JP2014071740A - Information processor, management method, and management program - Google Patents

Abstract

[PROBLEMS] To reduce recovery time.
In an information processing apparatus, a write command for a storage device is analyzed, and consistency check information including a read command for reading data written to the storage device by the write command and a read expected value is acquired. An instruction analysis means; a storage means for storing the consistency check information obtained by the instruction analysis means in association with the write instruction; an execution state of the write instruction at the time of failure recovery of the storage device; and the storage means Instruction execution means for confirming the consistency of the data written in the storage device based on the consistency information stored in the storage device.
[Selection] Figure 2

Description

  The present application relates to an information processing apparatus, a management method, and a management program.

  Conventionally, when a failure occurs in a management system equipped with a server, database, etc., a person in charge such as an administrator analyzes the cause based on his / her experience while referring to log information during operation. Have been restored. Here, conventionally, for example, when a failure occurs in a monitored server, a failure response system that generates a schedule that combines necessary response procedures to recover the failure so that the failure can be recovered in time for the deadline Exists. Conventionally, a batch job that has been interrupted due to the occurrence of a system failure can be retrieved from a database that stores the status of the batch job, or a recovery job can be resumed for a batch job that has been interrupted due to the occurrence of a system failure. There is a method for starting up (see, for example, Patent Documents 1 and 2).

  Conventionally, when a failure occurs while accessing the database and writing data, the consistency between the data stored in the database and the data written to the database is confirmed, resulting in inconsistencies. In some cases, it is necessary to restore the data.

JP 2008-210148 A JP 2008-181735 A

  However, in the conventional method described above, when checking the consistency of the data stored in the database, it takes time because the person in charge or the like manually and visually confirms the input data. In addition, in the conventional method described above, it is necessary to interrupt the writing operation to the database for a long time in order to check all the data in the database regardless of whether there is inconsistent data stored in the database. Occurs. Furthermore, in the above-described conventional method, in order to eliminate data inconsistency and restore data, it is necessary to write data again after returning to the state before data writing using backup data or the like. Therefore, the conventional method requires a considerable recovery time.

  In view of such a problem, the disclosed technique aims to shorten the recovery time.

  An information processing apparatus according to one aspect of the disclosure analyzes a write command to a storage device, and acquires consistency check information including a read command and a read expected value for reading data written to the storage device by the write command Instruction analyzing means, storage means for storing the consistency check information obtained by the instruction analyzing means in association with the write instruction, execution status of the write instruction at the time of failure recovery of the storage device, and the storage Instruction executing means for confirming the consistency of data written in the storage device based on the consistency information stored in the means.

  According to the disclosed technology, the recovery time can be shortened.

It is a figure which shows the schematic structural example of a data access management system. It is a figure which shows an example of a function structure of an access management apparatus. It is a figure which shows an example of the hardware constitutions which can implement | achieve access management processing. It is a figure which shows the example of data memorize | stored in a command memory | storage means. It is a figure which shows the example of data memorize | stored in ID state memory | storage means. It is a figure for demonstrating the specific example of the memory content of a database. It is a flowchart which shows an example of a data writing process. It is a figure for demonstrating the outline | summary of the data write process at the normal time. It is a flowchart which shows an example of a failure recovery process. It is a figure for demonstrating the outline | summary of a failure recovery process.

  Hereinafter, embodiments will be described in detail with reference to the accompanying drawings.

<Example of schematic configuration of data access management system>
FIG. 1 is a diagram illustrating a schematic configuration example of a data access management system. A data access management system 10 illustrated in FIG. 1 schematically includes an access management device 11 as an example of an information processing device, a database 12 as an example of a storage device, and one or a plurality of terminals 13-1 to 13-n. (Hereinafter collectively referred to as “terminal 13” as necessary). In the data access management system 10 of FIG. 1, the access management apparatus 11 and the terminal 13 are connected in a state where data can be transmitted and received through the communication network 14. In the present embodiment, the access management apparatus 11 and the database 12 may be connected via the communication network 14 in a state where data can be transmitted and received.

  The access management device 11 performs management of data access such as reading and writing of data with respect to the database 12 based on an instruction from the terminal 13. For example, the access management apparatus 11 performs consistency check information (for example, for checking the consistency between data written in the database 12 and data written to the database 12 as pre-processing for writing data to the database 12). A read command for confirmation and an expected read value) are generated and stored in the storage means or the like in association with the write command. Further, the access management apparatus 11 stores state information for execution such as the above-described preprocessing and the start and end of data writing to the database 12 in a storage unit or the like. In addition, when a failure or the like occurs, the access management apparatus 11 determines a data writing interruption point based on the above-described consistency confirmation information, status information, and the like. Therefore, the access management apparatus 11 can quickly resume data writing from the write interruption point when the failure of the database 12 is recovered. The failure is, for example, a failure due to a failure of each device in the data access management system 10 or a system failure such as a network failure or a power failure, but is not limited thereto.

  In other words, in the present embodiment, the data writing operation is determined for the data whose writing has been interrupted due to a failure, and after the failure is recovered, the writing is resumed from the interruption point. Reduce interruption time. Also, in this embodiment, when data inconsistency occurs, the process from the occurrence of failure to recovery is simplified without performing the process of returning to the state before data storage and inputting the data again. The consistency can be eliminated.

  Here, the access monitoring apparatus 11 may be, for example, a personal computer (PC), a notebook PC, a server, or the like, but is not limited thereto.

  The database 12 stores, for example, various setting information for executing a predetermined process in business, input / output data for execution of the predetermined process, and the like. Further, the database 12 can read and write data requested from the access management apparatus 11 that has received an instruction from the terminal 13 at a predetermined timing, for example.

  Here, the database 12 is a collection of various types of information, and is systematically configured so that the information can be searched and extracted using, for example, keywords. The database 12 in the present embodiment may be a database server, a data center composed of a plurality of databases, or a part or all of a database designed by cloud computing. .

  The terminal 13 is a computer used by each user. The terminal 13 sends a data read (input) request and a data write (output) request after processing to the database 12 which are necessary when each terminal 13 performs a predetermined process corresponding to each job. And input / output data. In the data access management system 10 shown in the example of FIG. 1, the request information from the terminal 13 is transmitted to the access management apparatus 11 via the communication network 14 and transmitted to the database 12 via the access management apparatus 11. Is done.

  Here, the terminal 13 may be, for example, a PC, a notebook PC, or the like, and may be various communication terminals such as a smartphone, a tablet terminal, or a mobile terminal, a game device, a music playback device, or the like, but is not limited thereto. Is not to be done.

  The communication network 14 has a network form represented by, for example, the Internet or a local area network (LAN), but is not limited thereto.

  In the data access management system 10 described above, for example, even when a failure occurs during data writing to the database 12 and the data becomes inconsistent, the access management device 11 resumes writing from the write interruption point after recovery from the failure. Thus, the data inconsistency is resolved. Therefore, in this embodiment, it is not necessary for an administrator or the like to confirm the possibility of data inconsistency at the time of failure recovery.

  Further, in the present embodiment, after the failure is recovered, the user who uses the terminal 13 enters the “input completion state” for the data to be input to the database 12, so that the trouble of rewriting the data that was not written due to the occurrence of the failure is reduced. Does not occur. Thereby, in this embodiment, since the process from failure occurrence to recovery can be simplified and the recovery time can be shortened, for example, the interruption time for data input to the database can be greatly shortened. Note that the recovery time means, for example, the time from the occurrence of a failure until the recovery is completed (until the state before the failure occurs), but is not limited thereto.

  Conventionally, for example, when a failure occurs while data is being written to the database 12, it is not known how far the data has been written, and the same data may be written twice. Therefore, conventionally, it has been necessary to temporarily return to the state before data input using backup data or the like, and input data again.

  However, according to the present embodiment, after the execution state of the write command is confirmed by the access management device 11, if the state is not completed, the value can be written by checking the value of the database 12 with the read command for confirmation. It is possible to quickly confirm at which position the interruption is made. Therefore, in this embodiment, data writing can be resumed from an interrupt point where writing has not been completed, and the processing from the occurrence of a failure to recovery can be simplified to shorten the recovery time.

<Example of Functional Configuration of Access Management Device 11>
Next, the access management apparatus 11 described above will be specifically described. FIG. 2 is a diagram illustrating an example of a functional configuration of the access management apparatus. The access management apparatus 11 shown in FIG. 2 includes an input unit 21, an output unit 22, an ID management unit 23, a command analysis unit 24, a command conversion unit 25, a command execution unit 26, a storage unit 27, and transmission / reception. Means 28 and control means 29 are provided. The instruction execution means 26 includes a write instruction execution means 31 and a read instruction execution means 32. The storage unit 27 includes an instruction storage unit 41 and an ID state storage unit 42 as an example of a state storage unit.

  The input unit 21 accepts various inputs such as start and end of various instructions and setting input from an administrator or the like who uses the access management apparatus 11. Specifically, the input unit 21 receives instructions such as an ID management instruction, a command analysis instruction, a command conversion instruction, a command execution instruction, a storage instruction, and a transmission / reception instruction in the present embodiment.

  Input of information acquired by the input means 21 may be input using an input interface such as a keyboard or a mouse, a touch panel type input using a screen, or input using an operation key or the like. Furthermore, the input unit 21 may include a voice input unit that inputs voice using, for example, a microphone.

  The output unit 22 outputs the content input by the input unit 21 and the content executed based on the input content. The output means 22 may have a display means such as a display or a monitor when outputting by screen display, for example, and may have an audio output means such as a speaker when outputting by sound. . The input unit 21 and the output unit 22 may be integrated with input / output, such as a touch panel.

  The ID management means 23, for example, with respect to one or a plurality of instruction groups for performing predetermined processing on the database 12 according to instructions from the instruction analysis means 24 etc., has unique identification information that does not overlap with other instruction groups ( IDentification (ID) is assigned and managed. In addition, although a predetermined | prescribed process shows one transaction process etc. which are performed with a certain terminal 13, for example, it is not limited to this. Transaction processing refers to a series of information processing units that cannot be separated. That is, the ID may be set for each instruction, or may be set for each of a plurality of instruction groups (for example, 5 instructions, 10 instructions) for executing the transaction processing described above. By managing in units of transaction processing, for example, data access management to the database 12 can be realized for each processing unit including a plurality of instructions. Further, the predetermined processing described above may be processing performed by each terminal 13 or processing performed by the access management apparatus 11.

  The command analysis unit 24 extracts one or a plurality of write command groups from a plurality of requests obtained from each terminal 13, acquires an ID from the ID management unit 23 for the extracted write command groups, and writes the write command group One ID is assigned to. The write command group can be extracted based on, for example, a command type (for example, WRITE) corresponding to the write command, but is not limited thereto.

  Further, the command analysis unit 24 performs preprocessing for each ID, for example. Here, the preprocessing refers to acquiring information (consistency confirmation information) for confirming data consistency before actually executing a request (command) for the database 12 obtained from the terminal 13 or the like, for example. It means a process of storing the acquired information in the instruction storage means 41.

  Specifically, in the preprocessing, for example, all write commands received from the terminal 13 or the like are analyzed, and the table, record, column (item), etc. of the database 12 to be written are specified. In the pre-processing, for example, a write command is converted by the command conversion means 25, and a read command (hereinafter referred to as “read command for confirmation”) for reading data written by execution of the write command, a read expected value, etc. Consistency check information is generated. In the preprocessing, the generated consistency check information is stored in the instruction storage unit 41 in association with the write instruction.

  Further, the instruction analysis unit 24 stores the state information for each ID in the ID state storage unit 42. Here, the state information is information related to the execution state of the process for each ID. For example, when the above-described preprocessing is started for a certain ID, the state information corresponding to the ID becomes “preprocessing”. . The state information corresponds to, for example, the state information corresponding to the case where the execution of the write command from the terminal 13 or the like to the database 12 is started by the command execution unit 26 for each ID after the pre-processing described above or when the execution is completed. Means “start” and “end”, but is not limited thereto.

  The instruction analysis unit 24 stores the “pre-processing” state information in the ID state storage unit 42. The above-described “start” and “end” state information is stored in the ID state storage unit 42 by the instruction execution unit 26. Further, the instruction analysis unit 24 stores “abnormal end” status information in the ID status storage unit 42 when an error or the like occurs during the execution of the above-described preprocessing.

  When the command analysis unit 24 extracts a write command from a plurality of commands, the command conversion unit 25 converts the extracted write command and stores the write destination (for example, table, record, column) of the database 12. A read command for confirmation for reading data is generated. In addition, the command conversion unit 25 generates an expected read value from the write command, and outputs consistency confirmation information including the generated read expected value and the confirmation read command described above to the command analysis unit 24.

  The command execution unit 26 executes a write command in the write command execution unit 31, executes a read command in the read command execution unit 32, and the like. Note that the instruction execution unit 26 is not limited to the above-described execution of writing or reading, and executes various instructions by, for example, an interrupt instruction or a predetermined control command (for example, sorting or merging data in the database 12). Etc. can be performed. The read command in the read command execution means 32 includes a normal read command obtained from the terminal 13 or the like.

  Here, the instruction execution unit 26 causes the write instruction execution unit 31 to store, for example, “ID”, “date”, “time”, “state”, etc. in the ID state storage unit 42. As the “state”, for example, state information such as “start”, “complete”, and “abnormal end” is stored, but is not limited thereto.

  For example, the write command execution unit 31 stores the state information in the ID state storage unit 42 as “start” before the execution of the write command at the head of each ID. The write command execution means 31 reads the write command from the information stored in the command storage means 41, performs the process for each ID, and executes the process for all the write instructions of all IDs. For example, when all the write commands within the same ID are completed, the write command execution unit 31 stores the state information in the ID state storage unit 42 as “completed”. The write command execution unit 31 sets the status information to “abnormal end” and stores it in the ID status storage unit 42 when, for example, a write error occurs, and stops the subsequent write command.

  Further, the read command execution means 32 reads a value stored in a predetermined area of the database 12 in order to confirm whether or not the database 12 is normally activated or operating at the time of failure recovery, for example. Here, as the value to be read, it is preferable to store a predetermined value that is not deleted in a predetermined area (for example, a table, a record, or a column) of the database 12 at the time of system construction, for example. Is not to be done. The predetermined value may be, for example, identification information of the database 12 or dummy data that cannot be changed or deleted.

  The read command execution means 32 determines that the state of the database 12 is normal when the read predetermined value matches the value previously stored in the storage means 27 or the like. Further, when the predetermined value cannot be read or the read predetermined value does not match the value stored in advance in the storage means 27 or the like, the read command execution means 32 does not start or operate normally. And error processing (for example, error message output) is performed.

  In the present embodiment, the state of the database 12 is confirmed by reading a predetermined value for the database 12 described above, but the present invention is not limited to this. For example, in the present embodiment, the predetermined value described above is not read, but, for example, the instruction storage unit 41 or the ID state storage unit 42 is read, and error processing may be performed when the read is not possible.

  Further, when the predetermined value described above can be read correctly, the read command execution means 32 reads the state of each ID from the ID state storage means 42 and acquires an ID that satisfies the predetermined condition. The predetermined condition may be, for example, a case where the status information is “start” and not “completed”, but is not limited thereto. Further, as the ID reading order, for example, the date and time (“date”, “time”) of the ID state storage means 42 may be referred to, for example, the date may be read in order from the oldest date, or the record with the latest date and time may be used. Data that has been set back in advance for a predetermined period (for example, one week, one month, one year) may be read. Further, for an ID whose status information is “pre-processing”, this means that the status has not been written to the database 12, so a message to that effect may be output from the output means 22 or the like. .

  Further, when there is an ID that satisfies the above-mentioned predetermined condition, the read command execution means 32 checks the consistency between the data written in the database 12 by the write command corresponding to the ID and the data written to the database 12. Check.

  Specifically, the read command execution unit 32 reads the confirmation read command and the read expected value from the command storage unit 41 and executes the read confirmation read command. Further, the read command execution means 32 determines whether or not the value read by the check read command matches the expected read value, and checks the data consistency. Here, the read command execution means 32 can determine that the data stored in the database 12 is inconsistent when the value read by the confirmation read command does not match the expected read value.

  Therefore, the instruction execution unit 26 executes the corresponding write instruction by the write instruction execution unit 31 when the read value does not match the read expected value. In this embodiment, after the execution of the above-described write command, the confirmation read command may be executed again to compare the read value with the read expected value.

  Note that if the value read by the read command execution unit 32 is the same as the expected read value, the instruction execution unit 26 assumes that the values are consistent and does not execute the write command described above.

  Further, the instruction execution means 26 sets the state to “completed” when all the read instructions within the same ID (including the write instruction if the corresponding write instruction is necessary) are completed, and stores them in the ID state storage means 41. Remember me. Further, the command execution unit 26 sets the status to “abnormal end” when a write error is generated by the write command execution unit 31 or a read error occurs by the read command execution unit 32, and is stored in the ID status storage unit 42. Store it and stop the subsequent processing.

  The storage unit 27 stores various types of information necessary in the present embodiment. Specifically, the storage unit 27 stores various information stored in the instruction storage unit 41 and the ID state storage unit 42, execution progress, execution results, log information, and the like.

  The storage unit 27 can read and write various stored information at a predetermined timing as necessary. The storage means 27 is a collection of various types of information as described above, and functions as a database that is systematically configured so that such information can be searched and extracted using, for example, keywords. You may have. The storage means 27 is composed of, for example, a hard disk or a memory.

  The transmission / reception unit 28 is a communication unit for transmitting / receiving data to / from an external device such as the terminal 13 or the database 12 via the communication network 14. The transmission / reception means 28 can receive various information already stored in the external device or the like, and can transmit the result processed by the access management device 11 to the external device or the like via the communication network 14. it can.

  The control unit 29 controls the entire components of the access management apparatus 11. Specifically, the control unit 29 performs each control related to access management to the database 12 or the like based on, for example, an instruction from a user using the terminal 13 or an instruction from the input unit 21 by an administrator or the like. . Here, each control includes, for example, ID management for each instruction group in the ID management means 23, instruction analysis in the instruction analysis means 24, instruction conversion in the instruction conversion means 25, and the like. Each control includes execution of various instructions in the instruction execution means 26 and storage of various information in the storage means 27, but is not limited thereto. These controls may be performed based on execution of a program, execution of a predetermined event, command, or the like according to an instruction from an administrator or the like, but is not limited thereto.

<Hardware Configuration Example of Access Management Device 11>
Here, in the access management apparatus 11 described above, an execution program (management program) capable of causing a computer to execute each function is generated, and the execution program is installed in, for example, a general-purpose PC or server. The access management process for the database 12 in the embodiment can be realized. Here, a hardware configuration example of a computer capable of realizing the access management process in the present embodiment will be described with reference to the drawings.

  FIG. 3 is a diagram illustrating an example of a hardware configuration capable of realizing the access management process. 3 includes an input device 51, an output device 52, a drive device 53, an auxiliary storage device 54, a main storage device 55, a central processing unit (CPU) 56, and a network connection device 57. These are connected to each other via a system bus B.

  The input device 51 includes a pointing device such as a keyboard and a mouse operated by an administrator of the access management device 11 and a voice input device such as a microphone. The input device 51 inputs a program execution instruction from the administrator or the like, various operation information, information for starting up software, and the like.

  The output device 52 has a display for displaying various windows and data necessary for operating the computer main body for performing the processing in the present embodiment. Can be displayed. In addition, the output device 52 can print the above processing result or the like on a print medium such as paper and present it to an administrator or the like.

  Here, the execution program installed in the computer main body in this embodiment is provided by, for example, a universal recording bus 58 such as a Universal Serial Bus (USB) memory, a CD-ROM, or a DVD. The recording medium 58 on which the program is recorded can be set in the drive device 53, and an execution program included in the recording medium 58 is transmitted from the recording medium 58 via the drive device 53 on the basis of a control signal from the CPU 56. To be installed.

  The auxiliary storage device 54 stores an execution program in this embodiment, a control program provided in the computer, an execution process, an execution result, and the like based on a control signal from the CPU 56. The auxiliary storage device 54 can read and write necessary information from each stored information based on a control signal from the CPU 56 and the like.

  The auxiliary storage device 54 includes, for example, a hard disk drive (HDD), a solid state drive (SSD), and the like, and corresponds to the storage unit 27 described above, for example.

  The main storage device 55 stores an execution program or the like read from the auxiliary storage device 54 by the CPU 56. The main storage device 55 is composed of a read only memory (ROM), a random access memory (RAM), and the like.

  The CPU 56 controls processing of the entire computer such as various operations and data input / output with each hardware component based on a control program such as an operating system and an execution program stored in the main storage device 55. Each process can be realized. Various information necessary during the execution of the program can be acquired from the auxiliary storage device 54, and the execution result and the like can also be stored.

  Specifically, the CPU 56 executes a management program installed in the auxiliary storage device 54 on the basis of, for example, a program execution instruction obtained from the input device 51, thereby executing processing corresponding to the program on the main storage device 55. I do.

  For example, by executing the management program, the CPU 56 executes ID management of each instruction group by the ID management means 23, instruction analysis by the instruction analysis means 24, instruction conversion by the instruction conversion means 25, and execution of various instructions by the instruction execution means 26. Perform execution control. Further, the CPU 56 performs execution control such as writing and reading of various information in the storage unit 27. In addition, the processing content in CPU56 is not limited to the content mentioned above. The contents (execution progress and execution results) executed by the CPU 56 can be stored in the auxiliary storage device 54 as necessary.

  The network connection device 57 acquires an execution program, software, various instructions, and the like from an external device or the like connected to the communication network 14 by connecting to the communication network 14 or the like based on a control signal from the CPU 56. Further, the network connection device 57 can provide an execution result obtained by executing the program or the execution program itself in the present embodiment to an external device or the like.

  With the hardware configuration described above, it is possible to execute the access management process for the database in the present embodiment. Further, by installing the program, the access management process in the present embodiment can be easily realized on a general-purpose PC or the like.

<Data example>
Next, various data examples in the present embodiment will be described with reference to the drawings. FIG. 4 is a diagram showing an example of data stored in the instruction storage means.

  The items of data stored by the command storage means 41 shown in FIG. 4 include, for example, “ID”, “write command”, “read command for confirmation”, “read expected value”, and the like, but are not limited thereto. It is not a thing. Note that the data stored in the instruction storage unit 41 is information generated and stored as a pre-process in which an actual process (command) is performed.

  The data example shown in FIG. 4 is, for example, a command for a write command (for example, a WRITE command) for the database 12 or a read command for confirmation (for example, a READ command), and other commands are not stored. The present invention is not limited, and all operations for performing a predetermined process may be stored.

  “ID” is a value set to identify one or a plurality of instruction groups for performing a predetermined process. In this embodiment, records with the same ID are stored in one or a plurality of write command groups corresponding to a predetermined process. The predetermined process indicates, for example, one transaction process executed by a certain terminal 13, but is not limited thereto. Transaction processing refers to a series of information processing units that cannot be separated. In other words, in this embodiment, the ID management unit 23 sets the same ID for a group of instructions that writes data to one or more positions in the database 12 by, for example, one transaction process. Note that the ID setting method is not limited to this.

  The “write command” writes the value of a command (WRITE command) as to what value is written in which column (item) of which record of which table in the database 12 as an actual command. This command is a write command for actual processing requested from the terminal 13 or the like, for example.

  The “read command for confirmation” stores a command for reading data written in the database 12 by executing the write command. More specifically, the “read command for confirmation” converts the write command to generate a read command for confirmation.

  In addition, the “read expected value” is set to an expected value of what value is read if the instruction (READ instruction) is executed with the contents stored in the “read instruction for confirmation” if it is normal. Is done. Specifically, the “read expected value” is set as the read expected value by the write value in the write command. The “confirmation read command” and the “read expected value” described above are examples of the consistency confirmation information.

  For example, in the example of FIG. 4, the “write command” of the first record whose “ID” is “1000” is “write 1000 to column C2 of the record of column C1 = 100 of table X”. In this case, in this embodiment, the command conversion means 25 converts the write command described above to generate a “read command for confirmation”. Specifically, an instruction “read the value of column C2 of the record of column C1 = 100 from table X” is stored. As the expected read value, a result of “C2 = 1000” is stored in association with the content of the write command.

  In the present embodiment, it is possible to write to a plurality of columns or the like by one write command. For example, in the example of FIG. 4, the “write command” of the first record whose “ID” is “1001” writes “A” to column C2 of the record of column C1 = 100 of table A and 1100 to column C3. "Write". In this case, the “read command for confirmation” is “reads the values of the columns C2 and C3 of the record of the column C1 = 100 from the table A”, and the “read expected value” is “C2 = 'A', C3 = 1100. "Become.

  FIG. 5 is a diagram illustrating an example of data stored in the ID state storage unit. The items of data stored by the ID state storage unit 42 illustrated in FIG. 5 include, for example, “ID”, “date”, “time”, and “state”, but are not limited thereto. For example, the ID state storage unit 42 may store “date” and “time” as the same item “date and time”.

  The ID state storage means 42 stores a state for each ID and a date and time when the state information is stored. The state types include “pre-processing”, “start”, “complete”, “abnormal end”, and the like, but are not limited thereto.

  “Pre-processing” means, for example, processing for generating data stored in the above-described instruction storage means 41 before actually executing all received write instructions, but is not limited thereto. Further, the status information of “preprocessing” is stored together with the ID, date, and time at the start of the preprocessing.

  “Start” means, for example, the start of data writing actually corresponding to the write command after the end of “preprocessing”, and the “start” status information includes the ID at the start of the write command for each ID, Stored with date and time.

  “End” means, for example, that data writing has actually been executed, and that execution has ended normally, and the “end” status information includes the ID, date, Stored with time.

  “Abnormal termination” means, for example, that an error has occurred with respect to preprocessing or a write error has occurred during execution of a write command, and the status information of “abnormal termination” includes the ID, date, and date when an error occurs for each ID. , Stored with time. In the case of abnormal termination, a message to that effect (such as an error message) may be output by the output means 22 such as a screen.

  In the data example stored in the ID state storage means 42, preprocessing, start, completion, etc. for the same ID composed of one or a plurality of instructions are managed. Therefore, the record of “complete” is stored when the execution of the write command for all the same IDs is normally completed.

  Here, FIG. 6 is a diagram for explaining a specific example of the stored contents of the database. In the example of FIG. 6, columns (items) C1, C2, C3, C4,... Are provided as the table X existing in the database 12. Further, it is assumed that identification information for identifying each record of the table X shown in FIG. 6 is set in the column C1.

  Here, for example, in the data example in the instruction storage means 41 shown in FIG. 4 described above, in the command for writing the first record with ID = 1000, “1000 is written in column C2 of the record in column C1 of table X = 100” is designated. Has been. In the example of FIG. 6, when the above write command is executed, 1000 is written to the column C2 of the record of C1 = 100 of the table X. In this embodiment, for example, values can be written to a plurality of columns by one write command.

<Access management process for database>
Next, a specific example of the access management process for the database according to the present embodiment will be described with reference to a flowchart. In the following processing, in order to describe an example of the access management processing in the present embodiment, the description will be divided into normal data writing processing and failure recovery processing when a system failure or the like occurs during data writing.

<Normal data writing process>
FIG. 7 is a flowchart illustrating an example of the data writing process. FIG. 8 is a diagram for explaining the outline of the normal data writing process. In the example of FIG. 8, each configuration of the access management apparatus 11 used in the normal data writing process shown in FIG. 7 is schematically shown.

  In the example of FIG. 7, the data write process receives one or a plurality of write command groups to the database 12 from each terminal 13-1 to 13-n or the like (S01). In addition, instructions other than the write instruction (for example, a read instruction or the like) are also transmitted from each terminal 13-1 to 13-n to the database 12. Therefore, the process of S01 includes a process of extracting a write instruction group from a plurality of instructions obtained from the terminals 13-1 to 13-n.

  Next, in the data writing process, as shown in FIG. 8, the command analysis unit 24 acquires the ID for the write command group from the ID management unit 23, and sets the acquired ID to the corresponding write command (S02). For example, the same ID is set for one or a plurality of write command groups generated in units of transaction processing executed individually by the terminal 13, but the ID is not limited to this.

  Next, in the data writing process, the processes of S04 to S11 shown below are repeated for the set number of all IDs (S03).

  In the data writing process, the instruction analysis unit 24 stores the ID, date, time, and state in the ID state storage unit 42 (S04). At this time, “pre-processing” is stored in the “state” in the ID state storage unit 42 as state information (ID state).

  Next, in the data writing process, the command analysis unit 24 analyzes the write command, the command conversion unit 25 converts the write command, and the information included in the consistency check information includes the above-described check read command, read expected value. Is generated (S05). In the data writing process, the ID, the write command, the confirmation read command, and the read expected value are stored in the command storage unit 41 (S06). Note that the processes in S05 and S06 correspond to the pre-processing described above.

  Next, in the data writing process, the ID, date, time, and state are stored in the ID state storage unit 42 by the write command execution unit 31 as shown in FIG. 8 (S07). At this time, “start” is stored in the “state” in the ID state storage unit 42 as state information (ID state). In the data writing process, the following processes of S09 and S10 are repeated for the number of write instructions with the same ID (S08).

  In the data writing process, as shown in FIG. 8, the write command execution unit 31 reads the write command from the command storage unit 41 (S09). In the data write process, the write command read out by the write command execution means 31 in the process of S09 is executed (S10).

  Further, in the data writing process, after the writing command execution means 31 finishes writing the same number of write instructions as the same ID, the ID state storage means 42 stores the ID, date, time, and state (S11). At this time, “complete” is stored in the “state” in the ID state storage means 42.

  Note that, for example, when the write command is not normally completed due to an error due to a hardware failure of the database 12 or a communication error, the retry process may be performed a predetermined number of times or within a predetermined time. Further, when the write command is not normally completed even in the retry process within a predetermined number of times or within a predetermined time, the “state” of the ID state storage means 42 is in the “abnormal end” state as the state information (ID state). Information may be stored.

  In the data writing process shown in the example of FIG. 7, an example in which “preprocessing” to “completion” are sequentially performed for each ID is shown, but the present invention is not limited to this. For example, in this embodiment, even if the state of processing for a certain ID has not yet been “completed”, the processing from “preprocessing” to “completed” for other IDs can be performed in parallel.

<Disaster recovery processing>
Next, an example of failure recovery processing will be described using a flowchart. FIG. 9 is a flowchart illustrating an example of the failure recovery process. FIG. 10 is a diagram for explaining an outline of the failure recovery processing. In addition, in the example of FIG. 10, each structure of the access management apparatus 11 used in the failure recovery process shown in FIG. 9 is shown schematically.

  In the example of FIG. 9, in the failure recovery process, the database state is read by the read command execution means 32 as shown in FIG. 10 at the time of failure recovery (S21). Specifically, in the process of S21, for example, predetermined data that is stored in advance in the database 12 and is not changed is read. That is, in the processing of S21, for example, in order to check whether an error due to a hardware failure of the database 12, a communication error, or the like has occurred, the database 12 is accessed and data is read.

  Here, in the failure recovery process, it is determined whether or not the reading in the process of S21 has been normally performed (reading = OK?) (S22). In the process of S22, the data read out in the process of S21 is compared with the data stored in advance in the storage means 27, and if the data match, the data can be read normally (reading = OK). to decide. If the data do not match, it can be determined that the data has not been read normally (reading = NG), but the determination method is not limited to this.

  In the failure recovery process, when the reading is normally performed (reading = OK) (YES in S22), the state for each ID is read from the ID state storage unit 42 by the read command execution unit 32 as shown in FIG. 10 (S23). ). In the process of S23, for example, by reading all the status information from the head of the data stored in the ID status storage means 42, the processing status for all IDs can be confirmed.

  In the process of S23, state information for a predetermined number (for example, for the past one week or one month) may be read from the latest data stored in the ID state storage unit 42. It is highly probable that processing will be interrupted when a failure occurs, because the write command was executed immediately before the failure occurred, so recovery processing can be performed by reading status information from recent data. The necessary processing can be found quickly.

  Next, the failure recovery process repeats the following processes of S25 to S35 for the number of IDs read from the ID state storage means 42 (S24). In the failure recovery process, it is determined whether or not the state of the ID read from the ID state storage means 42 is “start” (state = “start”?) (S25). In the process of S25, specifically, when there is a record whose ID state read from the ID state storage means 42 is “start” and there is no “complete” record corresponding to the record of “start”. Means. For example, in the example of FIG. 5 described above, the ID “1001” is in a state where “start” is present and “completion” is not present. For IDs having “start” and “complete” records, the normal write processing is normally completed before the failure occurs, and is excluded from the target of the failure recovery processing.

  In the failure recovery process, when the state of the read record is “start” (YES in S25), the read command execution means 32 reads and reads one or more confirmation read instructions corresponding to the ID from the instruction storage means 41. The expected value is read (S26). In the failure recovery process, the following processes of S28 to S33 are repeated for the number of confirmation read instructions (S27).

In the failure recovery processing, the confirmation read command read by the read command execution means 32 is executed (S28), and whether or not the read data matches the value stored in the read expected value (read data ≠ read) Is it an expected value?) (S29). In the failure recovery processing, when the read data does not match the expected read value (YES in S29), an instruction is sent from the read command execution means 32 to the write command execution means 31. The write command execution unit 31 reads the corresponding ID write command from the command storage unit 41 (S30).
In the failure recovery process, the write instruction executed by the write instruction execution means 31 is executed in S30 (S31). In this embodiment, after the execution of the write command in S31, the confirmation read command may be executed again to compare the read value with the read expected value. Thereby, the consistency of data can be confirmed more appropriately.

  In the failure recovery process, it is determined whether or not it is the last write command in the same ID (S32). If the failure recovery process is the last write command (YES in S32), the ID, date, time, and the like when the write process is performed on the ID state storage unit 42 by the read command execution unit 32 as shown in FIG. The state is stored (S33). At this time, in the “state” of the ID state storage means 42, for example, “complete” is stored as the state information (ID state). However, the present invention is not limited to this, and information such as “recovery complete” is stored. It may be stored. Thereby, in this embodiment, it can be grasped | ascertained easily whether the recovery was completed with reference to the ID state memory | storage means 42. FIG.

  Further, in the failure recovery process, when the state of the read record is not “start” in the process of S25 described above (NO in S25), it is determined whether or not the state is “preprocess” (S34). . In the process of S34, the status is “pre-process”, for example, there is a record indicating the status information of “pre-process”, and a record indicating the status information of “start” or “complete” for the ID. Means that does not exist.

  In the failure recovery processing, when the state is “preprocessing” (YES in S34), the writing processing for the ID is not started, and all data may not be stored in the instruction storage unit 41 in some cases. . Therefore, the failure recovery process outputs a message to that effect (such as an error message) on the screen or the like as an error process (S35). In the failure recovery process, if the state is not “pre-process” in S34 (NO in S34), the state of the ID is “complete”, and the write command for the ID has been completed normally. Judge that.

  Further, in the failure recovery process, when the database state cannot be read in the process of S22 described above (reading = NG) (NO in S22), a message to that effect (such as an error message) is displayed on the screen as an error process. (S36), and the process is terminated. In this case, since the database 12 cannot be accessed and the recovery process itself cannot be performed, a message to that effect is output and the process ends. In this embodiment, the state of the ID state storage means 42 may be “abnormal end”, but in this case, there is a possibility of an error such as a hardware failure that cannot be recovered. Excluded from failure recovery processing.

  According to the embodiment described above, the recovery time can be shortened by simplifying the processing from the occurrence of a failure to the recovery. So far, if a failure occurs while writing data to the database, it is not known how far the data has been written, and there is a risk of writing the same data twice. I had to go back and write the data again. However, in this embodiment, after confirming the execution state of the write command, the database value is read by a read command for confirmation for a process that has not been completed, and the data consistency is confirmed by comparing it with the expected read value. can do.

  Further, in the present embodiment, by using the above-described consistency confirmation information or the like, it is possible to appropriately acquire a portion where data is inconsistent (interruption point) in units of predetermined processing (for example, transaction processing). . In the present embodiment, the write process can be quickly restarted from the acquired write interruption point by using the write command corresponding to the consistency confirmation information. Therefore, in this embodiment, the interruption time due to rewriting of data due to a failure can be significantly reduced.

  In the above-described embodiment, an example of data access management using a database as an example of a storage device has been described. However, the present invention is not limited to this, and may be, for example, a computer that reads and writes data. . Therefore, this embodiment can be applied to, for example, data access management for a PC or the like, and can also be applied to data access management for a plurality of servers or the like in a cloud environment. In this case, the above-described PC or server is an example of a storage device.

  Each embodiment has been described in detail above. However, the present invention is not limited to the specific embodiment, and various modifications and changes other than the above-described modification are possible within the scope described in the claims. .

In addition, the following additional remarks are disclosed regarding the above Example.
(Appendix 1)
A command analysis means for analyzing a write command to the storage device and acquiring consistency check information including a read command and a read expected value for reading data written to the storage device by the write command;
Storage means for storing the consistency confirmation information obtained by the instruction analysis means in association with the write instruction;
Instruction execution means for confirming the consistency of data written in the storage device based on the execution state of the write command and the consistency information stored in the storage means when the storage device is recovered from a failure; An information processing apparatus comprising:
(Appendix 2)
The instruction execution means is
The information processing apparatus according to appendix 1, wherein a write command stored in the storage unit is executed when a value obtained by executing the read command does not match the expected read value.
(Appendix 3)
The storage means includes instruction storage means for storing one or a plurality of write instructions for performing the predetermined processing as the same identification information, and state storage means for storing the execution state of the write instruction for each of the identification information. And
The instruction execution unit reads an execution state of a write instruction for each identification information from the state storage unit when the storage device is recovered from a failure, and based on the read execution state, the identification stored in the instruction storage unit The information processing apparatus according to appendix 1 or 2, wherein a read command corresponding to information is executed, and the execution result is compared with the expected read value.
(Appendix 4)
The instruction execution means is
The supplementary note 3 is characterized in that when reading the execution state of the write command for each identification information, the information is read in order from the information with the oldest date or time, or the data dated back from the latest information for a predetermined period is read. Information processing device.
(Appendix 5)
The instruction execution means is
6. The information processing apparatus according to any one of appendices 1 to 5, wherein error processing is performed when reading for confirming the state of the storage device is not possible.
(Appendix 6)
An instruction analysis step of analyzing a write instruction to the storage device and acquiring consistency check information including a read instruction and a read expected value for reading data written to the storage device by the write instruction;
A storage step of storing in the storage means the consistency check information obtained by the instruction analysis step in association with the write instruction;
An instruction execution step of confirming the consistency of data written in the storage device based on the execution state of the write command and the consistency information stored in the storage means when the storage device is recovered from a failure; The management method characterized by having.
(Appendix 7)
Analyzing a write command to the storage device, obtaining consistency check information including a read command and a read expected value for reading data written to the storage device by the write command,
The acquired consistency check information is stored in a storage means in association with the write command,
A process for checking the consistency of data written in the storage device based on the execution state of the write command and the consistency information stored in the storage means when the storage device is recovered from a failure. Management program to make it run.

10 Data Access Management System 11 Access Management Device (Information Processing Device)
12 Database (storage device)
DESCRIPTION OF SYMBOLS 13 Terminal 14 Communication network 21 Input means 22 Output means 23 ID management means 24 Instruction analysis means 25 Instruction conversion means 26 Instruction execution means 27 Storage means 28 Transmission / reception means 29 Control means 31 Write instruction execution means 32 Read instruction execution means 41 Instruction storage means 42 ID state storage means (state storage means)
51 Input Device 52 Output Device 53 Drive Device 54 Auxiliary Storage Device 55 Main Storage Device 56 CPU
57 Network connection device 58 Recording medium

Claims (6)

A command analysis means for analyzing a write command to the storage device and acquiring consistency check information including a read command and a read expected value for reading data written to the storage device by the write command;
Storage means for storing the consistency confirmation information obtained by the instruction analysis means in association with the write instruction;
Instruction execution means for confirming the consistency of data written in the storage device based on the execution state of the write command and the consistency information stored in the storage means when the storage device is recovered from a failure; An information processing apparatus comprising: The instruction execution means is
2. The information processing apparatus according to claim 1, wherein when a value obtained by executing the read command does not match the expected read value, the write command stored in the storage unit is executed. The storage means includes instruction storage means for storing one or a plurality of write instructions for performing the predetermined processing as the same identification information, and state storage means for storing the execution state of the write instruction for each of the identification information. And
The instruction execution unit reads an execution state of a write instruction for each identification information from the state storage unit when the storage device is recovered from a failure, and based on the read execution state, the identification stored in the instruction storage unit The information processing apparatus according to claim 1, wherein a read instruction corresponding to information is executed, and the execution result is compared with the expected read value. The instruction execution means is
4. The information processing apparatus according to claim 1, wherein error processing is performed when reading for confirming a state of the storage device is not possible. 5. An instruction analysis step of analyzing a write instruction to the storage device and acquiring consistency check information including a read instruction and a read expected value for reading data written to the storage device by the write instruction;
A storage step of storing in the storage means the consistency check information obtained by the instruction analysis step in association with the write instruction;
An instruction execution step of confirming the consistency of data written in the storage device based on the execution state of the write command and the consistency information stored in the storage means when the storage device is recovered from a failure; The management method characterized by having. Analyzing a write command to the storage device, obtaining consistency check information including a read command and a read expected value for reading data written to the storage device by the write command,
The acquired consistency check information is stored in a storage means in association with the write command,
A process for checking the consistency of data written in the storage device based on the execution state of the write command and the consistency information stored in the storage means when the storage device is recovered from a failure. Management program to make it run.

Images