JP2000081992A - Computer performance checking system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a computer performance checking system capable of quickly and automatically inspecting a cause of system performance deterioration. SOLUTION: The performance (excess swapping, disk I/O and system management table) sampling data 24 to 26 of a computer system (10 to 12) are sampled and compared with thresholds stored in a system performance threshold management table 27 previously set up in the system corresponding to these data. The cause of performance deterioration and its treatment are displayed on a system console 11 based on the compared result so as to notify an operator of the displayed contents and they are registered in a fixed disk 12.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a computer performance check system for automatically checking the performance of a computer, particularly an office computer.

[0002]

2. Description of the Related Art Conventionally, when inspecting or investigating the performance of a computer such as an office computer, each sampling data relating to collected system performance (sampling data relating to a system memory, sampling data relating to a disk cache memory, and a system memory table). The system administrator analyzed the sampling data, etc., and investigated and judged whether the system was operating in a normal tuning state. In addition, when performance degradation has occurred in the system, the system administrator has also specified the cause of the performance degradation. As a conventional example related to computer performance check, see Japanese Patent Application Laid-Open
No. 9, JP-A-63-14289 and JP-A-6-14
It is disclosed in Japanese Patent Application Laid-Open No. 2-271151.

In the first conventional example, the information processing apparatus receives the evaluation information transmitted by the evaluation information transmitting means, sets it by the evaluation information setting means, and executes the test instruction by the instruction executing means. The execution result is sent to the automatic test equipment,
This is compared with an expected value created in advance. If the comparison result is determined to be within the correct range, the result is evaluated as “normal”, and if it is determined to be outside the correct range, as “error”, and each result is output by the output means. It is an interactive test system.

In the second prior art, the first and second input / output means, the alarm means, and the data processing means are connected by a data bus,
In addition, an automatic analyzer, a host computer, a workstation, and a printer are provided. The analysis data means is transferred to the computer of the clinical test system, and outputs a report or the like according to the instruction of the station. At this time, a check is performed by inserting a quality control sample whose accuracy is controlled for each test item between normal samples. This is a data collection processing device that generates an alarm if the result is defective, and re-analyzes the automatic analyzer by specifying it again.

In the third prior art, when an abnormality is detected in a test item corresponding to a test program, the test execution control program updates an execution status corresponding to the test item to NG by an abnormal end message, The updated test execution control table is restored to the external storage device. Further, the test execution program outputs an error message to the operation device and notifies the operator of the abnormality detection. Upon receiving the error message, the operator stops the system, performs a process according to the content of the fault, removes the cause of the fault, and starts the test execution program again. At this time, this is a computer main body automatic test method in which only the test items for which the execution status of the test execution control table is confirmed to be unexecuted are selected as execution targets.

[0006]

However, in these prior arts, when examining the system performance of an office computer, it is necessary to analyze in detail various sampling data relating to the system performance of the office computer. Further, when it is determined that performance degradation has occurred in the system, there is a disadvantage that it takes a long time to investigate the cause, and as a result, a great deal of investigation time is required. Further, there is a problem that since the investigation is arbitrarily performed by the system administrator, there is a possibility that an operation error or the like may occur due to human operation.

SUMMARY OF THE INVENTION An object of the present invention is to provide a computer performance check system which enables a system administrator and an operator to easily identify the cause of performance deterioration of a computer system and to promptly take measures after a problem occurs.

[0008]

In order to solve the above-mentioned problems, a computer performance check system according to the present invention employs the following characteristic configuration.

(1) In a computer performance check method for checking the performance of a computer system having a CPU, a memory, a system console, and a fixed disk, sampling data relating to the performance of the computer system is collected, and the sampling data is collected in accordance with the sampling data. Comparing with a performance threshold preset for the system, analyzing and identifying the cause of performance degradation of the system based on the comparison result, and analyzing and identifying the cause of the performance degradation and the action of the identified system to the system console. Computer performance check method to be displayed.

(2) The computer performance check method according to the above (1), wherein the cause and the measure of the performance deterioration are registered in the fixed disk.

(3) The computer performance check method according to the above (1) or (2), wherein a system administrator can change the performance threshold of the computer system.

(4) The computer performance check method according to (1), (2) or (3), wherein each of the steps is executed by software.

(5) The sampling data collected from the computer system includes any one of the above-mentioned items (1), (2), and (3) including any one of excessive swapping sampling data, disk I / O sampling data, and system management table sampling data. ) Or (4).

[0014]

Embodiments of the present invention will be described below with reference to the accompanying drawings.

FIG. 1 is a system configuration diagram of a preferred embodiment of a computer performance check system according to the present invention. FIG. 2 is an example of a software configuration of the computer performance check method of the present invention. FIG. 3 is a schematic flowchart of the computer performance check method of the present invention. 4 to 6 show a software basic flowchart of the computer performance check method of the present invention. FIG.
9 is a flowchart of a system performance threshold change in the computer performance check method of the present invention.

In the system configuration diagram of FIG.
(Central processing unit) and computer body 1 including memory
0, a system console 11 and a sampling data management module 12.

The computer performance check method of the present invention has a software configuration shown in FIG. That is, the sampling data management module 23, the excessive swapping sampling data I / O management module 24, the disk I / O, the sampling data I / O management module 2
5. System management table It has a sampling data I / O management module 26, a system performance threshold management table 27, and a system performance threshold control module 28.

When the end of the sampling period is confirmed by the sampling data management module 23 (process 1 in FIG. 3), various sampling data relating to the collected system managed by the modules 24 to 26 are designated by the system in advance. This is compared with each threshold managed in the system performance threshold management table (process 2 in FIG. 3).
These data are sampling data relating to excessive swapping, sampling data relating to disk I / O, and sampling data relating to the system management table.

The sampling data management module 23 shown in FIG. 2 uses the average number of memory pages SWP1 which can be used by a user process as sampling data relating to excessive swapping as shown in FIG. 4 and the system performance shown in FIG. The SWP1 threshold value 1 managed in the threshold value management table 27 is compared with the threshold value. If the SWP1 is smaller, the SWP flag 409 in FIG. next,
Number of page failures in address translation per second S
WP2 is compared with SWP2 threshold value 2 which is set in advance in the system and managed by the above-described system performance threshold value management table 27. If SWP2 is large, 2 is set to the SWP flag 410 of FIG.

Next, a comparison is made between the number of transfers SWP3 from the memory to the swap area per second and the SWP3 threshold value 3 managed in the system performance threshold value management table 27 set in the system. I do. If SWP3 is large, 3 is set to the SWP flag 411 in FIG.

When the CPU is in the idle state, SWP4
Is compared with the SWP4 threshold value 4 managed in the system performance threshold value management table 27 preset in the system. If SWP4 is large, the SWP flag 41 of FIG.
Set 2 to 4.

As described above, when data (other than 0) is set in the SWP flag at the end of an arbitrary sampling period (process 3 in FIG. 3), the sampling data management module 23 determines that the system memory is insufficient. Detect.
Then, the system console 11 in FIG. 1 is displayed on the system console 11 in FIG. 1 to indicate that the system memory is insufficient and the system memory needs to be expanded (processing 5 in FIG. 3) to call attention to an operator or the like. In addition, each detailed event (SWP1 to SWP
Regarding WP4), at the end of the sampling period, the sampling data management module 23 in FIG. 2 records it in an arbitrary file (fixed disk device) (process 4 in FIG. 3),
Each detail is displayed on the system console 11 on another screen.

The sampling data management module 23
Is the sampling time for the disk I / O shown in FIG.
The SK1 is compared with the DISK1 threshold 1 managed in the system performance threshold management table 27 of FIG. If DISK1 is large, 1 is set to the DISK flag 414 in FIG.

Next, a comparison is made between DISK2, which is the ratio of the time spent by the disk for the service for the transfer request, and DISK2 threshold 2, which is previously set in the system and managed in the system performance threshold management table 27. I do. If DISK2 is large, DISK flag 4 in FIG.
Set 2 to 15.

DISK3, which is the ratio of the logical read detected in the system buffer, and D managed by the system performance threshold management table 27 set in advance in the system.
Compare with ISK3 threshold value 3. If DISK3 is large, 3 is set to the DISK flag in FIG.

Thereafter, DISK4, which is the ratio of the logical write detected in the system buffer, is compared with DISK4 threshold 4, which is managed in the system performance threshold management table 27 designed in advance by the system. DISK4
Is larger, 4 is set to the DISK flag 417 in FIG.

As described above, if the data (other than 0) is set in the DISK flag at the end of an arbitrary sampling period (process 3 in FIG. 3), the sampling data management module 23 in FIG. Is detected, the fact that the DISK cache memory needs to be extended is displayed on the system console 11 of FIG. 1 (FIG. 3).
5) to call attention to an operator or the like. Also DI
Each detailed event (DI
SK1 to DISK4) are recorded in an arbitrary file (fixed disk device) at the end of the sampling period, and details of each are displayed on the system console of FIG. 1 on a separate screen.

Next, the sampling data management module 23 shown in FIG. 2 stores BUF1, which is the number of overflows of the process entry table used or allocated in the OS kernel, which is sampling data relating to the system management table shown in FIG. A comparison is made with the BUF1 threshold 1 managed by the system performance threshold management table 27 of FIG. 2 set in the system. If BUF1 is large, 1 is set to the BUF flag 419 in FIG.

Further, BUF2, which is the number of overflows of the file management table used in the OS kernel, is compared with the BUF2 threshold value 2 managed in the system performance threshold management table 27 set in advance in the system. I do.
If BUF2 is large, the BUF flag 420 in FIG.
Is set.

As described above, the sampling data management module 23 in FIG. 2 determines that the data (other than 0) is set in the BUF flag at the end of an arbitrary sampling period (process 3 in FIG. 3). Detect shortage of management table. As a result, the fact that the system management table needs to be extended indicates that the system console 11 of FIG.
(Process 5 in FIG. 3) to alert the operator or the like.

Each detailed event (BUF1, BUF2) that has become insufficient in the system management table is recorded in a file (fixed disk device) at the end of this sampling period (process 4 in FIG. 3), and is displayed on another screen. Details of each are displayed on the system console 11 of FIG. The system performance thresholds managed in the above-described system performance threshold management table 27 are the events obtained during this sampling period and any file (fixed disk 1 in FIG. 1).
Judging from 2), it can be changed by the system performance threshold control module 28 of FIG.

The preferred embodiment of the computer performance check system of the present invention has been described above in detail. However, it goes without saying that various modifications can be made without departing from the spirit of the present invention.

[0033]

As is apparent from the above description, according to the computer performance check method of the present invention, each sampled data sampled by sampling is compared with a corresponding threshold value set in advance, and this comparison analysis is performed. The cause of the performance degradation of the computer system is identified based on the result. The computer system console further includes software for displaying the cause of system performance deterioration and any countermeasures against the deterioration. As a result, the system administrator and the operator can easily and quickly identify the cause of the performance degradation of the computer system, and can immediately take a corrective action after the occurrence of the failure. Further, by changing the system performance threshold value to each system performance threshold value adapted to the user operation for each system, it becomes possible to perform appropriate system performance tuning.

[Brief description of the drawings]

FIG. 1 is a configuration example of a computer system to which a computer performance check method of the present invention is applied.

FIG. 2 is a software configuration diagram for executing the computer performance check method of the present invention.

FIG. 3 is a schematic flowchart of a computer performance check method according to the present invention.

FIG. 4 is a software flowchart relating to excessive swapping according to the computer performance check method of the present invention.

FIG. 5 is a flowchart similar to that of FIG. 4 for a disk I / O.

FIG. 6 is a flowchart related to a system management table similar to FIG.

FIG. 7 is a flowchart of changing a system performance threshold value in the computer performance check method of the present invention.

[Explanation of symbols]

 Reference Signs List 10 Computer main body 11 System console 12 Fixed disk 23 Sampling data management module 27 System performance threshold management table

Claims (5)

[Claims] 1. A computer performance check method for checking the performance of a computer system having a CPU, a memory, a system console, and a fixed disk, wherein sampling data related to the performance of the computer system is collected, and the system is sampled in accordance with the sampling data. Is compared with a preset performance threshold value, and the cause of performance degradation of the system is analyzed and identified based on the comparison result, and the analyzed and identified cause of performance degradation of the system and the action are displayed on the system console. A computer performance check method. 2. The computer performance check method according to claim 1, wherein the cause of the performance deterioration and the action are registered in the fixed disk. 3. The computer performance check method according to claim 1, wherein a system administrator can change the performance threshold of the computer system. 4. The computer performance check method according to claim 1, wherein each of the steps is executed by software. 5. The system according to claim 1, wherein the sampling data collected from the computer system includes any of excessive swapping sampling data, disk I / O sampling data, and system management table sampling data. Or the computer performance check method according to 4.