JP2005141555A - Backup method of database, and online system using same - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To facilitate recovery from a system failure and shorten the recovery time by recovering a database partly when data in the database are damaged. <P>SOLUTION: A database storage part 9 is configured in divisions of work corresponding to transaction data. A work terminal 4 comprises work processing means 4a for adding to data a destination specifying a database and an identifier specifying a work division. A store server 7 comprises sorting means 7b for sorting the data by the destination, transferring means 7d for transfer to the destination, work sorting means 7f for storing the transferred data, and backup means 7i for saving the sorted data, saving a replication of the database 9 periodically, and if data in the database storage part 9 are damaged, copying a work division area corresponding to the damaged part from a backup storage part 8 to the damaged database storage part 9 and updating the database storage part 9 with the saved data. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to an online system including a database, and more particularly to a method and system for converting data input from a data input terminal into transaction data and backing up a database updated by the transaction data.

  Conventionally, an online system has a configuration shown in FIG. 7, for example. This system consists of a business terminal that processes input data, converts it into transaction data, and transmits transaction data using a communication line, and a server that receives this transaction data and updates the corresponding database. A plurality of server systems are connected via a network.

  Specifically, a store-side server system 81 and a headquarters-side server system 82 are connected via a network 83 formed of the Internet or a dedicated line, and the store-side server system 81 has two store operations. A terminal 84, two POS terminals 85 for managing product accounting, a store server 87 connected to the console 86, and a store database storage unit 88 connected to the store server 87. A backup storage unit 89 that periodically stores the data in the store database storage unit 88.

  The server system 82 on the headquarters side is connected to one headquarters business terminal 90 via a communication line, a headquarter server 92 connected to the console 91, and a headquarter database storage unit 93 connected to the headquarter server 92. And a backup storage unit 94 for periodically storing data in the headquarter database storage unit 93.

FIG. 8 is a diagram showing a flow of transaction data by each means provided in each device described in the online system of FIG.
The store operation terminal 84 is provided with store terminal operation means 84a, and performs operations such as ordering, returns, inventory, and sales management from this terminal based on data input from a keyboard or a barcode reader (not shown). Can do.
Further, the POS terminal 85 is provided with a POS means 85a, and price lookup (reference) of commodity unit price, accounting processing, and sales totaling operations can be performed from the POS terminal 85.

  On the other hand, the store server 87 is provided with store server business means 87a that accepts input from a console (not shown) and displays the result on the console, and can perform the same processing as the store terminal business means 84a. .

Further, the store server 87 is provided with store database management means 87b that exclusively controls reading and writing to the store database storage unit 88 for each terminal, and stores the store database according to the transaction data transmitted from each business means. Processing for updating the data in the section 88 is performed.
Further, the store server 87 is provided with a backup unit 87c between the store database management unit 87b and the store database storage unit 88, and the update data of the database passing therethrough is sequentially stored as the backup update unit. 89.
On the other hand, the data in the store database storage unit 88 is replicated to the backup storage unit 89 at regular intervals by the backup unit 87c, and is stored as a storage database together with the storage update data corresponding to the above-described predetermined period. .

  If the data in the store database storage unit 88 is damaged for some reason, the backup unit 87c overwrites the store database storage unit 88 with the data of the storage database copied in the backup storage unit 89, and then saves and updates. By sequentially updating the store database storage unit 88 using data, the data in the store database storage unit 88 is restored to the state before being damaged.

On the other hand, each business means and POS means can access not only the store database storage unit 88 of the store server 87 but also the headquarter database storage unit 93 of the server system 82 on the headquarters side via the network 83.
Since the operation of the server system 82 on the headquarters side is the same as that of the server system 81 on the store side, detailed description is omitted. The main difference is the processing of transaction data based on the work required at the headquarters side, specifically, the totaling processing of each store's data, such as the totaling processing of each store's sales, order data, and general inventory management. .
Each means described above is realized by software operating on a personal computer or a server.

  Each business means in the headquarters and stores can access each other's database as well as in its own server system. Accordingly, each business means converts the currently input data into transaction data, determines whether it is in the same server system or an external server system, and selects and processes a communication method corresponding to the server system.

  In such an online system, the data in the database is very important. Due to some trouble, for example, an error in program creation, abnormal data was written and the data in the database was damaged (abnormal data). If it has been rewritten, immediate repair is required.

In the case of a large-scale system such as a bank system, a method of preparing two databases that are separated from each other in advance for disaster countermeasures and constantly updating the two databases with the same contents for one transaction data Is adopted (see, for example, Patent Document 1). However, since such a system is provided with two systems having the same configuration, the cost is very high, and there is a problem that it cannot be adopted for a medium-scale or small-scale system.
Therefore, medium-scale and small-scale systems give up immediate recovery in the event of a data rewrite failure. For example, on the night of the failure, a method is used to restore the database to the state before the failure using the method described above. ing.

However, in conventional medium-scale and small-scale systems, transaction data is updated to the business terminal side, that is, at the initiative of the business means. In addition, since it is impossible to determine what type of business it is, when the transaction data is saved as saved update data by the backup means, only the update data itself is saved.
Therefore, when the database is restored, the entire store database storage unit is overwritten with the stored database data, and then the stored database is sequentially read from the beginning of the stored update data to re-update the entire database. A great deal of time was spent recovering the damage.

Japanese Patent Laid-Open No. 9-259023 (page 3-4, FIG. 1)

  The present invention solves the above-described problems, and when a failure occurs and the data in the database is damaged (rewritten), the entire database is not restored, but partially restored, The purpose is to facilitate the recovery of system failures and shorten the recovery time.

In order to solve the above-described problems, the present invention processes input data and converts it into transaction data, and updates the corresponding database by receiving the transaction data and transmitting the transaction data through a communication line. In addition, in a database backup method performed in an online system in which a plurality of server systems configured with a server that backs up the database are connected via a network,
The business terminal transmits the transaction data by adding a destination for identifying the database and an identifier for identifying the business category, and the server distributes the transaction data according to the destination, and a database corresponding to the destination. Transferred to the server system comprising:
The server that has received the distributed transaction data updates the database corresponding to the destination, classifies the transaction data according to the identifier, and stores it as backup data.

In addition, the database is configured to be divided for each business corresponding to the transaction data, and the server replicates the database every predetermined period and stores it as a backup database,
When the data of the database is damaged, the database stores the database by copying the area of the business division corresponding to the damaged part from the backup database to the damaged database, and updating the database using the backup data. To restore.

Alternatively, a business terminal that processes input data to convert it into transaction data, and transmits the transaction data through a communication line, and a server that receives the transaction data and updates a corresponding database and backs up the database In an online system in which multiple server systems configured with
The business terminal comprises a business processing means for adding a destination for identifying the database, an identifier for identifying the business classification to the transaction data, and a transaction data transmission means for transmitting the transaction data.
The server includes a distribution unit that distributes the transaction data according to the destination, a database update unit that updates a database corresponding to the destination in the server with the transaction data, and a database corresponding to the destination Transfer means for transferring to the system, business distribution means for classifying the transferred transaction data, and storing the classified transaction data as backup data, and replicating the database every predetermined period as a backup database And a backup means for storing.

  Further, the database is configured to be divided for each business corresponding to the transaction data, and when the data of the database is damaged, the backup unit backs up the area of the business classification corresponding to the damaged portion. In addition to copying the database from the database to the damaged database, the system includes a restoring means for restoring the database by updating the database using the backup data.

By using the above means, according to the database backup method and the online system using the same according to the present invention,
In the invention according to claim 1, the business terminal transmits the transaction data by adding a destination for identifying the database and an identifier for identifying the business classification, and the server distributes the transaction data according to the destination and creates a database corresponding to the destination. Transfer to the server system
The server that receives the distributed transaction data updates the database corresponding to the destination, classifies the transaction data according to the identifier, and stores it as backup data.
Since each transaction data is classified and stored according to the store where the transaction occurred and the target business, it becomes easy to select transaction data necessary for recovery, and fine data recovery and recovery time can be shortened.

The invention according to claim 2 is configured by dividing the database for each business corresponding to the transaction data, and the server duplicates the database every predetermined period and stores it as a backup database,
When the data in the database is damaged, the area of the business division corresponding to the damaged part is copied from the backup database to the damaged database, and the database is updated with the backup data to restore the database.
Since only the area of the business division corresponding to the damaged part of the database is partially recovered, it is possible to easily recover the system failure and shorten the recovery time.

The invention according to claim 3 comprises business processing means for adding a destination to which the business terminal specifies the database and an identifier for specifying the business classification to the transaction data, and transaction data transmission means for transmitting the transaction data.
The server distributes the transaction data according to the destination, a database update unit that updates a database corresponding to the destination in the server with the transaction data, a transfer unit that transfers to a server system including a database corresponding to the destination, By providing business distribution means for classifying the transferred transaction data and backup means for storing the classified transaction data as backup data and copying the database every predetermined period and saving it as a backup database,
Since each transaction data is classified and stored according to the store where the transaction occurred and the target business, it becomes easy to select transaction data necessary for recovery, and fine data recovery and recovery time can be shortened.

The invention according to claim 4 is configured such that the database is divided for each business corresponding to the transaction data, and when the database data is damaged, the backup means backs up the business classification area corresponding to the damaged portion. By providing a restoration means for restoring the database by copying the database from the database to the damaged database and updating the database using the backup data,
Since only the area of the business division corresponding to the damaged part of the database is partially recovered, it is possible to easily recover the system failure and shorten the recovery time.

  DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, embodiments of the present invention will be described in detail as examples based on the attached drawings. In addition, about the same part as a prior art example, while using the same number, detailed description is abbreviate | omitted.

FIG. 1 is an explanatory diagram showing the flow of transaction data according to the transaction processing method of the present application.
The feature of this application is that when the data in the database is damaged for some reason (rewritten to abnormal data) or the writing is not executed normally, it is possible to perform a fine restoration and a short recovery operation. .
For this reason, the access control function to the database provided for each business means is separated, and the transaction distribution means operating on the store server or the head office server is used to transfer the transaction data to the database to be processed. To. Then, the target database is updated with the distributed transaction data, and at the same time, the transaction data is classified and stored for each store where the data occurred and the target business, and the database in which the failure occurs corresponds to this classification target. Only part was restored.

Next, the configuration and operation of each means of the present application will be described with reference to FIG. Each means is controlled by using software corresponding to each means.
The store business terminal 4 is provided with business processing means 4a, and operations such as ordering, return, inventory, and sales management can be performed from this terminal. Further, the business processing means 4a adds to the transaction data generated in each business, the database destination where this data is to be processed, and the identifier indicating the type of each business and the origin of the data (occurrence time, generating store, terminal, slip) Number).

  Further, the store business terminal 4 is provided with transaction data transmission means 4b having an FTP (File Transfer Protocol) function for transferring a file, and the transaction data created by the business processing means 4a as one file. It is configured to transmit to the store server 7.

  Further, the POS terminal 5 is provided with business processing means 5b, and price lookup and accounting processing of product unit price and sales totaling business can be performed from this POS terminal 5. Further, the business processing means 5b adds transaction data generated in each business to a database destination where this data is to be processed, an identifier indicating the type of each business and the origin of the data (occurrence time, generated store, terminal, slip) Number etc.) is added.

  Further, the POS terminal 5 is provided with transaction data transmission means 5a having an FTP function for transferring a file, and the transaction data created by the business processing means 5b is transmitted to the store server 7 as one file. It is the composition to do.

On the other hand, the store server 7 is provided with business processing means 7a that receives an input from a console (not shown) and displays the result on the console, and performs the same processing as the business processing means 4a of the store terminal 4. be able to.
The store server 7 is provided with distribution means 7b that receives transaction data transmitted from the transaction data transmission means 4b and 5a, and distributes and transfers the transaction data to the destinations described therein.

  Each transaction data received by the allocating means 7b updates the data in the store database storage unit 9 in the same store server system 1 with the destination, and another server system via the network 83 via the transfer means 7d. For example, it is distributed to the one that updates data in the headquarter database storage unit 13 of the headquarter server system 2. At this time, if there are a plurality of destinations, transaction data is duplicated and distributed by that number.

  As described above, transaction data is replicated and the database is accessed by the store server 7, so that the load on each business processing means is reduced, and when the LAN is used from the business terminal to the server, the amount of communication data is reduced. As a result of the reduction, it is possible to reduce the data collision in the communication line and improve the usage efficiency of the LAN.

  The transaction data for updating the data in the store database storage unit 9 of the store server system 1 is, for each business, for example, ordering or purchasing, according to the business identifier added to the transaction data by the business distribution means 7f. The transaction data distributed for each business and distributed for each business update the data in the store database storage unit 9 by the operation of the database management means 7g.

  On the other hand, the transaction data for updating the database of the other server system via the network 83 is transferred to the headquarter server system 2, for example, via the network 83 by the transfer means 7d. The data is received by the receiving unit 12c, and distributed by the distribution unit 12b so as to update the data in the headquarter database storage unit 13 which is a destination added to the transaction data.

  The distributed transaction data is distributed by the store distribution means 12e for each store, for example, for each store A, store B, and store C, according to the business identifier added to the transaction data.

  The transaction data distributed for each store is distributed by the business distribution means 12f for each business, for example, for each order, purchase, or receivable, according to the business identifier added to the transaction data, and is distributed for each business. The transaction data is updated in the central database storage unit 13 by the operation of the database management unit 12g.

In this way, each business processing means can access not only the store database storage unit of the store server via the transfer means, but also the headquarter database storage unit of the server system on the headquarters side via the network.
Since the operation of the server system on the headquarters side is the same as that on the store side, a detailed description thereof is omitted. The main difference is the processing of transaction data based on the work required at the headquarters side, specifically, the totaling processing of each store's data, such as the totaling processing of each store's sales, order data, and general inventory management. .

  The store server system 1 and the headquarter server system 2 are provided with backup means 7h and 12h, respectively, and each server has a series of transaction data processed by the database management means 7g and 12g of each server. The backup storage units 8 and 14 are stored for a certain period.

  In addition, the store database storage unit 9 and the headquarter database storage unit 13 are configured by a database for each business (store, ordering, purchasing, and accounts receivable) corresponding to the identifier added to the transaction data. The means 7h and 12h can read and write for each partition, and the entire contents of the database storage units 9 and 13 are stored in the corresponding backup storage units 8 and 14 at regular intervals, for example, every week. It also has a function to save.

  Further, the backup means 7h and 12h are provided with restoring means 7i and 12i for recovering data corruption in the store database storage unit 9 and the headquarter database storage unit 13, respectively. When the data contents in a certain partition are partially damaged (rewritten with abnormal data), the restoration means can restore the normal partition corresponding to the partially damaged database storage unit from the backup storage unit. Read and rewrite the corresponding section of the database storage unit. Then, from the transaction data stored in the backup storage unit, the data group for one week corresponding to the partially damaged partition is sequentially read, and only the corresponding database partition in the database storage unit is read with the read data. Update. By such processing, a database whose data content is partially damaged is recovered.

In this way, each transaction data is classified and stored according to the store where the transaction occurred and the target business, making it easy to select the transaction data required for recovery, enabling detailed data recovery and shortening the recovery time. It becomes.
In addition, since only the area of the business division corresponding to the damaged part of the database is partially recovered, it is possible to easily recover the system failure and shorten the recovery time.

  Next, the procedure for backing up / restoring the database will be described with reference to the explanatory diagram of FIG. 2 showing the data flow. In addition, since the store server 7 and the headquarter server 12 differ only in the presence or absence of the store distribution means 12e, in this description, the headquarter server 12 will be described and the description of the store server 7 will be omitted.

  Transaction data whose destination is described in the data is sent to the distribution means from the reception means of the headquarter server, the transaction data transmission means of the headquarters business terminal, and the business processing means of the headquarters terminal. Then, the allocating means distributes the data according to the destination into one that goes to another server system via the transfer means of the headquarter server and one that goes into the system of the headquarter server.

Next, the processing of the distribution means will be described in detail. Each holder to be described is stored in a storage unit (not shown) in the server.
The distribution means manages transaction data, so that an input holder for storing received transaction data, an output holder (for FTP) for storing transaction data sent to another server after distributing the transaction data, and a transaction An output holder (for database) that stores transaction data for updating a database in the same server after data distribution is provided.

By the operation of this distribution means, transaction data whose destination is described in the data is sequentially input to the input holder from the reception means of the headquarter server, the transaction data transmission means of the headquarters business terminal, and the business processing means of the headquarters terminal. Stored.
At the same time, the distribution means retrieves transaction data one by one from the input holder and confirms the destination in this data. If there are multiple destinations, duplicate that number of transaction data. If the destination is a database in the same headquarters, store the transaction data in the output holder (for database). If the destination is another server system Transaction data is stored in an output holder (for FTP).

At the same time, the output holder (for FTP) and the output holder (for database) are respectively examined, and if transaction data exists, one is taken out.
Then, the transaction data taken out from the output holder (for FTP) is transmitted to the corresponding server system via the network by the operation of the transfer means.
Further, the transaction data taken out from the output holder (for database) is delivered to the store distribution means.

  In accordance with the code of the data generation store in the identifier added to the delivered data, the store distribution means stores, for example, three unprocessed folders of stores A, B, and C (stored in the storage unit of the headquarter server not shown). Sort and store.

  For example, as described in the section of the target file shown in FIG. 3B, the transaction data is composed of a combination of several identifiers, and the data generation date, data generation time, data generation store / Floor, terminal number, business type, slip number can be specified. Accordingly, the store distribution means performs a process of storing the file name in the folder corresponding to the data generation store code.

  Next, as shown in FIG. 2, the business distribution means retrieves the transaction data in the unprocessed folder distributed for each store one by one and, for example, places an order according to the business type code in the identifier added to this data. , Purchase, and accounts receivable are sorted and stored in three unprocessed folders (stored in the storage unit of the headquarter server not shown). At the same time, the retrieved data is stored in the corresponding folder of the processed folder group assigned to each store.

  Next, the database management means retrieves the transaction data in the unprocessed folder distributed for each job one by one, and updates the corresponding file in the headquarter database. At the same time, the retrieved data is stored in the corresponding folder of the processed folder group assigned to each job.

  On the other hand, when the headquarters server system operates, the backup means first reads a profile definition file that defines a data backup method. This file has the content shown in FIG. 3A, for example, and can be specified for each item. This file is stored in a storage unit (not shown) in the headquarters server.

The item “Backup Execution Timing” defines when transaction data is backed up, and here, immediately before the headquarter server system is stopped is specified. In this way, for example, when the system administrator instructs the system to end, the backup process is automatically performed before the system is closed.
The “backup data storage period” is a period during which transaction data is stored, and is stored for 7 days here.
The “backup target folder” is a target folder that is stored at the timing specified in “backup execution timing”, and here, stores A to C, ordering, purchasing, and accounts receivable are specified.
“Existence of external backup medium” designates whether or not data whose “backup data storage period” has elapsed is further moved to the external backup medium. “Drive name of external backup medium” is the drive name of the migration destination.
“Processing (moving / deleting) data whose retention period has elapsed” defines whether data whose “backup data retention period” has elapsed is to be deleted or moved to an external backup medium.
“Backup timing of entire database” defines a period during which the entire data of the head office database is copied to the backup storage unit. If this period is short, the recovery time will be short, but the backup work amount will increase each time. If the period is long, the recovery time will be long, but the backup work amount will be reduced.

  In FIG. 2, the backup means automatically stores the “backup target folder”, that is, each processed folder of stores A to C, in the store in the backup storage unit when the timing of “backup execution timing” comes according to the above definition. Store as completed data. At the same time, each processed folder for ordering, purchasing, and accounts receivable is automatically stored as business processed data in the backup storage unit. At the same time, the backup progress is recorded in the backup log.

  As shown in FIG. 3B, this backup log records the target file name, the storage source, the storage destination, the storage date / time, the deletion / movement date / time, and the history of each deletion / movement processing each time. It can be displayed by an instruction from a console (not shown). In this case, the target file name, storage source, storage destination, and storage date / time are newly recorded.

  Then, as shown in FIG. 2, the backup means next determines whether the corresponding data exists in the store-processed data of the backup data or the business-processed data according to the definition of “backup data retention period” according to the above definition. If the corresponding data exists, the data whose “backup data retention period” has passed is deleted or moved to an external backup medium according to the definition of “processing of data whose retention period has passed (movement / deletion)” Determine whether or not to execute.

  Further, the backup means determines whether “the backup timing of the entire database” has come according to the above definition, and if it is the execution timing, reads the entire current headquarter database and writes it as the previous headquarter database backup data in the backup storage section. At this time, the previous corresponding headquarter database backup data is overwritten and deleted.

  Further, the restoration means provided in the backup means performs a recovery operation when the data in the database is damaged in accordance with an instruction from the system administrator. When receiving the recovery instruction from the console of the headquarter server, the restoring means displays the contents of the backup log on the console and waits for a detailed recovery instruction from the system administrator. Then, the restoring means receives a detailed recovery instruction from the console, for example, an instruction to restore which business section with which transaction data from which day to what day.

  Then, based on this instruction, the restoring means reads out the business section corresponding to the instruction from the previous headquarter database backup data and writes this data in the corresponding section of the headquarter database. The transaction data corresponding to the recovery instruction is searched and extracted from the store-processed data and the business-processed data in the backup data. Update the database for the appropriate partition.

  As described above, only the corresponding section of the headquarter database can be restored, so that the restoration time can be shortened. Also, by using the backup log, complex operations such as searching and extracting transaction data required for recovery can be performed by simply specifying various conditions (data occurrence date, occurrence store, business type, etc.) that make up the file name. Can be done easily. Furthermore, if the contents of the backup work are described as a profile definition file only once at the beginning, the backup work is automatically performed every time, so that there is no omission and accurate work can be carried out without any trouble.

  FIG. 4 is a flowchart showing the procedure of the program used in the distribution means. This program includes the input process shown in FIG. 4A, the FTP output process shown in FIG. 4B, the database output process shown in FIG. 4C, and the distribution shown in FIG. 4D. The four programs for processing are configured as tasks and operate simultaneously.

Although this description has been given by taking the store server system as an example, the distribution means in the headquarter server system is also configured with the same idea, so this description is omitted.
The number following ST represents the process step number. N represents No and Y represents Yes.

  When the input process is started in FIG. 4A, it is confirmed whether there is received transaction data (ST1). If not received, it waits (ST1-N). If received, the received data is stored in the input holder (ST2). Then jump to ST1.

  When the FTP output process is started in FIG. 4B, it is confirmed whether data exists in the output holder (for FTP) (ST11). If it does not exist (ST11-N), it waits. If it exists (ST11-Y), the data is extracted (ST12) and transferred to the transfer means (ST13). Then jump to ST11.

  When the database output process is started in FIG. 4C, it is confirmed whether data exists in the output holder (database) (ST21). If it does not exist (ST21-N), it waits. If it exists (ST21-Y), the data is taken out (ST22) and delivered to the store distribution means (ST23). Then jump to ST21.

  When the sorting process is started in FIG. 4D, it is confirmed whether data exists in the input holder (ST31). If it does not exist (ST31-N), it waits. If it exists (ST31-Y), data is extracted (ST32), and it is confirmed whether there are a plurality of destinations (ST33). If there are a plurality (ST33-Y), data is copied by the number of destinations (ST34). If not plural (ST33-N), the process jumps to ST35.

  It is confirmed whether the destination of the extracted data or the replicated data is another server system (ST35). If it is another server system (ST35-Y), the corresponding data is stored in the output holder (for FTP) (ST37). If the destination is the database of the own server system (ST35-N), the corresponding data is stored in the output holder (for database) (ST36).

Although not shown, when there are a plurality of destinations, either ST36 or ST37 is performed for each data depending on the destination for each data.
Then, the processed corresponding data in the input holder is deleted (ST38), and the process jumps to ST31.

  Next, referring to the flowchart of FIG. 5, store distribution processing (A) for realizing store distribution means, business distribution processing (B) for realizing business distribution means, and database management for realizing database management means Process (C) will be described. Each process is executed according to a program constituting each means. Each program is composed of tasks, and each program is started at the same time as the system operates.

  When the store distribution process is started, it is confirmed whether there is received transaction data (ST40), and if there is no data (ST40-N), it waits. If there is data (ST40-Y), the received data is sorted for each store according to the store number described in the file name of the transaction data, and stored in the corresponding store unprocessed folder (ST41). Then, the process jumps to ST40.

When the business distribution process is started, it is confirmed whether there is data in the unprocessed folder for each store (ST42), and if there is no data (ST42-N), it waits. If there is data (ST42-Y), data is extracted one by one from the unprocessed folder for each store (ST43), and the extracted data is distributed according to the business type described in the file name of the transaction data, and the corresponding business Store in the unprocessed folder (ST44).
Then, the take-out data is written into the processed folder for each store (ST45). Then, the process jumps to ST42.

  When the database management process is started, it is confirmed whether there is a read request from the backup means (ST50). If there is a request (ST50-Y), the corresponding data is read from the database and transferred to the backup means (ST56).

  If there is no read request (ST50-N), it is confirmed whether there is a write request from the backup means (ST51), and if there is a request (ST51-Y), the database is updated with the write data (ST57).

  If there is no write request (ST51-N), it is confirmed whether there is data in the unprocessed folder for each job (ST52), and if there is no data (ST52-N), jump to ST50 and if there is data (ST52-Y). Then, transaction data is extracted one by one from the unprocessed folder for each business (ST53), and the database is updated according to this data (ST54). Then, the extracted data is written into the processed folder for each job (ST55). Then, the process jumps to ST50.

  Next, backup processing for realizing backup means will be described using the flowchart of FIG. This process is executed according to a program constituting the backup means. In addition, this program is composed of tasks, and the program is started at the same time as the system operates. Note that steps ST71 to ST74 in FIG. 6 are steps for realizing the restoring means.

  When the backup process is started, it is confirmed whether there is a database recovery instruction from the console (ST60). If there is no instruction (ST60-N), the contents of the profile file are read (ST61), and it is determined whether the backup execution timing defined in the profile file is reached (ST62). If it is not the defined backup execution timing (ST62-N), the process jumps to ST60.

  If it is the defined backup execution timing (ST62-Y), the entire processed folder for each store is saved as backup data, and the save target file name, save source, save destination, and save time at this time are stored in the backup log file. Write as history (ST63), then save the entire processed folder for each job as backup data, and write the save target file name, save source, save destination, and save time to the backup log file as history (ST64) .

  Next, it is determined whether it is the deletion timing defined in the profile file (ST65). In other words, the processing of data whose retention period has elapsed is defined as “delete”, and a search is performed to determine whether a file whose retention date and time recorded in the backup log file has passed is present in all backup data. If it is the defined deletion timing (ST65-Y), the corresponding transaction data is deleted, and the deletion / move time is deleted in the corresponding deletion / movement date / time item of the corresponding save target file name of the backup log file. “Delete” is written as a history in the item (ST69). Then, the process jumps to ST60.

  Next, it is determined whether the movement timing is defined in the profile file (ST66). In other words, the processing of data whose retention period has passed is defined as “move”, and a search is performed to determine whether a file whose retention date and time recorded in the backup log file has passed is present in all backup data. If it is the defined move timing (ST66-Y), the corresponding transaction data is moved to the storage medium of the designated drive name, and the corresponding deletion / movement date / time item of the corresponding save target file name of the backup log file “Move” is written in the delete / move item as the history (ST70). Then, the process jumps to ST60.

  Next, it is determined whether it is the backup timing of the entire database defined in the profile file (ST67). If it is not the defined timing (ST67-N), the process jumps to ST60. If it is the defined timing (ST67-Y), the entire database is read and stored as a backup database (ST68). Then, the process jumps to ST60.

  On the other hand, if the backup process is started and there is a database recovery instruction from the console (ST60-Y), the contents of the backup log file are displayed on the console (ST71), and a recovery data instruction input from the console is accepted (ST72).

  Then, from the previous database backup data, the database file of the partition corresponding to the data of the partition to be restored instructed by the console is read and written to the corresponding partition of the database to be recovered (ST73).

  Next, the stored transaction data corresponding to the data of the section to be restored instructed on the console is searched from the store processed data or the business processed data, and the database to be recovered using the searched data The partition is updated again (ST74).

  The transaction data stored in this way is classified for each store or business, and the data generation date / time, store number, business type, slip number, etc. are added to each transaction data as a file name, and a backup log file Since the backup history is stored, transaction data necessary for database recovery can be quickly retrieved and retrieved according to the cause of the failure.

For example, if there is an abnormality in data from store A, only the data of store A that occurred at the relevant time is searched, or if there is a failure factor in the ordering work, only the order business data that occurred at the relevant time is searched. The cause of the failure can be identified in a short time.
In addition, the database is partitioned for each business, and the recovery can be shortened because the partition can be recovered as a minimum unit.

It is a block diagram which shows the flow of transaction data by this invention. It is explanatory drawing which shows the flow of the transaction data by this invention in detail. It is explanatory drawing which shows an example of the file content by this invention, (A) shows the profile definition file, (B) shows the content of the backup log. It is a flowchart for demonstrating the process of the distribution means by this invention. It is a flowchart for demonstrating each process of the store distribution means (A) by this invention, a business distribution means (B), and a database management means (C). It is a flowchart for demonstrating the process of the backup means by this invention. It is a system diagram which shows the structure of the conventional online system. It is a block diagram which shows the flow of a transaction of the conventional online system.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Store server system 2 Headquarters server system 4 Store business terminal 4a Business processing means 4b Transaction data transmission means 5 POS terminal 5a Transaction data transmission means 5b Business processing means 7 Store servers 7a, 12a Business processing means 7b, 12b Distribution means 7d Transfer Means 7f Business distribution means 7g Database management means 7h Backup means 7i, 12i Restoration means 8, 14 Backup storage section 9 Store database storage section 12 Central server 7c, 12c Receiving means 12e Store distribution means 12f Business distribution means 12g Database management Means 13 Headquarter database storage 83 Network

Claims (4)

A business terminal that processes input data and converts it into transaction data, and transmits the transaction data through a communication line; and a server that receives the transaction data and updates a corresponding database and backs up the database. In a database backup method performed in an online system in which a plurality of configured server systems are connected via a network,
The business terminal transmits the transaction data by adding a destination for identifying the database and an identifier for identifying the business category, and the server distributes the transaction data according to the destination, and a database corresponding to the destination. Transferred to the server system comprising:
The server receiving the distributed transaction data updates the database corresponding to the destination, classifies the transaction data according to the identifier, and stores the data as backup data. . The database is divided and configured for each business corresponding to the transaction data, and the server replicates the database every predetermined period and stores it as a backup database,
When the data of the database is damaged, the database stores the database by copying the area of the business division corresponding to the damaged part from the backup database to the damaged database, and updating the database using the backup data. The database backup method according to claim 1, wherein the database is restored. A business terminal that processes input data and converts it into transaction data, and transmits the transaction data through a communication line; and a server that receives the transaction data and updates a corresponding database and backs up the database. In an online system in which a plurality of configured server systems are connected via a network,
The business terminal comprises a business processing means for adding a destination for identifying the database, an identifier for identifying the business classification to the transaction data, and a transaction data transmission means for transmitting the transaction data.
The server includes a distribution unit that distributes the transaction data according to the destination, a database update unit that updates a database corresponding to the destination in the server with the transaction data, and a database corresponding to the destination Transfer means for transferring to the system, business distribution means for classifying the transferred transaction data, and storing the classified transaction data as backup data, and replicating the database every predetermined period as a backup database An on-line system comprising backup means for saving.   The database is configured to be divided for each business corresponding to the transaction data, and when the data in the database is damaged, the backup unit is configured to identify the area of the business division corresponding to the damaged portion from the backup database. 4. The on-line system according to claim 3, further comprising a restoring means for copying the damaged database and restoring the database by updating the database using the backup data.

Images