JP2003280939A - Process pair execution control method and process pair execution control program in fault tolerant system, and fault tolerant system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To utilize an open system without requiring CPU resource double. <P>SOLUTION: A checkpoint collecting part 15 conducts a reading-out D of the condition of a primary process 11 every time when a checkpoint collection time comes, and executes a copy E to a backup process 12. If this backup process 12 is in an execution state, the checkpoint obtaining part 15 stops the process 12 by an instruction F to a backup process execution state control part 17. A process-issue system call detecting part 19 restarts the process 12 from the latest obtained checkpoint, and brings it into an execution by an instruction G to the backup process execution state control part 17 if the backup process 12 is stopped when a system call is issued from the primary process 11. <P>COPYRIGHT: (C)2004,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a fault tolerant technique which enables a computer executing a process to continue to execute the process by utilizing another computer even if a failure occurs in the computer. In particular, the present invention relates to a process pair execution control method, a process pair execution control program and a fault tolerant system in a fault tolerant system, which is suitable when the technique is applied to a scientific and technological calculation program such as CAD or simulation.

[0002]

2. Description of the Related Art Even if a computer that is executing a process fails, a process is used as a typical fault-tolerant technique that makes it possible to continue execution of the process by using another computer. The pair method is known.

In the process pair system, a process is composed of two processes, a primary process and a backup process,
This is a method of arranging both processes on different computers. Conventionally, there are first and second methods described below in this process pair method.

(1) First method The first method is the one described in the document "" The Process Group Approach t.
o Reliable Distributed Computing, ”K. Birman, Tech
nical Report, Computer Science Department, Cornel U
niversity, July 1991 "(hereinafter referred to as the first document)
It is described in.

In this first document, a process pair is called a process group, and processing by two or more processes is multiplexed. Here, as a form of the process pair system in which the number of processes is limited to two, it is referred to as a first system.

FIG. 9 is a diagram for explaining the first method. As shown in FIG. 9, in the first method, one process is configured as a process pair 91 including a primary process and a backup process. When the process pair 91 communicates with another process pair 92, communication between process pairs 93 to 96 and the like are performed. In the process pair-to-process communication, a function is provided in which the sender primary process and the sender backup process send and receive messages to the receiver primary process and the receiver backup process in a consistent state.

Note that the meaning of transmitting and receiving a message in a consistent state means that both the primary process and the backup process send or receive only one message. Conversely, it means that only the primary process receives the message and the backup process does not receive the message.

In the first method, fault1,
Alternatively, at the time indicated by fault2, even if the computer on which the primary process is running fails, the backup process running on another computer continues processing and replaces the role of the primary process. As a result, the process pair 91 can continue processing.

In the first method shown in FIG. 9, exactly the same processing is executed by two processes (a primary process and a backup process), so that twice the CPU resource is required.

(2) Second Method The second method is described in the document "Fault Tolerant System", Gray et al., Translated by Eiichi Watanabe, McGraw-Hill Publishing "(hereinafter referred to as the second document). Has been done.
In this second document, the process pair is directly called a process pair.

FIG. 10 is a diagram for explaining the second method. As shown in FIG. 10, in the second method, one process is configured as a process pair 101 including a primary process and a backup process. When the process pair 101 communicates with another process pair 102, communication between process pairs is performed.

In the process pair-to-process communication, from the sender-side primary process and the sender-side backup process to the receiver-side primary process and the receiver-side backup process,
Provides a function to send and receive messages in a consistent state.

In the first method, the backup process also executes the same processing as the primary process.
However, in the second method, the backup process exists as a process but does not execute the actual processing,
Checkpoint collection (ckp1, ckp2, ckp
3, ckp4), copy the state of the primary process to the backup process.

In the second method, the fault shown in FIG.
At the time indicated by 1 or fault2, if the computer on which the primary process is running fails, the backup process on the other computer is re-reset.
The process is restarted from the time point of the checkpoint taken at the end, which is indicated by start1 or restart2, and replaces the role of the primary process. For this reason,
Processing can continue as a process pair.

Unlike the first method, the second method does not perform exactly the same processing in two processes, a primary process and a backup process. Therefore, the second method does not require twice the CPU resource.

[0016]

In the conventional fault-tolerant technology called the process pair method described above, for example, the first method, even if the computer in which the primary process is executed fails, another computer can execute it. The backup process being executed continues processing, and by replacing the role of the primary process, processing can be continued as a process pair. However, in the first method, since exactly the same processing is executed by two processes (primary process and backup process), C
There is a problem that PU resources are required twice.

On the other hand, unlike the first method, in the second method, exactly the same processing is not executed by the two processes, and therefore the CPU resource need not be doubled. However, the second method has another problem described below.

First, in the second method, a backup process that is not actually executed, that is, a backup process that is in a stationary state unless a failure occurs on the primary process side, is restarted from the time of checkpoint collection. First, copy the state of the primary process to the backup process. For address space and context, this is fine.

However, in the second method, since the backup process is constantly in a stopped state, the OS is executed by executing a system call by simply copying the state of the primary process to the backup process.
Unable to restore the state of the service received from (operating system). The status of the service received from the OS is, for example, which file is opened by which descriptor, its seek pointer, or the like. Therefore, in the second method, such an OS
It employs a unique OS that has the function of saving and restoring the status of services received from.

Therefore, in the second method, the open system widely used in the industrial world cannot be used, and it is necessary to independently develop all the applications, which causes a problem of low productivity. is there.

The present invention has been made in consideration of the above circumstances, and an object thereof is to enable the processing to be continued even when a failure occurs, without needing to double the CPU resources, and to use an open system. It is to provide a process pair execution control method, a process pair execution control program, and a fault tolerant system in a fault-tolerant system that can be performed.

[0022]

According to one aspect of the present invention, a fault tolerant system in which a process pair composed of a primary process and a backup process is executed, which can continue processing even when a failure occurs A process pair execution control method in is provided.
This process pair execution control method includes a step of setting both the primary process and the backup process to the running state at the time of starting the process pair, and a step of copying the state of the primary process to the backup process at each checkpoint collection time, If the backup process is in the running state at the time of checkpoint collection, the step to put the backup process in the stopped state, and if the backup process is in the stopped state when a system call is issued from the process pair, the backup process is in the stopped state. Is restarted from the checkpoint (the last checkpoint) taken most recently and is put into the execution state.

In the process pair execution control method according to the first aspect of the present invention, both the primary process and the backup process operate in the running state immediately after the program (process pair) is started. Therefore, both processes are in a state of being provided with services from the OS. After that, the backup process is stopped due to the arrival of the first checkpoint collection time,
CPU resource consumption on the backup process side is suppressed. Also, if a system call is issued while the backup process is stopped, the backup process resumes processing from the last checkpoint. That is, the backup process also operates in the running state, and the primary process and the backup process are again in the state of being provided with the service from the OS.

As described above, in the process pair execution control method according to the first aspect of the present invention, the backup process is different from the second method described in the section of the prior art,
It will be in the running state immediately after the program is started, and if a system call is issued while the backup process is stopped after that. On the other hand, as in the case of the second method, the primary process remains in the running state until the program ends unless a failure occurs. That is, in the process pair execution control method according to the first aspect, unlike the first method described in the section of the related art, the primary process and the backup process do not always execute exactly the same processing, and Unlike the method of 2, the backup process is not constantly in the stopped state, and there is a period in which both processes are in the running state. During this period, both processes are in a state of receiving service from the OS.

Therefore, in the process pair execution control method according to the first aspect of the present invention, there is a function capable of saving / restoring the state of the service received from the OS, as in the second method. It is possible to use an open system widely used in the industrial world without having to adopt a proprietary OS that the company has. Further, since there is a period during which the backup process is in a stopped state, it is not necessary to double the CPU resource. Further, the process pair execution control method according to the first aspect of the present invention is particularly suitable for a science and technology calculation program such as CAD and simulation for the following reasons. That is, in this kind of scientific and technological calculation program, first, input data is read with issuance of a system call, etc., and then the CPU operation is often repeated without issuing a system call. The period to be closed is significantly longer than the period involving issuance of a system call. Therefore, in the process pair execution control method according to the first aspect, both the primary process and the backup process are put into the running state immediately after the program is started and thereafter only during the period accompanied by the issuance of the system call, and other long time CP performed
By suspending the backup process during the U operation, the time taken for the CPU resource to execute the backup process is significantly reduced, and even if a general-purpose OS is used, the service received from the OS is Allows you to save and restore state.

If the backup process is in a stopped state when a failure occurs on the primary process side, if a step of restarting the backup process from the checkpoint taken most recently to put it in the execution state is added. For example, even if a failure occurs on the primary process side while the backup process is stopped, the backup process can continue the processing.

The checkpoint collection timing (first checkpoint collection timing) is set at the timing immediately before the backup process needs to be switched from the stopped state to the running state, for example, immediately before the system call is issued from the primary process. ) Is set, the time required for the processing after the restart is restarted when the primary process actually issues a system call to restart the backup process.

If a step for setting a checkpoint collection time (second checkpoint collection time) is added immediately after it becomes unnecessary to keep the backup process in the execution state, the backup process will be in the execution state. Can be kept to a minimum to prevent excessive use of CPU resources required to perform the backup process.

Also, checkpoints are collected at predetermined time intervals during the period during which the backup process is in a stopped state, excluding the period from the first checkpoint collection period to the next second checkpoint collection period. If a step for setting the time (third checkpoint collection time) is added, the checkpoint collection operation is performed during the period when both the primary process and the backup process are in the running state, and the backup process is stopped each time, It is possible to prevent the useless operation that delays the processing, that is, the backup process is restarted from the last checkpoint by the system call in the primary process thereafter. As a result, it is possible to improve processing efficiency and efficiently collect check points.

The present invention relating to the process pair execution control method described above also executes the method as an invention relating to a program (process pair execution control program) for causing a computer to execute each step constituting the method. The present invention can also be realized as an invention relating to a fault tolerant system.

[0031]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing the configuration of a fault tolerant system according to an embodiment of the present invention.
In FIG. 1, computers 1a and 1b are network 2
Interconnected by. A primary process 11 is arranged on (the storage device (not shown) of) the computer 1a, and a backup process 12 is arranged on (storage device thereof (not shown)) of the computer 1b. That is, computers 1a and 1
The group b realizes a process pair 13 including a primary process 11 and a backup process 12 that can continue processing even when a failure occurs.

Each of the computers 1a and 1b has an inter-process pair communication unit 14, a checkpoint collection unit 15, a process restart unit 16, a backup process execution state control unit 17, and a backup process execution state management unit 1.
8 and each functional element of the process issuing system call detection unit 19. Each of these units 14 to 19 is a library 20 that stores a function program or the like used from the primary process 11 or the backup process 12.
Is realized by

FIG. 2 is a functional block diagram showing the relationship between the process pair 13 including the primary process 11 and the backup process 12 and the units 14 to 19 realized by the library 20.

The inter-process pair communication unit 14 is the computer 1a.
Communication between the process pair 13 composed of the primary process 11 and the backup process 12 on 1b and another process pair is performed. When both the primary process 11 and the backup process 12 are alive (existing), the inter-process pair communication unit 14 arbitrates both processes and sends only one message to another process pair. On the contrary, when a message is sent from another process pair, the inter-process pair communication unit 14 sends the message to both the primary process 11 and the backup process 12 when both are alive.

The checkpoint sampling unit 15 is activated when the checkpoint sampling time comes. here,
The checkpoint collection unit 15 uses the primary process 1
It is activated by an instruction A for checkpoint collection by a system call from 1 (checkpoint collection instruction) A or a periodic interrupt (checkpoint collection interrupt) B from a timer TM. The checkpoint collection unit 15 includes a backup process execution state management unit 18
When it is determined that the primary process 11 is alive from the state of the primary process 11 managed by, the instruction C is issued to the primary process 11 and the state D of the primary process 11 is read. , Copy to Backup Process 12 E
Has a function of performing a checkpoint collection operation. The state of the process mentioned here is the address space and the context.

The checkpoint collection unit 15 also uses the backup process execution state management unit 18 to determine the state of the backup process 12 (execution state or stop state).
If it is in the execution state, an instruction F for stopping the execution of the backup process 12 is issued to the backup process execution state control unit 17.

When the process restart unit 16 detects a failure in the computer that is executing the primary process 11, the process restart unit 16 checks the state of the backup process 12 using the backup process execution state management unit 17, and if it is in the stopped state. , An instruction H for starting the execution of the backup process 12 is issued to the backup process execution state control unit 17.

Backup process execution status controller 17
When the process pair 13 including the primary process 11 and the backup process 12 is activated (when the execution of the program is started), the primary process 11 and the backup process 12 are both set to the execution state, and the backup process execution state management unit 18 Has the function to register. The backup process execution state control unit 17 also has a function of receiving the instruction F from the checkpoint collection unit 15, stopping the backup process 12, and registering the state in the backup process execution state management unit 18 as a stop state. . The backup process execution state control unit 17 also receives the instruction H from the process restart unit 16 or the process issuing system call detection unit 1
It also has a function of receiving the instruction G described later from 9 to execute the backup process 12 and registering the state in the backup process execution state management unit 18 as the execution state.

Backup process execution status management unit 18
Holds and manages the state of the backup process 12.

Process issuing system call detector 19
Detects that the process pair 13 (the primary process 11 or the backup process 12 constituting the process pair) has executed (issued) a system call. In addition, the process issuing system call detection unit 19 uses the primary process 1
When it is detected that 1 has executed the system call, the state of the backup process 12 is checked using the backup process execution state management unit 18. The process issuing system call detection unit 19 uses the backup process 12
Is stopped, an instruction G for starting the execution of the backup process 12 is issued to the backup process execution state control unit 17.

On the computers 1a and 1b, the inter-process pair communication units 14 each other, the checkpoint collection units 15 each other, the process restart units 16 each other, the backup process execution state control units 17 each other, the backup process execution state management unit 18 respectively. And the process issuing system call detecting units 19 are connected to each other via the network 2
By communicating via, it operates as if it were one.

FIG. 3 is a state transition diagram. FIG. 3A is a state transition diagram showing possible states of the primary process 11, and FIG. 3B is a state transition diagram showing possible states of the backup process 12. is there.

First, the primary process 11 is shown in FIG.
As shown in (a), it takes one of a stopped state and a running state. The primary process 11 transits from the stopped state to the running state when the program execution starts a1.
The primary process 11 maintains the execution state until the program execution end a2.

Next, the backup process 12 is also shown in FIG.
As shown in (b), it takes one of a stopped state and an execution state. The backup process 12 transits from the stopped state to the running state when the program execution starts b1. The backup process 12 is
When the checkpoint collection b2 is executed, the state transits to the stopped state. Further, the backup process 12 transitions to the execution state when the system call issuance b3 in the primary process 11 is performed in the stopped state.

Next, the operation of this embodiment will be described with reference to FIGS.
Will be described as appropriate. Note that FIG. 4 shows process pair 1
3 is a timing chart for explaining the overall operation of the primary process 11 and the backup process 12, FIG. 5 is a flowchart for explaining the operation of the checkpoint sampling unit 15, and FIG. 6 is the operation of the process restart unit 16. 7 is a flow chart for explaining the operation of the process issuing system call detection unit 19, and FIG. 8 is a flow chart for explaining the relationship between the states of the primary process 11 and backup process 12 and the checkpoint collection timing. 2 is a timing chart of.

First, immediately after the start of the process, that is, immediately after the start of program execution a1 and b1, both the primary process 11 on the computer 1a and the backup process 12 on the computer 1b transit from the stopped state to the running state as shown in FIG. To do.

Now, when the primary process 11 and the backup process 12 are in the execution state, for example, as shown in FIG.
It is assumed that the primary process 11 issues a checkpoint collection system call A to the checkpoint collection unit 15 at the time of ckp1.

In this case, the checkpoint collection unit 15 is activated, the backup process execution state management unit 18 is inquired about the states of the primary process 11 and the backup process 12 managed by the management unit 18, and both processes are executed. It is determined whether both are alive (step S1). If primary process 1
If at least one of 1 and the backup process 12 is not alive, the checkpoint collection unit 15 ends the operation as it is.

On the other hand, if both the primary process 11 and the backup process 12 are alive,
The checkpoint collection unit 15 is the primary process 11
To execute the operation D for reading the state of the primary process 11 and the operation E for copying the read state to the backup process 12 (step S2).

When the checkpoint collecting unit 15 executes the processing of step S2 (checkpoint collecting operation), the checkpoint collecting unit 15 inquires of the backup process execution state management unit 18 about the state of the backup process 12, and the backup process 12 becomes the execution state. It is determined whether there is any (step S3). If backup process 1
If 2 is in a stopped state, the checkpoint collection unit 15
Ends the operation as it is. On the other hand, if the backup process 12 is in the execution state, the checkpoint collection unit 15 determines that the backup process execution state control unit 1
By issuing the instruction F to 7, the backup process 12 is transited from the running state to the stopped state (step S4). The new state (stop state) of the backup process 12 is registered in the backup process execution state management unit 18 by the backup process execution state control unit 17.

In the example of FIG. 4, a checkpoint is taken even after ckp1 at ckp2, ckp3, and ckp4, which are determined by the timing of the periodic interrupt B from the timer TM, for example. At this time, backup process 1
No. 2 remains in a stopped state as is clear from the above description.

After that, for example, in order to write the output data accompanying the processing in the primary process 11, the primary process 11 performs the send processing 41 shown in FIG. This send processing 4
It is assumed that 1 is a system call. In this case, the process issuing system call detection unit 19 detects the system call (transmission process). Then, the process issuing system call detection unit 19 inquires of the backup process execution state management unit 18 about the states of the primary process 11 and the backup process 12, and determines whether both processes are alive (step S2).
1). If at least one of the primary process 11 and the backup process 12 is not alive, the process issuing system call detection unit 19 ends the operation as it is.

On the other hand, if both the primary process 11 and the backup process 12 are alive,
The process issuing system call detection unit 19 uses the backup process execution state management unit 18 to determine whether the backup process 12 is in the execution state (step S22).

If the backup process 12 is not in the execution state, that is, if the backup process 12 is in the stopped state, the process issuing system call detection unit 19 issues an instruction G to the backup process execution state control unit 17. Then, the backup process 12 is transited to the execution state, and the backup process 12 is restarted from the last (most recent) checkpoint based on that time point (step S23). In the example of FIG. 4, the last checkpoint taken in view of the timing of the transmission process 41 is ckp4. In this case, the backup process 12 restarts the process from the checkpoint ckp4.

Process issuing system call detector 19
When the backup process 12 is in the stopped state (NO in step S22), after the backup process 12 is set to the running state and restarted as described above (step S23), the process proceeds to step S24. If the backup process 12 is already in the execution state (YES in step S22), the process issuing system call detection unit 19 proceeds directly to step S24. In step S24, the process issuing system call detecting unit 19 synchronizes the primary process 11 and the backup process 12 and ends the operation. That is, the backup process 12 is synchronized by the inter-process pair communication unit 14 when the send process 41 is performed after the restart.

Next, it is assumed that a failure has occurred in the computer 1a executing the primary process 11 and the primary process 11 has stopped. It is also assumed that the failure of the computer 1a has occurred at the time of fault1 in FIG. This computer 1a failure is caused by the process restart unit 16
Detected by.

When a computer failure is detected, the process restart section 16 determines whether the failed computer is the backup process 12 side computer (step S11). If a failure occurs in the computer on the backup process 12 side, that is, in the computer 1b, the primary process 11 on the computer 1a can be executed, so the process restart unit 16 ends the operation as it is.

On the other hand, when a failure occurs in the computer on the primary process 11 side, that is, the computer 1a, the process restart unit 16 inquires of the backup process execution state management unit 18 about the state of the backup process 12, It is determined whether the backup process 12 is alive and in a stopped state (step S
12, S13). If the backup process 12 is not alive, or if it is alive and is still in the running state, the process restart unit 16 ends the operation as it is.

On the other hand, when the backup process 12 is alive and is in a stopped state, the process restart unit 16 issues an instruction H to the backup process execution state control unit 17 to cause the backup process 12 to be executed. Transition to the running state, at that time, that is, faul
Based on the time point of t1, the backup process 12 is restarted from the last check point (here, ckp4 as apparent from FIG. 3) (step S14). This allows the backup process 12
In the example of FIG. 4, the process restarts from the checkpoint ckp4, which is the time point of restart1.

Generally, in science and technology calculation programs such as CAD and simulation, the input data accompanied by the issuance of a system call is first read, and then the CPU operation is often repeated without issuing the system call. Then, at the end of the CPU operation for a long time, output data is written with issuance of a system call. If the backup process 12 is also in the execution state only during the period in which the system call is issued, even if only the primary process 11 is executed for the period in which the CPU operation is repeated without issuing the system call, it is received from the OS. Unique O to save / restore the service state
It is not necessary to adopt S, and an open system widely used in industry can be used.

Here, it is assumed that the fault tolerant system of FIG. 1 is applied to the execution of the scientific and technological calculation program described above. In this case, for example, as shown in FIG. 8, immediately after the start of the science and technology calculation program, that is, immediately after the start of the process pair 13 including the primary process 11 and the backup process 12, the input data read 8
During the first period in which a system call is issued with 1 or the like, both the primary process 11 and the backup process 12 operate and are in a state (execution state) in which the OS provides service.

During the subsequent CPU operation for a long time, the backup process 12 is in the stopped state after the time of the first checkpoint ckp1, as is apparent from the operation of the previous checkpoint sampling unit 15. For this reason, during the CPU operation for a long time, the backup process 12 does not consume the CPU resource for executing the scientific and technological calculation program. And finally write output data 82
While the system call is issued, etc., the primary process 11 and the backup process 12 operate together again and are in a state of receiving the service from the OS.

In the present embodiment, the process issuance system call detection unit 19 detects the issuance of the system call in the primary process 11 and automatically restarts the backup process 12 from the last checkpoint taken. You are processing a call. In this case, if a long time has passed since the last checkpoint was taken, the processing after the restart will take time.

Therefore, in the present embodiment, checkpoint collection is explicitly instructed in the program to prevent the system call from being executed after a long time has elapsed since the last checkpoint was taken. There is. In the example of FIG. 8, checkpoints ckp5 and ckp10 correspond to this, and are designated to the checkpoint collection unit 15 by the instruction A by the system call from the primary process 11. On the other hand, check point c in FIG.
kp1, ckp2, ckp3, ckp4 and ckp7,
ckp8 and ckp9 are designated by the periodic interrupt B from the timer TM while the backup process 12 is in the stopped state.

By the way, as described above, when the checkpoint collection time comes while the backup process 12 is in the execution state, the backup process 12 is stopped. When a system call is executed in this state, it is necessary to restart the backup process 12 from the last checkpoint taken based on that point. Therefore, it is not preferable in terms of processing efficiency that the checkpoint collection time comes during a period in which the system call needs to be repeated.

Therefore, the period during which the checkpoint collection unit 15 is activated by the periodic interrupt B from the timer TM is limited to the period during which the backup process 12 is in the stopped state, and the period during which the backup process 12 is in the executed state is limited. The interrupt B may be ignored (treated as invalid) by the checkpoint collection unit 15.
In addition, the checkpoint sampling instruction A by the system call from the primary process 11 is always processed as valid by the checkpoint sampling unit 15.

Further, even immediately after the end of the period in which the system call needs to be repeated, that is, immediately after the backup process 12 does not need to be kept in the active state, it is preferable to explicitly instruct the program to take checkpoints. By doing so, the backup process 12 is stopped immediately after the end of the period in which the system call needs to be repeated, and thereafter, it becomes possible to instruct the checkpoint collection by the periodic interrupt B from the timer TM.
As a result, checkpoints can be taken efficiently. In the example of FIG. 8, ckp6 is a checkpoint explicitly designated in the program so as to be immediately after the end of the period in which the system call needs to be repeated.

Similarly, the backup process 12
It is also possible to explicitly state that is to be executed in the program.

As described above, according to the present embodiment, the processing can be continued when a failure occurs, while (1) CP
It is possible to realize a fault-tolerant system by the process pair method applicable to the open system (2) without using twice the U resource.

In the above embodiment, the primary process 11 and the backup process 12 which compose the process pair 13 are operated on different computers 1a and 1b, respectively. However, in a fault-tolerant system that only needs to consider a program failure, the primary process 11 and the backup process 12 may be operated on the same computer. However, if a failure occurs in only one computer on which the primary process 11 and the backup process 12 operate, the processing cannot be continued.

The present invention is not limited to the above embodiment, and can be variously modified at the stage of implementation without departing from the spirit of the invention. Furthermore, the embodiments include inventions at various stages, and various inventions can be extracted by appropriately combining a plurality of disclosed constituent elements. For example, even if some constituent elements are deleted from all the constituent elements shown in the embodiment, the problem described in the section of the problem to be solved by the invention can be solved, and the effect described in the section of the effect of the invention can be solved. If you get
A configuration in which this component is deleted can be extracted as an invention.

[0073]

As described above in detail, according to the present invention, the backup process is executed when the system call is issued during the period immediately after the start of the program and after that when the backup process is stopped. On the other hand, if the backup process is in the execution state when the checkpoint is taken, the backup process is stopped, so that it is possible to continue the process even when a failure occurs, and the CPU resource is reduced. You don't need to do it twice, and you can save and restore the status of services received from the OS without using your own OS, so you can use an open system.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a fault tolerant system according to an embodiment of the present invention.

FIG. 2 is a process pair 13 including a primary process 11 and a backup process 12 in FIG.
The functional block block diagram which shows the relationship with each functional element implement | achieved by the library 20.

FIG. 3 is a state transition diagram showing possible states of a primary process 11 and a backup process 12.

FIG. 4 is a primary process 11 and a backup process 1 forming a process pair 13 in the same embodiment.
2 is a timing chart for explaining the overall operation of 2.

FIG. 5 is a checkpoint sampling unit 1 according to the same embodiment.
6 is a flowchart for explaining the operation of No. 5.

FIG. 6 is a process restart unit 16 in the same embodiment.
6 is a flowchart for explaining the operation of FIG.

FIG. 7 is a flowchart for explaining the operation of the process issuance system call detection unit 19 in the same embodiment.

FIG. 8 is a timing chart for explaining the relationship between the states of the primary process 11 and the backup process 12 and the checkpoint collection timing in the same embodiment.

FIG. 9 is a diagram for explaining a first method of a conventional process pair.

FIG. 10 is a diagram for explaining a second method of a conventional process pair.

[Explanation of symbols]

1a, 1b ... Calculator 2 ... Network 11 ... Primary process 12 ... Backup process 13 ... Process pair 14 ... Process pair communication unit 15 ... Checkpoint collection section 16 ... Process restart section 17 ... Backup process execution status control unit 18 ... Backup process execution status management unit 19 ... Process issuing system call detector 20 ... Library

Claims (9)

[Claims] 1. A process pair execution control method in a fault tolerant system in which a process pair including a primary process and a backup process, which can continue processing even when a failure occurs, is provided. The step of putting both the primary process and the backup process into the running state at the time of start-up, the step of copying the state of the primary process to the backup process at each checkpoint collection time, and the backup process at the checkpoint collection If the backup process is in the stopped state, if the system call is issued from the process pair, and if the backup process is in the stopped state, the backup process is stopped. Process pair execution control method in a fault tolerant system, characterized by comprising the steps of: in the execution state is resumed from the most recently taken checkpoint to click up process. 2. If the backup process is in a stopped state when a failure occurs on the primary process side, the method further comprises the step of restarting the backup process from the checkpoint taken most recently to put it in the running state. The process pair execution control method in a fault tolerant system according to claim 1, wherein 3. The fault tolerant system according to claim 1, further comprising the step of setting a checkpoint sampling time immediately before the need to switch the backup process from the stopped state to the running state. Control method for process pair execution. 4. The process in a fault tolerant system according to claim 3, further comprising the step of setting a checkpoint collection timing immediately after the backup process does not need to be kept in an active state. Pair execution control method. 5. A step of setting a first checkpoint collection time at a timing immediately before it is necessary to switch the backup process from a stopped state to an execution state, and the backup process must be kept in the execution state. A step of setting a second checkpoint sampling time immediately after the first checkpoint sampling time, and a step of setting the second checkpoint sampling time from the first checkpoint sampling time
Setting a third checkpoint collection time at a predetermined time interval during the period in which the backup process is in a stopped state except the period up to the checkpoint collection time. The process pair execution control method in the fault tolerant system according to claim 1. 6. A process pair execution control program for a fault tolerant system, which executes a process pair composed of a primary process and a backup process, capable of continuing processing even when a failure occurs, the computer comprising: When the process pair is started, both the primary process and the backup process are set to the running state, the state of the primary process is copied to the backup process at each checkpoint collection time, and the checkpoint collection is performed. Sometimes if the backup process is in the running state, the step of bringing the primary process into the stopped state, and if the backup process is in the stopped state when a system call is issued from the process pair. Process Pair execution control program for executing a step of the execution state by resuming the backup process from the most recently taken checkpoint. 7. A fault-tolerant system that executes a process pair consisting of a primary process and a backup process, which can continue processing even when a failure occurs, in a communication between the process pair and another process pair. And a backup process execution state control unit for controlling the execution state of the backup process, which executes both the primary process and the backup process when the process pair is activated. State control means and checkpoint collection means for collecting checkpoints by copying the state of the primary process to the backup process each time the checkpoint collection time comes. If the backup process is in the execution state at the time of taking, checkpoint collection means for bringing the backup process into the stopped state by the backup process execution state control means, and a process issuing system call for detecting a system call issued by the process pair When the system call is detected, if the backup process is in a stopped state, the backup process is restarted from the checkpoint most recently taken by the backup process execution state control unit and executed. A fault tolerant system, comprising: a process issuing system call detecting means for bringing the system into a state. 8. If the backup process is in a stopped state at the time of occurrence of a failure on the primary process side, the backup process execution state control means restarts the backup process from the checkpoint most recently taken. 8. The method according to claim 7, further comprising a process restart means for bringing the process into an execution state.
The described fault tolerant system. 9. A backup process execution state management unit for managing the execution state of the backup process, wherein the checkpoint collection unit, the process issuing system call detection unit and the process restart unit are included in the backup process. 9. The fault tolerant system according to claim 8, wherein the state is determined by inquiring the backup process execution state management means.