JPH07271611A - Automatic process restarting and processing system - Google Patents

Abstract

PURPOSE:To prevent mutual interference resulting from the presence of a continued process and restart the processing by forcibly ending also its subordinate process when a main processing constituting a job becomes abnormal. CONSTITUTION:This system has an abnormality detection part 21 which detects abnormality of at least one process, a collecting processing part 101 which forcibly ends remaining normal processes and collects information on the processes, and a restart processing part 16 which restarts at least a process that becomes a main process among the processes newly in a memory space. Consequently, if the main process having subordinate processes becomes abnormal, the subordinate processes which are normal are forcibly ended by the collecting processing part 101, so that only the subordinate processes which are not abnormal, operate to prevent discordance with the main process.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a process automatic restart process for automatically repairing a computer system for process control when an intermittent failure occurs.

[0002]

2. Description of the Related Art The importance of computer systems tends to increase year by year, and there is a demand for 24-hour no-down systems. Therefore, operating system (OS)
Requires a function to confirm recovery when an error occurs, and to prevent the system from going down even if an error occurs, the system's fault tolerance (Fault Toreran)
t) is becoming necessary.

In view of such a point, the present inventor prepares a standby process by automatically restarting a process in which an abnormality has occurred into a new address space, as disclosed in Japanese Patent Laid-Open No. 2-158843. We propose a technology that eliminates the need for.

The outline of this prior art is briefly described as shown in FIG. That is, when the abnormality detecting unit in the job scheduler detects an abnormality in the main process, the restart processing unit is instructed to restart and the job is restarted in the newly established address space.

[0005]

However, in the above-mentioned conventional technique, since a job constituted by a single process is targeted, it is not possible to cope with a failure occurring in a job constituted by a plurality of processes. That is, FIG.
As shown in 0, there is a problem when a job is composed of a main process and a plurality of processes.

In this case, when a failure occurs in the sub-process, no problem occurs because the sub-process having the abnormality can be recovered by using the normally operating main process, but the main process itself. If an error is detected by the error detection unit, the main process will be restarted while the sub-process is still operating, and the main process and the sub-process will be restarted. It could be inconsistent and dangerous.

The present invention has been made in view of the above problems, and an object of the present invention is to create a mutual process caused by the existence of continuous processes in a system including a plurality of processes in which a plurality of jobs do not share a memory. It is to propose a system capable of preventing interference and restarting a process.

[0008]

The first means of the present invention is to:
As shown in FIG. 1, which is a principle diagram, in a computer system including a plurality of processes in which jobs do not share memory with each other, an abnormality detection unit (21) for detecting an abnormality in at least one process, and a process constituting the job Of these, a recovery processing unit (101) for forcibly ending the remaining process in which no abnormality has occurred and recovering the process information
And a restart processing section (16) for restarting at least a main process among the above processes on a new memory space.

The second means is, as shown in FIG. 2, that the restart processing unit (16) determines whether or not the restarted process is forcibly terminated.
2) is provided.

[0010]

According to the first means of the present invention, when an abnormality occurs in the main process having the sub-process, the recovery processing unit forcibly terminates the sub-process in which no abnormality has occurred. It is possible to prevent only the sub-process that does not operate and cause a conflict with the main process.

According to the second means, after the main process is restarted in the new memory space, the job itself is provided with a function capable of aborting the restart, so that the job can be executed even when a failure occurs. Flexible response by the program is possible.

[0012]

First Embodiment The present invention will be described below with reference to the drawings. [System Configuration of the Present Embodiment] FIG. 3 shows the system configuration of a computer which is an embodiment of the present invention.

In FIG. 3, 10 is a computer system having a CPU and memory, and 11 and 12 are processes which are processing units that use computer resources. Of these, a child process group is a main process and a plurality of sub-processes. It consists of and.

Reference numeral 14 is a process control unit for controlling process activation, 15 is a recovery processing unit, 16 is a restart processing unit, 1
Reference numeral 7 denotes a session control unit that controls a session, which is a unit of a processing request from a user, 18 denotes a recovery information storage unit, 1
9 is a non-volatile memory, 20 is a kernel part which is the core of the OS,
Reference numeral 21 is an abnormality detection unit, 22 is a space management unit that manages an address space, 23 is a space management table that stores management information of the address space, and A0 to A2 are address spaces that are individual virtual storage spaces. Further, 101 generated in the address space A1 represents a collection processing unit. [Outline of System Operation State] In FIG. 3, the session control unit 17 opens a session based on a processing request from a user or a processing request from a procedure including a command sequence. In this session, for example, the address space A is acquired and the process 11 is activated. Process 1
1 is a processing unit (job) that uses computer resources,
It forms the execution environment of the program. It is also possible to acquire other address spaces A2 to An and start other processes 12, ... (A plurality of main processes and sub-processes are possible) operating in that space, if necessary.

The recovery information storage unit 18 managed by the session control unit 17 has destination information of a message used for a processing request for each of the started processes 11 and 12, a process identifier for uniquely identifying a process in the system, It stores information such as an address space identifier that uniquely identifies the address space, and a process identifier of the starting source. When a process starts another process, the session control unit 17 is notified of these pieces of information, and the session control unit 1
7 has a function of storing this content in the recovery information storage unit 18 and saving it in the non-volatile memory 19.

The non-volatile memory 19 is a part of hardware, and guarantees the contents of the recovery information storage unit 18 even when an abnormality occurs in a process or an address space.

Anomaly detection section 2 provided in kernel section 20
1 is for detecting a program check interrupt or other abnormalities of the program, and when an abnormality is detected,
It has a function of notifying the address space A1 of the occurrence of an abnormality.

The process control section 14 for controlling the process has a recovery processing section 15, a restart processing section 16 and a recovery processing section 101. Upon receiving the abnormality occurrence notification from the abnormality detection unit 21, the recovery processing unit 15 requests the session control unit 17 for the related contents of the recovery information storage unit 18, and collects the resources of the process 12 started by the own process 11. Then, the address spaces A2 to An are extinguished and the execution environment is organized to restore the system state to a consistent state.

The recovery processing unit 101 has a function of forcibly ending the sub-process group when a sub-process exists in the abnormal process (main process).

Further, the restart processing unit 16 requests the space management unit 22 to create a new address space after the main process and the sub process are forcibly terminated, and restarts the process 12 in the address space. It has a function to do. [Normal Operating State of System] Next, the operating state when the system is normally started will be described with reference to FIG.

4 and 5, the same reference numerals as those in FIG. 3 have the same functions. Reference numerals 30 and 31 are processors, 32 and 33 are supervisors corresponding to the kernel unit 20 shown in FIG. 3, 40 is a restarted process, and 41 is a process control unit.

Here, from address space A1 to process 1
The flow of the process when starting 2 will be described according to (1) to (6) (indicated by circled numbers in the figure). (1) When the process 11 is started, the address space A1 issues a process start macro command with automatic restart set by a parameter. (2) By issuing this macro instruction, the process control unit 1
The process starting unit 3 requests the supervisor 32 to create an address space. (3) The supervisor 32 creates an address space, and returns the destination of the message to the new space and the address space identifier to the request source. (4) In the newly opened address space Ax, the process activation unit of the process control unit 13 operates, and the supervisor 32 creates a program execution unit, that is, a CPU.
Request the acquisition of resources. (5) In the process control unit 13, the process control unit 13 receives the information (destination of the new address space, new address space identifier) returned from the supervisor 32, its own process identifier, and the process identifier of the new process 12 in the session control unit. Notify 17. (6) The session control unit 17 registers the notified information in the recovery information storage unit 18 so that when the address space A0 of its own is crashed, the information is not lost. The information is saved in the nonvolatile memory 19.

Although the session control unit 17 is also a type of process, in the present embodiment, the session control unit 17 is provided with a standby process so that a failure can be quickly dealt with.

In the state shown in FIG. 4, the flow of processing when an abnormality occurs in the address space A2 is as shown in FIG. 5 below. [Operating state at the time of failure] Follow the procedure below (1)-
(8) (Indicated by circled numbers in the figure). (1) Assume that the abnormality is a program check (main process) in address space A2. If there is a program check interrupt from the hardware, the supervisor will detect this anomaly. (2) When the supervisor 32 receives a space creation request, it stores information about which space created which space. Therefore, the originator of the crashed address space A2 is checked to see if an error has occurred. Notice. (2) ′ The recovery processing unit 101 requests the supervisor 32 to recover the space of the sub process. This space recovery is performed for each sub-process, and the recovery processing unit 1
A space end notification is issued to 01 from the supervisor 32. This space recovery processing is repeated by the number (n) of sub-processes expanded in the address spaces A2 to An. (3) The recovery processing unit 15 in the activation-source address space A1 requests the session control unit 17 to notify process information, and acquires the contents of the corresponding recovery information storage unit 18. (4) Next, the recovery processing unit 15 knows the address spaces A2 to An to be deleted based on the acquired information, requests the supervisor to delete the addresses, and collects the address spaces A2 to An. (5) The restart processing unit 16 requests the supervisor 32 to create the address space A3 in order to start the process 11 as a new process 40. (6) The supervisor 32 creates a new address space A3 in response to the request, and returns its destination and address space identifier. (7) The restart processing unit 16 again notifies the session control unit 17 of 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 registers the information in the recovery information storage unit 18. (8) Program control unit 41 in address space A3
The process activation unit of (1) requests the supervisor 32 to create a process execution body (acquires CPU).

Due to the abnormality, the process performed in the process 12 can be taken over by the process 40 operating in the new space A3, and the state including the sub-process can be restored to the state at the time of the abnormality.

[0026]

Second Embodiment FIG. 6 shows the configuration of the second embodiment. The figure is almost the same as FIG.
3 is different in that a restart determination unit 102 is provided in addition to the restart processing unit 16.

The restart determination unit 102 has a function of determining whether or not the restart operation is terminated based on the notification from the restarted main process after the process (8) of the first embodiment. There is.

That is, in the process (8), after the main process is restarted in the new memory space A3, when the restart abort declaration is notified from the main process, this is accepted and the main process abort control is performed thereafter. Does not restart.

FIG. 7 conceptually shows this.
That is, the job scheduler 701 activates the main process 703 and the sub-processes 704 and 705 in the job 702. If an abnormality is detected by the main process 703 during this process, the recovery process unit 101 recovers the sub processes 704 and 705. When the end of the entire job 702 is recognized, the job 702 ′ is secured in a new memory space (A3 in FIG. 6) and the main process 703 ′ is restarted.
When the restart determination unit 102 receives the restart termination declaration from the main process 703 ', the main process 703' is terminated and the subsequent restart operation is not performed.

FIG. 8 shows the processing procedure for restarting the job (main process). In the figure, a job 70
When the end of the main process 703 constituting the second process is detected (801), the end state of the main process 703 is saved (802). Next, it is determined whether or not there is another main process that constitutes the job (803), and if there is, the user is notified of the end of the job (805). If they do not exist, the sub-processes 704 and 705 other than the main process 703 forming the job are forcibly terminated (80
4).

Next, it is judged whether or not there is a termination instruction from the job (806), but this branches to the positive branch when the main process 703 'is restarted.
If there is no termination instruction from the job, the main process 703 'constituting the job is restarted using the saved end status of the main process 703 as the startup information (807).

As described above, according to the second embodiment, by incorporating the logic for aborting the restart in the job itself,
The job program can flexibly support the system and contribute to the automatic operation of the system.

[0033]

According to the present invention, when an abnormality occurs in the main process that constitutes a job, the sub-process is forcibly terminated, so that a part of the program in which the abnormality occurs continues to operate. It is possible to prevent malfunction due to mutual interference.

[Brief description of drawings]

FIG. 1 is a first principle diagram of the present invention.

FIG. 2 is a second principle diagram of the present invention.

FIG. 3 is an explanatory diagram showing a system configuration of a computer according to the first embodiment of the present invention.

FIG. 4 is an explanatory diagram showing the operation of the computer according to the first embodiment.

FIG. 5 is an explanatory diagram showing the operation of the computer according to the first embodiment.

FIG. 6 is an explanatory diagram showing the operation of the computer according to the second embodiment.

FIG. 7 is an explanatory diagram conceptually showing the operation of the second embodiment.

FIG. 8 is a flowchart showing a processing procedure in restarting a job according to the second embodiment.

FIG. 9 is an explanatory diagram of a conventional technique.

FIG. 10 is an explanatory diagram of a conventional technique.

[Explanation of symbols]

 10 ... Computer system 11, 12 ... Process 13, 14 ... Process control unit 15 ... Recovery processing unit 16 ... Reboot processing unit 17 ... Session control unit 18 ... Recovery information storage unit 19 ... Nonvolatile Memory 20 Kernel unit 21 Abnormality detection unit 22 Space management unit 23 Space management table 101 Recovery processing unit 102 Restart determination unit 701 Job scheduler 702 Job 703 Main Process 704, 705 ... Sub process

─────────────────────────────────────────────────── ───

[Procedure amendment]

[Submission date] April 5, 1994

[Procedure Amendment 1]

[Document name to be corrected] Drawing

[Name of item to be corrected] Figure 3

[Correction method] Change

[Correction content]

[Figure 3]

[Procedure Amendment 2]

[Document name to be corrected] Drawing

[Name of item to be corrected] Fig. 4

[Correction method] Change

[Correction content]

[Figure 4]

[Procedure 3]

[Document name to be corrected] Drawing

[Name of item to be corrected] Figure 5

[Correction method] Change

[Correction content]

[Figure 5]

[Procedure amendment 4]

[Document name to be corrected] Drawing

[Name of item to be corrected] Figure 6

[Correction method] Change

[Correction content]

[Figure 6]

Claims (2)

[Claims] 1. A computer system comprising a plurality of processes in which jobs do not share memory with each other. An abnormality detection unit (21) for detecting an abnormality of at least one process, and an abnormality occurs among processes constituting a job. A recovery processing unit (101) for forcibly terminating the remaining unprocessed process and recovering process information, and a restart processing unit for restarting at least the main process of the processes on a new memory space. (16) A process automatic restart processing method including: 2. The restart processing section (16) is provided with a restart determination section (102) for determining whether or not the process restarted in the above is forcibly terminated. The process automatic restart processing method according to claim 1.