JPH05342058A - Process abnormality detection system - Google Patents

Abstract

PURPOSE:To provide the high reliability of a system by performing processing while detecting the abnormal end of processes. CONSTITUTION:An application program (process) group 3 is constituted so as to be activated from a managing process 1 and further, a monitor process 2 activated from the managing process 1 is constituted so as to monitor the managing process 1 and to easily detect the abnormal end of the managing process 1 itself as well. Thus, since the abnormal end of processes can be detected and processing can be performed even after the abnormal end, the high-reliability system can be provided.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a system for detecting abnormal termination of a process in, for example, UNIX system (UNIX is a registered trademark of AT & T Corporation).

[0002]

2. Description of the Related Art FIG. 3 is a diagram showing a process configuration when an application program is created in a UNIX system. The application process group 4 is a set of processes activated by the system. Next, the operation will be described. When the system detects an error for each application process group 4,
The system interrupts the process group 4 by a forced termination (in UNIX systems, a sig called SIGKILL).
nal) is generated. The process that received the interrupt for forced termination cannot catch the interrupt, so it is terminated in the form of abnormal termination. Abnormal termination of this process cannot be detected by the application, so processing when the process is terminated cannot be performed.Therefore, if the system is operated as it is, malfunction will occur in the system constructed by the application. There is a risk of causing this.

[0003]

In the conventional system, when a process is forcibly terminated, the abnormal termination of the process cannot be detected. Therefore, the abnormal processing at the end of the process cannot be performed. There was a problem that normal operation of the system could not be guaranteed. The present invention has been made in order to solve the above problems, and is a process abnormality detection method capable of performing process management capable of detecting abnormal termination of a process and improving reliability with respect to abnormal process termination. Is intended to provide.

[0004]

In the process abnormality detection method according to the present invention, for example, a management process activates each application program (process) to detect an abnormal termination, performs abnormal processing, and abnormally terminates the management process. It is equipped with a monitoring process that can detect even the following, and has the following elements. (A) a management process having the following elements, (a1) a starting means for starting the process, (a2) a detecting means for detecting an abnormal end of the process started by the starting means,
(A3) Post-processing means for performing processing after the occurrence of an abnormality in the process detected by the detection means, and (b) a monitoring process which is started by the starting means and monitors the management process.

[0005]

In the process abnormality detecting method according to the present invention, by starting the process by the management process, the abnormal end of each process can be detected, and the processing after the abnormal end can be performed. Further, the abnormal termination of the management process can be detected by the monitoring process. As a result, abnormal termination of a process can be immediately detected and abnormal processing can be performed, so that a highly reliable system can be obtained.

[0006]

【Example】

Example 1. FIG. 1 is a diagram showing an embodiment of a process configuration diagram according to the present invention. 1 is a management process, 2 is a monitoring process for monitoring abnormal termination of the management process 1, and 3 is created by each application. It is a group of processes.

The management process 1 activates the monitoring process 2 when the system is activated. In general,
In the UNIX system, a process activated by a certain process has a parent-child process relationship. In this case, the started process becomes the parent process,
The started process becomes a child process. Therefore, FIG.
In, all the processes started by the management process 1 are parent processes. That is, the processes A and B in the process group 3 started by the management process 1 are child processes having the management process 1 as a parent. The monitoring process 2 started by the management process 1 when the system is started is also a child process having the management process 1 as a parent.

Next, the operation will be described with reference to FIG. FIG. 2 shows the processing of each process shown in FIG. 1 over time. The dotted line represents the operation from the parent process to the child process or from the child process to the parent process. First, the management process 1 activates the system, and first activates the monitoring process 3. After that, the monitoring process 3 periodically uses a system call called getppid, which can confirm the existence of the parent process, and confirms that the management process 1, which is the parent process, has not terminated abnormally.

Next, the management process activates process A and subsequently process B. Here, assuming an example in which the process A abnormally ends, the process B performs the update processing only when the process A ends normally. The process A notifies the management process 1 of the abnormal termination. The management process 1 that has detected the abnormal termination notifies the process B to perform post-processing. The process B that has received the notification performs post-processing (for example, processing such as returning the update result to the original), completes normally, and notifies the management process 1 to that effect. Meanwhile, the monitoring process 2 confirms that the management process 1 is operating.

When a process in the application process group 3 activated by the management process 1 is forcibly terminated in this way, the child process always notifies the parent process of its termination at the time of termination. Abnormal termination of each process in the application process group 3 is always notified to the management process 1, and the abnormal termination can be detected. Since the abnormal end can be detected, the processing after the abnormal end occurs can be performed at a proper timing. Further, when the management process 1 ends abnormally, the monitoring process 2 can detect the abnormal end of the management process 1.

Embodiment 2. In the first embodiment, the UN
Although the case of the IX system has been described as an example, the present invention is not limited to the UNIX system, but can be applied to any system in which a process can detect abnormal termination of another process.

[0012]

As described above, according to the present invention, since the process is excited by the management process, the abnormal termination of the application process can be detected, and the abnormal termination of the management process itself can be detected. Since it is configured so that it can be easily detected, it is possible to easily establish high reliability of the system against abnormal termination.

[Brief description of drawings]

FIG. 1 is a diagram showing a process configuration according to a first embodiment of the present invention.

FIG. 2 is a diagram showing the operation of the first embodiment of the present invention.

FIG. 3 is a diagram showing a conventional process configuration.

[Explanation of symbols]

 1 Management process 2 Monitoring process 3 Application process group

Claims (1)

[Claims] 1. A process abnormality detection method having the following elements: (a) a management process having the following elements, (a1) an activation means for activating a process, (a2) an abnormal termination of a process activated by the activation means. Detection means to detect,
(A3) Post-processing means for performing processing after occurrence of an abnormality in the process detected by the detection means, (b) Monitoring process started by the starting means and monitoring the management process.