JPH04241633A - Process managing system - Google Patents

Abstract

PURPOSE:To easily control processes synchronously to each other at the time of start-up and to easily collect information in a system fault. CONSTITUTION:A process managing process 1 manages the operating states of processes with a process management table 2. In the case of receiving a process generating request, the process 1 generates the requested process. In addition, when a process start-up request is received from a host program, the process 1 generates all processes to be generated next by setting the value adding '1' to the number of processes to be generated to a variable semaphore and, when the semaphore value goes to '1' by subtracting the number of generated processes from the set value, the processing of each waiting process is allowed to start by returning the semaphore value to '1'. In the case of receiving a system termination notice or fault notice, the process 1 transmits a preclosing notice to all processes during execution and, when a termination notice is received from each process against the preclosing notice, the process 1 forcibly terminates the process which does not transmit the termination notice after fixed waiting time has elapsed.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a process management method for building a multitasking system using a plurality of processes on the UNIX operating system.

0002

2. Description of the Related Art Conventional process management systems on UNIX do not have the concept of centrally managing multiple processes from generation to termination, and are as follows.

(A) There is no one who centrally manages information on all processes.

(B) Process generation is performed by the process that becomes necessary.

(C) Synchronous control of each process at the time of process startup has been performed by devising the process creation order among the processes.

(D) Processing to terminate all processes at the time of system failure or termination was forcibly performed.

[0007]

SUMMARY OF THE INVENTION The conventional process management system described above has the following problems.

(A) Since information on all processes is not managed centrally, program logic is wasted. Furthermore, it becomes difficult to collect information when a system failure occurs.

(B) Since process generation is performed and managed by the process that becomes necessary, process management information is distributed.

(C) Synchronous control of each process at the time of process startup relies on ingenuity between each process, which makes it difficult to cope with the increase in the number of processes and the complexity of processing.

(D) At the time of system failure or termination, all process termination processing is forcibly performed without pre-blocking, so each process is terminated even in the middle of processing, and post-processing cannot be performed. Otherwise, there is a risk of processing inconsistency occurring at the next restart.

0012

[Means for solving the problem] The first invention is based on UNIX
The above process management method for a system that uses multiple processes and operates in multitasking is equipped with a process management table that is loaded from the disk to memory at system startup and centrally manages the operation information of the processes. When a generation request is received, the requested process is generated and the operating status of the process in the process management table is set to ``generating.'' When a process startup request is received from the upper program, the process is set in a semaphore, which is a variable prepared in advance. Set a value that adds 1 to the number of processes to be generated in the management table, then generate all the processes to be generated, and the semaphore value is decremented by 1 by the number of processes for each of the generated processes. When the semaphore value becomes 1, the semaphore value is reset to 0, and the processes of the processes that were waiting until the semaphore value becomes 0 are started all at once, and when a system termination notification or failure notification is received, all the processes that are running are restarted. When a post-processing completion notification in preparation for the next restart is received from each process that has received the pre-blockage notification, the process that does not send the termination notification after waiting for a certain period of time is It is characterized by having a process management process that is forcibly terminated.

[0013]

Embodiments Next, embodiments of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram showing an embodiment of the process management system of the present invention. As shown in FIG. 1, the process management table 2 exists on a disk and is loaded into memory at system startup. Therefore, process information can be easily registered, changed, etc. using an editor. The process management process 1 uses this process management table 2 to centrally manage all processes.

FIG. 2 is an operational explanatory diagram showing one embodiment of the process generation method. As shown in Figure 2, process A (3
) generates process B(4), requests process management process 1 to generate process B(4), and process management process 1 generates process B(4) and manages the status of all processes.

FIG. 3 is an explanatory diagram of the operation of synchronization processing at process startup. As shown in FIG. 3, at process startup, the process management process 1 first sets the semaphore value to the number of generated processes+1. Next, create all necessary processes at system startup. Each created process performs initial settings and sets the semaphore value to -1.
Wait until it becomes 0. If all production processes reduce the semaphore value, the value becomes 1. When the above processing is completed, the process management process 1 returns the semaphore value to 0, and the waiting processes start processing all at once.

FIG. 4 is an explanatory diagram of the operation of all process termination processing. As shown in FIG. 4, when the process management process 1 receives a system termination notification or a failure notification, it sends a pre-blockage notification to all running processes. Each process that receives the pre-blockage notification performs post-processing in preparation for the next restart, and sends an end notification. The process management process 1 forcibly terminates a process that does not send a termination notification after waiting for a certain period of time.

[0018]

[Effects of the Invention] As explained above, the present invention enables synchronized control of each process at process startup by centrally managing multiple processes, and facilitates information collection in the event of a system failure. By enabling post-processing in each process when a system failure occurs, it is possible to prevent processing inconsistencies from occurring at the next restart.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of the process management method of the present invention.

FIG. 2 is an operation explanatory diagram showing an example of process generation processing according to the present embodiment.

FIG. 3 is an explanatory diagram of the operation of synchronization processing at process startup in this embodiment.

FIG. 4 is an explanatory diagram of the operation of all process termination processing in this embodiment.

[Explanation of symbols]

1 Process management process 2 Process management table 3 Process A 4 Process B

Claims (1)

[Claims] 1. In a process management method for a system that uses multiple processes and operates in multitasking on UNIX, a process management table that is loaded from a disk to memory at system startup and centrally manages operation information of the processes is provided. When a process creation request is received from the process, the requested process is created, and the operating state of the process in the process management table is set to "generating," and when a process startup request is received from the higher-level program, the previously prepared variable is created. Set a value equal to the number of processes to be generated in the process management table plus 1 to the semaphore, then generate all the processes to be generated, and set the semaphore value to each of the processes by each of the generated processes. When the number of processes is subtracted by 1 and becomes 1, the semaphore value is reset to 0, and the processing of each process that was waiting until the semaphore value becomes 0 is started all at once, and when a system termination notification or failure notification is received, all When a pre-blocking notification is sent to the process that is currently being executed, and a completion notification of post-processing in preparation for the next restart is received from each process that received the pre-blocking notification, the completion notification is sent after waiting for a certain period of time. 1. A process management method comprising: a process management process for forcibly terminating said process that does not send a message.