JP2017068668A - Database system, replication controller, replication method and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To suppress reduction of throughput, and create a rest point of a database in which a start time of business processing is reference.SOLUTION: A database system 10 comprises: a master database 12 (hereinafter, described a master DB); a replica database 14 (hereinafter, described a replica DB); a serial number numbering part 16; an update log creation part 18; a non-synchronous processing part 20; and a control part 22. The database system 10 executes replication processing in which a start time of business processing is reference. Therefore, in a non-synchronous to data update which is performed on the master DB 12, the data update can be reflected on the replica DB 14. The database system 10 stops the replication processing when a prescribed condition is satisfied.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a database system, a replication control device, a replication method, and a program.

  2. Description of the Related Art Conventionally, systems for managing business process data (hereinafter referred to as business data) performed online using a database are known. In such a system, a master database and a backup replica database are often used to store business data. For example, when a failure occurs in the master database due to a disk failure or the like, the master database is disconnected from the system, and the replica database is incorporated into the system as a new master database. As a result, the system can be immediately restored. During normal operation, the update contents of the business data in the master database are reflected in the backup replica database as needed or periodically.

  By the way, when business data is managed using a database, there is often a demand to create a quiesce point of the database at a certain point in time. For example, when statistical data of business data for the previous day is required, it is conceivable to create a quiesce point of the database at the timing when the date changes. In this case, for example, the statistical data of the previous day's work can be acquired from the quiesce point of the database by batch processing.

  A method for creating a stationary point of a database is disclosed in, for example, Patent Document 1. In the system disclosed in Patent Document 1, the update of the database is stopped by temporarily interrupting the online business process. Thereby, a quiesce point of the database can be created. However, in the above method, since it is necessary to interrupt the online business process, the availability of the system is lowered.

  On the other hand, Patent Document 2 discloses a method capable of creating a quiesce point of a database without stopping online business processing. In the system disclosed in Patent Document 2, the database is composed of a plurality of storage areas. In this system, the latest data is stored in one of the plurality of storage areas, and data at a certain point in the past is stored in the other one. For this reason, in the system of patent document 2, it is not necessary to stop the update of a database in order to create a stationary point. Therefore, it is possible to create a quiesce point of a database without stopping online business processing.

Japanese Patent Laid-Open No. 5-314071 JP-A-10-143411

By the way, when managing a plurality of business processes, statistical processing of business data may be performed based on the start time of each business process. In order to perform such statistical processing, it is necessary to create a quiesce point of the database on the basis of the start time of business processing, as will be described later with reference to FIG. On the other hand, in the conventional system as disclosed in Patent Document 2, the quiesce point of the database is created based on the update time of the database. For this reason, in the conventional system as described above, the problem described below arises when a database quiesce point is created based on the start time of business processing.

  FIG. 12 is a diagram for explaining a problem that occurs when a conventional system tries to create a quiesce point of a database based on the start time of business processing. FIG. 12 illustrates a case where the processing (transaction) of business A, business B, and business C is performed online. Note that business A, business B, and business C are processed independently of each other.

  In FIG. 12, the time required for the processing of the business A, the business B, and the business C is indicated by the length of the arrow. Therefore, in the example of FIG. 12, the processing time of the business A is the longest and the processing time of the business C is the shortest. Further, in the example of FIG. 12, the processing of the business A is started on January 1, and the processing of the business B and C is started on January 2. The process for business C is started after the process for business B. The business A process is completed after the business B process, and the business B process is completed after the business C process. Therefore, when the operations A, B, and C are arranged in time series based on the start time of the operation processing, the operations A, operations B, and operations C are performed in this order. On the other hand, when the operations A, B, and C are arranged in time series based on the time when the operation data is updated in the database, the operations C, operation B, and operation A are performed in this order. In other words, in the example of FIG. 12, the processing A, B, and C processing times are different from each other, and the processing A, B, and C processing data start order is different from the business A, B, and C business data update orders. ing.

  Here, for example, it is assumed that there is a request to acquire a quiesce point of a database reflecting all the update contents of the business process started on January 1st. In this case, in the example of FIG. 12, it is necessary to reflect only the updated content of the business A and not the updated content of the business B and C in the stationary point.

  However, as described above, in the conventional system, a quiesce point is created based on the update time of the database. For this reason, the conventional system can create a quiesce point that indicates the state of data at a certain point in the database, but cannot create a quiesce point that satisfies the above requirements. Specifically, in the case of the example in FIG. 12, it is possible to create a quiesce point that reflects only the updated content of the business B or a quiesce point that reflects only the updated content of the business B and C. However, it is not possible to create a quiesce point that reflects only the updated contents of job A.

  In order to create a quiesce point that reflects only the update contents of the business A by the conventional system, the business data of the business B and C are updated after the business data of the business A is updated. It is necessary to control the order of the update timing. However, in this case, the processes B and C cannot be completed until the process A is completed, and the system throughput decreases.

  An example of an object of the present invention is to provide a database system, a replication control apparatus, and a replication method capable of creating a quiesce point of a database on the basis of the start time of business processing without solving the above problem and reducing the throughput And providing a program.

In order to achieve the above object, a database system according to an aspect of the present invention includes a master database in which data is updated based on a plurality of transactions, a replica database serving as a replication destination of the master database, and start of the plurality of transactions. Generating a plurality of update logs corresponding to the plurality of transactions, and a plurality of update logs and the serial number generation unit, Corresponding to the plurality of update logs based on the plurality of update logs generated by the update log generation unit and the update log generation unit for associating serial numbers assigned to a plurality of transactions According to the serial number assigned, the master data An asynchronous processing unit that asynchronously updates the data in the replica database, and a control unit that stops data updating in the replica database by the asynchronous processing unit when a predetermined condition is satisfied. Prepare.

  In order to achieve the above object, a replication control apparatus according to one aspect of the present invention includes a master database in which data is updated based on a plurality of transactions, and a replica database serving as a replication destination of the master database. The replication control device performs data update of the replica database asynchronously with data update of the master database. The replication apparatus generates a plurality of update logs corresponding to the plurality of transactions, a serial numbering unit for assigning serial numbers to the plurality of transactions based on start times of the plurality of transactions, An update log generator that associates the update log with the serial numbers assigned to the plurality of transactions by the serial number generator, and the update log generator based on the plurality of update logs generated by the update log generator In accordance with the serial numbers associated with the plurality of update logs by the log generation unit, an asynchronous processing unit that performs data update of the replica database asynchronously with data update of the master database, and when a predetermined condition is satisfied And stop the update of data in the replica database by the asynchronous processing unit And a control unit.

  In order to achieve the above object, a replication method according to one aspect of the present invention is a database system including a master database in which data is updated based on a plurality of transactions, and a replica database serving as a replication destination of the master database. In this replication method, the data in the replica database is updated asynchronously with the data update in the master database. The replication method includes (a) a step of assigning serial numbers to the plurality of transactions on the basis of start times of the plurality of transactions, and (b) generating a plurality of update logs corresponding to the plurality of transactions, and A step of associating the plurality of update logs with the serial numbers assigned to the plurality of transactions in the step (a), and (c) based on the plurality of update logs generated in the step (b). And, in accordance with the serial numbers associated with the plurality of update logs in the step (b), updating the data in the replica database asynchronously with the data update in the master database, and (d) a predetermined value When the condition is met, the replica database data update is stopped. Tsu and a flop.

Furthermore, in order to achieve the above object, a program according to one aspect of the present invention is a database system including a master database in which data is updated based on a plurality of transactions, and a replica database serving as a replication destination of the master database. A program for causing a computer to execute data update of the replica database asynchronously with data update of the master database. The program includes: (a) assigning serial numbers to the plurality of transactions on the basis of start times of the plurality of transactions; and (b) generating a plurality of update logs corresponding to the plurality of transactions. And a step of associating the plurality of update logs with the serial numbers assigned to the plurality of transactions in the step (a), and (c) the plurality of update logs generated in the step (b). And updating the data in the replica database asynchronously with the data update in the master database according to the serial numbers associated with the plurality of update logs in the step (b), and (d) ) Update data in the replica database when a predetermined condition is met And a step of stopping.

  It is possible to provide a database system, a replication control device, a replication method, and a program capable of creating a quiesce point of a database on the basis of the start time of business processing without reducing the throughput.

FIG. 1 is a block diagram showing a schematic configuration of a database system according to an embodiment of the present invention. FIG. 2 is a block diagram specifically showing the configuration of the database system. FIG. 3 is a diagram illustrating an example of a serial number management table stored in the serial number management unit. FIG. 4 is a diagram illustrating an example of a processing data storage table stored in the processing data storage unit. FIG. 5 is a flowchart showing an example of the basic operation of the database system. FIG. 6 is a diagram showing the processing time of the business application and the asynchronous process control application, and the serial numbering timing by the serial numbering unit. FIG. 7 is a diagram illustrating an example of processing contents (asynchronous processing data) stored in the processing data storage table. FIG. 8 is a flowchart showing an example of the recovery operation of the database system. FIG. 9 is a diagram showing a modification of the database system. FIG. 10 is a diagram illustrating an example of processing contents registered in the processing data storage table by the control unit. FIG. 11 is a block diagram illustrating an example of a computer that implements the replication control apparatus according to the embodiment of the present invention. FIG. 12 is a diagram for explaining a problem that occurs when a conventional system tries to create a quiesce point of a database based on the start time of business processing.

(Summary of Invention)
The present invention can be used, for example, when data update to a database (master database) used in online business processing is replicated to a replica database for backup. Specifically, in an example of the present invention, replication processing to the replica database is executed in the order of serial numbers determined by the transaction processing unit of each business, not the order of updating the business data of each business to the master database. Thereby, a quiesce point based on the start time of the business process can be acquired from the replica database. In an example of the present invention, replication processing can be stopped from a business application. Thereby, the replication process can be stopped at an arbitrary timing desired by the user. As a result, an arbitrary quiesce point desired by the user can be acquired from the replica database.

(Embodiment)
Hereinafter, a database system, a replication control device, a replication method, and a program according to an embodiment of the present invention will be described with reference to FIGS.

[System configuration]
FIG. 1 is a block diagram showing a schematic configuration of a database system according to an embodiment of the present invention. Referring to FIG. 1, a database system 10 according to the present embodiment includes a master database (hereinafter referred to as a master DB) 12, a replica database (hereinafter referred to as a replica DB) 14, a serial number numbering unit 16, an update. A log generation unit 18, an asynchronous processing unit 20, and a control unit 22 are provided.

  Although details will be described later, the database system 10 can execute the replication process based on the start time of the business process (transaction). Thereby, the database system 10 can reflect the data update in the replica DB 14 asynchronously with the data update performed in the master DB 12. Further, the database system 10 stops the replication process when a predetermined condition is satisfied (for example, when a replication process stop request is received). Therefore, for example, when the user requests the database system 10 to stop the replication process at an arbitrary timing, an arbitrary quiesce point desired by the user is created. Hereinafter, each component of the database system 10 will be described.

  The master DB 12 stores business data. In the present embodiment, for example, business data in the master DB 12 is referred to and updated by a plurality (three in this embodiment) of business applications 40a, 40b, and 40c. The business applications 40a, 40b, and 40c are user applications, for example, and execute a transaction for each business event performed by the user. In the present embodiment, the business applications 40a, 40b, and 40c receive a business event event message that occurs online and execute a transaction. The business applications 40a, 40b, and 40c operate in parallel. In the present embodiment, the business applications 40a, 40b, and 40c update different records.

  In the present embodiment, the business applications 40a, 40b, and 40c make a numbering request to the serial numbering unit 16 when executing a transaction, for example. Each of the business applications 40a, 40b, and 40c transmits the contents of all data updates executed by each transaction to the update log generation unit 18 together with the transaction numbers assigned by the sequence numbering unit 16.

  The replica DB 14 is a database for backup of the master DB 12, and is also a database for creating a quiesce point of the database.

  The serial number assignment unit 16 receives serial number requests from the business applications 40a, 40b, and 40c, and assigns serial numbers to a plurality of transactions. Further, the serial number numbering unit 16 notifies the business application 40a, 40b, 40c that executes the transaction of the serial number assigned to the transaction. In the present embodiment, the serial number numbering unit 16 assigns serial numbers to a plurality of transactions based on the start time of each transaction. For example, as shown in FIG. 12, when transactions are started in the order of business A, business B, and business C, the serial numbering unit 16 performs the business A transaction, business B transaction, and business C transaction in this order. Give a serial number.

  In the present embodiment, the “transaction start time” includes a transaction start time in the database area, and further a time suggested to start a transaction in the database area. For example, it is assumed that the business application once returns a message “accepted” to the request source in the middle of processing, thereby confirming the acceptance of the transaction. In this case, the “transaction start time” is the time before the actual processing is executed in the database, that is, when the acceptance is confirmed.

Referring to FIG. 1, the update log generator 18 includes business applications 40a, 40b, 40c.
An update log corresponding to each transaction is generated in accordance with the contents of the data update for each transaction transmitted from. In addition, the update log generation unit 18 associates the update log with the serial numbers transmitted from the business applications 40a, 40b, and 40c. Specifically, the update log generation unit 18 associates the plurality of update logs corresponding to the plurality of transactions with the serial numbers assigned to the plurality of transactions by the serial number numbering unit 16. For example, as shown in FIG. 12, when business A, business B, and business C transactions are executed, the update log generation unit 18 performs three updates corresponding to the business A, business B, and business C transactions, respectively. Generate logs. The update log generation unit 18 also includes an update log corresponding to the transactions of the business A, the business B, and the business C, and the serial numbers assigned to the transactions of the business A, the business B, and the business C by the serial numbering unit 16 Associate.

  Referring to FIG. 1, the asynchronous processing unit 20 updates data in the replica DB 14 asynchronously with data update in the master DB 12. Specifically, the asynchronous processing unit 20 uses the replica DB 14 based on the plurality of update logs generated by the update log generation unit 18 and according to the serial numbers associated with the plurality of update logs by the update log generation unit 18. Update the data. For example, as shown in FIG. 12, when transactions of business A, B, and C are started, the asynchronous processing unit 20 updates the data of business A, B, and C performed in the master DB 12 with business A, business B, And in the order of business C, it is reflected in the replica DB 14.

  With reference to FIG. 1, the control unit 22 stops the data update of the replica DB 14 by the asynchronous processing unit 20 when a predetermined condition is satisfied. In this embodiment, for example, when the control unit 22 receives a stop request from the asynchronous processing control application 50, the control unit 22 controls the asynchronous processing unit 20 and updates the data in the replica DB 14 assuming that the predetermined condition is satisfied. To stop.

  The asynchronous process control application 50 is a business application for controlling the processing of the asynchronous processing unit 20. The asynchronous process control application 50 makes a stop request to the control unit 22 when receiving an event message of a stop event that occurs irregularly or periodically. For example, the asynchronous process control application 50 makes a stop request to the control unit 22 in response to an event message of a stop event performed by the system administrator using the operation terminal. Further, for example, the asynchronous process control application 50 receives an event message of a stop event scheduled to occur at a predetermined time (for example, midnight) and makes a stop request to the control unit 22. In addition to the event intended only to stop the replication process, the stop event may cause the asynchronous processing unit 20 to execute a process other than the replication process (for example, a batch process in a modified example described later). The target event is also included. Therefore, in the above stop request, in addition to a request only for stopping the replication processing by the asynchronous processing unit 20, a request for causing the asynchronous processing unit 20 to execute processing other than the replication processing (for example, , An execution request for batch processing in a modified example to be described later).

  As described above, in the database system 10 according to the present embodiment, the data update performed in the master DB 12 is reflected in the replica DB 14 with reference to the start time of the business process (transaction). Further, the database system 10 stops the replication process to the replica DB 14 when a stop request is made to the control unit 22. As a result, it is possible to create a quiesce point based on the start time of a transaction without performing order control on the timing of database data update by a plurality of transactions. For example, as shown in FIG. 12, processing is started in the order of business A, business B, and business C, and further, a stop request is sent to the control unit 22 after the processing of business A and before the processing of business B starts. When it is performed, only the data update of the job A is reflected in the replica DB 14.

  Further, according to the database system 10 according to the present embodiment, it is possible to create a quiesce point of a database in the replica DB 14 without stopping processing for the master DB 12. Therefore, it is possible to create a database quiesce point based on the start time of business processing without reducing the throughput. In addition, according to the database system 10, for example, a user or the like can create an arbitrary quiesce point desired by the user or the like by generating a stop event online at any time when the quiesce point of the database is desired to be created. .

  Next, the configuration of the database system 10 according to the present embodiment will be specifically described. FIG. 2 is a block diagram specifically showing the configuration of the database system 10.

  Referring to FIG. 2, the database system 10 includes the master DB 12 described above, the replica DB 14 described above, and the replication control device 30. In addition to the serial number numbering unit 16, the update log generation unit 18, the asynchronous processing unit 20, and the control unit 22, the replication control device 30 further includes a serial number management unit 24, a processing data storage unit 26, and an asynchronous processing start instruction unit. 28. In addition, among the descriptions regarding each component of the database system 10, the business applications 40a, 40b, and 40c, and the asynchronous processing control application 50, the description of the contents described with reference to FIG. 1 is omitted.

  In this embodiment, the asynchronous process control application 50 makes a numbering request to the serial numbering unit 16 when an event message of a stop event is received. The serial numbering unit 16 receives a numbering request from the asynchronous process control application 50 and assigns a serial number to the stop event. Further, the serial number assignment unit 16 notifies the asynchronous process control application 50 of the serial number assigned to the stop event. The asynchronous process control application 50 transmits the serial number assigned by the serial number assignment unit 16 to the control unit 22 together with the stop request.

  The serial number management unit 24 stores a serial number management table. FIG. 3 is a diagram illustrating an example of a serial number management table stored in the serial number management unit 24. As shown in FIG. 3, the serial number management table manages numbered serial numbers and processed serial numbers. Although the details will be described later, the numbered serial number indicates the newest serial number among the serial numbers that the serial numbering unit 16 has assigned so far. The numbered serial number is updated by the serial numbering unit 16. The processed serial number indicates the serial number corresponding to the newest process among the processes executed by the asynchronous processing unit 20. The processed serial number is updated by the asynchronous processing unit 20.

  With reference to FIG. 2, in this embodiment, the serial number numbering part 16 refers to the serial number management table stored in the serial number management part 24 and assigns serial numbers to the above-described stop event and transaction. Specifically, the serial number numbering unit 16 acquires the serial number that has been numbered in the serial number table, and numbers the acquired serial number plus 1 as the serial number of the above-described stop event or transaction. Thereafter, the numbered serial number is updated to the numbered serial number.

  The update log generation unit 18 registers the update log in a processing data storage table (to be described later) of the processing data storage unit 26 together with a serial number associated with the update log. Further, the control unit 22 generates control information for stopping the processing of the asynchronous processing unit 20 based on the stop request transmitted from the asynchronous processing control application 50. Further, the control unit 22 registers the generated control information in the processing data storage table of the processing data storage unit 26 together with the serial number transmitted from the asynchronous processing control application 50. In the present embodiment, the update log and the control information are registered in the processing data storage table of the processing data storage unit 26 as the processing content (asynchronous processing data) to be executed by the control unit 22.

The processing data storage unit 26 stores a processing data storage table. The contents of the process to be executed by the asynchronous processing unit 20 are registered in the process data storage table. FIG. 4 is a diagram illustrating an example of a processing data storage table stored in the processing data storage unit 26. As shown in FIG. 4, the serial number, type, and processing content are registered in the processing data storage table by the update log generation unit 18 and the control unit 22. The type indicates the type of data registered in the processing content. The type “DB update” indicates that the process is registered by the update log generation unit 18, and the type “control” indicates that the process is registered by the control unit 22.

  When the business applications 40a, 40b, and 40c roll back the data update for the master DB 12, the business applications 40a, 40b, and 40c transmit data indicating no processing to the update log generation unit 18. In this case, as illustrated in FIG. 4, the update log generation unit 18 registers data indicating “no processing” in the processing data storage table as the processing content.

  With reference to FIG. 2, the asynchronous processing unit 20 refers to the serial number management table (see FIG. 3) stored in the serial number management unit 24 and confirms whether there is a process that has not been executed. If there is a process that has not been executed, the asynchronous processing unit 20 acquires the data of the process that has not been executed from the process data storage table (see FIG. 4) stored in the process data storage unit 26. Thereafter, the asynchronous processing unit 20 executes processing based on the data acquired from the processing data storage unit 26.

  For example, when the serial number management table is in the state shown in FIG. 3 and the processing data storage table is in the state shown in FIG. 4, the asynchronous processing unit 20 uses the processing data storage table as data for processing that has not been executed. Acquire data indicating "stop asynchronous processing". Thereafter, the asynchronous processing unit 20 stops the processing to the replica DB 14 based on the data indicating the asynchronous processing stop. That is, the replication process to the replica DB 14 is stopped. The data indicating “stop of asynchronous processing” is data registered by the control unit 22 based on a stop request from the asynchronous processing control application 50.

  When receiving a start request from the operation terminal 60, the asynchronous process start instruction unit 28 instructs the asynchronous process unit 20 to start processing. For example, when the processing to the replica DB 14 is stopped as described above, when the asynchronous processing start instruction unit 28 transmits a start instruction, the asynchronous processing unit 20 resumes the processing according to the serial number. The start instruction from the asynchronous process start instruction unit 28 includes, for example, a “normal start” instruction for resuming the replication process and a “recovery start” instruction for the database recovery process.

  The operation terminal 60 is a terminal operated by a user and an administrator of the database system 10, for example. As described above, the user and the administrator of the database system 10 can make a start request from the operation terminal 60 to the asynchronous process start instruction unit 28. In addition, the user and the administrator can acquire a database quiesce point from the replica DB 14 via the operation terminal 60. Although detailed description is omitted, the event message of the stop event may be transmitted from the operation terminal 60 to the asynchronous process control application 50.

[System operation]
Next, the operation of the database system 10 according to the present embodiment will be described. In this embodiment, the replication method is performed by operating the database system 10. Therefore, the description of the replication method in the present embodiment is replaced with the operation description of the database system 10 below. In the following description, FIGS. 2 to 4 are referred to as appropriate.

FIG. 5 is a flowchart showing an example of the basic operation of the database system 10. Referring to FIG. 5, first, when a numbering request is received from business application 40 a, 40 b, 40 c, serial number numbering unit 16 assigns serial numbers to a plurality of transactions and serial number management stored in serial number management unit 24. Update the table (see FIG. 3). And the asynchronous process part 20 acquires a numbered serial number and a processed serial number from a serial number management table (step S1).

  Next, the asynchronous processing unit 20 determines whether or not the numbered serial number matches the processed serial number (step S2). If the numbered serial number matches the processed serial number, it means that there is no processing to be executed by the asynchronous processing unit 20. Therefore, after waiting for a certain time, the asynchronous processing unit 20 executes the process of step S1 again.

  If the numbered serial number does not match the processed serial number in step S2, the asynchronous processing unit 20 refers to the processing data storage table (see FIG. 4) stored in the processing data storage unit 26, and performs processing. It is determined whether or not serial number data obtained by adding 1 to the completed serial number is registered (step S3).

  If no data is registered in step S3, the asynchronous processing unit 20 waits for a predetermined time and then performs the process of step S3 again. If no data is registered in step S3, it means that the business applications 40a, 40b, and 40c are being processed, and the update log of the process is not yet registered in the process data storage table. .

  When data is registered in step S3, the asynchronous processing unit 20 acquires data indicating the processing content from the processing data storage table (step S4). Next, the asynchronous processing unit 20 determines whether or not the type of processing content is “DB update” based on the acquired data (step S5).

  In step S5, when the type of processing content is “DB update”, the asynchronous processing unit 20 updates the data in the replica DB 14 according to the processing content (step S6). Thereby, the data update performed in the master DB 12 is reflected in the replica DB 14. If the processing content is “no processing” (see, for example, serial number 23 in FIG. 4), the asynchronous processing unit 20 does not update the data in the replica DB 14 in step S6.

  Next, the asynchronous processing unit 20 adds 1 to the processed serial number of the serial number management table (see FIG. 3) stored in the serial number management unit 24, and the processed serial number is the serial number of the processing content executed in step S6. (Step S7). Thereafter, the database system 10 returns to the process of step S1.

  In step S5, when the type of processing content is not “DB update”, that is, when the type of processing content is “control” (see, for example, serial number 20 in FIG. 4), the asynchronous processing unit 20 uses the serial number management table ( 1 is added to the processed serial number in FIG. 3), and the processed serial number is updated to the serial number of the data acquired in step S4 (step S8).

  Next, the asynchronous processing unit 20 determines whether or not the above-described “normal start” instruction is received from the asynchronous processing start instruction unit 28 (step S9). When there is no “normal start” instruction from the asynchronous process start instructing unit 28, the asynchronous processing unit 20 waits until there is a “normal start” instruction. Thereby, the asynchronous processing by the asynchronous processing unit 20 is stopped. That is, the data update of the replica DB 14 is stopped.

  If there is an instruction of “normal start” in step S9, the database system 10 returns to step S1.

  As described above, in the database system 10, the quiesce point of the database can be acquired from the replica DB 14 by stopping the data update from the asynchronous processing unit 20 to the replica DB 14. By referring to the replica DB 14 from the operation terminal 60, the user and the system administrator can perform statistical processing at the stationary point. Further, the processing of the business applications 40a, 40b, and 40c can be continued even while the processing by the asynchronous processing unit 20 is stopped. The update log is accumulated in the processing data storage table of the processing data storage unit 26. When processing (statistical processing, etc.) at a quiesce point is completed, a request for “normal start” is sent from the operation terminal 60 to the asynchronous processing start instructing unit 28, so that the asynchronous processing start instructing unit 28 sends to the asynchronous processing unit 20 A start instruction is sent. As a result, the asynchronous processing unit 20 resumes the processing.

  Here, the effect of the database system 10 will be described by taking as an example a case where a plurality of tasks having different processing times are executed in parallel and the task start order is different from the database update order.

  FIG. 6 is a diagram showing the processing time of the business applications 40a, 40b, 40c and the asynchronous process control application 50, and the serial numbering timing by the serial numbering unit 16. In this example, the business applications 40a, 40b, and 40c operate in parallel, and update the master DB 12. The business applications 40a, 40b, and 40c update the master DB 12 and register the updated contents in the update log generation unit 18 at the timing of the tip of the arrow shown in FIG.

  The business application 40a starts business processing on January 1 and completes data update of the master DB 12 on January 2. The business application 40b and the business application 40c start business processing on January 2. In this example, it is assumed that a quiesce point of a database reflecting all the updates of the business application 40a that started processing during January 1 is reflected. Therefore, the asynchronous process control application 50 starts processing at the timing when the date changes to January 2.

  An example of the processing contents (asynchronous processing data) stored in the processing data storage table of the processing data storage unit 26 by the business applications 40a, 40b, 40c and the asynchronous processing control application 50 is shown in FIG. With reference to FIGS. 6 and 7, in this example, serial number 25 is assigned to the process of business application 40a, serial number 26 is assigned to the process of asynchronous process control application 50, and A serial number 27 is assigned to the process, and a serial number 28 is assigned to the process of the business application 40c. However, referring to FIG. 7, the order in which the processing contents are registered in the processing data storage table is not the same as the order of the serial numbers. In this example, the processing content of serial number 26, the processing content of serial number 28, the processing content of serial number 27, and the processing content of serial number 25 are registered in the processing data storage table.

  In this example, the asynchronous processing unit 20 refers to the processing data storage table (see FIG. 7) of the processing data storage unit 26 and executes the processing contents of the serial number 25 in order. Since the serial number 25 has the type “DB update”, the asynchronous processing unit 20 updates the replica DB 14 according to the processing content. Since the serial number 26 has the type “control”, the asynchronous processing unit 20 stops the asynchronous processing (data update of the replica DB 14). As a result, the data update of the business application 40a that started processing on January 1 is reflected, and the quiesce point of the database that does not reflect the data update of the business applications 40b and 40c that started processing on January 2 is reflected. Can be created in the replica DB 14. A user of the database system 10 can use the operation terminal 60 to refer to the replica DB 14 and perform statistical processing of business data generated on January 1st.

Next, operations up to system recovery when a failure occurs in the master DB 12 will be described.
When a failure occurs in the master DB 12, processing of all business applications 40a, 40b, and 40c that update the master DB 12 is stopped. In this case, for example, the system administrator makes a “recovery start” start request from the operation terminal 60 to the asynchronous process start instruction unit 28. As a result, the asynchronous process start instruction unit 28 transmits a “recovery start” start instruction to the asynchronous process unit 20.

  When the asynchronous processing unit 20 receives an instruction to start “recovery start”, the asynchronous processing unit 20 reflects all update logs registered in the processing data storage table of the processing data storage unit 26 to the replica DB 14 in accordance with a processing flow (FIG. 8) described later. To do. When the reflection of the update log to the replica DB 14 is completed, the data state immediately before the failure of the master DB 12 is restored in the replica DB 14. Next, the master DB 12 is disconnected from the database system 10 and the replica DB 14 is used instead of the master DB 12. Thereafter, the processing to the replica DB 14 by the business applications 40a, 40b, and 40c is resumed. As a result, the business can be recovered by using the replica DB 14 as a new master DB. Hereinafter, the recovery operation of the database system 10 will be specifically described with reference to a flowchart.

  FIG. 8 is a flowchart showing an example of the recovery operation of the database system 10. Referring to FIG. 8, when asynchronous processing unit 20 receives an instruction of “recovery start” from asynchronous processing start instruction unit 28, serial numbering unit 16 stores a serial number management table (FIG. 3) stored in serial number management unit 24. The numbered serial number and the processed serial number are acquired from (refer to).

  Next, the asynchronous processing unit 20 determines whether or not the numbered serial number matches the processed serial number (step A2). If the numbered serial number and the processed serial number match, all the update logs to be reflected are reflected in the replica DB 14, and the recovery process is terminated.

  In step A2, if the numbered serial number and the processed serial number do not match, the asynchronous processing unit 20 refers to the processing data storage table (see FIG. 4) stored in the processing data storage unit 26 and performs processing. It is determined whether or not the data corresponding to the serial number obtained by adding 1 to the completed serial number is registered (step A3).

  If no data is registered in step A3, the asynchronous processing unit 20 proceeds to the process of step A7 described later, updates the processed serial number, and then returns to step A1. If the processing contents are not registered in step A3, serial numbers are assigned to the processing (transactions) of the business applications 40a, 40b, and 40c, but the data update of the master DB 12 by the processing is not completed. It means that. When the recovery operation is being performed, a failure has occurred in the master DB 12, so that the data update of the master DB 12 is not performed by the process corresponding to the serial number. Accordingly, the update log corresponding to the serial number need not be reflected in the replica DB 14. For this reason, processing in steps A4 to A6 described later can be skipped.

  When the data is registered in step A3, the asynchronous processing unit 20 acquires data indicating the processing content from the processing data storage table (step A4). Next, the asynchronous processing unit 20 determines whether or not the type of processing content is “DB update” based on the acquired data (step A5). When the type is not “DB update”, that is, when the type is “control”, it means that the processing content is not related to the recovery processing. Therefore, the asynchronous processor 20 proceeds to the process of step A7 described later, updates the processed serial number, and then returns to step A1.

When the type of processing content is “DB update” in step A5, the asynchronous processing unit 20 updates data in the replica DB 14 according to the processing content (step A6). Thereby, the data state of the replica DB 14 can be changed to the data state of the master DB 12 before the failure occurs.

  Next, the asynchronous processing unit 20 adds 1 to the processed serial number of the serial number management table stored in the serial number management unit 24, and updates the processed serial number to the serial number of the processing content executed in step A6 (step S6). A7). Thereafter, the database system 10 returns to the process of step A1.

(Modification)
In the above-described embodiment, the case where the business applications 40a, 40b, and 40c update different records has been described. If a plurality of business applications that update the same record in the master DB 12 operate in parallel, it is preferable to update the data in the master DB 12 in the order in which serial numbers are assigned. For example, it is preferable to control the order of the data update timings of a plurality of business applications so that the data in the master DB 12 is updated in the order in which the serial numbers are assigned. Thereby, it is possible to prevent the data update order in the replica DB 14 from being reversed with respect to the data update order in the master DB 12. As a result, it is possible to prevent the record status from being inconsistent between the master DB 12 and the replica DB 14.

  In the above-described embodiment, it is assumed that the system administrator or the like can execute an arbitrary operation on the replica DB 14 while the quiesce point of the database can be acquired. Therefore, it is assumed that the processing of the asynchronous processing unit 20 that has stopped once is resumed at an arbitrary timing by the system administrator operating the operation terminal 60, for example. However, for example, the database system according to the present invention can be used even when the processing to be executed on the replica DB 14 in a state where the quiesce point of the database can be acquired is determined in advance as batch processing. A brief description is given below.

  FIG. 9 is a diagram showing a modification of the database system. The database system 10a shown in FIG. 9 can be suitably used when, for example, statistical processing of business data is performed on a date basis by batch processing to create statistical data. The statistical processing is an example, and the database system 10a can be suitably used for any batch processing that is desired to be executed at a quiesce point of the database.

  Referring to FIG. 9, the database system 10 a is different from the database system 10 described above in that a replication control device 30 a is provided instead of the replication control device 30. The replication control device 30a differs from the replication control device 30 described above in that a statistical data acquisition unit 32 is further provided.

  In the database system 10 a, for example, the asynchronous process control application 50 makes a batch process execution request to the control unit 22. When the control unit 22 receives the execution request for the batch processing, the control information for causing the statistical data acquisition unit 32 to execute the batch processing together with the serial number assigned by the serial number numbering unit 16 is processed in the processing data storage unit 26. Register in the data storage table. FIG. 10 is a diagram illustrating an example of processing contents registered in the processing data storage table by the control unit 22 of the database system 10a. In the example of FIG. 10, the processing content of the serial number 31 (asynchronous processing data) is data registered by the control unit 22.

  For example, the asynchronous processing unit 20 acquires the data of the serial number 31 from the processing data storage table illustrated in FIG. Since the data of the serial number 31 is “control” and the processing content is “statistical data acquisition unit execution”, the asynchronous processing unit 20 instructs the statistical data acquisition unit 32 to start processing. In accordance with the start instruction from the asynchronous processing unit 20, the statistical data acquisition unit 32 refers to the replica DB 14 that is a stationary point and creates statistical data.

  The statistical data acquisition unit 32 notifies the asynchronous processing unit 20 of the completion of the processing after the statistical data is created. When receiving the completion notification from the statistical data acquisition unit 32, the asynchronous processing unit 20 updates the processed serial number of the serial number management table stored in the serial number management unit 24, and the next serial number (in the example of FIG. 10, serial number 32). Perform the process. Thereby, the replica DB 14 is not updated while the statistical data acquisition unit 32 is operating. Therefore, according to the database system 10a, batch processing is possible at a stationary point of the database. Further, according to the database system 10a, the replication processing to the replica DB 14 can be restarted promptly after the batch processing is completed.

[program]
The program in the embodiment of the present invention may be a program that causes a computer to execute the processes of steps S1 to S9 shown in FIG. 5 and / or steps A1 to A7 shown in FIG. By installing and executing this program on a computer, the replication control apparatuses 30 and 30a and the replication method in the present embodiment can be realized. In this case, the CPU (Central Processing Unit) of the computer functions as a serial numbering unit 16, an update log generation unit 18, an asynchronous processing unit 20, a control unit 22, an asynchronous processing start instruction unit 28, and a statistical data acquisition unit 32. Perform processing.

  In the present embodiment, the master DB 12, the replica DB 14, the serial number management unit 24, and the processing data storage unit 26 store data files constituting them in a storage device such as a hard disk provided in the computer. Alternatively, it is realized by mounting a recording medium storing this data file on a reading device connected to a computer.

[Physical configuration]
Here, a computer that realizes the replication control apparatuses 30 and 30a by executing the program according to the present embodiment will be described with reference to FIG. FIG. 11 is a block diagram illustrating an example of a computer that implements the replication control apparatuses 30 and 30a according to the embodiment of this invention.

  As shown in FIG. 11, the computer 110 includes a CPU 111, a main memory 112, a storage device 113, an input interface 114, a display controller 115, a data reader / writer 116, and a communication interface 117. These units are connected to each other via a bus 121 so that data communication is possible.

  The CPU 111 performs various calculations by developing the program (code) in the present embodiment stored in the storage device 113 in the main memory 112 and executing them in a predetermined order. The main memory 112 is typically a volatile storage device such as a DRAM (Dynamic Random Access Memory). Further, the program in the present embodiment is provided in a state of being stored in a computer-readable recording medium 120. Note that the program in the present embodiment may be distributed on the Internet connected via the communication interface 117.

  Specific examples of the storage device 113 include a hard disk drive and a semiconductor storage device such as a flash memory. The input interface 114 mediates data transmission between the CPU 111 and an input device 118 such as a keyboard and a mouse. The display controller 115 is connected to the display device 119 and controls display on the display device 119.

The data reader / writer 116 mediates data transmission between the CPU 111 and the recording medium 120, and reads a program from the recording medium 120 and writes a processing result in the computer 110 to the recording medium 120. The communication interface 117 mediates data transmission between the CPU 111 and another computer.

  Specific examples of the recording medium 120 include general-purpose semiconductor storage devices such as CF (Compact Flash (registered trademark)) and SD (Secure Digital), magnetic storage media such as a flexible disk, or CD- An optical storage medium such as ROM (Compact Disk Read Only Memory) can be used.

  Part or all of the above-described embodiment can be expressed by (Appendix 1) to (Appendix 15) described below, but is not limited to the following description.

(Appendix 1)
A master database that updates data based on multiple transactions;
A replica database as a replication destination of the master database;
With reference to the start times of the plurality of transactions, a serial number numbering unit that assigns serial numbers to the plurality of transactions,
An update log generation unit that generates a plurality of update logs corresponding to the plurality of transactions, and associates the plurality of update logs with the serial numbers assigned to the plurality of transactions by the serial numbering unit;
Based on the plurality of update logs generated by the update log generation unit and according to the serial numbers associated with the plurality of update logs by the update log generation unit, asynchronous with data update of the master database. An asynchronous processing unit for updating data in the replica database;
A database system comprising: a control unit that stops data update of the replica database by the asynchronous processing unit when a predetermined condition is satisfied.

(Appendix 2)
When the control unit receives a request to stop the asynchronous processing unit, the control unit generates control information for stopping the data update by the asynchronous processing unit, assuming that the predetermined condition is satisfied,
The serial number numbering unit includes the plurality of update logs and the control information so that the plurality of update logs and the control information are arranged in time series with reference to a start time of the plurality of transactions and a generation time of the control information. And give a serial number to
The asynchronous processing unit updates the data in the replica database based on the plurality of update logs and the control information, and according to the serial numbers assigned to the plurality of update logs and the control information, and the data The database system according to attachment 1, wherein the updating is stopped.

(Appendix 3)
The database system according to appendix 2, further comprising an asynchronous processing start instruction unit that instructs the asynchronous processing unit to resume data update of the replica database.

(Appendix 4)
When the control unit receives a request for execution of batch processing, the control unit generates control information for causing the previous asynchronous processing unit to execute batch processing, assuming that the predetermined condition is satisfied,
The serial number numbering unit includes the plurality of update logs and the control information so that the plurality of update logs and the control information are arranged in time series with reference to a start time of the plurality of transactions and a generation time of the control information. And give a serial number to
The asynchronous processing unit is configured to update the data in the replica database and perform batch processing based on the plurality of update logs and the control information and according to the serial numbers assigned to the plurality of update logs and the control information. The database system according to attachment 3, wherein:

(Appendix 5)
In a database system comprising a master database in which data is updated based on a plurality of transactions and a replica database as a replication destination of the master database, data in the replica database is updated asynchronously with data update in the master database A replication control device,
With reference to the start times of the plurality of transactions, a serial number numbering unit that assigns serial numbers to the plurality of transactions,
An update log generation unit that generates a plurality of update logs corresponding to the plurality of transactions, and associates the plurality of update logs with the serial numbers assigned to the plurality of transactions by the serial numbering unit;
Based on the plurality of update logs generated by the update log generation unit and according to the serial numbers associated with the plurality of update logs by the update log generation unit, asynchronous with data update of the master database. An asynchronous processing unit for updating data in the replica database;
A replication control apparatus comprising: a control unit that stops data update of the replica database by the asynchronous processing unit when a predetermined condition is satisfied.

(Appendix 6)
When the control unit receives a request to stop the asynchronous processing unit, the control unit generates control information for stopping the data update by the asynchronous processing unit, assuming that the predetermined condition is satisfied,
The serial number numbering unit includes the plurality of update logs and the control information so that the plurality of update logs and the control information are arranged in time series with reference to a start time of the plurality of transactions and a generation time of the control information. And give a serial number to
The asynchronous processing unit updates the data in the replica database based on the plurality of update logs and the control information, and according to the serial numbers assigned to the plurality of update logs and the control information, and the data The replication control device according to appendix 5, which performs updating stop.

(Appendix 7)
The replication control device according to appendix 6, further comprising an asynchronous processing start instruction unit that instructs the asynchronous processing unit to resume data update of the replica database.

(Appendix 8)
When the control unit receives a request for execution of batch processing, the control unit generates control information for causing the previous asynchronous processing unit to execute batch processing, assuming that the predetermined condition is satisfied,
The serial number numbering unit includes the plurality of update logs and the control information so that the plurality of update logs and the control information are arranged in time series with reference to a start time of the plurality of transactions and a generation time of the control information. And give a serial number to
The asynchronous processing unit is configured to update the data in the replica database and perform batch processing based on the plurality of update logs and the control information and according to the serial numbers assigned to the plurality of update logs and the control information. The replication control device according to attachment 7, wherein:

(Appendix 9)
In a database system comprising a master database in which data is updated based on a plurality of transactions and a replica database as a replication destination of the master database, data in the replica database is updated asynchronously with data update in the master database A replication method,
(A) assigning serial numbers to the plurality of transactions on the basis of start times of the plurality of transactions;
(B) generating a plurality of update logs corresponding to the plurality of transactions, and associating the plurality of update logs with the serial numbers assigned to the plurality of transactions in the step (a);
(C) Data in the master database based on the plurality of update logs generated in the step (b) and according to the serial numbers associated with the plurality of update logs in the step (b) Updating the replica database data asynchronously with the update; and
(D) a replication method comprising a step of stopping data update of the replica database when a predetermined condition is satisfied.

(Appendix 10)
The step (d) generates control information for stopping the data update by the asynchronous processing unit, assuming that the predetermined condition is satisfied when a request for stopping the asynchronous processing unit is received,
The step (a) includes the plurality of update logs and the control so that the plurality of update logs and the control information are arranged in time series with reference to a start time of the plurality of transactions and a generation time of the control information. Give a serial number to the information,
The step (c) includes updating the data in the replica database based on the plurality of update logs and the control information, and according to the serial numbers assigned to the plurality of update logs and the control information. The replication method according to appendix 9, wherein the data update is stopped.

(Appendix 11)
The replication method according to appendix 10, further comprising an asynchronous processing start instruction step for instructing the asynchronous processing unit to resume data update of the replica database.

(Appendix 12)
The step (d) generates control information for causing the pre-asynchronous processing unit to execute batch processing, assuming that the predetermined condition is satisfied when a batch processing execution request is received,
The step (a) includes the plurality of update logs and the control so that the plurality of update logs and the control information are arranged in time series with reference to a start time of the plurality of transactions and a generation time of the control information. Give a serial number to the information,
The step (c) includes updating the data in the replica database based on the plurality of update logs and the control information, and according to the serial numbers assigned to the plurality of update logs and the control information. The replication method according to attachment 11, wherein batch processing is performed.

(Appendix 13)
In a database system comprising a master database in which data is updated based on a plurality of transactions, and a replica database that is a replication destination of the master database, the data in the replica database is asynchronous with the data update in the master database. A program for executing an update,
In the computer,
(A) assigning serial numbers to the plurality of transactions on the basis of start times of the plurality of transactions;
(B) generating a plurality of update logs corresponding to the plurality of transactions, and associating the plurality of update logs with the serial numbers assigned to the plurality of transactions in the step (a);
(C) Data in the master database based on the plurality of update logs generated in the step (b) and according to the serial numbers associated with the plurality of update logs in the step (b) Updating the replica database data asynchronously with the update; and
(D) A program that executes a step of stopping data update of the replica database when a predetermined condition is satisfied.

(Appendix 14)
The step (d) generates control information for stopping the data update by the asynchronous processing unit, assuming that the predetermined condition is satisfied when a request for stopping the asynchronous processing unit is received,
The step (a) includes the plurality of update logs and the control so that the plurality of update logs and the control information are arranged in time series with reference to a start time of the plurality of transactions and a generation time of the control information. Give a serial number to the information,
The step (c) includes updating the data in the replica database based on the plurality of update logs and the control information, and according to the serial numbers assigned to the plurality of update logs and the control information. 14. The program according to appendix 13, which stops the data update.

(Appendix 15)
15. The program according to appendix 14, further comprising an asynchronous process start instruction step for instructing the asynchronous processor to resume data update of the replica database.

(Appendix 16)
The step (d) generates control information for causing the pre-asynchronous processing unit to execute batch processing, assuming that the predetermined condition is satisfied when a batch processing execution request is received,
The step (a) includes the plurality of update logs and the control so that the plurality of update logs and the control information are arranged in time series with reference to a start time of the plurality of transactions and a generation time of the control information. Give a serial number to the information,
The step (c) includes updating the data in the replica database based on the plurality of update logs and the control information, and according to the serial numbers assigned to the plurality of update logs and the control information. The program according to appendix 15, which performs batch processing.

  The present invention can be suitably used for, for example, an online system such as a billing system that operates for 24 hours and periodically performs aggregation processing.

10,10a Database system 12 Master DB
14 Replica DB
16 serial number numbering unit 18 update log generation unit 20 asynchronous processing unit 22 control unit 24 serial number management unit 26 processing data storage unit 28 asynchronous processing start instruction unit 30, 30a replication control device 32 statistical data acquisition unit 40a, 40b, 40c business application 50 Asynchronous processing control application 110 Computer 111 CPU
112 Main Memory 113 Storage Device 114 Input Interface 115 Display Controller 116 Data Reader / Writer 117 Communication Interface 118 Input Device 119 Display Device 120 Recording Medium 121 Bus

Claims (7)

A master database that updates data based on multiple transactions;
A replica database as a replication destination of the master database;
With reference to the start times of the plurality of transactions, a serial number numbering unit that assigns serial numbers to the plurality of transactions,
An update log generation unit that generates a plurality of update logs corresponding to the plurality of transactions, and associates the plurality of update logs with the serial numbers assigned to the plurality of transactions by the serial numbering unit;
Based on the plurality of update logs generated by the update log generation unit and according to the serial numbers associated with the plurality of update logs by the update log generation unit, asynchronous with data update of the master database. An asynchronous processing unit for updating data in the replica database;
A database system comprising: a control unit that stops data update of the replica database by the asynchronous processing unit when a predetermined condition is satisfied. When the control unit receives a request to stop the asynchronous processing unit, the control unit generates control information for stopping the data update by the asynchronous processing unit, assuming that the predetermined condition is satisfied,
The serial number numbering unit includes the plurality of update logs and the control information so that the plurality of update logs and the control information are arranged in time series with reference to a start time of the plurality of transactions and a generation time of the control information. And give a serial number to
The asynchronous processing unit updates the data in the replica database based on the plurality of update logs and the control information, and according to the serial numbers assigned to the plurality of update logs and the control information, and the data The database system according to claim 1, wherein updating is stopped.   The database system according to claim 2, further comprising an asynchronous processing start instruction unit that instructs the asynchronous processing unit to resume data update of the replica database. When the control unit receives a batch processing execution request, the control unit generates control information for causing the asynchronous processing unit to execute batch processing, assuming that the predetermined condition is satisfied,
The serial number numbering unit includes the plurality of update logs and the control information so that the plurality of update logs and the control information are arranged in time series with reference to a start time of the plurality of transactions and a generation time of the control information. And give a serial number to
The asynchronous processing unit is configured to update the data in the replica database and perform batch processing based on the plurality of update logs and the control information and according to the serial numbers assigned to the plurality of update logs and the control information. The database system according to claim 3, wherein: In a database system comprising a master database in which data is updated based on a plurality of transactions and a replica database as a replication destination of the master database, data in the replica database is updated asynchronously with data update in the master database A replication control device,
With reference to the start times of the plurality of transactions, a serial number numbering unit that assigns serial numbers to the plurality of transactions,
An update log generation unit that generates a plurality of update logs corresponding to the plurality of transactions, and associates the plurality of update logs with the serial numbers assigned to the plurality of transactions by the serial numbering unit;
Based on the plurality of update logs generated by the update log generation unit and according to the serial numbers associated with the plurality of update logs by the update log generation unit, asynchronous with data update of the master database. An asynchronous processing unit for updating data in the replica database;
A replication control apparatus comprising: a control unit that stops data update of the replica database by the asynchronous processing unit when a predetermined condition is satisfied. In a database system comprising a master database in which data is updated based on a plurality of transactions and a replica database as a replication destination of the master database, data in the replica database is updated asynchronously with data update in the master database A replication method,
(A) assigning serial numbers to the plurality of transactions on the basis of start times of the plurality of transactions;
(B) generating a plurality of update logs corresponding to the plurality of transactions, and associating the plurality of update logs with the serial numbers assigned to the plurality of transactions in the step (a);
(C) Data in the master database based on the plurality of update logs generated in the step (b) and according to the serial numbers associated with the plurality of update logs in the step (b) Updating the replica database data asynchronously with the update; and
(D) a replication method comprising a step of stopping data update of the replica database when a predetermined condition is satisfied. In a database system comprising a master database in which data is updated based on a plurality of transactions, and a replica database that is a replication destination of the master database, the data in the replica database is asynchronous with the data update in the master database. A program for executing an update,
In the computer,
(A) assigning serial numbers to the plurality of transactions on the basis of start times of the plurality of transactions;
(B) generating a plurality of update logs corresponding to the plurality of transactions, and associating the plurality of update logs with the serial numbers assigned to the plurality of transactions in the step (a);
(C) Data in the master database based on the plurality of update logs generated in the step (b) and according to the serial numbers associated with the plurality of update logs in the step (b) Updating the replica database data asynchronously with the update; and
(D) A program that executes a step of stopping data update of the replica database when a predetermined condition is satisfied.

Images