JP2008186208A - Cpu load reducing method and monitoring/control server - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To solve the problem that an important program can not be quickly executed, that is, a new application can not be started because an application with a low priority is active when many undesired programs are executed. <P>SOLUTION: The load of a CPU is monitored, and when the load of the CPU exceeds a threshold, an application with a low priority is terminated. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a technique for reducing a load on a CPU in a computer that activates / operates a plurality of applications.

  In a system that can execute multiple programs, by limiting the number of programs to be executed or monitoring the execution status of the programs, the application programs can be processed with high reliability and a stable operation management function of the computer system is provided. There is a system (for example, Patent Document 1).

JP-A-8-328880

  In Patent Document 1, since the number of programs that can be executed is limited, in order to execute a new program, it is necessary to terminate the program being executed. Therefore, when many unnecessary programs are executed, there is a problem that an important program cannot be executed immediately. That is, a situation has arisen in which a new application cannot be activated because an application with a low priority is activated.

  In order to solve the above problem, the present invention monitors the CPU load, and terminates an application with a lower priority when the CPU load exceeds a threshold. The lower priority application includes an application having a lower priority than the predetermined priority.

  According to the present invention, it is possible to reduce the load on the CPU, provide a more stable operation state for a high priority application, and remove a failure to start a new application.

Hereinafter, embodiments of the present invention will be described in detail.
FIG. 1 shows an overall configuration diagram of a system that restricts the operation from the monitoring / control server in accordance with the operation status of the application and the login status of the user.

  The main components of this system are the user terminals (70, 80, 90) that users operate to perform business processing, the network (60) that connects each component of this system, and application programs for business processing. It is an application server (30) to be operated, and a monitoring / control server that is the key of the system of the present invention.

  Here, there may be a plurality of application servers, and the application server and the monitoring / control server may be integrated. Moreover, although the Example has described the example in which a some client uses an application, the case where one client uses a some application may be sufficient. In addition, the client PC, the application server, and the monitoring / control server may be integrated.

  In the monitoring / control server (10), under the OS unit (12), a startup request processing unit (14) that performs processing for a login request from a user and startup processing of an application program installed in the application server, A monitoring processor (16) that understands and manages the user's login status and application server operating status, and a control process that forcibly logs out the user terminal, stops the application program, and deletes it from memory And a monitoring control information DB (20) for storing data indicating the state and reference data for performing control.

  First, the basic processing flow will be described with reference to the overall system configuration diagram of FIG. Usually, business processing is performed through the routes of the user terminal (70, 80, 90), the network (60), the monitoring / control server (10), and the application server group (30, 40, 50).

  On the other hand, the installed monitoring / control server (10) periodically grasps the status of the application server by the monitoring processing unit (12) mounted on the server, and the control processing unit (18) limits the operation of the application server. Is calculated based on the state information and the reference data preset in the system, and the monitoring control information DB (20) is updated.

  Next, the flow of processing for starting an application program will be described below. The user terminal << 70, 80, 90 >> sends a request for starting an application program installed in the application server to the monitoring / control server (10).

  Before starting the requested application program, the monitoring / control server (10) determines the application program based on the processing result of periodic status grasp in the startup request processing unit << 14 >> and the control processing unit (18). Determines whether or not to start.

  The monitoring / control server (10) makes an activation request to the application server in order to execute the application program requested to be activated only after it is determined that “it is possible to activate”.

  If the monitoring / control server (10) determines that "it is not possible to start", it notifies the user terminal user that the application program requested to be used is not available and the reason for this. To do.

  The above is the basic processing flow of this system. Next, the processing contents will be described in more detail using the flowcharts of FIGS. The flow in FIGS. 2 to 4 is performed by the monitoring / control server.

  FIG. 2 shows processing when the user logs in or launches an application. In S21, activation request processing is started. In S22, a user login is accepted from the client PC. That is, the user ID is received from the client PC, the user ID is stored in the user login table in S24, the current time is acquired, and the current time is stored as the login time.

  In S25, an application activation request (including the ID of the application to be activated) is received from the client PC. In S26, it is determined whether or not there is room for the CPU to start the application. Specifically, the CPU load monitoring table is referred to and it is determined whether or not the designated application can be activated from the load factor.

  If it is determined in S27 that the application cannot be started in S26, the message with message ID 1 is extracted from the message list table and transmitted to the client PC.

  If it is determined in S28 that the application can be started in S26, an instruction to start the application is issued to the application server, and a record of the task operation monitoring table is added. Here, the client PC and the application server perform communication processing via the transmission path, and the client PC can execute the application stored in the application server.

  In S29, the system control processing unit is activated to start the flow shown in FIG.

  The processing in FIG. 3 is the operation of the system monitoring processing unit. It is assumed that the system monitoring processing unit operates periodically. In S31, the CPU load factor of the application server is collected. In S32, the collected load factor is stored in the CPU load monitoring table.

  In S33, the user ID and task activation time are collected from the task operation monitoring table. In S34, the task operation time is calculated from the task activation time and written to the task operation monitoring table. In the example shown in FIG. 5, it is determined whether the task operation time is less than 30 minutes or more than 30 minutes, and the result is written.

  In S35, the rank of the scale and occurrence frequency is calculated, and the priority is calculated by the scale rank × occurrence frequency rank. In the example of FIG. 5, the smaller the priority value, the higher the priority. Here, the occurrence frequency is, for example, a log of the number of activations, and the higher the number of activations, the higher the importance as an application that is often used. The priority calculation is an example, and other calculation methods may be used.

  In S36, the calculated result is stored in the task priority table. In S37, the login status (usage time) of the user login is collected. In S38, the login status is written into the user login table. S31 to S38 are periodically repeated.

  Next, processing of the system control processing unit in FIG. 4 will be described. In S41, the task operation monitoring table is read. In S42, the CPU load monitoring table is read.

  In S43, it is determined whether or not the load factor read in S42 falls within the threshold range defined in the threshold table. If it falls within the threshold range, it is judged which threshold (threshold 1, threshold 2, or threshold 3) corresponds to. If it does not fall within the threshold range, the process returns to S41. In the example shown in FIG. 5, since the load factor is 60%, it corresponds to the threshold 3 in the threshold table.

  In S44, the threshold priority table is read, and the priority corresponding to the threshold extracted in S43 is extracted. In S45, the task corresponding to the priority is extracted from the task priority table. For example, in FIG. 5, since the threshold value is 3, priority 6 or higher corresponds. Next, tasks with a priority of 6 or higher are extracted from the task priority table. In the example of FIG. 5, AP3 has a priority of 9, which corresponds to a priority of 6 or higher.

  In S46, the task extracted in S45 is referred to the task operation monitoring table to determine whether or not the task operation time is 30 minutes or longer. When the task operation time is less than 30 minutes, even if the priority is lower than that extracted from the task priority table, the process for terminating the corresponding AP (application program) is not performed (the corresponding AP You may perform processing to continue the activation of. As a result, it is possible to prevent a situation where an AP with a low priority has been standing up for a long time.

Here, it is also possible to determine whether the AP is to be terminated based on whether the user login time is longer than a certain time, not the task operation time. For example, AP3 corresponds to the task priority table, and the user using AP3 is the user with the user ID 0003 based on the task operation monitoring table. However, when the user login table is determined, the usage time is the user ID 0003. Since the number of users is 60 minutes or more, the process of terminating the AP after S47 is performed.
If it is determined in S46 that it is “less than”, the process returns to S41.

In S47, the work status of the corresponding AP is saved. The storage destination may be an application server, a monitoring / control server, or an individual client PC.
In S48, the corresponding AP is terminated.

In S49, the message list table is read. For example, if it is determined in S46 that the task is 30 minutes or longer, message ID3 is read. If it is determined in S45 that the login is 60 minutes or longer, message ID2 is read.
In S50, the message read in S49 is transmitted to the client.
In S51, the task operation monitoring table record and task are deleted.

System configuration block diagram showing an embodiment of the present invention The flowchart which shows the procedure of the process of the starting request | requirement process part in FIG. The flowchart which shows the procedure of the process of the system monitoring process part in FIG. The flowchart which shows the procedure of the process of the system control process part in FIG. Tables referenced by the activation request processing unit, system monitoring processing unit, and system control processing unit in FIG.

Explanation of symbols

10 Monitoring / Control Server 12 OS Unit 14 Startup Request Processing Unit 16 Monitoring Processing Unit 18 Control Processing Unit 20 Monitoring Control Information DB
30, 40, 50 Application server group 31, 41, 51 Monitoring control information collection unit 60 Optical network 70, 80, 90 User terminal

Claims (5)

In a CPU load reduction method using a monitoring / control server that reduces the CPU load of an application server that operates multiple applications,
When multiple applications are running,
Monitoring the CPU load,
When the CPU load is equal to or greater than a predetermined threshold,
A method for reducing CPU load, wherein at least one of the activated applications is terminated. The CPU load reducing method according to claim 1,
Calculate the priority of the running application,
The CPU load reducing method, wherein the application to be terminated is an application having a lower priority than a predetermined priority. The CPU load reducing method according to claim 2,
The monitoring / control server includes:
Depending on the CPU load stage, it holds a table that defines the priority of the application to be terminated.
As a result of monitoring the CPU load,
Extracting the predetermined priority from the table;
A CPU load reduction method, wherein an application having a priority lower than the extracted priority is terminated. In the monitoring / control server that reduces the CPU load of the application server that operates multiple applications,
Monitoring means for monitoring the load on the CPU when a plurality of the applications are activated; and at least one of the activated applications when the load on the CPU exceeds a predetermined threshold. A monitoring / control server comprising: an ending unit for ending the connection. The monitoring / control server according to claim 4,
Having calculation means for calculating the priority of the application that is running;
The monitoring / control server characterized in that the termination means terminates an application having a lower priority than a predetermined priority.

Images