JP2011022959A - Process execution device and computer program, and process execution method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To normally continue processing of a process by overcoming faults, even if there is a time difference from the occurrence of cause of the fault, until the actualization thereof. <P>SOLUTION: An execution unit 110 executes the process. A state memory unit 123 stores a state of the execution unit 110 about a plurality of intermediate points, respectively. When a fault occurs in the execution unit 110, a state-rewinding unit 133 returns a state of the execution unit 110 to a state in the intermediate point. When a fault further occurs in the execution unit 110, the state rewinding unit 133 returns a state of the execution unit 110 to a state, in further previous intermediate points. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a process execution apparatus that copes with a failure that occurs during process execution.

  There is a technique to prepare for a case where a failure occurs in the preceding system, with the subsequent system executing the process by duplicating the process execution and chasing the preceding system processing with a slight delay.

JP 2003-15900 A

If the cause of the failure has occurred before the occurrence of the failure, there is a possibility that the cause of the failure has already occurred in the subsequent system where the failure has not yet been realized. In that case, since the same failure occurs in the subsequent system, the failure cannot be overcome.
The present invention has been made to solve the above-described problems, for example, and even when there is a time difference between the occurrence of the failure and the manifestation of the failure, the failure is overcome and processed normally. The purpose is to be able to continue.

The process execution device according to the present invention includes a storage device that stores data, a processing device that processes data, an execution unit, a state storage unit, and a state rewind unit,
The execution unit executes a process using the processing device,
The state storage unit stores the state of the execution unit at the intermediate time point for each of a plurality of intermediate time points while the execution unit is executing the process using the storage device,
The state rewinding unit sets the state of the execution unit to a state at any intermediate point in the state stored in the state storage unit when a failure occurs in the execution unit using the processing device. In the event that a failure further occurs in the execution unit, the state of the execution unit is changed to a state at an intermediate time point before the intermediate point in time when the state of the execution unit is returned among the states stored in the state storage unit. It is characterized by returning.

  According to the process execution device of the present invention, when a failure occurs again even if processing is performed again by returning the state of the execution unit to the state at the intermediate point, the state of the execution unit is changed to the state at the intermediate point before that. If the cause of the failure occurs at an early stage, the process can be restarted from an intermediate point before that, overcoming the failure and processing normally. You can continue.

1 is an overall configuration diagram illustrating an example of an overall configuration of a process execution device 100 according to Embodiment 1. FIG. FIG. 3 is a perspective view illustrating an example of an appearance of a process execution device 100 according to the first embodiment. FIG. 3 is a diagram illustrating an example of hardware resources of the process execution device 100 according to the first embodiment. FIG. 3 is a block configuration diagram illustrating an example of a functional block configuration of the process execution device 100 according to the first embodiment. FIG. 5 is a flowchart showing an example of a flow of failure processing S510 in the first embodiment. FIG. 3 is a schematic diagram illustrating an example of the operation of the process execution device 100 according to the first embodiment. FIG. 3 is a schematic diagram illustrating an example of the operation of the process execution device 100 according to the first embodiment. FIG. 3 is a schematic diagram illustrating an example of the operation of the process execution device 100 according to the first embodiment. FIG. 3 is a schematic diagram illustrating an example of the operation of the process execution device 100 according to the first embodiment. FIG. 3 is a schematic diagram illustrating an example of the operation of the process execution device 100 according to the first embodiment. FIG. 4 is a block configuration diagram showing an example of a functional block configuration of a process execution device 100 according to a second embodiment. FIG. 9 is a flowchart showing an example of a flow of failure processing S510 in the second embodiment. FIG. 10 is a schematic diagram illustrating an example of an operation of the process execution device 100 according to the second embodiment. FIG. 10 is a schematic diagram illustrating an example of an operation of the process execution device 100 according to the second embodiment. FIG. 10 is a schematic diagram illustrating an example of an operation of the process execution device 100 according to the second embodiment. FIG. 10 is a schematic diagram illustrating an example of an operation of the process execution device 100 according to the second embodiment. FIG. 9 is a schematic diagram illustrating an example of an operation of a process execution device 100 according to a third embodiment. FIG. 10 is a schematic diagram illustrating an example of an operation of a process execution device 100 according to a fourth embodiment. FIG. 6 is an overall configuration diagram illustrating an example of an overall configuration of a process execution device 100 according to a fifth embodiment. FIG. 20 is a flowchart showing an example of a flow of execution processing S550 in the fifth embodiment. FIG. 20 is a flowchart showing an example of a flow of system switching processing S590 in the fifth embodiment. FIG. 20 is a flowchart showing an example of a flow of execution processing S550 in the sixth embodiment. FIG. 20 is a flowchart showing an example of a flow of system switching processing S590 in the sixth embodiment. FIG. 20 is a flowchart showing an example of a flow of failure processing S510 in the sixth embodiment. FIG. 10 is a schematic diagram illustrating an example of an operation of a process execution device 100 according to a sixth embodiment. FIG. 10 is a schematic diagram illustrating an example of an operation of a process execution device 100 according to a sixth embodiment. FIG. 10 is a schematic diagram illustrating an example of an operation of a process execution device 100 according to a sixth embodiment.

Embodiment 1 FIG.
The first embodiment will be described with reference to FIGS.

FIG. 1 is an overall configuration diagram showing an example of the overall configuration of the process execution apparatus 100 in this embodiment.
The process execution device 100 is a device that executes a predetermined process. The process execution device 100 includes an execution unit 110 and a failure processing unit 120.
The execution unit 110 is a main part of the process execution apparatus 100 and is a part that actually executes a process. The execution unit 110 inputs parameters necessary for executing the process from the outside, executes the process based on the input parameters, and outputs the execution result to the outside.
The failure processing unit 120 monitors the execution of the process in the execution unit 110, and performs processing for dealing with the failure when a failure occurs.

FIG. 2 is a perspective view showing an example of the appearance of the process execution apparatus 100 in this embodiment.
The process execution device 100 includes a system unit 910, a display device 901 having a CRT (Cathode / Ray / Tube) or LCD (Liquid Crystal) display screen, a keyboard 902 (Key / Board: K / B), a mouse 903, an FDD 904 (Flexible). (Disk / Drive), compact disk device 905 (CDD), printer device 906, scanner device 907, and other hardware resources, which are connected by a cable or a signal line.
The system unit 910 is a computer, and is connected to the facsimile machine 932 and the telephone 931 with a cable, and is connected to the Internet 940 via a local area network 942 (LAN) and a gateway 941.

FIG. 3 is a diagram illustrating an example of hardware resources of the process execution device 100 according to this embodiment.
The process execution apparatus 100 includes a CPU 911 (also referred to as a central processing unit, a central processing unit, a processing unit, an arithmetic unit, a microprocessor, a microcomputer, and a processor) that executes a program. The CPU 911 is connected to the ROM 913, the RAM 914, the communication device 915, the display device 901, the keyboard 902, the mouse 903, the FDD 904, the CDD 905, the printer device 906, the scanner device 907, and the magnetic disk device 920 via the bus 912, and the hardware. Control the device. Instead of the magnetic disk device 920, a storage device such as an optical disk device or a memory card read / write device may be used.
The RAM 914 is an example of a volatile memory. The storage media of the ROM 913, the FDD 904, the CDD 905, and the magnetic disk device 920 are an example of a nonvolatile memory. These are examples of a storage device or a storage unit. A communication device 915, a keyboard 902, a scanner device 907, an FDD 904, and the like are examples of an input unit and an input device.
Further, the communication device 915, the display device 901, the printer device 906, and the like are examples of an output unit and an output device.

The communication device 915 is connected to a facsimile machine 932, a telephone 931, a LAN 942, and the like. The communication device 915 is not limited to the LAN 942, and may be connected to the Internet 940, a WAN (wide area network) such as ISDN, or the like. When connected to a WAN such as the Internet 940 or ISDN, the gateway 941 is unnecessary.
The magnetic disk device 920 stores an operating system 921 (OS), a window system 922, a program group 923, and a file group 924. The programs in the program group 923 are executed by the CPU 911, the operating system 921, and the window system 922.

The program group 923 stores programs that execute functions described as “˜units” in the description of the embodiments described below. The program is read and executed by the CPU 911.
The file group 924 includes information, data, signal values, variable values, and parameters that are described as “determination results of”, “calculation results of”, and “processing results of” in the description of the embodiments described below. Are stored as items of “˜file” and “˜database”. The “˜file” and “˜database” are stored in a recording medium such as a disk or a memory. Information, data, signal values, variable values, and parameters stored in a storage medium such as a disk or memory are read out to the main memory or cache memory by the CPU 911 via a read / write circuit, and extracted, searched, referenced, compared, and calculated. Used for CPU operations such as calculation, processing, output, printing, and display. Information, data, signal values, variable values, and parameters are temporarily stored in the main memory, cache memory, and buffer memory during the CPU operations of extraction, search, reference, comparison, operation, calculation, processing, output, printing, and display. Is remembered.
In addition, the arrows in the flowcharts described in the following description of the embodiments mainly indicate input / output of data and signals. The data and signal values are the RAM 914 memory, the FDD 904 flexible disk, the CDD 905 compact disk, and the magnetic field. The data is recorded on a recording medium such as a magnetic disk of the disk device 920, another optical disk, a mini disk, and a DVD (Digital Versatile Disk). Data and signals are transmitted online via a bus 912, signal lines, cables, or other transmission media.

  In the description of the embodiments described below, what is described as “to part” may be “to circuit”, “to device”, and “to device”, and “to step” and “to”. “Procedure” and “˜Process” may be used. That is, what is described as “˜unit” may be realized by firmware stored in the ROM 913. Alternatively, it may be implemented only by software, only hardware such as elements, devices, substrates, wirings, etc., or a combination of software and hardware, and further a combination of firmware. Firmware and software are stored as programs in a recording medium such as a magnetic disk, a flexible disk, an optical disk, a compact disk, a mini disk, and a DVD. The program is read by the CPU 911 and executed by the CPU 911. That is, the program causes the computer to function as “to part” described below. Alternatively, the procedure or method of “to part” described below is executed by a computer.

FIG. 4 is a block configuration diagram showing an example of a functional block configuration of the process execution device 100 according to this embodiment.
The execution unit 110 includes an initialization execution unit 111, an internal state storage unit 112, a parameter input unit 113, a process execution unit 114, and a result output unit 115. The failure processing unit 120 includes a checkpoint determination unit 121, a state acquisition unit 122, a state storage unit 123, a failure determination unit 131, a rewind time determination unit 132, and a state rewind unit 133.

The initialization execution unit 111 uses the CPU 911 to execute initialization processing. The initialization execution unit 111 executes initialization processing, for example, when the process execution device 100 is started up or restarted. In the initialization process, the initialization execution unit 111 initializes the data stored in the internal state storage unit 112, for example.
The internal state storage unit 112 uses the RAM 914 to store data representing the internal state of the execution unit 110. Examples of the internal state of the execution unit 110 include an internal variable value and a stack state.
The parameter input unit 113 uses the CPU 911 to input process parameters.
The process execution unit 114 uses the CPU 911 to execute process processing. The process execution unit 114 executes the process based on the parameters input by the parameter input unit 113, the internal state represented by the data stored in the internal state storage unit 112, and the like. The internal state storage unit 112 stores the progress of the process executed by the process execution unit 114 and data to be held for the next process execution.
The result output unit 115 uses the CPU 911 to output the result of the process executed by the process execution unit 114 to the outside.

  The checkpoint determination unit 121 uses the CPU 911 to determine whether the process of the execution unit 110 has reached the checkpoint. A checkpoint is a predetermined time (intermediate time) during the execution of a process. For example, when a process includes a plurality of steps, a checkpoint is provided every time one step is completed. Thereby, for example, even when the execution of the process is stopped halfway due to some trouble, it is possible to know to which process the process has progressed.

The state acquisition unit 122 acquires all the data stored in the internal state storage unit 112 using the CPU 911 when the checkpoint determination unit 121 determines that the process of the process in the execution unit 110 has reached the checkpoint.
The state storage unit 123 stores data acquired by the state acquisition unit 122 using the magnetic disk device 920. Thereby, the internal state of the execution unit 110 at the checkpoint is recorded. If the execution of the process is stopped halfway due to some trouble, the state stored in the state storage unit 123 can be written back to the internal state storage unit 112 to return the state of the execution unit 110 to the state at the time of the checkpoint. Can be resumed from that point. Further, the state storage unit 123 stores the data acquired by the state acquisition unit 122 for each check point. As a result, if the process is resumed by returning the state of the execution unit 110 to the state at the time of a certain checkpoint, but the operation is not successful, the state of the execution unit 110 is returned to the state at the time of another checkpoint. be able to.

  The failure determination unit 131 uses the CPU 911 to monitor the execution unit 110 and determines whether the processing of the execution unit 110 has stopped midway due to some failure. For example, the failure determination unit 131 monitors the data stored in the internal state storage unit 112 to determine whether the value is a normal value. If the value is an abnormal value, a failure has occurred in the execution unit 110. It is determined that Alternatively, the failure determination unit 131 monitors the interval at which the result output unit 115 outputs the result, and if the next result is not output even if a predetermined time has elapsed since the previous result output, the failure determination unit 131 sends the result to the execution unit 110. Determine that a failure has occurred.

  The rewind time determination unit 132 uses the CPU 911 to determine to which time the state of the execution unit 110 is to be returned from among the states at the plurality of check points stored in the state storage unit 123.

  When the failure determination unit 131 determines that a failure has occurred in the execution unit 110, the state rewinding unit 133 uses the CPU 911 to indicate the state of the execution unit 110 at the time point determined by the rewind time determination unit 132. Data to be represented is read from the state storage unit 123 and written to the internal state storage unit 112.

FIG. 5 is a flowchart showing an example of the flow of the failure processing S510 in this embodiment.
In failure processing S510, the failure processing unit 120 records the state of the execution unit 110 at the checkpoint, and when a failure occurs, returns the state of the execution unit 110 to the state at the checkpoint and attempts recovery. The failure process S510 includes a failure determination step S511, a checkpoint determination step S512, a state deletion step S521, a state acquisition step S522, a state storage step S523, a state determination step S524, a rewind time advance step S525, a rewind time point determination step S531, It includes a process interruption step S532, a state rewinding step S533, and a rewinding point retracting step S534.

In the failure determination step S511, the failure determination unit 131 uses the CPU 911 to determine whether a failure has occurred in the execution unit 110.
If it is determined that a failure has occurred, the process proceeds to the rewind time determination step S531.
If it is determined that no failure has occurred, the process proceeds to checkpoint determination step S512.

In the checkpoint determination step S512, the checkpoint determination unit 121 uses the CPU 911 to determine whether the process of the execution unit 110 has reached the checkpoint.
If it is determined that the check point has not yet been reached, the process returns to the failure determination step S511.
If it is determined that the last check point has been reached, the process proceeds to the state erasing step S521. It is assumed that the last checkpoint is provided when the process execution is completed.
When it is determined that the other check points have been reached, the process proceeds to state acquisition step S522.

In the state erasing step S521, the state storage unit 123 uses the CPU 911 to delete the stored data. Since the last checkpoint is provided when the process execution is completed, the fact that the last checkpoint has been reached means that the execution of the process has been completed normally. Since it is no longer necessary to return the state of the execution unit 110 to the state at the previous checkpoint, the state storage unit 123 erases the stored data, frees up the storage area, and prepares for the next process execution.
Thereafter, the failure process S510 is terminated.

In the state acquisition step S522, the state acquisition unit 122 uses the CPU 911 to acquire all the data stored in the internal state storage unit 112.
In the state storage step S523, the state storage unit 123 uses the magnetic disk device 920 to determine that the data acquired by the state acquisition unit 122 in the state acquisition step S522 has been reached by the checkpoint determination unit 121 in the checkpoint determination step S512. The data is stored as data corresponding to the determined checkpoint.

In the state determination step S524, the rewinding point determination unit 132 uses the CPU 911 to store data corresponding to the checkpoint determined by the checkpoint determination unit 121 in the checkpoint determination step S512. It is determined whether or not it has been stored. As a result, in the process of the process in the execution unit 110 this time, it is the first time that the check point has been reached, or after passing the check point, a failure occurs and the state of the execution unit 110 is returned. As a result, you can see if the checkpoint has been reached again.
If it is determined that the check point has been reached for the first time, the process proceeds to the rewind time advance step S525.
If it is determined that the checkpoint has been reached before, the process returns to the failure determination step S511.

In the rewind time advance step S525, the rewind time determination unit 132 uses the CPU 911 to determine the checkpoint determined by the checkpoint determination unit 121 in the checkpoint determination step S512 as the rewind time. The rewind time determination unit 132 stores the determination result using the magnetic disk device 920.
Thereafter, the process returns to the failure determination step S511.

In the rewind time determination step S531, the rewind time determination unit 132 uses the CPU 911 to determine the checkpoint determined as the rewind time based on the stored determination result.
If there is no checkpoint determined to be the rewind time, the process proceeds to the process interruption step S532. This is because, for example, when a failure occurs in the execution unit 110 before processing of the process in the execution unit 110 reaches the first checkpoint, or when the failure occurs, the execution unit 110 enters the state at the first checkpoint. This is the case when the processing is resumed by returning the status, but a failure occurs again. That is, there is no checkpoint to which the state of the execution unit 110 should be returned.
If there is a checkpoint determined as the rewinding time, the process proceeds to the state rewinding step S533.

In the processing interruption step S532, the state rewinding unit 133 rewrites the data stored in the internal state storage unit 112 using the CPU 911, and interrupts the current process processing in the execution unit 110. This is because there is no checkpoint for returning the state of the execution unit 110, and thus the processing of the execution unit 110 cannot be continued any more. Thereafter, the process proceeds to the state erasing step S521.
The execution unit 110 discards the parameter input by the parameter input unit 113 and ends the process. The process execution unit 114 outputs a result of abnormal termination. Thereafter, when the parameter input unit 113 inputs the next parameter, the execution unit 110 starts processing the next process.

In the state rewinding step S533, the state rewinding unit 133 uses the CPU 911 to acquire the data stored in the state storage unit 123 for the check points determined by the rewinding time determining unit 132 in the rewinding time determining step S531. . The state rewinding unit 133 uses the CPU 911 to write the acquired data to the internal state storage unit 112. Thereafter, the process proceeds to the rewinding time step S534.
The execution unit 110 returns to the state at the time of the checkpoint when the data stored in the internal state storage unit 112 is rewritten, and resumes the processing of the process from that state.

In the rewinding time retraction process S534, the rewinding time determination unit 132 uses the CPU 911, and among the data stored in the state storage unit 123, the state rewinding unit 133 is stored in the internal state storage unit 112 in the state rewinding step S533. It is determined that the check point immediately before the check point corresponding to the written data is the rewinding time. When the checkpoint corresponding to the data written to the internal state storage unit 112 by the state rewinding unit 133 in the state rewinding step S533 is the first checkpoint, there is no previous checkpoint, so the rewinding time point The determination unit 132 determines that there is no check point that is the rewinding point. The rewind time determination unit 132 stores the determination result using the magnetic disk device 920.
Thereafter, the process returns to the failure determination step S511.

  FIG. 6 is a schematic diagram showing an example of the operation of the process execution apparatus 100 in this embodiment.

  At the start of the process 601, the parameter input unit 113 inputs parameters, and the process execution unit 114 starts processing the process. The storage content 611 of the state storage unit 123 at this time is empty. That is, the state storage unit 123 does not store any data representing the state at the checkpoint.

When the process of the process execution unit 114 reaches the first check point 602, the check point determination unit 121 determines that, and the state acquisition unit 122 acquires the data stored in the internal state storage unit 112, and the first check The state storage unit 123 stores the data corresponding to the point 602. The storage content 612 of the state storage unit 123 at this time is only the state 652 at the first check point 602.
In this process, since the process of the process execution unit 114 has reached the first check point 602 for the first time, the rewind time determination unit 132 determines the first check point 602 as the rewind time.

When the process of the process execution unit 114 reaches the second check point 603, the check point determination unit 121 determines that, the state acquisition unit 122 acquires the data stored in the internal state storage unit 112, and the second check point The state storage unit 123 stores the data corresponding to the point 603. The stored contents 613 of the state storage unit 123 at this time are two states, a state 652 at the first check point 602 and a state 653 at the second check point 603.
In this process, since the process of the process execution unit 114 has reached the second check point 603 for the first time, the rewind time determination unit 132 determines the second check point 603 as the rewind time.

When the process of the process execution unit 114 reaches the third checkpoint 604, the checkpoint determination unit 121 determines that, and the state acquisition unit 122 acquires the data stored in the internal state storage unit 112, and the third checkpoint The state storage unit 123 stores the data corresponding to the point 604. The stored contents 614 of the state storage unit 123 at this time are three states: a state 652 at the first check point 602, a state 653 at the second check point 603, and a state 654 at the third check point 604.
In this process, since the process of the process execution unit 114 has reached the third check point 604 for the first time, the rewind time determination unit 132 determines the third check point 604 as the rewind time.

  As described above, as the processing of the processing execution unit 114 proceeds smoothly, the state storage unit 123 accumulates data representing the state of the execution unit 110 at each checkpoint. Also, the check point that the rewind time determination unit 132 determines as the rewind time progresses as the process proceeds.

  Here, it is assumed that a failure has occurred in the execution unit 110 before the processing of the process execution unit 114 reaches the fourth check point 605. The failure determination unit 131 determines that, and the state rewinding unit 133 returns the state of the execution unit 110 to the state 654 at the third checkpoint 604 based on the determination result of the rewinding time determination unit 132. The process execution unit 114 resumes the process from the third check point 604.

  FIG. 7 is a schematic diagram showing an example of the operation of the process execution device 100 in this embodiment.

  When the process execution unit 114 resumes the process from the third check point 604, the storage content 615 of the state storage unit 123 is the same as the storage content 614 before the process is resumed. The rewind time determination unit 132 determines the previous check point, that is, the second check point 603, as the rewind time.

  It is assumed that a failure has occurred again before the processing of the processing execution unit 114 reaches the fourth check point 605. The failure determination unit 131 determines that, and the state rewinding unit 133 returns the state of the execution unit 110 to the state 653 at the second checkpoint 603 based on the determination result of the rewinding time determination unit 132. The process execution unit 114 resumes the process from the second check point 603.

Although the failure has occurred again despite returning to the third checkpoint 604 and the processing is started again, the failure has become apparent after the third checkpoint 604 has passed. It may have occurred before the third checkpoint 604 was reached.
When the failure occurs again, the process execution device 100 returns to the previous checkpoint instead of the previous checkpoint, and performs the process again. Thereby, even if the cause of the failure has occurred before the arrival of the immediately preceding checkpoint, the failure can be overcome and processing can be continued normally.

  FIG. 8 is a schematic diagram showing an example of the operation of the process execution apparatus 100 in this embodiment.

  When the process execution unit 114 resumes the process from the second check point 603, the storage content 616 of the state storage unit 123 is the same as the storage content 615 before the process is resumed. The rewind time determination unit 132 determines that the previous check point, that is, the first check point 602 is the rewind time.

When the process of the process execution unit 114 reaches the third checkpoint 604 again, the checkpoint determination unit 121 determines that, and the state acquisition unit 122 acquires the data stored in the internal state storage unit 112, and the third The state storage unit 123 stores the data corresponding to the check point 604. The stored contents 617 of the state storage unit 123 at this time are three states: a state 652 at the first check point 602, a state 653 at the second check point 603, and a state 654 at the third check point 604. Among these, the state 654 at the third check point 604 is a state at the second arrival time.
In this process, since the process of the process execution unit 114 has reached the third check point 604 for the second time, the rewind time determination unit 132 cannot advance the rewind time. Therefore, the rewinding point remains the first check point 602.

  It is assumed that a failure has occurred three times before the process of the process execution unit 114 reaches the fourth check point 605. The failure determination unit 131 determines that, and the state rewinding unit 133 returns the state of the execution unit 110 to the state 652 at the first checkpoint 602 based on the determination result of the rewinding time determination unit 132. The process execution unit 114 resumes the process from the first check point 602.

  As described above, when a failure repeatedly occurs at the same time and the process does not proceed, the process execution device 100 returns the checkpoint for restarting the process one by one. Thus, if the processing can be resumed from the check point before the cause of the failure, the failure can be overcome and the processing can be continued normally.

  FIG. 9 is a schematic diagram showing an example of the operation of the process execution device 100 in this embodiment.

It is assumed that the process execution unit 114 resumes the process from the second check point 603, passes the third check point 604, and then reaches the fourth check point 605 without any failure at the same place. The check point determination unit 121 determines that, the state acquisition unit 122 acquires the data stored in the internal state storage unit 112, and the state storage unit 123 stores the data as data corresponding to the third check point 604. The stored contents 618 of the state storage unit 123 at this time point include the state 652 at the first check point 602, the state 653 at the second check point 603, the state 654 at the third check point 604, and the state at the fourth check point 605. 655 and four.
In this process, since the process of the process execution unit 114 has reached the fourth check point 605 for the first time, the rewind time determination unit 132 determines the fourth check point 605 as the rewind time.

  It is assumed that a failure has occurred again before the processing of the processing execution unit 114 reaches the final check point 606. The failure determination unit 131 determines that, and the state rewinding unit 133 returns the state of the execution unit 110 to the state 655 at the fourth checkpoint 605 based on the determination result of the rewinding time determination unit 132. The process execution unit 114 resumes the process from the fourth check point 605.

  In this way, even if a failure occurs repeatedly, if the processing has progressed beyond the point at which the previous failure occurred, the cause of the previous failure is removed and a new failure It seems that the cause of this occurred. In this case, the process execution device 100 does not return to the previous checkpoint, but returns to the previous checkpoint and resumes processing.

  FIG. 10 is a schematic diagram showing an example of the operation of the process execution device 100 in this embodiment.

  Assume that after passing through the fourth checkpoint 605, the final checkpoint 606 has been reached without any failure. The check point determination unit 121 determines that, and the state storage unit 123 erases the stored data. Thereby, the storage content 619 of the state storage unit 123 becomes empty.

The process execution device 100 in this embodiment includes a storage device (magnetic disk device 920) for storing data, a processing device (CPU 911) for processing data, an execution unit 110, a state storage unit 123, and a state rewind. Part 133.
The execution unit 110 executes a process using the processing device.
The state storage unit 123 uses the storage device to indicate the state of the execution unit 110 at the intermediate time point for each of a plurality of intermediate time points (checkpoints) while the execution unit 110 is executing the process. Remember.
When the failure occurs in the execution unit 110 using the processing device, the state rewinding unit 133 moves the execution unit into a state at any intermediate point among the states stored in the state storage unit 123. 110, the state at the intermediate time point before the intermediate time point at which the state of the execution unit 110 is returned among the states stored in the state storage unit 123 when the execution unit 110 further fails. The state of the execution unit 110 is returned.

  According to the process execution device 100 in this embodiment, when a failure occurs again even if the state of the execution unit 110 is returned to the state at the intermediate time and the process is performed again, the state of the execution unit 110 is changed to the previous state. Since it returns to the state at the intermediate point and redoes the process, even if the cause of the failure occurred at an early stage, the process can be redone from the intermediate point before that, overcoming the obstacle, Processing can be continued normally.

  In the process execution device 100 according to this embodiment, the state rewinding unit 133 uses the processing device to return the state of the execution unit 110 when the execution unit 110 further fails. When the state storage unit 123 does not store the state at the intermediate time point before the intermediate time point, the execution unit 110 interrupts the execution of the process.

  According to the process execution device 100 in this embodiment, since the processing of the execution unit 110 is interrupted when the state storage unit 123 does not store the state at the previous intermediate time point, the parameter input by the execution unit 110 If a failure occurs no matter how many times the process is re-executed, it is possible to abandon the continuation of the process and move on to the next process.

  The process execution device 100 in this embodiment is a computer program that causes a computer having a storage device (magnetic disk device 920) for storing data and a processing device (CPU 911) for processing data to function as the process execution device 100. Can be realized.

  According to the computer program of this embodiment, when a failure occurs again even if the state of the execution unit 110 is returned to the state at the intermediate point and the process is performed again, the state of the execution unit 110 is changed to an intermediate point before that. The process is restarted and the process is restarted, so even if the cause of the failure occurs at an early stage, the process is restarted from the intermediate point before that, overcoming the failure and continuing normally. The process execution device 100 can be obtained.

A process execution method in which the process execution apparatus 100 according to this embodiment executes a process includes the following steps.
The said processing apparatus performs the said process (execution process).
The storage device stores a state at the intermediate time point for each of a plurality of intermediate time points (check points) in the middle of execution of the process by the processing device (state storage step).
When a failure occurs in the execution of the process, the processing device returns the state to a state at any intermediate point among the states stored in the storage device (first state rewinding step).
When the processing device further fails in the execution of the process, the processing device returns the state to the state at the intermediate time point before the intermediate time point when the state is returned among the states stored in the storage device (second State unwinding step).

  According to the process execution method in this embodiment, when a failure occurs again even if the state is returned to the state at the intermediate point and the process is performed again, the state is returned to the state at the intermediate point before that and the process is performed. Since the process is redone, even if the cause of the failure has occurred at an early stage, the process can be restarted from an intermediate point before that to overcome the failure and continue processing normally.

  The system restart apparatus (process execution apparatus 100) described above includes a checkpoint detection apparatus (checkpoint determination unit 121) that detects a plurality of checkpoints in the middle of process processing, and detects that a failure has occurred. A failure detection device (failure determination unit 131) is provided, a process restart device (state rewinding unit 133) that restarts the system (execution unit 110) when a failure occurs, and the system is restarted when the failure occurs. Is returned to the state of the most recent checkpoint and a restart is attempted. When the restart fails, the checkpoint is further returned to the previous state, and after that, a restart attempt device (failure processing unit) that tries to restart until the restart is successful. 120).

  The system restart device includes a process processing unit (execution unit 110) that executes process processing, a checkpoint processing unit (state acquisition unit 122) that collects a snapshot of memory contents at a checkpoint, and a snap that accumulates the collected snapshot. Shot accumulation unit (state storage unit 123), failure detection unit (failure determination unit 131) for detecting occurrence of a failure during process processing, restart instruction unit (rewind time determination) for instructing restart of the process processing unit when a failure occurs Unit 132) and a state return instruction unit (state rewind unit 133) for writing back the snapshot data of the snapshot storage unit to the process processing unit.

In the process processing of the process processing unit, check points are embedded at arbitrary processing execution intervals. When the process processing unit executes the process processing and reaches the checkpoint, the checkpoint processing unit is called. The checkpoint processing unit dumps the contents of the internal memory (internal state storage unit 112) of the process processing execution unit (processing execution unit 114), and stores this as a snapshot in the snapshot storage unit. Thus, the internal memory information of the process processing execution unit is stored in the snapshot storage unit in chronological order.
When a failure occurs during execution of the process processing and the process processing of the process processing unit goes down, the failure detection unit detects the occurrence of the failure. The failure detection unit that has detected the failure instructs the restart instruction unit to restart the process processing unit. Upon receiving the restart instruction, the restart instruction unit instructs the state return instruction unit to write back the latest snapshot to the process processing unit. Upon receiving the instruction, the state return instruction unit takes the latest snapshot from the snapshot storage unit and writes it in the memory inside the process processing unit. Thereafter, the restart instruction unit issues a restart instruction to the process processing unit. Upon receiving the restart instruction, the process processing unit restarts the process processing based on the internal memory contents received from the state return instruction unit.
At this time, if a failure occurs again in the process processing execution unit, the restart instruction is issued again according to the above procedure, but the restart instruction unit further adds 1 to the state return instruction unit from the latest. An instruction is issued to write back the previous snapshot to the process processing unit, and the process processing unit is instructed to restart. Thereafter, the snapshot is returned to the previous one in order until the process processing unit operates normally.

  As described above, when a failure occurs in the process processing, the memory contents are returned to the state at or near the time of the restart, and then the process processing is restarted. Thus, it is possible to return to a normal processing state without performing processing, and to shorten the restart time.

  When the process reaches the checkpoint, the system restart device (process execution device 100) described above sucks information such as the memory state and CPU state of the system into the snapshot storage unit (state storage unit 123). A snapshot management unit (state storage unit 123) is provided that continues to accumulate until the processing of the process is completed, and discards the information at the same time as the end of the process.

  The system restart device includes a process processing unit (execution unit 110) that executes process processing, a checkpoint processing unit (state acquisition unit 122) that collects a snapshot of memory contents at a checkpoint, and a snap that accumulates the collected snapshot. Shot accumulation unit (state storage unit 123), failure detection unit (failure determination unit 131) for detecting occurrence of a failure during process processing, restart instruction unit (rewind time determination) for instructing restart of the process processing unit when a failure occurs Section 132), a state return instruction section (state rewind section 133) for writing back the snapshot data of the snapshot storage section to the process processing section, a process processing snapshot storage instruction, or a snapshot is no longer necessary Snapshot management unit (state storage unit 1) for instructing to discard a snapshot Having three).

Checkpoints are inserted at arbitrary processing intervals in the process processing of the process processing unit. When the process process reaches this checkpoint, the snapshot management unit operates and performs the following operations.
When the checkpoint is at the beginning of the process processing, a snapshot of the process state is acquired and newly registered in the snapshot management unit using the input information as a key.
If the checkpoint is in the middle of process processing, a snapshot of the process state is acquired, and additional registration is performed in the snapshot management unit using the input information as a key.
When the checkpoint is at the end of process processing, all snapshots registered with the input information as a key are discarded.

  As described above, by registering and deleting a snapshot of the process state when passing through a checkpoint provided during the process of the process, only the snapshot directly associated with the process can be accumulated. The amount of data stored in can be reduced.

  When the system restart device (process execution device 100) described above attempts to restart by returning to the state of all checkpoints stored in the snapshot storage unit (state storage unit 123), An input data judging device (state rewinding unit 133) is provided that discards the input data itself as invalid and restarts the process processing with the next input data.

  The system restart device includes a process processing unit (execution unit 110) that executes process processing, a checkpoint processing unit (state acquisition unit 122) that collects a snapshot of memory contents at a checkpoint, and a snap that accumulates the collected snapshot. Shot accumulation unit (state storage unit 123), failure detection unit (failure determination unit 131) for detecting occurrence of a failure during process processing, restart instruction unit (rewind time determination) for instructing restart of the process processing unit when a failure occurs Unit 132), a state return instruction unit (state rewind unit 133) for writing back the snapshot data of the snapshot storage unit to the process processing unit, and input when an attempt to restart using all snapshots fails An input data discard determination unit (rewind time determination unit 132 that discards data and restarts with the next input data. Having.

  When the process in the process processing unit fails and the restart instruction unit restarts while restoring the system status with the snapshot, the restart fails repeatedly, and the last snapshot stored in the snapshot storage unit Use to reboot. At this time, if the process of the process processing unit is executed up to the next check point, it is determined that the process restart is successful, and normal operation is performed. If the restart fails by the next check point, the state return instruction unit notifies the input data determination unit of the restart failure. Upon receiving the notification, the input data discard determination unit instructs the restart instruction unit to discard the current input data and restart the process with the next input data. Upon receiving the process restart instruction, the restart instruction unit discards the input data currently being passed to the process processing unit, passes the next input data to the process processing unit, and restarts the process. The process processing unit discards the current input data and executes the process with the next input data.

  As mentioned above, even if the process restart is attempted using all snapshots and fails, the current input data is discarded and processing continues with the next data, so the process restart time can be shortened and efficient. The system can be operated.

Embodiment 2. FIG.
The second embodiment will be described with reference to FIGS.
In addition, about the part which is common in Embodiment 1, the same code | symbol is attached | subjected and description is abbreviate | omitted.

FIG. 11 is a block configuration diagram showing an example of a functional block configuration of the process execution device 100 according to this embodiment.
The process execution apparatus 100 does not include the rewind time determination unit 132 described in the first embodiment, but includes a state deletion unit 134 instead.

  When the failure determination unit 131 determines that a failure has occurred in the execution unit 110, the state rewinding unit 133 sets the data corresponding to the newest checkpoint among the data stored in the state storage unit 123. Is read from and written to the internal state storage unit 112. As a result, the state of the execution unit 110 returns to the state at the time of the checkpoint, and the process execution unit 114 resumes the process from the checkpoint.

When the state rewinding unit 133 writes the data stored in the state storage unit 123 to the internal state storage unit 112 and returns the state of the execution unit 110 to the state at the time of a certain checkpoint, the state deletion unit 134 is the CPU 911. The data stored in the internal state storage unit 112 by the state storage unit 123 is deleted from the data stored in the state storage unit 123.
When the failure occurs again in the execution unit 110, the data that has been written back once is deleted, so the state rewinding unit 133 does not return the state of the execution unit 110 to the same state again.

FIG. 12 is a flowchart showing an example of the flow of the failure process S510 in this embodiment.
The failure process S510 does not include the state determination step S524, the rewind time advance step S525, and the rewind time retreat step S534 described in the first embodiment, but has a state deletion step S535 instead.

In the rewinding time determination step S531, the state rewinding unit 133 uses the CPU 911 to determine whether or not the state storage unit 123 stores data representing the state at any checkpoint.
When the state storage unit 123 determines that no data is stored, the process proceeds to the process interruption step S532.
When it is determined that the state storage unit 123 stores data, the process proceeds to the state rewinding step S533.

In the state rewinding step S533, the state rewinding unit 133 uses the CPU 911 to read data corresponding to the last checkpoint among the data stored in the state storage unit 123 and write the data to the internal state storage unit 112. .
Then, it progresses to state deletion process S535.

In the state deletion step S535, the state deletion unit 134 uses the CPU 911 to delete the data read by the state rewinding unit 133 in the state rewinding step S533 among the data stored in the state storage unit 123.
Thereafter, the process returns to the failure determination step S511.

  FIG. 13 is a schematic diagram showing an example of the operation of the process execution device 100 in this embodiment.

At the time of starting the process 601, the storage content 621 of the state storage unit 123 is empty.
When the process of the process execution unit 114 reaches the first check point 602, the state 652 at the first check point 602 is added to the storage content 622 of the state storage unit 123.
When the process of the process execution unit 114 reaches the second check point 603, the state 653 at the second check point 603 is added to the storage content 623 of the state storage unit 123.
When the process of the process execution unit 114 reaches the third check point 604, the state 654 at the third check point 604 is added to the storage content 624 of the state storage unit 123.
Thus, as long as the process of the process execution unit 114 is proceeding smoothly, the content of the data stored in the state storage unit 123 is the same as in the case of the first embodiment.

  It is assumed that a failure has occurred before the process of the process execution unit 114 reaches the fourth check point 605. Since the last data among the data stored in the state storage unit 123 is data corresponding to the third check point 604, the state rewind unit 133 returns the state of the execution unit 110 to the state 654 at the third check point 604. . The process execution unit 114 resumes the process from the third check point 604.

  FIG. 14 is a schematic diagram showing an example of the operation of the process execution device 100 in this embodiment.

  When the process execution unit 114 resumes processing from the third check point 604, the state deletion unit 134 deletes data corresponding to the third check point 604 from the data stored in the state storage unit 123. The storage contents 625 of the state storage unit 123 are in two states: a state 652 at the first check point 602 and a state 653 at the second check point 603.

  It is assumed that a failure has occurred again before the processing of the processing execution unit 114 reaches the fourth check point 605. Since the last data among the data stored in the state storage unit 123 is data corresponding to the second check point 603, the state rewinding unit 133 returns the state of the execution unit 110 to the state 653 at the second check point 603. . The process execution unit 114 resumes the process from the second check point 603.

  In this way, since the state deletion unit 134 deletes the data at the time of restarting the process from the state storage unit 123, if a failure occurs again at the same point, the process execution unit 114 returns to the previous checkpoint. To resume processing. As a result, even if the cause of the failure has occurred before the previous checkpoint, the process can be restarted from the point before the failure, Processing can be continued normally.

  FIG. 15 is a schematic diagram showing an example of the operation of the process execution device 100 in this embodiment.

  When the process execution unit 114 resumes the process from the second check point 603, the state deletion unit 134 deletes the data corresponding to the second check point 603 from the data stored in the state storage unit 123. The storage content 626 of the state storage unit 123 is only the state 652 at the first check point 602.

  When the process of the process execution unit 114 reaches the third check point 604 again, the state storage unit 123 stores the data stored in the internal state storage unit 112 as new data corresponding to the third check point 604. The stored contents 627 of the state storage unit 123 are two states: a state 652 at the first check point 602 and a new state 654 at the third check point 604.

  It is assumed that a failure has occurred three times before the process of the process execution unit 114 reaches the fourth check point 605. Since the last data among the data stored in the state storage unit 123 is data corresponding to the third check point 604, the state rewind unit 133 returns the state of the execution unit 110 to the state 654 at the third check point 604. . The process execution unit 114 resumes the process from the third check point 604.

  Even if a failure occurs at the same time, the cause of the failure may be different because the state at the previous checkpoint is different from the previous one. Therefore, it is worth trying again from the previous checkpoint. The process execution device 100 returns to the previous checkpoint and redoes the process. When the state at the checkpoint is updated after passing through the checkpoint after that, the process execution device 100 returns to the checkpoint with the updated state. And redo the process. As a result, there is a case where it is possible to overcome the obstacle and continue processing normally without returning to the previous checkpoint.

  FIG. 16 is a schematic diagram showing an example of the operation of the process execution device 100 in this embodiment.

  When the process execution unit 114 resumes the process from the third check point 604 again, the state deletion unit 134 deletes data corresponding to the third check point 604 from the data stored in the state storage unit 123. The storage content 628 of the state storage unit 123 is only the state 652 at the first check point 602.

  It is assumed that a failure has occurred four times before the processing of the processing execution unit 114 reaches the fourth check point 605. Since the last data among the data stored in the state storage unit 123 is data corresponding to the first check point 602, the state rewind unit 133 returns the state of the execution unit 110 to the state 652 at the first check point 602. . The process execution unit 114 resumes the process from the first check point 602.

  If you return to a checkpoint with an updated state and resume processing, but a failure occurs at the same point, the cause of the failure is at an earlier point in time than if a new cause of failure occurred It is more reasonable to think. The process execution device 100 returns to a checkpoint that is earlier than the checkpoint that was returned to and returns to the previous checkpoint. As a result, if the processing can be resumed from the time before the occurrence of the cause of the failure, the failure can be overcome and the processing can be continued normally.

In the process execution device 100 according to this embodiment, the state rewinding unit 133 uses the processing device (CPU 911) to store the state stored in the state storage unit 123 when a failure occurs in the execution unit 110. Of these, the state of the execution unit 110 is returned to the state of the latest intermediate point (checkpoint).
The state storage unit (state deletion unit 134) uses the processing device to delete a state at an intermediate time point when the state rewinding unit 133 returns the state of the execution unit 110 among the stored states.

  According to the process execution device 100 in this embodiment, the state storage unit (state deletion unit 134) deletes the state at the intermediate point when the state rewinding unit 133 returns the state of the execution unit 110. Time point management becomes easy. In addition, since the number of states stored in the state storage unit 123 is reduced for a process in which a failure repeatedly occurs, the storage area used by the state storage unit 123 decreases.

  The system restart device (process execution device 100) described above attempts to restart the system (execution unit 110) by returning the system (execution unit 110) to the state of the most recent check point when a failure occurs. A checkpoint information discard determination unit (state deletion unit 134) that discards the acquired information as invalid is provided.

  The system restart device includes a process processing unit (execution unit 110) that executes process processing, a checkpoint processing unit (state acquisition unit 122) that collects a snapshot of memory contents at a checkpoint, and a snap that accumulates the collected snapshot. Shot accumulation unit (state storage unit 123), failure detection unit (failure determination unit 131) for detecting occurrence of a failure during process processing, restart instruction unit (state rewinding unit) for instructing restart of the process processing unit when a failure occurs 133), a state return instructing unit (state rewinding unit 133) for writing back the snapshot data of the snapshot accumulating unit to the process processing unit, and a checkpoint for instructing to discard the snapshot when a snapshot is no longer necessary when a failure occurs An information discard determination unit (state deletion unit 134) is included.

  In response to a failure detected by the failure detection unit, the state return instruction unit retrieves a snapshot from the snapshot storage unit to return the state of the process processing unit and registers it as a snapshot to be restarted in the checkpoint discard determination unit. . In this state, the process is restarted in accordance with an instruction from the restart instruction unit, and the determination that the process has normally operated when the process reaches the next checkpoint is made from the state return instruction unit to the checkpoint information discard determination unit And the checkpoint discard determination unit instructs the snapshot storage unit to hold the snapshot as valid. If a failure occurs in the process without reaching the next checkpoint when the process is restarted, the state return instruction unit notifies the checkpoint discard determination unit that the restart has failed. The checkpoint discard determination unit determines that the snapshot has a problem and discards the snapshot in the snapshot storage unit.

  As described above, invalid snapshot information is no longer stored in the snapshot storage unit, and even if a failure occurs multiple times and restarts while passing through a new checkpoint, the snapshot is stored. From this section, it is always necessary to take out snapshots in the reverse order of time series, and it is possible to efficiently take out snapshots at the time of restart.

Embodiment 3 FIG.
The third embodiment will be described with reference to FIG.
Since the configuration of the process execution device 100 in this embodiment is the same as that of the process execution device 100 described in the first embodiment or the second embodiment, the description thereof will be omitted and only different points will be described.

When the execution unit 110 starts processing a process, the checkpoint determination unit 121 uses the CPU 911 to determine that the process processing in the execution unit 110 has reached a checkpoint. That is, the first checkpoint is provided at the start of the process.
When the execution unit 110 starts processing the process, the checkpoint determination unit 121 determines that the process processing in the execution unit 110 has reached the checkpoint, so the state acquisition unit 122 uses the data stored in the internal state storage unit 112. Acquired and stored in the state storage unit 123.

  FIG. 17 is a schematic diagram showing an example of the operation of the process execution device 100 in this embodiment.

  When the process execution unit 114 starts processing the process, the checkpoint determination unit 121 determines that the process of the execution unit 110 has reached the first checkpoint 602, and the state acquisition unit 122 uses the data stored in the internal state storage unit 112. Is acquired and stored in the state storage unit 123. Therefore, at the time of starting the process, the state storage unit 123 stores data representing the state 652 at the start of the process as the storage content 631.

  It is assumed that a failure has occurred before the process of the process execution unit 114 reaches the second check point 603. Since the state storage unit 123 stores the state 652 at the start of the process, the rewind time determination unit 132 returns the state of the process execution unit 114 to the state 652 at the start of the process. The process execution unit 114 resumes the process from the first check point 602 (process start time).

  The process execution device 100 according to the first embodiment or the second embodiment returns the state of the execution unit 110 when a failure occurs before the process of the process execution unit 114 reaches the first checkpoint, such as immediately after the start of the process. Since there is no checkpoint to be processed, the processing of the process is interrupted. In the process execution device 100 according to this embodiment, even if a failure occurs immediately after the start of the process, there is a checkpoint (at the start of the process) for returning the state of the execution unit 110. The state of 110 can be returned to resume the processing.

Embodiment 4 FIG.
The fourth embodiment will be described with reference to FIG.
Since the configuration of the process execution device 100 in this embodiment is the same as that of the process execution device 100 described in the first to third embodiments, only the differences will be described.

The state acquisition unit 122 uses the CPU 911 in advance to store data stored in the internal state storage unit 112 immediately after the initialization execution unit 111 executes the initialization process (hereinafter referred to as “initialization state data”). Keep getting. The state storage unit 123 stores, in advance, the initialization state data acquired by the state acquisition unit 122 as data corresponding to the first time point of the process using the magnetic disk device 920.
Further, the state storage unit 123 uses the CPU 911 to erase the stored data when the process execution is completed or when the process execution is interrupted. However, the state storage unit 123 leaves only the initialization state data and deletes other data using the CPU 911.

  FIG. 18 is a schematic diagram showing an example of the operation of the process execution device 100 in this embodiment.

Before the execution unit 110 starts processing the process, the state storage unit 123 stores data representing the initialization state 651 as the storage content 632.
Therefore, even if a failure occurs immediately after the start of the process, there is a checkpoint (at the time of initialization execution) to which the state of the execution unit 110 should be returned. Can be resumed.
Further, if a failure occurs at a later time and the failure cannot be overcome even if the state of the execution unit 110 is returned and the process is restarted, the cause of the failure is caused in the internal state of the execution unit 110 at the time when the process processing is started. There is a possibility. In such a case, by executing the initialization process and then restarting the process, the failure can be overcome and the process can be continued normally.

Further, as in the third embodiment, when the process is started, if the checkpoint determination unit 121 determines that the execution unit 110 has reached the checkpoint, the process execution unit 114 starts processing the process. Thus, the state storage unit 123 stores, as the storage content 633, data representing the initialization state 651 and the state 652 at the start of the process.
Therefore, a failure occurs immediately after the start of the process, and the state of the execution unit 110 is returned to the state at the start of the process and the processing is resumed. However, even if a failure occurs again, the state of the execution unit 110 is still changed. Since there is a checkpoint to be returned (at the time of initialization execution), the state of the execution unit 110 can be returned to that state, and the process can be resumed again.

  In the process execution device 100 in this embodiment, the state storage unit 123 uses the storage device (magnetic disk device 920) as the state of the execution unit 110 at the first intermediate point (checkpoint). The state where the unit 110 is initialized is stored.

  According to the process execution device 100 in this embodiment, if the failure cannot be overcome even if the state of the execution unit 110 is returned to the state at the intermediate time and the process is performed again, the state of the execution unit 110 is returned to the initialized state. Therefore, even if there is a cause of failure in the state of the execution unit 110 at the start of the process, the failure can be overcome and processing can be continued normally.

  It should be noted that the state storage unit 123 does not store the state 651 immediately after the initialization execution unit 111 executes the initialization process, but instead stores the execution unit in the data stored in the state storage unit 123. The configuration may be such that the initialization execution unit 111 executes the initialization process when there is no corresponding checkpoint to which the state of 110 is to be returned, and then the process execution unit 114 resumes the process.

  When the system restart device (process execution device 100) described above attempts to restart by returning to the state of all checkpoints stored in the snapshot storage unit (state storage unit 123), An internal data initialization device (initialization execution unit 111) is provided that determines that there is an abnormality in the data held in the process, initializes all internal data, and resumes process processing with the same input data.

  The system restart device includes a process processing unit (execution unit 110) that executes process processing, a checkpoint processing unit (state acquisition unit 122) that collects a snapshot of memory contents at a checkpoint, and a snap that accumulates the collected snapshot. Shot accumulation unit (state storage unit 123), failure detection unit (failure determination unit 131) for detecting occurrence of a failure during process processing, restart instruction unit (state rewinding unit) for instructing restart of the process processing unit when a failure occurs 133), a state return instructing unit (state rewinding unit 133) for writing back the snapshot data of the snapshot storage unit to the process processing unit, and internal information when restart is attempted using all snapshots but fails. Has an internal data initialization unit (initialization execution unit 111) that initializes and restarts.

  When the process in the process processing unit fails and the restart instruction unit restarts while restoring the system status with the snapshot, the restart fails repeatedly, and the last snapshot stored in the snapshot storage unit Use to reboot. At this time, if the process of the process processing unit is executed up to the next check point, it is determined that the process restart is successful, and normal operation is performed. If the restart fails by the next checkpoint, the state return instruction unit notifies the internal data initialization unit of the restart failure. Upon receiving the notification, the internal data initialization unit instructs the process initialization to set all internal information used for process processing to initial values. Thereafter, the restart instruction unit is instructed to restart the process. Upon receiving the process restart instruction, the restart instruction unit gives the current input data to the process processing unit and instructs the process start. Upon receiving the process start instruction, the process processing unit receives input data from the restart instruction unit and starts processing the process.

  As mentioned above, even if you try to restart the process using all snapshots and fail, it is possible to restart the process without bringing down the system by initializing the internal information of the process itself The restart time of the entire system can be shortened, and the system can be operated efficiently.

Embodiment 5 FIG.
The fifth embodiment will be described with reference to FIGS.
In addition, about the part which is common in Embodiment 1- Embodiment 4, the same code | symbol is attached | subjected and description is abbreviate | omitted.

FIG. 19 is an overall configuration diagram showing an example of the overall configuration of the process execution device 100 in this embodiment.
The process execution apparatus 100 includes two execution units 110, two failure processing units 120, and a system switching unit 180.
The system switching unit 180 uses the CPU 911 to set one of the two execution units 110 as a regular system and the other as a standby system. The system switching unit 180 uses the CPU 911 to determine which execution unit 110 is the regular system, and notifies the execution unit 110 and the failure processing unit 120 of the determination result.
The system switching unit 180 uses the CPU 911 to input parameters and the like from the outside, and notifies the execution unit 110 of both the normal system and the standby system. In addition, the system switching unit 180 acquires the process execution result output by the regular system execution unit 110 and outputs the result to the outside.

Each failure processing unit 120 uses the CPU 911 to monitor the execution of the process in the corresponding execution unit 110, and performs processing to deal with the failure when a failure occurs. The handling of the failure differs depending on whether the execution unit 110 is a regular system or a standby system. The failure processing unit 120 uses the CPU 911 to determine whether the corresponding execution unit 110 is a normal system or a standby system based on the notification from the system switching unit 180, and determines a failure based on the determination result. deal with.
Further, the failure processing unit 120 notifies the system switching unit 180 of the occurrence status of the failure in the execution unit 110 using the CPU 911. The system switching unit 180 uses the CPU 911 to determine which execution unit 110 is the regular system based on the notification from the failure processing unit 120.

FIG. 20 is a flowchart showing an example of the flow of the execution process S550 in this embodiment.
In the execution process S550, the execution unit 110 executes a process. The execution process S550 includes a system determination step S551, an initialization step S552, and a process execution step S553.

In the system determination step S551, the execution unit 110 uses the CPU 911 to determine whether it is a regular system or a standby system based on the notification from the system switching unit 180.
When it is determined that the system is the regular system, the process proceeds to the process execution step S553.
If it is determined that the system is a standby system, the process proceeds to the initialization step S552.

In the initialization step S552, the execution unit 110 (initialization execution unit 111) executes an initialization process using the CPU 911.
In the process execution step S553, the execution unit 110 (process execution unit 114) executes a process using the CPU 911.

  In this way, the normal execution unit 110 takes over the state at the end of the previous process and executes the next process, whereas the standby execution unit 110 executes the initialization process, Perform the following process: If a failure occurs in the active execution unit 110 due to an abnormality in the state at the end of the previous process, the standby execution unit 110 can execute the process normally.

FIG. 21 is a flowchart showing an example of the flow of the system switching process S590 in this embodiment.
In the system switching process S590, the system switching unit 180 manages the active system and the standby system, and performs system switching based on the failure occurrence status in the execution unit 110. The system switching process S590 includes a parameter input process S591, a regular system status determination process S592, a switching process S593, an old standby system status determination process S594, and a result output process S595.

In the parameter input step S591, the system switching unit 180 uses the CPU 911 to input parameters from the outside. The system switching unit 180 uses the CPU 911 to notify the two execution units 110 of the input parameters regardless of whether the system is a normal system or a standby system.
Thus, the regular execution unit 110 takes over the state at the end of the previous process and starts executing the process. The standby execution unit 110 starts executing the process after executing the initialization process. The standby execution unit 110 may be configured to execute an initialization process in advance and start the execution of the process immediately after inputting the parameters.

In the regular system status determination step S592, the system switching unit 180 uses the CPU 911 to determine the execution status of the process in the execution unit 110 that is the regular system.
If it is determined that the process has been completed, the process proceeds to a result output step S595.
When it is determined that the execution of the process is interrupted, the process proceeds to the switching step S593.
If it is determined that the process has not been completed or interrupted, the routine status determination step S592 is repeated.

For example, as described in the first embodiment, when a failure has occurred in the execution unit 110, but there is no checkpoint to be returned, the failure processing unit 120 interrupts the execution of the process in the execution unit 110. The failure processing unit 120 notifies the system switching unit 180 of this using the CPU 911. The system switching unit 180 uses the CPU 911 to determine that the execution of the process in the execution unit 110 is interrupted based on the notification from the failure processing unit 120.
Further, when the execution of the process in the execution unit 110 is completed, the execution unit 110 outputs the result of the process execution. The system switching unit 180 uses the CPU 911 to acquire the result output by the execution unit 110 and determines that the execution of the process in the execution unit 110 has ended.

In the switching step S <b> 593, the system switching unit 180 uses the CPU 911 to switch the standby execution unit 110 to the regular system, and switches the regular system execution unit 110 to the standby system.
The old standby execution unit 110 has already started executing the process at this point. In some cases, the execution of the process has already been completed and the result has been output. In addition, when a failure occurs even after executing the initialization process, such as when there is an abnormality in the input parameters, the execution of the process may already be interrupted.

In the old standby system status determination step S594, the system switching unit 180 uses the CPU 911 to determine the execution status of the process in the execution unit 110 that is the new regular system (= old standby system).
If it is determined that the execution of the process has been completed, the process proceeds to a result output step S595.
If it is determined that the execution of the process is interrupted, the process returns to the parameter input step S591. That is, abandon the continuation of the process and prepare for the next process.
If it is determined that the execution of the process has not been terminated or interrupted, the old standby system status determination step S594 is repeated.

In the result output step S595, the system switching unit 180 uses the CPU 911 to acquire the result output by the execution unit 110 that is the regular system. Using the CPU 911, the system switching unit 180 outputs the acquired result to the outside.
Thereafter, the process returns to the parameter input step S591 to prepare for the next process.

As described above, when a failure occurs due to an abnormality in the state at the time when the execution unit 110 that is the active system starts executing the process, the execution unit 110 that is the standby system performs the process after the initialization process is executed. Is the output of the process execution apparatus 100. Thereby, obstacles can be overcome.
In addition, since the execution unit 110, which is a standby system, executes the process in advance, the delay due to the occurrence of a failure can be reduced. That is, the process execution device 100 can output the result without much delay.
In addition, when a parameter itself input by the process execution apparatus 100 is abnormal and a failure occurs due to the failure, a failure occurs in both the normal system and the (old) standby system, and the process cannot be continued. Even in such a case, if the execution unit 110, which is a standby system, executes the process in advance, it can be quickly detected that a failure has occurred. Therefore, it is possible to abandon the process early and move on to the next process. This is particularly effective when there is a lot of traffic and a process execution queue is generated.

  In this example, a configuration has been described in which a plurality of failure processing units 120 corresponding to a plurality of execution units 110 are provided, and each failure processing unit 120 manages a corresponding execution unit 110. The unit 120 may manage the plurality of execution units 110.

The process execution device 100 in this embodiment further includes a plurality of the execution units 110 and a system switching unit 180.
Among the execution units, the first execution unit 110 executes the process as a regular system using the processing device.
Of the execution units, the second execution unit 110 executes the process as a standby system using the processing device.
The state storage unit 123 stores the state of the execution unit 110 at the intermediate time point for each of the plurality of execution units 110 using the storage device.
The system switching unit 180 uses the processing device to switch the second execution unit 110 to a normal system when a failure occurs in the first execution unit 110, and switches the first execution unit 110. Switch to the standby system.

  According to the process execution device 100 in this embodiment, when a failure occurs in the execution unit 110 that is the active system, the execution unit 110 that is the standby system is switched to the normal system, so that the failure can be dealt with early. it can.

In the process execution device 100 according to this embodiment, the second execution unit 110 executes the process after executing initialization processing as a standby system using the processing device.
The system switching unit 180 is a case where a failure has occurred in the first execution unit 110 using the processing device, and the state rewinding unit 133 should return the state of the first execution unit 110. When the state storage unit 123 does not store the state at the intermediate time point, the second execution unit 110 is switched to the regular system, and the first execution unit 110 is switched to the standby system.

  According to the process execution device 100 in this embodiment, the execution unit 110 that is the standby system executes the process after executing the initialization process before a failure occurs in the execution unit 110 that is the active system. If the initialization process is executed, such as when a failure occurs due to an abnormality in the internal state at the start of the process, the failure can be quickly overcome when the process can be executed normally.

  The dual system (process execution device 100) provided with the system restart device described above performs normal processing in the normal system and simultaneously processes input data in a process in which internal data is initialized in the standby system. An output data switching device (system) that makes the output data of the standby system valid data when the restart from all checkpoint states accumulated in the snapshot accumulation unit (status storage unit 123) fails when a failure occurs A switching unit 180) is provided.

  The system restart device includes a process processing unit (execution unit 110) that executes normal process processing, a checkpoint processing unit (checkpoint determination unit 121) that collects a snapshot of memory contents at a normal checkpoint, A snapshot storage unit (state storage unit 123) for storing snapshots collected by the system, a failure detection unit (failure determination unit 131) for detecting occurrence of a failure during normal system process processing, and process processing when a failure occurs in the normal system A restart instructing unit (rewind time determining unit 132) for instructing restart of the unit, a state return instructing unit (state rewinding unit 133) for writing back the snapshot data of the regular snapshot storage unit to the process processing unit, Notify system switchover when failed to restart using all snapshots of regular system but failed A switching notification unit, a checkpoint processing unit (state acquisition unit 122) that collects a snapshot of memory contents at a checkpoint of a standby system, a snapshot storage unit (state storage unit 123) that accumulates snapshots collected by a standby system, A system switching receiving unit for receiving a system switching notification, a system switching unit (system switching unit 180) for notifying the standby system of a system switching when a system switching notification is received from the active system, and a process in the standby system An internal data initialization unit (initialization execution unit 111) for instructing initialization of internal information is included.

When the regular process processing unit performs processing with input data as input, the standby process processing unit similarly performs processing with the same input data. However, the normal process processing unit executes the process processing unit while retaining the internal data as usual, whereas the standby process processing unit inputs the input data after initializing the internal data. The difference is that the process is executed.
The last snapshot that is accumulated in the snapshot accumulation unit after a failure occurs in the active process processing unit and the restart instruction unit restarts while restoring the system status with the snapshot, but the restart fails repeatedly Use to restart. At this time, if the process of the process processing unit is executed up to the next check point, it is determined that the process restart is successful, and normal operation is performed. If the restart fails before the next checkpoint, the failure detection unit abandons the continuation of processing in the normal system, notifies the system switching unit of system switching and makes itself a standby system. The system switching unit that has received the system switching notification notifies the standby system switching receiving unit to shift to the regular system. The standby system switching reception unit notifies the received system switching notification to the process processing unit, activates a processing unit necessary for the active system, and shifts to the normal system.

  As described above, when the process is executed with the internal information of the process processing unit initialized in advance on the standby system, the process restarts using all snapshots in the normal system and fails. However, by switching the system, processing equivalent to initializing the internal information of the process itself can be performed quickly, so the restart time of the entire system can be shortened and the system can be operated efficiently. .

Embodiment 6 FIG.
The sixth embodiment will be described with reference to FIGS.
Since the configuration of the process execution apparatus 100 in this embodiment is the same as that of the fifth embodiment, the description thereof will be omitted and only different points will be described.

FIG. 22 is a flowchart showing an example of the flow of the execution process S550 in this embodiment.
The execution process S550 includes a parameter acquisition step S561, a standby step S562, a step execution step S563, a checkpoint determination step S564, and a result notification step S565.

In the parameter acquisition step S561, the parameter input unit 113 acquires parameters from the system switching unit 180 using the CPU 911.
In the standby step S562, the process execution unit 114 uses the CPU 911 to determine whether or not the next step of the process may be executed.
If it is determined that the process can proceed to the next step, the process proceeds to step execution step S563.
When it is determined that the process should wait without proceeding to the next step, the standby process S562 is repeated.

There are several cases where the next step should not be performed. For example, when the process of the process execution unit 114 reaches a checkpoint and the state acquisition unit 122 reads out the data stored in the internal state storage unit 112, the execution of the next step during reading is started. Since the data stored in the internal state storage unit 112 changes, it is necessary to wait until the reading is completed.
Further, when a failure occurs and the state rewinding unit 133 is writing data to the internal state storage unit 112, it is necessary to wait until the writing is completed.

Further, in the process execution apparatus 100 in this embodiment, the standby execution unit 110 waits at the previous checkpoint until the regular execution unit 110 reaches the next checkpoint. The standby execution unit 110 determines whether or not the regular execution unit 110 has reached the next check point based on the notification from the system switching unit 180.
Since the standby execution unit 110 waits for the normal execution unit 110 to reach the next check point without any problem, the standby execution unit 110 executes the process thereafter. No failure occurs.

In step execution step S563, the process execution unit 114 uses the CPU 911 to execute the next step of the process.
In checkpoint determination step S564, the process execution unit 114 uses the CPU 911 to determine whether the execution of the process has reached the checkpoint.
If it is determined that the check point has not yet been reached, the process returns to the step execution step S563 to execute the next step.
If it is determined that the last checkpoint has been reached, the process proceeds to a result notification step S565.
If it is determined that another check point has been reached, the process returns to the standby step S562, and waits for the state acquisition unit 122 to finish reading the data stored in the internal state storage unit 112.

In the result notification step S565, the result output unit 115 uses the CPU 911 to notify the system switching unit 180 of the result of the process execution unit 114 executing the process.
Then, it returns to parameter acquisition process S561 and prepares for the next process.

When the failure determination unit 131 determines that a failure has occurred in the execution unit 110 during the execution of the step execution step S563 or the checkpoint determination step S564, the state rewinding unit 133 forces the execution unit 110 to wait for the standby step S562. Return to. The execution unit 110 waits for the state rewinding unit 133 to finish writing data in the internal state storage unit 112.
In addition, when one of the steps is being executed, the failure determination unit 131 determines that a failure has occurred in the execution unit 110, and the state rewinding unit 133 determines that the execution of the process in the execution unit 110 is to be interrupted. The rewinding unit 133 forcibly returns the execution unit 110 to the parameter acquisition step S561.

FIG. 23 is a flowchart showing an example of the flow of the system switching process S590 in this embodiment.
The system switching process S590 includes a parameter input process S591, a regular system status determination process S592, a switching process S593, and a result output process S595.

In the regular system status determination step S592, the system switching unit 180 uses the CPU 911 to determine the execution status of the process in the execution unit 110 that is the regular system.
If it is determined that the process has been completed, the process proceeds to a result output step S595. The process execution device 100 ends the execution of the process normally.
If it is determined that the process execution has been interrupted, the process returns to the parameter input step S591. The process execution device 100 abandons execution of the process and prepares for the next process.
If it is determined that a failure has not occurred in the process execution, the process proceeds to the switching step S593.
When it is determined that the process has been normally executed out of the cases where the execution of the process has not been terminated or interrupted, the regular system status determination step S592 is repeated.

In the switching step S593, the system switching unit 180 uses the CPU 911 to switch the execution unit 110 that is the active system to the standby system, and switches the execution unit 110 that is the standby system to the normal system.
Thereafter, the process returns to the regular system status determination step S592.

FIG. 24 is a flowchart showing an example of the flow of the failure process S510 in this embodiment.
The failure processing S510 includes a system determination step S541 and a standby determination step S542 in addition to the steps described in the first embodiment.

In the system determination step S541, the failure processing unit 120 uses the CPU 911 to determine whether the corresponding execution unit 110 is the normal system or the standby system based on the notification from the system switching unit 180.
When it is determined that the system is the regular system, the process proceeds to the failure determination step S511.
When it determines with it being a standby system, it progresses to standby determination process S542.

In the standby determination step S542, the failure processing unit 120 uses the CPU 911 to determine whether or not the corresponding execution unit 110 (which is a standby system) can execute the process. The failure processing unit 120 uses the CPU 911 to notify the corresponding execution unit 110 of the determined result. The execution unit 110 determines whether to execute the next step or wait based on the notification from the failure processing unit 120.
If it is determined that the corresponding execution unit 110 can execute the process, the process proceeds to a checkpoint determination step S512.
If it is determined that the corresponding execution unit 110 is to be put on standby, the process returns to the system determination step S541.

In the state determination step S524, the rewinding time determination unit 132 uses the CPU 911 to determine whether each of the two execution units 110 has reached the check point before that. For example, the rewind time determination unit 132 determines whether or not the state storage unit 123 of each of the two failure processing units 120 has stored data corresponding to the checkpoint so far. It is determined whether or not the checkpoint has been reached.
If neither of the two execution units 110 has reached the checkpoint before that time, and determines that this is the first time, the process proceeds to the rewind time advance step S525.
If any of the two execution units 110 determines that the checkpoint has been reached before that, the process returns to the system determination step S541.

In the rewind time advance step S525, the rewind time determination unit 132 uses the CPU 911 to rewind the checkpoint immediately before the checkpoint determined that the checkpoint determination unit 121 has reached in the checkpoint determination step S512. It is determined that the return time. If the arrived checkpoint is the first checkpoint, there is no previous checkpoint, so the rewind time determination unit 132 determines that there is no checkpoint that will be the rewind time. The rewind time determination unit 132 stores the determination result using the magnetic disk device 920.
In addition, the rewinding time determination unit 132 of the other failure processing unit 120 uses the CPU 911 to determine the previous checkpoint as the rewinding time. If the reached checkpoint is the first or second checkpoint, there is no checkpoint two prior to that, so the other rewind time determination unit 132 still has a checkpoint that is the rewind time. Judge that it does not exist. Another rewind time determination unit 132 stores the determination result using the magnetic disk device 920.
Thereafter, the process returns to the system determination step S541.

In the rewinding time retraction process S534, the rewinding time determination unit 132 uses the CPU 911, and among the data stored in the state storage unit 123, the state rewinding unit 133 is stored in the internal state storage unit 112 in the state rewinding step S533. It is determined that the check point two before the check point corresponding to the written data is the rewinding point. When the checkpoint corresponding to the data written by the state rewinding unit 133 to the internal state storage unit 112 in the state rewinding step S533 is the first or second checkpoint, there is no checkpoint two prior to that. Therefore, the rewind time determination unit 132 determines that there is no check point that is the rewind time. The rewind time determination unit 132 stores the determination result using the magnetic disk device 920.
Thereafter, the process returns to the system determination step S541.

  FIG. 25 is a schematic diagram showing an example of the operation of the process execution device 100 in this embodiment.

  At the process start 601, in the regular execution unit 110, the parameter input unit 113 inputs parameters, and the process execution unit 114 starts processing the process. In the standby execution unit 110, the process execution unit 114 waits without starting the process until the processing of the regular process execution unit 114 reaches the first checkpoint 602.

  When the processing of the regular processing execution unit 114 reaches the first check point 602, the standby processing execution unit 114 starts processing the process. At this time, the storage content 641a of the regular state storage unit 123 is only the state 652a at the first checkpoint 602. In addition, the storage content 641b of the standby state storage unit 123 is empty. This is the first time that any one of the process execution units 114 has reached the first check point 602, but the first check point 602 is the first check point. It is determined that the return point does not exist yet. The standby rewinding time determination unit 132 also determines that the rewinding time does not exist yet.

  When the processing of the standby processing execution unit 114 reaches the first check point 602 and the normal processing execution unit 114 has not yet reached the second check point 603, the standby processing execution unit 114 Waits until the processing of the regular processing execution unit 114 reaches the second check point 603. The storage content 642b of the standby state storage unit 123 is only the state 652b at the first checkpoint 602. Since the normal process execution unit 114 has already reached the first checkpoint 602, the standby system rewind time determination unit 132 determines that the rewind time does not exist yet because the rewind time determination unit 132 does not advance. Keep it as it is.

  When the process of the regular process execution unit 114 reaches the second check point 603, the standby process execution unit 114 resumes the process. At this time, the storage contents 642a of the regular state storage unit 123 are two states, a state 652a at the first check point 602 and a state 653a at the second check point 603. Since the process of any of the process execution units 114 has reached the second check point 603 for the first time, the regular system rewind time determination unit 132 determines that the previous first check point 602 is the rewind time. In addition, the standby rewinding time determination unit 132 determines that there is no rewinding time yet because there is no checkpoint immediately before the first checkpoint 602.

  Assume that the processing of the standby processing execution unit 114 reaches the third check point 604 earlier than the processing of the standby processing execution unit 114 reaches the second check point 603. The storage contents 643 a of the regular state storage unit 123 includes three states: a state 652 a at the first check point 602, a state 653 a at the second check point 603, and a state 654 a at the third check point 604. Since the process of any one of the processing execution units 114 has reached the third check point 604, the regular rewind time determination unit 132 determines that the previous second check point 603 is the rewind time. . In addition, the standby rewinding time determination unit 132 determines that the first checkpoint 602 immediately before the second checkpoint 603 is the rewinding time.

  When the processing of the standby processing execution unit 114 reaches the second check point 603, the processing of the normal processing execution unit 114 passes through the third check point 604. Does not wait and continues the process. The stored content 643b of the standby state storage unit 123 is in two states: a state 652b at the first check point 602 and a state 653b at the second check point 603. Since the normal system process execution unit 114 has already reached the first check point 602, the standby system rewind time determination unit 132 does not advance the rewind time, and rewinds the first check point 602. It keeps the state as judged.

It is assumed that a failure has occurred before the processing of the regular processing execution unit 114 reaches the fourth checkpoint 605. The system switching unit 180 switches between the normal system and the standby system.
The process execution unit 114 of the old standby system (new service system) continues the process processing as it is.
The old regular system (new standby system) state rewinding unit 133 returns the state of the execution unit 110 to the state 653 a at the second checkpoint 603 based on the determination result of the rewinding time determination unit 132. The process execution unit 114 of the old regular system (new standby system) resumes the process from the second checkpoint 603.

  FIG. 26 is a schematic diagram showing an example of the operation of the process execution device 100 in this embodiment.

  When the process execution unit 114 of the new standby system (old service system) resumes the processing from the second check point 603, the rewind time determination unit 132 does not have a check point two prior to the second check point 603. Therefore, it is determined that there is no rewinding point.

  When the processing of the new regular system (old standby system) process execution unit 114 reaches the third checkpoint 604, the storage content 644b of the new regular system status storage unit 123 is changed to the state 652b at the first checkpoint 602 and the first checkpoint 602. There are three states: a state 653b at the second checkpoint 603 and a state 654b at the third checkpoint 604. Since the processing execution unit 114 of the old service system (new standby system) has already reached the third checkpoint 604, the rewind time determination unit 132 of the new service system (old standby system) determines the rewind time point. The first check point 602 is kept in a state where the rewind time is determined without proceeding.

It is assumed that a failure occurs again before the processing of the processing execution unit 114 of the new normal system (old standby system) reaches the fourth checkpoint 605. The system switching unit 180 switches between the normal system and the standby system.
The process execution unit 114 of the new / new service system (new standby system / old service system) continues the process.
The state rewinding unit 133 of the new / new standby system (new service system / old standby system) returns the state of the execution unit 110 to the state 652b at the first check point 602 based on the determination result of the rewinding time determination unit 132. . The process execution unit 114 of the new / new standby system (new service system / old standby system) restarts the process from the first checkpoint 602.

  FIG. 27 is a schematic diagram showing an example of the operation of the process execution device 100 in this embodiment.

At the time when the processing execution unit 114 of the new and new standby system resumes processing from the first check point 602, the rewind time determination unit 132 does not have a check point two prior to the first check point 602. It is determined that the time does not exist.
At this time, the storage content 645a of the state storage unit 123 of the new / new use system includes three states: a state 652a at the first check point 602, a state 653a at the second check point 603, and a state 654a at the third check point 604. One. In addition, the storage contents 645b of the state storage unit 123 of the new / new standby system include three states: a state 652b at the first check point 602, a state 653b at the second check point 603, and a state 654b at the third check point 604. is there.

Therefore, the rewind time determination unit 132 cannot advance the rewind time until the process of any of the process execution units 114 reaches the fourth check point 605.
If, before that, a failure occurs in the processing execution unit 114 of the new and new active system, the system switching unit 180 switches between the active system and the standby system. The process execution unit 114 of the new / new standby system (new / new use system) interrupts the process of the process because there is no checkpoint whose state should be returned.
Further, when a failure occurs in the processing execution unit 114 of the new / new / usual system (new / standby system), the system switching unit 180 switches between the normal system and the standby system. The process execution unit 114 of the new / new / new standby system (new / new / usual system / new / new standby system) interrupts the process of the process because there is no checkpoint whose state should be returned. Since the process execution unit 114 of the new / new / new-use system (new / new / standby system / new / new-use system) has already suspended the process, the system switching unit 180 abandons the continuation of the process, Prepare for the next process.

  It is assumed that no failure has occurred and the processing of the new / new normal processing execution unit 114 has reached the fourth checkpoint 605. The storage contents 646a of the new / usual state storage unit 123 include a state 652a at the first checkpoint 602, a state 653a at the second checkpoint 603, a state 654a at the third checkpoint 604, and a fourth checkpoint 605. There are four states with state 655a in FIG. Since the process of any one of the process execution units 114 has reached the fourth check point 605 for the first time, the rewind time determination unit 132 of the new and new service system sets the previous third check point 604 as the rewind time. judge. Further, the rewind time determination unit 132 of the new and new standby system determines that the second check point 603 immediately before the third check point 604 is the rewind time.

In this way, in preparation for the occurrence of a failure, the standby execution unit 110 follows the normal execution unit 110 and executes the processing of the process. Therefore, if a failure occurs in the normal execution unit 110, the standby system By switching to the regular system, it is possible to quickly cope with the failure.
In particular, even when a failure occurs repeatedly and returns to two or more previous checkpoints and resumes processing, the standby execution unit 110 always keeps one checkpoint behind the regular execution unit 110. Since it is chasing, only one checkpoint delay occurs, and the delay due to re-processing can be minimized.

In the process execution device 100 in this embodiment, the second execution unit 110 uses the processing device as a standby system, and the state storage unit 123 stores the state of the first execution unit 110 in the middle. Wait at an intermediate time before the time (checkpoint).
When the system switching unit 180 switches the first execution unit 110 from the regular system to the standby system using the processing device, the state rewinding unit 133 causes the state storage unit 123 to The state of the first execution unit 110 is returned to the state at the intermediate time point before the intermediate time point at which the state is stored for the execution unit 110.

  According to the process execution apparatus 100 in this embodiment, the standby execution unit 110 waits at an intermediate time before the normal execution unit 110, so that when the failure occurs, the system switching unit 180 causes the normal system and the standby system to It is possible to deal with a failure promptly by switching between. In addition, since the former active system execution unit 110 switched to the standby system returns to the intermediate point before that and performs the process again, the old standby system execution unit 110 switched to the normal system also fails. When this occurs, the system switching unit 180 switches between the normal system and the standby system again, so that the failure can be quickly dealt with.

  In the dual system (process execution device 100) provided with the system restart device described above, the standby system always waits at the previous checkpoint while the normal system is performing normal processing. When a failure occurs, a system that has been switched from the standby system to the normal system performs processing, and the system that has switched to the standby system is further provided with a system switching device (system switching unit 180) that returns to the previous checkpoint and waits.

  The system restart device includes a process processing unit (execution unit 110) that executes normal process processing, a checkpoint processing unit (state acquisition unit 122) that collects a snapshot of memory contents at a normal checkpoint, and a normal system A snapshot storage unit (state storage unit 123) for storing the snapshots collected, a failure detection unit (failure determination unit 131) for detecting a failure during normal process processing, and a process processing unit when a failure occurs in the normal system A restart instructing unit (rewind time determining unit 132) for instructing restart of the state, a state return instructing unit (state rewinding unit 133) for writing back the snapshot data of the regular snapshot accumulation unit to the process processing unit, A system switchover notification that notifies the system switchover when an attempt to restart using all snapshots of the system fails. A checkpoint processing unit (state acquisition unit 122) that collects a snapshot of memory contents at a checkpoint of a standby system, a snapshot storage unit (state storage unit 123) that accumulates snapshots collected by a standby system, A system switching receiving unit that receives a switching notification, a system switching unit (system switching unit 180) that notifies the standby system of a system switching notification when there is a system switching notification from the active system, and is delayed from the active system and waits A processing request delay unit for executing the process in the system;

When the regular process processing unit performs processing with input data as input, the standby process processing unit similarly performs processing with the same input data. However, the normal process processing unit starts processing immediately when input data is given, whereas the standby process processing unit starts with the first check when the input data is given. The difference is that the processing is executed after a certain period of time that passes the point is delayed.
The normal process processing unit and the standby process processing unit store snapshots in the normal system snapshot storage unit and the standby system snapshot storage unit, respectively, when they pass the checkpoint. Here, when a failure occurs in the normal process processing unit and the failure detection unit detects a failure, the failure detection unit notifies the system switching notification unit of the system switching and starts up the processing unit necessary for the standby system. Shift their own system to the standby system. The system switching notifying unit notifies the system switching receiving unit via the system switching unit to switch the standby system to the active system, and the system switching receiving unit processes the standby system process processing unit as the normal system. Make a notification. At this time, the process processing unit checks the internal information and the input data, initializes the internal information if there is a problem with the internal information, and discards the input data if there is a problem with the input data. For the system switched from the regular system to the standby system, two snapshots are restored to restore the process state, and the internal information and input data are checked in the same manner as described above, and then the process processing is resumed.

  As described above, the standby process is delayed by about one checkpoint, and when a failure occurs, the process is restarted by switching the system and checking the input data and internal information. The system can be started quickly, and the standby system is restarted after returning one more checkpoint. Therefore, even if a failure occurs repeatedly, the process can be quickly restarted by switching the system again. You can start up.

  100 process execution device, 110 execution unit, 111 initialization execution unit, 112 internal state storage unit, 113 parameter input unit, 114 process execution unit, 115 result output unit, 120 fault processing unit, 121 checkpoint determination unit, 122 status acquisition Unit, 123 state storage unit, 131 failure determination unit, 132 rewind time determination unit, 133 state rewind unit, 134 state deletion unit, 180 system switching unit, 601 process start time, 602 first check point, 603 second check Point, 604 third check point, 605 fourth check point, 606 final check point, 611-646 stored contents, 651-655 status, 901 display device, 902 keyboard, 903 mouse, 904 FDD, 905 CDD, 906 printer device, 907 CANA device, 910 system unit, 911 CPU, 912 bus, 913 ROM, 914 RAM, 915 communication device, 920 magnetic disk device, 921 OS, 922 window system, 923 program group, 924 file group, 931 telephone, 932 facsimile machine 940 Internet, 941 Gateway, 942 LAN.

Claims (10)

A storage device for storing data, a processing device for processing data, an execution unit, a state storage unit, and a state rewinding unit;
The execution unit executes a process using the processing device,
The state storage unit stores the state of the execution unit at the intermediate time point for each of a plurality of intermediate time points while the execution unit is executing the process using the storage device,
The state rewinding unit sets the state of the execution unit to a state at any intermediate point in the state stored in the state storage unit when a failure occurs in the execution unit using the processing device. In the event that a failure further occurs in the execution unit, the state of the execution unit is changed to a state at an intermediate time point before the intermediate point in time when the state of the execution unit is returned among the states stored in the state storage unit. Is a process execution device. The state rewinding unit sets the state of the execution unit to the state of the newest intermediate time among the states stored in the state storage unit when a failure occurs in the execution unit using the processing device. Return,
The said state memory | storage part deletes the state of the intermediate | middle time point in which the said state rewinding part returned the state of the said execution part among the memorize | stored states using the said processing apparatus. Process execution device.   The said state memory | storage part memorize | stores the state which initialized the said execution part as the state of the said execution part in the first intermediate | middle time using the said memory | storage device. Process execution device.   The state rewinding unit is a state where the execution unit further fails, and the state at the intermediate time point before the intermediate time point when the state of the execution unit is returned is The process execution apparatus according to claim 1, wherein when the storage unit does not store, the execution unit interrupts execution of the process. The process execution device further includes a plurality of the execution units and a system switching unit.
The first execution unit of the execution units executes the process as a regular system using the processing device,
The second execution unit among the execution units executes the process as a standby system using the processing device,
The state storage unit stores the state of the execution unit at the intermediate time point for each of the plurality of execution units using the storage device.
The system switching unit switches the second execution unit to a normal system and switches the first execution unit to a standby system when a failure occurs in the first execution unit using the processing device. The process execution apparatus according to claim 1, wherein the process execution apparatus is provided. The second execution unit waits at an intermediate time point before the intermediate time point when the state storage unit stores the state of the first execution unit as a standby system using the processing device,
When the system switching unit switches the first execution unit from the regular system to the standby system, the state storage unit uses the processing device to state the state storage unit with respect to the second execution unit. The process execution device according to claim 5, wherein the state of the first execution unit is returned to the state of the intermediate time point before the intermediate time point in which is stored. The second execution unit executes the process after executing the initialization process as a standby system using the processing device,
The system switching unit is a state at an intermediate time when the state rewinding unit should return the state of the first execution unit when a failure occurs in the first execution unit using the processing device. When the state storage unit does not store, the second execution unit is switched to the regular system, and the first execution unit is switched to the standby system. The process execution device described.   The said state memory | storage part deletes the state of the said said execution part memorize | stored, when the said execution part complete | finishes execution of the said process using the said processing apparatus. The process execution apparatus in any one.   A computer program for causing a computer having a storage device for storing data and a processing device for processing data to function as the process execution device according to any one of claims 1 to 8. In a process execution method in which a process execution device having a storage device for storing data and a processing device for processing data executes the process,
The processing device executes the process,
The storage device stores the state at the intermediate time point for each of a plurality of intermediate time points while the processing device is executing the process,
When the processing device fails in the execution of the process, the processing device returns the state to the state at any intermediate point among the states stored in the storage device,
When the processing device further fails in the execution of the process, the processing device returns the state to a state at an intermediate time point before the intermediate time point when the state is returned among the states stored in the storage device. Process execution method.

Images