determination. The following four types of screen selection messages 116 can be selected. Option 1) “Ignore” Option 2) “Countermeasure processing” Option 3) “Countermeasure with log processing” so that the user can manually correct later Option 4) “Immediate manual correction” 2 ".

In step A17, after the system administrator inputs the screen selection message 116 on the console 30 and the processing contents of the restoration processing means 27 are completed, if there is no problem there, The recovery processing unit 27 reports a completion message 117 to the update log unit 22. This step A17 is referred to as “troubleshooting mode end processing 1”.

In step A18, the update log means 22 in the server function means 1 issues a recovery completion message 110A to the distributed log management means 23 in the server function means. Distributed log management means 23 in the server function means
When the recovery completion message 110A is obtained, the distributed log management unit 7 in the standby client function unit 3
Then, a recovery completion message 110B is notified by simultaneous broadcast. When the distributed log management unit 7 in the standby client function unit 3 receives the recovery completion message 110B, the recovery completion message 110 is delivered to the failure recovery function processing unit 16. Step A18 is referred to as "failure-handling mode end processing 2".

In step A19, the processing means 16 of the failure recovery function sets the standby client function means 3
In order to delete the temporary update log 98, a delete command 111 is issued to reconnect to the database realizing means 20 in the server function means 1 and return to the state before the failure occurred. Step A19 is referred to as “failure-handling mode end processing 3”.

[0122]

As described above, the effects achieved by the present invention are as follows.

First, the client function means in which the standby system database realizing means are arranged in the present invention does not need to have the same reliability as the server function means, so the expensive server function means is multiplexed and introduced. do not have to. Therefore, the same function can be provided by arranging the standby database realizing means, which is disassembled on a small scale, in the inexpensive client functional means.

Second, since it is not necessary to multiplex and introduce the expensive server function means, there is an effect that the operation environment is simplified and the operation becomes easy.

[Brief description of the drawings]

FIG. 1 is a block diagram of an embodiment of a distributed duplex failure countermeasure device of the present invention.

FIG. 2 is a block diagram showing a processing procedure when a transfer file is transferred from a server function unit.

FIG. 3 is a block diagram illustrating a processing procedure when a response file is transferred from a standby client function unit.

FIG. 4 is a block diagram showing a processing procedure when a confirmation file is transferred from the server function means side.

FIG. 5 is a block diagram illustrating a processing procedure when a confirmation file is received by the standby client function unit.

FIG. 6 is a block diagram illustrating a processing procedure of a database recovery method including functional elements.

FIG. 7 is a flowchart of one embodiment of a transfer method according to the present invention.

FIG. 8 is a flowchart of one embodiment of a database recovery method according to the present invention.

FIG. 9 is a block diagram showing a conventional failure countermeasure device for a distributed processing information system.

FIG. 10 is a block diagram showing a conventional failure countermeasure device for a distributed processing information system.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 Server function means 2 Communication means 3 Standby client function means 4 General client function means 5 System input / result display function means 6 Standby database implementation means 7 Standby client function means distributed log management means 8 Standby client function means Update log means 9 Task processing means in client function means 10 Time synchronization processing means in standby client function means 11 Operation management processing means 12 Transfer monitoring processing means in standby client function means 13 Administrator processing means in standby client function means 14 Standby Recovery processing means in the standby client function means 15 loader processing means in the standby client function means 16 failure recovery function processing means in the standby client function means 17 file transfer processing means in the standby client function means 18 system Clock function means 19 Business processing means in server function means 20 Database realizing means in server function means 21 Administrator processing means in server function means 22 Update log means in server function means 23 Distributed log management means in server function means 24 Time in server function means Synchronization processing means 25 Terminal management processing means 26 File transfer resident processing means in server function means 27 Recovery processing means in server function means 28 Failure prevention function processing means 29 Transfer monitoring processing means in server function means 30 Server function means console 31 Cumulative operation Situation data file 32 Transfer file 33 Horizontally distributable table 34 Master table 35 File creation semaphore 36 Error directory 37 Transfer directory 38 Transfer file in error directory 39 Response file 39A Response file 40 Confirmation File 43 Magnetic Medium 44 Input Request Primitive 45 Processing Outgoing Request Primitive 46 Result Answer Primitive 47 Transaction Request Primitive 48 Transaction Result Primitive 49 Update Content Primitive 50 Result Answer Primitive 51 Distributed Storage Request Primitive 52 Additional Log Information Data 53 Revision Request Primitive 54 Synchronization message 55 Synchronization request primitive 56 Operation report message 57 Result data 58 Server function means accumulated operation time data 59 Server function means MTBF data 60 Standby client function means accumulated operation time data 61 Standby client function means MTBF data 62 Export request primitive 63 Database implementation means content data 64 On message 65 Off message Page 66 Error notification primitive 67 Error response primitive 68 Reset response primitive 69 Normal response primitive 70 Drop command 71 Delete command 72 Import request primitive 73 Creation command 74 Error notification primitive 75 Error content primitive 76 Delete command 77 Response file Binary data 77A Transfer file Binary contents 78 Normal response primitive 79 Error response primitive 80 All binary data on transfer file side 80A All binary data on response file 81 Transfer file side delete command 81A Response file side delete command 82 Normal value response primitive 83 Confirmation file name data 84 Database implementation means Delete request primitive 85 Normal value primitive 86 Database Implementing means deletion request primitive 87 Drop command 88 Delete command 89 Clear request primitive 90 Log clear request primitive 91 Error information primitive 92 Startup command 93 Request command 93A Request command 94 Log information data until immediately before failure of server function means 94A Magnetic media Log information data written until immediately before the failure of the server function means 95 Communication line failure notification message 97 All transaction request primitives 98 Extraordinary update log 99 Creation command 100 Processing commission request primitive 101 Record message 102 Log inside the administrator means 103 Update request Primitive 104 Temporary update log collection message 105 Temporary update log collection message 106 Temporary update log information message 107 Update log content data 108 Restriction request primitive 109 Warning message 110 Recovery completion message 110A Recovery completion message 110B Recovery completion message 111 Command to delete the temporary update log 112 Start command 114 Modified transaction request primitive 115 Message 116 Screen selection message 117 Completion message 118 Wait Business processing means in system client function means 119 Transfer file name 151 Server function means 151A Server function means 152 Communication means 153 Client function means 154 Business processing means 155 Business processing means 155A Business processing means 156 Database realizing means 156A Database realizing means 157 Replica Creation function means 157A Replica creation function means 158 Separate business processing means 159 Input request primitive 160 Processing commission request primitive 161 Transaction request primitive 171 Server function unit 172 Communication unit 173 Client function unit 174 Standby database realizing unit 174A Standby database realizing unit replica 175 Database realizing unit in server functional unit 175A State immediately before storage 176 input request primitive 177 request primitive 178 business processing means 179 result response 180 transfer processing means 181 periodic database realization means storage data reading means 182 database realization means data 183 database realization means creating means 184 database realization means Comparison means 185 database realization means update 186 database realization means Contents

Continuation of the front page (51) Int.Cl. ⁶ Identification code FI G06F 12/00 531 G06F 12/00 531J 12/16 310 12/16 310J 13/00 357 13/00 357Z 15/16 470 15/16 470J ( 58) Fields investigated (Int.Cl. ⁶ , DB name) G06F 15/00 320 G06F 1/00 370 G06F 11/20 310 G06F 11/34 G06F 12/00 531 G06F 12/16 310 G06F 13/00 357 G06F 15/16 470

Claims (3)

(57) [Claims] (A) a database realizing means; (b) a server function means for holding and managing the database realizing means; and (c) a database realizing means in the server functional means.
Standby client that functions as a terminal when accessing the stage
And Ant functional means, the data contents included in; (d) a database implementing means one
The same data content as that of the standby
Standby database realizing means located in the remote function means
If, (e) the server function unit and communication means for providing access path between the standby client unit, (f) the arranged server function unit the standby database data content of the database implementation means A failure avoidance function processing means for automatically transferring data to the realization means via the communication means periodically; and (g) matching between an update request to the database realization means and an update request to the standby database realization means. Update log means arranged in each of the standby client function means and the server function means, and (h) system time managed in the standby client function means and in the server function means. In order to synchronize with the system time to be managed, in the standby client function means and in the server function means (I) grasping the operating status information of the standby client function unit disposed in the standby client function unit and reporting the operation status information to the server function unit via the communication unit; Operation management processing means; (j) calculating the cumulative operation time of all of the standby client function means after receiving the operation status information from the operation management processing means via the communication means and arranged in the server function means; And (k) a cumulative operation status data management means arranged in the server function means for centrally managing the cumulative operation time information and the reliability information of the standby client function means and the server function means. (L) log information managed by the update log means is distributed and managed, and the database realizing means and the standby system data are managed. And a distributed log management means arranged in each of the database realization means. 2. A database implementation means, the server function unit that manages to hold the database implementation means, the
Access to the database realization means in the server function means
Standby client function that functions as a terminal when running
Means and a data content included in the database realizing means.
Save the same data content as a part of the standby client
A standby database realization method that is placed in the
And stage, said communication and communication means for providing an access path, the arranged server function unit the data contents of the database implementation means in the standby database implementation means between said standby client unit and the server function unit A failure avoiding function processing means for automatically transferring data periodically through the means; and a standby client which ensures consistency between an update request to the database realizing means and an update request to the standby database realizing means. In order to synchronize the update log means arranged in each of the function means and the server function means with the system time managed in the standby client function means and the system time managed in the server function means, Times allocated in the standby client function unit and in the server function unit, respectively Terminating function means, operation management processing means arranged in the standby client function means, grasping operation status information of the standby client function means, and reporting to the server function means via the communication means, and the server function Terminal management processing means arranged in the means for calculating the cumulative operation time of all of the standby client function means after receiving the operation status information from the operation management processing means via the communication means, and in the server function means A cumulative operation status data management unit that centrally manages the cumulative operation time information and the reliability information of the standby client function unit and the server function unit, and distributes log information managed by the update log unit. Distributed log pipes managed and arranged in each of the database realizing means and the standby database realizing means Means for restoring a database for a distributed duplex failure countermeasure device, comprising: (a) a "log save transfer" step of copying the log content of the server function means to a standby client function means; and (b) A "system time synchronization" step for synchronizing the system time in the server function means with the system time in the standby client function means; and (c) log information until immediately before the failure of the server function means.
The data is transferred to the standby client function means or the
“Get automatic log” step obtained from server function
In the “Automatic log acquisition” step
If the hand fails, the same as the “automatic log acquisition” step
`` Manual log acquisition '' step to perform one function manually
And flop, and; (d) "automatic log shelf" step or the "Manual Log obtain" the log information data obtained in step overwrites the data of the standby database implementation means "troubleshooting mode preprocessing 1" step (E) a "failure-handling mode pre-processing 2" step in which a job is processed by the standby client function unit when a failure has occurred in the server function unit; and (f) a failure recovery of the server function unit. "Server recovery processing 1" that sometimes makes the data contents of the server function means sound
(G) a "temporary update log collection request 1" step of creating a collection request for the log information data copied to the standby client function means; and (h) the "temporary update log collection request 1". A "temporary update log collection request 2" step of performing simultaneous broadcast in response to the collection request, and (i) a "temporary update log collection response 1" step of creating a collection request response to the collection request. A) "temporary update log collection response 2" step of transferring the collection request response; and (k) access to the database realizing means in the standby client function means during restoration of the server function means. to limit the cell Graphics and "extraordinary update limit" step, enforcement by the standby system in the client function means in the case that has become unnecessary extraordinary update log information data contained in the collection request response (l) "Temporary update log information deletion"step; and (m) "temporary update performed when the temporary update log information data is no longer needed, to rearrange the temporary update log information data in chronological order in the server function means. (N) a "system administrator judgment 1" step of displaying a message for asking the user for the editing performed in the "temporary update log editing"step; and (o) the "system management". A "system administrator decision 2" step in which a system administrator inputs decision data after the "user decision 1" step is completed; and (p) a completion message is sent to the server function when the "system manager decision 2" step is completed. A "troubleshooting mode end process 1" step issued in the means; and (q) the "troubleshooting mode end process 1" step. A "failure handling mode end process 2" step of simultaneously broadcasting an end message to the standby client function means; and (r) restarting the connection between the standby client function means and the database realizing means in the server function means. And a "failure handling mode end process 3" step. 3. A database implementation means, the server function unit that manages to hold the database implementation means, the
Access to the database realization means in the server function means
Standby client function that functions as a terminal when running
Means and a data content included in the database realizing means.
Save the same data content as a part of the standby client
A standby database realization method that is placed in the
And stage, said communication and communication means for providing an access path, the arranged server function unit the data contents of the database implementation means in the standby database implementation means between said standby client unit and the server function unit A failure avoiding function processing means for automatically transferring data periodically through the means; and a standby client which ensures consistency between an update request to the database realizing means and an update request to the standby database realizing means. In order to synchronize the update log means arranged in each of the function means and the server function means with the system time managed in the standby client function means and the system time managed in the server function means, Times allocated in the standby client function unit and in the server function unit, respectively Terminating function means, operation management processing means arranged in the standby client function means, grasping operation status information of the standby client function means, and reporting to the server function means via the communication means, and the server function Terminal management processing means arranged in the means for calculating the cumulative operation time of all of the standby client function means after receiving the operation status information from the operation management processing means via the communication means, and in the server function means A cumulative operation status data management unit that centrally manages the cumulative operation time information and the reliability information of the standby client function unit and the server function unit, and distributes log information managed by the update log unit. Distributed log pipes managed and arranged in each of the database realizing means and the standby database realizing means Means, in the data transfer method for distributed duplex type disorders countermeasure device having, (b) and "transfer data generation" retrieving As can distribute the data content from the database implementation means, (c) said database realized A "transfer data transfer" step of transferring the distributed data contents from the server function means including the means to the client function means including the standby database implementing means; and (d) "transfer data holding the data contents". (E) a "transfer data recovery" step which is performed when an error occurs in the "data transfer"step; and (f) a "standby" for creating the contents of the standby database realizing means from the data contents. (G) establishing the standby system. A "standby system construction recovery" step to be executed when an error occurs in (a), and (h) the data contents from the client function means including the standby database realizing means to the server function means including the database realizing means. and "create response data" step of pulling back the same data content, it is transferred in (i) the "response data creation" step
It is enforced when failed and switches to failure handling mode
And "Response Data Recovery" step, is transferred in (j) the "response data creation" step
It is enforced when it can be done, and in the "response data creation" step
"Response data post processing" to delete the created data contents
And management "step, have been described (k) and" commit test "step of confirming the reciprocating said data content, (l) enforced correctly processing request when the abnormality is not detected in the" commit test "step A "commit authorization transfer" step of transferring a confirmation file to the standby client function means; and (m) a confirmation file which is executed when an abnormality is detected in the "commit check" step and contains a database implementation means deletion request. And (n) transfer after commit of the “commit authorization transfer” step, which is performed after the completion of the “commit authorization transfer” step. (O) the "transferred data recovery" step and the "standby system construction" Cavalli "step and said" enacted after the response data recovery "step the" Create Transfer Data "
"Post-processing" to organize the environment to start over from the step
A data transfer method, comprising: