JP2629412B2 - Data extraction device for performance evaluation - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to a performance evaluation data extraction device used for operation management of a computer system.

[Conventional technology]

In order to operate the computer system efficiently, it is necessary to manage the operation of the system so that performance problems such as an abnormal increase in the usage rate of the central processing unit and the number of page faults of the main storage device do not occur. Further, when such a performance problem occurs, it is desired to improve it promptly.

In order to facilitate the operation management and operation improvement of such a computer system, the conventional computer system includes resource operation status data for each resource constituting the system, for example, the usage rate of the central processing unit, the main storage device. The number of page faults, the usage rate of the external storage device, etc. are periodically collected, and the job execution information data at the same time is also collected. After outputting the collected data to the system operation status storage data file, the contents are output. A function is provided to output the data as a list from the output device, or to collect them and display them on an output device such as a CRT at the same time.

The resource operation status data and the job execution information data of each resource obtained by using such a function accurately indicate the operation status of the computer system. Judgment is made for system performance diagnosis, such as determining whether a performance problem has occurred and presenting an improvement plan if a problem has been confirmed, and for situations such as a change in the equipment configuration of the computer system or a change in business. System performance prediction such as system performance prediction.

Furthermore, conventionally, a performance prediction / diagnosis apparatus that automatically performs the above-described system performance prediction and system performance diagnosis based on input system operation status storage data has been proposed. Have been.

[Problems to be solved by the invention]

By the way, generally, the system performance prediction is to predict a change in the operating status of the system when the target system has a device configuration change or a business addition change as described above. In this case, unless the utilization rate of the job executed in the target system is high to some extent and a performance problem occurs, it is not possible to know the limit of the processing capacity of the system unless it is predicted using data. There is no. Therefore, of the system operation status storage data including the resource operation status data and the job execution information data collected from the target system, data at the time when the job operation rate is high to some extent and a performance problem is occurring is extracted to estimate the performance. It is important to give the diagnostic device.

In the system performance diagnosis, as described above, it is determined whether or not a performance problem has occurred in the target system, and when the problem has been confirmed, an improvement plan is presented.
Therefore, of the system operation status record data including the resource operation status data and the job execution information data collected from the target system, data recognized as having a performance problem is extracted and provided to the performance prediction / diagnosis device. This is important for correctly and efficiently performing system performance diagnosis.

However, in the past, such data extraction had to be performed manually, and in general, this kind of data extraction work required a lot of time and expertise, so it was difficult for ordinary people to correctly As a result, system performance prediction and system performance diagnosis cannot be performed as a result, and there is a problem that it is difficult to design and efficiently operate a computer system for improvement.

The present invention has solved such a conventional problem.
The purpose is to provide a performance evaluation data extraction device that allows anyone to relatively easily extract data to be targeted for system performance prediction and system performance diagnosis from collected system operation status record data. is there.

[Means for solving the problem]

In order to achieve the above object, the present invention provides a system operation status record data including resource operation status data periodically collected for each resource constituting a computer system and job execution information data collected at the same time. An apparatus for extracting performance evaluation data to be subjected to performance evaluation from the data, wherein data that does not satisfy a preset performance reference value among data included in the system operation status recording data is distinguished from other data at a collection time. First means for displaying an evaluation target section selection screen arranged in order on the input / output device; and data included in the evaluation target section selected on the evaluation target section selection screen among the system operation status record data. Second means for extracting as performance evaluation data.

The data that does not satisfy the performance reference value is distinguished from other data by, for example, attaching a predetermined mark to the data.

[Action]

In the data extraction device for performance evaluation of the present invention, the first
Means for selecting an evaluation target section in which data that does not satisfy a predetermined performance reference value among the data included in the system operation status record data are arranged in the order of collection time by distinguishing them from other data by attaching a predetermined mark or the like. The screen is displayed on the input / output device, and when the user selects an evaluation target section while viewing the contents of the evaluation target section selection screen, the second means selects the selected one of the system operation status record data. The data included in the evaluation target section is extracted as performance evaluation data.

〔Example〕

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

FIG. 1 is a configuration diagram of one embodiment of the present invention. Referring to FIG. 1, a performance evaluation data extraction device 1 according to an embodiment of the present invention includes an external storage device 2 for storing system operation status record data 3 and an external storage device for storing performance reference value data 6.
5, an input / output device 12 having display means such as a CRT and input means such as a keyboard and a mouse, and a performance prediction / diagnosis device 13
And has a functional section including an evaluation target recommended section setting means 8 and a performance prediction / diagnosis device input data loading means 10; a copy 4 of system operation status record data; and performance reference value data. And a storage area (not shown) for storing a copy 7, edit output data 9 and performance prediction / diagnosis device input data 11.

System operation status record data 3 in external storage device 2
Includes resource operation status data periodically collected for each of the resources constituting the computer system, such as the central processing unit, main storage device, and external storage device, and job execution information data collected at the same time. Such resource operation status data and job execution information data are collected by a known operation status collection function provided in the computer system. That is, the operation status collection function collects the operation status of each resource and the execution information of the job together with the sampling time by sampling processing every elapse of a predetermined interval time, and collects them as one record constituting the system operation status record data 3. It is recorded and stored in the external storage device 2. The specific contents of the resource operation status data include the usage rate and the like for the central processing unit, and the number of page faults for each of the system domain side and the user domain side for the main storage unit.
For the external storage device, the usage rate, the queue length and the like are included. Also, as the specific contents of the job execution information data,
Data such as the number of jabs executed during the interval is included.

The performance reference value data 6 in the external storage device 5 is a threshold value used when a performance diagnosis expert evaluates the usage rate of the resources constituting the computer system, and shows an example of the performance reference value data 6. Referring to FIG. 2, the performance reference value of the central processing unit is “20% <use rate <90%”, the performance reference value of the main storage device is “system domain side page fault frequency <10 times / second”, [The number of page faults on the user domain side <the job multiplicity × 2 times / second], and the performance reference values of the external storage device are “usage rate <30%” and “average queue length <0.5”.

The copy 4 of the system operation status record data in the performance evaluation data extraction device 1 is copied from the system operation status record data 3 in the external storage device 2 to the central processing unit, main storage device, and external storage device by means not shown. And only records relating to job execution are extracted and sorted by sampling time. Further, the copy 7 of the performance reference value data is stored in the performance reference value data 6 in the external storage device 5.
Is a copy of

The evaluation target recommendation setting means 8 selects a copy 7 of the performance reference value data from the copy 4 of the system operation status record data.
Search for data that does not satisfy the conditions of the above, and if there is data that does not satisfy the performance standard value, add a predetermined mark, such as “★”, to the data, and add such a mark to other data Instead, the data is accumulated in the edit output data 9, and when all the data of the copy 4 of the system operation status record data have been processed, the evaluation target section selection screen created based on the edit output data 9 is completed. By providing the information to the input / output device 12, data that does not satisfy the preset performance reference value among the data included in the system operation status record data is distinguished from other data and is arranged in the order of the collection time to select the evaluation target section. This is a means for displaying a screen on the display device of the input / output device 12. The user selects an evaluation target section while viewing the evaluation target section selection screen displayed on the display device of the input / output device 12. At this time, the data marked with “★” is data that does not satisfy the performance reference value, that is, data of candidates to be subjected to system performance prediction and system performance diagnosis. Can be easily selected.

On the other hand, the performance prediction / diagnosis device input data loading means 10
When an evaluation target section is selected on the evaluation target section selection screen displayed on the input / output device 12, the external storage device 2
The data included in the selected evaluation target section is sequentially extracted from the system operation status record data 3 stored in the storage unit 3 and stored as performance prediction / diagnosis device input data 11.
When all data has been extracted, this is a means for loading the performance prediction / diagnosis device input data 11 into the performance prediction / diagnosis device 13 as input data.

Next, the operation of the performance evaluation data extraction device 1 of the present embodiment will be described focusing on the operation of the evaluation target recommended section setting means 8 and the performance prediction / diagnosis device input data loading means 10.

When the performance evaluation data extraction device 1 is started, it extracts only the central processing unit, the main storage device, the external storage device, and the records related to the job execution from the system operation status record data 3 in the external storage device 2. Of system operation status record data by sorting data by sampling time 4
Is generated, and a copy 7 of the performance reference value data 6 in the external storage device 5 is generated. Then, first, the evaluation target recommended section setting means 8 is activated.

When activated, the evaluation target recommended section setting means 8 starts the processing shown in FIG. 3, and first determines whether an unprocessed record remains in the copy 4 of the system operation status record data ( Step 31), if any, retrieve one record (step 32).

Next, if the retrieved record is a record indicating the operation status of the central processing unit, the information on the usage rate of the central processing unit indicated in this record is set in the copy 7 of the performance reference value data. Determine whether the performance standard value of the utilization rate of the central processing unit is satisfied (Step 3
3). When it is determined in step 33 that the usage rate of the central processing unit does not satisfy the performance standard value, a mark “★” indicating that the performance rate does not satisfy the performance standard value is added to the usage rate information, and the information is added. The data is transferred to the edit output data 9 (step 34). When the usage rate of the central processing unit satisfies the performance reference value, the information is directly used as the edit output data 9.
(Step 35). In a computer system including a plurality of central processing units, step 34 and step
In the transfer to the edit output data 9 in 35, the usage rate is averaged and overwritten.

If the record extracted in step 32 is a record indicating the operation status of the main storage device, information on the number of page faults (for each of the system domain side and the user domain side) of the main storage device indicated in this record It is determined whether or not satisfies the performance reference value of the number of page faults of the main storage device set in the copy 7 of the performance reference value data (step 36). If it is determined in step 36 that the number of page faults in the main storage device does not satisfy the performance reference value, a “★” mark indicating that the performance reference value is not satisfied is added to the information on the number of page faults. The information is transferred to the edit output data 9 (step 37).
If the number of page faults in the main storage device satisfies the performance reference value, the information is directly transferred to the edit output data 9 (step 38).

Further, when the record extracted in step 32 is a record indicating the operation status of the external storage device, the usage rate and the average queue length of the external storage device indicated in this record are stored in the copy 7 of the performance reference value data. It is determined whether or not the set performance ratios of the external storage device and the average queue length are satisfied (step 39). When it is determined in step 39 that the usage rate of the external storage device and the average queue length do not satisfy the performance standard value, a mark “★” indicating that the performance standard value is not satisfied is displayed with the usage rate and the average queue length. The information in the record is transferred to the edit output data 9 (step 3A). If the external storage device usage rate and average queue length satisfy the performance standard value,
The information is directly transferred to the edit output data 9 (step 3B). Since a plurality of external storage devices are usually connected to the system, the information is always overwritten with the information of the external storage device having a high usage rate at the time of transfer to the edit output data 9 in steps 3A and 3B.

If the record extracted in step 32 is a job execution information record, information on the number of jobs executed during the interval (job multiplicity) is transferred to the edit output data 9 (step 3C).

FIG. 4 shows an example of the contents of the editing output data 9 created as described above. It should be noted that the section in the figure is the sampling time of the resource operation status data and the job execution information data.

The evaluation target recommended section setting means 8 completes the check based on the performance reference value for all the records of the copy 4 of the system operation status record data (N in step 31).
O), by giving the evaluation target section selection screen information created based on the edit output data 9 to the input / output device 12, the evaluation target section selection screen as shown in FIG. Display on the display device (Step 3
D). If all data in the system operation status record data 3 is displayed, if the display amount becomes too large, the section where the “★” mark does not appear at all or appears sparsely is omitted, and “★” is omitted. Only the section where the mark appears densely, that is, the section centered on the evaluation target recommended section is displayed on the screen.

For the time being, all data are to be displayed, but the first section (in terms of time) in which the “★” mark appears densely is displayed on the input / output device 12 first. The contents of other sections can be referred to by switching the screen. Further, the page number or the like of the screen where "★" mark appears densely may be displayed.

Next, the user looks at the screen for selecting an evaluation target section as shown in FIG. 5 displayed on the display device of the input / output device 12, and places, for example, “S” in a portion desired to be the start time of the evaluation target section. Is replaced with the letter "E"
Is input and the evaluation target section is selected, the performance prediction / diagnosis device input data loading means 10 reads the data included in the selected evaluation target section from the system operation status record data 3, Is loaded into the performance prediction / diagnosis device 13 as the performance prediction / diagnosis device input data 11 (step 3E in FIG. 3).

According to the performance evaluation data extraction device 1 of this embodiment including the evaluation target recommended section setting means 8 and the performance prediction / diagnosis device input data loading means 10 which function as described above, each resource constituting the system Data candidate group to be a target of performance prediction or diagnosis is automatically found out of the system operation status record data 3 composed of data relating to the operation status of the system and job execution information data, and an evaluation target section in which the data candidates are specified. A selection screen is displayed on the input / output device 12, and when the section to be evaluated is selected by the user, the data of the system operation status record data 3 included in the selected section is input to the performance prediction / diagnosis device 13. This kind of data extraction operation can be performed easily and reliably without relying on performance evaluation specialists. It made.

Although the embodiments of the present invention have been described above, the present invention is not limited to the above embodiments, and various other modifications can be made. For example, it is possible to distinguish data that does not satisfy the performance reference value from other data by another method such as making the luminance on the screen different.

〔The invention's effect〕

As described above, the performance evaluation data extraction device of the present invention includes the resource operation status data periodically collected for each resource constituting the computer system and the job execution information data collected at the same time. From the system operation status record data, data that does not satisfy the preset performance reference value, that is, candidate data to be subjected to system performance prediction or system performance diagnosis is searched for, and these candidate data are marked by a predetermined mark or the like. Is displayed on the input / output device in order to distinguish the data from other data and arranged in the order of the collection time, thereby prompting the user to select the evaluation target section. Therefore, even a user who does not have specialized knowledge can easily select an evaluation target section to be subjected to system performance prediction and system performance diagnosis.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of one embodiment of the present invention, FIG. 2 is a diagram showing an example of performance reference value data, and FIG. 3 is a process of an evaluation target recommended section setting means and a performance prediction / diagnosis device input data loading means. FIG. 4 is a diagram showing an example of the contents of edit output data, and FIG. 5 is a diagram showing an example of an evaluation target section selection screen displayed on the input / output device. In the figure, 1... Performance evaluation data extraction device 2, 5... External storage device 3... System operation status record data 4... Copy of system operation status record data 6... Performance reference value data 7. Copying of value data 8 ... Evaluation target recommended section setting means 9 ... Edit output data 10 ... Performance prediction / diagnosis device input data loading means 11 ... Performance prediction / diagnosis device input data 12 ... I / O device 13 ... … Performance prediction / diagnosis device

Claims (2)

(57) [Claims] An object of performance evaluation is system performance record data including resource operation status data periodically collected for each resource constituting a computer system and job execution information data collected at the same time. An apparatus for extracting performance evaluation data to be evaluated, wherein an evaluation target section selection in which data not satisfying a preset performance reference value among data included in the system operation status record data is arranged in the order of collection time to be distinguished from other data. First means for displaying an application screen on an input / output device; and extracting data included in the evaluation target section selected on the evaluation target section selection screen from the system operation status record data as performance evaluation data. A data extraction device for performance evaluation, comprising: a second means. 2. The performance evaluation data extraction device according to claim 1, wherein the data that does not satisfy the performance reference value is distinguished from other data by attaching a predetermined mark to the data.