JPH02158843A - Automatic restart processing system for process - Google Patents

Abstract

PURPOSE:To omit the preparation of a waiting process by automatically restarting a process generating abnormality in a new address space. CONSTITUTION:The subject system is provided with a recovering information storage part 18 for storing at lest process identifying information relating to a started process, its address space information and the process identifying information of a starting process, an abnormality detecting part 21 for detecting the abnormality of a process and informing the generation of the abnormality to an address space starting source in which the process is driven, a recovery processing part 15 for executing recovery processing including resource recovery based upon information stored in the storage part 18 in accordance with the information of abnormality generation, and a restart processing part 16 for restarting the process driving the address space generating the abnormality in a new address space. When abnormality is generated in a system program or a general application program, the program is restarted in a new executing environment. Thereby, the preparation of a preceding waiting process can be omitted.

Description

[Detailed Description of the Invention] [Summary] A method for automatically restarting a process that performs automatic recovery when an intermittent failure occurs in a computer system that performs process control.
By automatically restarting into the new address space,
The purpose is to achieve fault tolerance.

Process identification information about the launched process.

a recovery information storage unit that stores at least the address space information and process identification information of the startup source process; and an abnormality detection unit that detects an abnormality in the process and notifies the startup source of the address space in which the process operates of the occurrence of the abnormality. Based on the information recorded in the recovery information storage unit tα by the error occurrence notification, the recovery processing unit performs recovery processing including resource recovery, and the process operating under the address space where the error occurred is moved to a new address. and a restart processing unit that restarts within the space.

[Industrial application field]

The present invention provides a process automatic restart processing computer system that performs automatic recovery when an intermittent failure occurs in a computer system that performs process control. A no-down time system is required. Therefore, the operating system (O3) requires a recovery function in the event of an error, and the system must be made fault tolerant to prevent the system from going down even if an error occurs. ing.

[Conventional technology] FIG. 4 shows an example of the conventional method.

In the conventional no-down computer system.

In addition to multiplexing the hardware, redundancy was also provided in the software.

By switching to the standby system in the event of a hardware failure or software abnormality.

We tried to make the system fault tolerant.

FIG. 4 shows an example of switching when a software abnormality occurs. In preparation for a failure, for the current process 53,
The operating environment of the standby process 54 having the same processing function as the active process 53 is prepared in advance.

In the failure processing unit 50 of the operating system, when the abnormality detection unit 51 detects that an abnormality has occurred in the active process 53, it calls the active/standby switching unit 52, and the active/standby switching unit 52 switches between the active process 53 and the active process 53. Thereafter, the standby process 54 is operated as the active process.

[Problem to be solved by the invention]

A process is a unit of processing that uses computer resources such as the CPU and memory, but in conventional methods, a standby process must be prepared in advance for each process, which can be considered useless under normal conditions. Therefore, there was a problem in that the amount of memory used for the operating environment for the standby process increased. Another problem is that the processing overhead for preparing the standby process is large.

The present invention aims to solve the above problems and automatically restarts a process in which an error has occurred in a new address space, eliminating the need to prepare a standby process and preventing faults.
The purpose is to make it tolerant.

[Means to solve the problem]

FIG. 1 shows an example of the configuration of the present invention.

In FIG. 1, 10 is a computer system equipped with a CPU, memory, etc., 11 and 12 are processes that are processing units that use computer resources, 13 and 14 are process control units that control process startup, etc., and 15 is a recovery unit. A processing unit, 16 is a restart processing unit, 17 is a session control unit that controls a session which is a unit of processing requests from a user, 18 is a recovery information storage unit, 19 is a non-volatile memory, and 20 is the core of the operating system. 21 is an abnormality detection unit, 22 is a space management unit that manages address space,
Reference numeral 23 represents a space management table that stores address space management information, and AO to A2 represent address spaces that are individual virtual storage spaces.

The session control unit 17 responds to a processing request from a user or a processing request from a procedure consisting of a command string.
Open a session. In that session, 9 acquires address space A1, for example, and starts process 11. The process 11 is a processing unit that uses computer resources, and forms an execution environment for a program. Multiple processes may exist in one address space. Also.

Acquire another address space A2 as necessary.

It is also possible to start other processes 12 that operate in that space.

The recovery information storage unit 18 managed by the session control unit 17 is stored for each activated process 11.12.

Message destination information used for processing requests, etc.

a process identifier that uniquely identifies a process in the system; an address space identifier that uniquely identifies an address space;
It stores information such as the process identifier of the startup source. When a certain process starts another process, these pieces of information are notified to the session control unit 17, and the session control unit 17 stores the contents in the recovery information storage unit 18 and saves them to the nonvolatile memory 19.

The nonvolatile memory 19 is a part of the hardware, and is designed to guarantee the contents of the recovery information storage section 18 even if an abnormality occurs in a process or address space.

The abnormality detection section 21 provided in the kernel section 20 is as follows.

It detects program check interrupts and other process abnormalities, and when an abnormality is detected, it notifies the activation source of the address space in which the process operates of the occurrence of the abnormality. Note that information regarding the address space activation source is provided at the time of space creation.

Since it is registered as a message destination in the 2-space management table 23 by the space management unit 22,
Refer to it when notifying the occurrence of an abnormality.

The process control unit 13 etc. that control the process B.

As a program, it can operate in all address spaces, including 1 recovery processing unit 15. It has a restart processing section 16.

When the recovery processing unit 15 receives an error occurrence notification from the error detection unit 21, it requests the session control unit 17 for the contents of the related recovery information storage unit 18, and recovers the resources of the process 12 started by its own process 11. By doing so, erasing the address space A2, and organizing the execution environment, the system state is restored to a consistent state.

Next, the restart processing unit 16 requests the space management unit 22 to create a new address space, and restarts the process 12 in that address space.

[Effect]

If redundancy is provided in the software and a standby process is always prepared for each active process, switching between the active and standby processes will be performed in the event of a failure.

Processing using the new process can be continued in a short period of time, which can be called instantaneous. However, in a normal process, the time required for switching.

A very short time is not required, and in the case of most programs, when a lit problem occurs, it is often enough to simply re-execute the program.

In the present invention, the recovery processing unit 15 guarantees system atomicity against intermittent faults such as bugs that occur at any timing, and the restart processing unit 16 allows the process in which the abnormality has occurred to be restarted in a new execution environment. Since it is started with , there is no need to prepare a standby process.

〔Example] 2 and 3 show two embodiments of the invention.

In particular, FIG. 2 shows the state when the process is started, and FIG. 3 shows the state when an abnormality occurs.

In FIGS. 2 and 3, the same symbols as in FIG. 1 correspond to those shown in FIG. 1, and 30.31 is a processor;
32 and 33 represent a supervisor corresponding to the kernel unit 20 shown in FIG. 1, 40 represents a restarted process, and 41 represents a process control unit.

The flow of processing when a certain process 11 starts up another process 12 will be explained according to ① to ② shown in FIG. 2.

(2) When the process 11 starts the process 12, it specifies automatic restart with a parameter and issues a process start macro instruction.

■ By issuing this macro instruction, the process control unit 13
The process startup section requests the supervisor 32 to create an address space.

■ The supervisor 32 creates an address space.

Returns the message destination and address space identifier for the new space to the requester.

■ The process starting section of the process control section 14 operates within the newly opened address space A2.

The supervisor 32 is requested to create a program execution body, that is, to acquire CPU resources.

■ The process startup unit of the process control unit 13 uses the information returned from the supervisor 32 (destination of the new address space,
new address space identifier) and your process identifier,
The process identifier of the new process 12 is set to the session control unit 1.
Notify 7.

■ The session control unit 17 registers the notified information in the recovery information storage unit 18, and stores the information in order to prevent the information from being lost when its own address space AO crashes. It is saved to the nonvolatile memory 19.

Nao, Seshigon control unit 17 is also a type of process, but 9
In this embodiment, the session control unit 17 is provided with a standby process so that it can quickly deal with the occurrence of a failure.

In the state shown in FIG. 2, the flow of processing when an abnormality occurs in the address space A2 is as shown in FIG. 3, for example. Hereinafter, explanation will be made according to (1) to (4) shown in FIG.

(2) Assume that the abnormality is a program check in address space A2. When a program check interrupt is generated from hardware, the supervisor 32 detects it.

■ When the supervisor 32 receives a space creation request, it remembers the information about which space created which space, so it checks the startup source of the crashed address space A2 and notifies it of the occurrence of an abnormality. .

■ The recovery processing unit 15 in the address space AI that is the startup source recovers the memory resources, CPU resources, etc. allocated to the processes that were started from within the crashed space, if they remain under the crashed space. Therefore, the session control unit 17 is requested to notify the process information, and the corresponding recovery information storage unit 1
Obtain the contents of 8.

(2) Next, the 2 recovery processing unit 15 learns the address space A2 to be deleted from the acquired information, requests the supervisor 32 to delete it, and recovers the address space A2.

(2) The restart processing unit 16 requests the supervisor 32 to create an address space A3 in order to start the process 12 as a new process 40.

(2) In response to the request, the supervisor 32 creates a new address space A3 and returns its destination and address space identifier.

■ The restart processing unit 16 re-enters the destination of the returned new address space, its address space identifier, its own process identifier, and the new process identifier.

The session control unit 17 is notified. The session control unit 17 registers the information in the recovery information storage unit 18.

(2) The process activation unit of the program control unit 41 in the address space A3 requests the supervisor 32 to create a process execution body (to obtain a CPU percentage).

With the above, the processing performed in process 12 is completed.

This can be taken over by the process 40 operating in the new space A3, and the state before the abnormality can be restored.

〔Effect of the invention〕

As explained above, according to the present invention, when an abnormality occurs in a system program, general application program, etc., the program is started again under a new execution environment.

There is no need to prepare a standby process in advance.
Memory resources and CPU are reduced by not creating a process in advance.
It becomes possible to save resources.

[Brief explanation of the drawing]

FIG. 1 shows a configuration example of the present invention. FIGS. 2 and 3 show embodiments of the present invention. FIG. 4 shows an example of the conventional method. In the figure, 10 is a computer system, 11.12 is a process,
13. 14 is a process control unit, 15 is a recovery processing unit, 16
17 is a restart processing unit, 17 is a session control unit, 18 is a recovery information storage unit, 19 is a non-volatile memory, 20 is a kernel unit,
21 is an abnormality detection section. 22 represents a space management unit, 23 represents a space management table, and AO to A2 represent address spaces.

Claims (1)

[Claims] In a computer system in which a process, which is a processing unit that uses computer resources, has a processing function to launch other processes operating in a new address space, process identification information regarding the launched process and its address. a recovery information storage unit (18) that stores at least space information and process identification information of the startup process; and an abnormality detection unit that detects an abnormality in the process and notifies the startup source of the address space in which the process operates of the occurrence of the abnormality. (2
1), a recovery processing unit (15) that performs recovery processing including resource recovery based on the information stored in the recovery information storage unit upon notification of the occurrence of an error; and a recovery processing unit (15) that performs recovery processing including resource recovery based on the information stored in the recovery information storage unit; 1. A process automatic restart processing method, comprising: a restart processing unit (16) that restarts a process that has been reactivated in a new address space.