JP4409498B2 - Management computer control method and server control method - Google Patents

Description

  The present invention calculates an allowable value for detecting load fluctuations according to server capabilities and processing contents when detecting server load fluctuations using an index value that can quickly detect load fluctuations. The present invention relates to a control method of a management computer capable of detecting the load, and a control method of a server that detects a load change using the index value and the allowable value.

In recent years, the performance of computers has been rapidly improved by the advancement of information technology. And the form which utilizes a some computer as a server for providing information processing service is also prevailing. In this case, various techniques for distributing the load on each server have been developed (see, for example, Patent Document 1).
JP 2001-216282 A

  However, it may be difficult to predict load fluctuations in advance, for example, when an information processing service is provided in response to a processing request from a client connected via the Internet. In such a case, it is desirable to be able to immediately detect and deal with any sudden increase in load.

  By the way, in general, the load on the server is often based on a CPU (Central Processing Unit) utilization rate. However, since the CPU utilization rate may vary greatly even with a slight load variation, it is necessary to perform leveling such as taking an average value within a predetermined time. Therefore, when the CPU usage rate is used as a load index, even when the load suddenly fluctuates, it cannot be detected immediately, and it is difficult to quickly cope with the load fluctuation. .

  On the other hand, the number of processing requests received from the client can also be used as an index of server load. In this case, it is possible to detect an increase in load immediately when the number of processing requests exceeds an allowable value. However, even if the number of processing requests sent from the client is the same, the load on the server varies depending on the content of the processing request and the server's capability, so the number of processing requests received from the client reflects the load on the server. Is not necessarily true.

  The present invention has been made in view of the above problems, and in the case of detecting a load fluctuation of a server using an index value capable of quickly detecting the load fluctuation, the load fluctuation corresponding to the server capability and the processing content is detected. The main object is to provide a control method of a management computer capable of calculating an allowable value for detection, and a control method of a server that detects a load variation using the index value and the allowable value.

In order to solve the above-mentioned problems, the present invention provides a management that is communicably connected to one or a plurality of servers that transmit a processing result for each processing request transmitted from one or a plurality of clients to the client. In the computer control method, the management computer calculates an absolute load allowable value that is a countable value indicating an absolute amount of load applied to the server for each of the one or a plurality of servers, and A setting adjustment program for transmitting the calculated absolute load allowable value to the server, and a value indicating a relative load according to the processing capacity of each of the one or more servers, the processing capacity of each server Regardless of the storage device that stores the relative load allowable value that can represent the degree of load with the same value, the setting adjustment program is read into the memory. And a processing device that performs communication with the one or more servers, and the processing device has the absolute load of the server for each of the one or more servers. And the relative load received from the server using the communication device, and the absolute load received from each server using the communication device for each of the one or a plurality of servers, and the storage device The absolute load tolerance value of each server is calculated from the relative load tolerance value stored in the server and the relative load value received from each server using the communication device. The control method of the management computer is characterized in that the absolute load allowable value calculated for each server is transmitted to the respective servers using the communication device to request setting .

  In addition, the problems disclosed by the present application and the solutions thereof will be clarified by the column of the best mode for carrying out the invention and the drawings.

  Management that can calculate an allowable value for detecting load fluctuations according to server capabilities and processing contents when detecting server load fluctuations using index values that can quickly detect load fluctuations It is possible to provide a computer control method and a server control method that quickly detects a load variation using the index value and the allowable value.

=== Example of Overall Configuration ===
FIG. 1 shows an overall configuration of a computer system according to the present embodiment.

  The computer system according to the present embodiment is an information processing system that is communicably connected to the client 100 via the first network 500 and provides various information processing services to the client 100. In the computer system according to the present embodiment, a management computer 200, a load distribution device 700, and a web server 600 (corresponding to the server described in the claims) are connected via a second network 510 so as to be communicable. It consists of

  The client 100 is a computer that transmits a processing request for receiving provision of an information processing service to the web server 600. FIG. 1 illustrates a client A100, a client B100, and a client C100, but there may be one client 100 or a plurality of clients 100.

  The web server 600 is a computer that transmits a processing result for each processing request transmitted from the client 100 to the client 100. Although FIG. 1 illustrates the web server A 600, the web server B 600, and the web server C 600, there may be one or more web servers 600.

  Each web server 600 stores an allowable value (also referred to as a determination value) of the number of processing requests received from the client 100 within a predetermined time. The allowable value may be an upper limit value or a lower limit value of the number of processing requests received from the client 100 within a predetermined time. As will be described in detail later, this allowable value is not limited to the allowable value of the number of processing requests received from the client 100 within a predetermined time, but the allowable value calculated based on the number of processing requests received from the client 100 Can be a value. For example, the number of processing requests currently being processed among the processing requests received from the client 100 may be used.

  When each web server 600 detects that the number of processing requests transmitted from the client 100 exceeds the upper limit value, the web server 600 transmits a message to that effect to the management computer 200. In this case, the management computer 200 performs a process such as adding a spare web server 600 to the computer system. Alternatively, for example, when each web server 600 detects that the number of processing requests has fallen below a lower limit, the web server 600 transmits a message to that effect to the management computer 200. In this case, the management computer 200 performs a measure such as reducing the web server 600 from the computer system, for example.

  The load balancer 700 is a computer that distributes and transmits each processing request transmitted from the client 100 to the web server 600 to each web server 600. Of course, when only one web server 600 is provided, the load balancer 700 may not be provided. Further, even when there are a plurality of web servers 600, a configuration in which the load distribution device 700 is not provided may be employed. As will be described in detail later, the load distribution apparatus 700 stores a distribution ratio when the load distribution apparatus 700 distributes and transmits each processing request transmitted from the client 100 to each web server 600. The load distribution apparatus 700 transmits each processing request transmitted from the client 100 to each web server 600 according to the distribution ratio. This distribution ratio can be updated by the management computer 200 repeatedly, for example, periodically.

  The management computer 200 is a computer for managing a computer system. For example, calculation and update of the distribution ratio stored in the load distribution device 700 described above, calculation and update of the allowable value of the number of processing requests received by each web server 600 from the client 100 within a predetermined time, and the like are performed. Also, processing such as addition or reduction of the web server 600 described above is performed. Details will be described later.

  The first network 500 is a communication network that connects the client 100 and the computer system in a communicable manner. The first network 500 may be a WAN (Wide Area Network) such as the Internet, or may be a communication network in a company.

  The second network 510 is a communication network that connects the load balancer 700, the web server 600, and the management computer 200 so that they can communicate with each other. In the present embodiment, the first network 500 and the second network 510 are separate networks, but they may be the same network.

=== Device configuration ===
Next, configurations of the client 100, the management computer 200, the web server 600, and the load distribution apparatus 700 will be described.

  The client 100, the management computer 200, the web server 600, and the load distribution device 700 are all computers, and the hardware configuration is basically the same. Therefore, these hardware configurations are summarized in one block diagram and shown in FIG. Further, control programs and tables for realizing these functions are shown in FIGS.

<Management computer configuration>
The management computer 200 includes a CPU 210, a memory 220, a port 230, a recording medium reading device 240, an input device 250, an output device 260, and a storage device 280.

  The CPU 210 is responsible for overall control of the management computer 200. The management computer control program 900 composed of codes for performing various operations according to the present embodiment stored in the storage device 280, a setting adjustment program (patent) Various functions as the management computer 200 are realized by reading the event handling program 920 into the memory 220 and executing it (910 corresponding to the program described in the claims). For example, the management computer control program 900 and the setting adjustment program 910 are executed by the CPU 210, and load tolerance value storage is performed by cooperating with hardware devices such as the memory 220, the port 230, the input device 250, the output device 260, and the storage device 280. A server information receiving unit, an allowable value calculating unit, a determination value calculating unit, an allowable value transmitting unit, and a determination value transmitting unit. Here, the allowable load value storage unit stores an allowable value of a value indicating the load of the web server 600. This allowable value is set by a system administrator, for example, and stored in the memory 220 of the management computer 200. As an allowable value of the value indicating the load of the web server 600, for example, the CPU usage rate can be within 80%. The server information receiving unit receives a value calculated based on the number of processing requests received by the web server 600 from the client 100 and a value indicating the load of the web server 600 from the web server 600. The value calculated based on the number of processing requests received by the web server 600 from the client 100 is an index value used for determining whether or not to send a message to the management computer 200. The index value can be, for example, the number of processing requests received from the client 100 within a predetermined time, or the number of processing cases currently being processed. The allowable value calculation unit calculates based on the number of processing requests based on the value calculated based on the number of processing requests, the allowable value of the value indicating the load, and the value indicating the load. The allowable value of the value to be calculated is calculated. This allowable value can be, for example, 150 or less processing requests received from the client 100 per minute. The determination value calculation unit is configured based on the value indicating the load of the web server 600, the index value, and the allowable value of the value indicating the load of the web server 600 stored in the allowable load value storage unit. A determination value used for determining whether or not the server 600 transmits a message to the management computer 200 is calculated. This determination value is the same as the allowable value calculated based on the number of processing requests. The allowable value transmission unit transmits the allowable value calculated based on the calculated number of processing requests to the web server 600. The determination value transmission unit transmits the calculated determination value to the web server 600. The memory 220 can be constituted by a semiconductor memory device, for example.

  The management computer control program 900 is a program for controlling the management computer 200 like an operating system, for example. Thereby, control of various hardware devices and software of the management computer 200 are performed. For example, various types of data can be communicated with the web server 600 and the load distribution apparatus 700.

  The setting adjustment program 910 is a program for calculating an allowable value of the number of processing requests received by each web server 600 from the client 100 within a predetermined time or calculating a distribution ratio stored in the load balancer 700. .

  The event handling program 920 is a program for performing processing according to various messages transmitted from the web server 600 to the management computer 200. For example, addition or reduction of the web server 600 described above is performed. The content of the message received from the web server 600 and the content of the processing are stored in an event handling list table 810 described later. An example of the event handling list table 810 is shown in FIG. An example of a message transmitted from the web server 200 is shown in FIG.

  As described above, in the present embodiment, the management computer 200 receives a message from the web server 200 and the event handling program 920 is executed, whereby autonomous control of the computer system is performed. Autonomous control is control that manages a computer system without specific instructions from a system administrator. As an example of autonomous control, there is control that autonomously responds to a failure or load increase that occurs in a computer system. For example, when an increase in a failure or load is detected in the web server 200, the cause of the failure or load increase is autonomously estimated, an appropriate measure is selected and executed based on the result, and the result is evaluated. To do. By autonomously carrying out these series of procedures, the system can autonomously handle failures and load increases that occur, and realizes management burden reduction, early cause investigation, and early recovery. Autonomous control is realized by the event handling program 920 being executed in cooperation with the setting adjustment program 910 and the management computer control program 900.

  The recording medium reading device 240 is a device for reading programs and data recorded on the recording medium 1000. The read program and data are stored in the memory 220 and the storage device 280. Therefore, for example, the management server control program 900, the setting adjustment program 910, and the event handling program 930 recorded on the recording medium 1000 are read from the recording medium 1000 by using the recording medium reading device 240 and stored in the memory 220 and the storage device 280. It can be memorized.

  As the recording medium 1000, a flexible disk, a magnetic tape, a compact disk, or the like can be used. The recording medium reading device 240 can be built in the management computer 200 or can be externally attached.

  The storage device 280 can be, for example, a hard disk device or a semiconductor storage device. The storage device 280 stores a management computer control program 900, a setting adjustment program 910, an event handling program 920, a monitoring target list table 800, an event handling list table 810, and an allowable value 830 indicating a load. This is shown in FIG.

  The monitoring target list table 800 is a table in which a list of information indicating each computer constituting the computer system is described. An example of the monitoring target list table 800 is shown in FIG. As shown in FIG. 7, the monitoring target list table 800 describes, for example, the IP addresses and types of the web server 600 and the load balancer 700. Thereby, the management computer 200 can grasp the web server 600 and the load distribution device 700 in the computer system and can communicate with the web server 600 and the load distribution device 700.

  The event handling list table 810 is a table in which the content of each message transmitted from the web server 600 is associated with the content of the action to be performed by the management computer 200 when each message is received. An example of the event handling list table 810 is shown in FIG.

  The allowable value 830 of the value indicating the load is an allowable value of the CPU usage rate of each web server 600. This allowable value can be set by an administrator of the computer system, for example. It is set to an appropriate value according to the operation policy of the computer system. For example, the allowable value of the CPU utilization rate can be set as 80%.

  The input device 250 is a device used for data input to the management computer 200 and functions as a user interface. For example, a keyboard or a mouse can be used as the input device 250.

  The output device 260 is a device for outputting information to the outside and functions as a user interface. For example, a display or a printer can be used as the output device 260.

  The port 230 is a device for performing communication. For example, communication with other computers such as the web server 600 and the load distribution device 700 performed via the second network 510 is performed via the port 230. Further, for example, the management computer control program 900, the setting adjustment program 910, and the event handling program 920 are received from another computer via the port 230 via the second network 510 and stored in the memory 220 or the storage device 280. You can also

<Web server configuration>
Next, the configuration of the web server 600 will be described. The web server 600 includes a CPU 610, a memory 620, a port 630, a recording medium reading device 640, an input device 650, an output device 660, and a storage device 680. The functions of these devices are the same as those of the devices included in the management computer 200 described above.

  As illustrated in FIG. 4, a web server control program 930 and an agent program 940 are stored in the storage device 680 included in the web server 600. Various functions as the web server 600 are realized by the CPU 610 executing the web server control program 930 and the agent program 940. For example, the CPU 610 executes a web server control program 930 and an agent program 940 and cooperates with hardware devices such as the memory 620, the port 630, the input device 650, the output device 660, and the storage device 680. A processing request number storage unit, a determination value storage unit, an index value calculation unit, a transmission determination unit, a message transmission unit, and a load detection unit are realized. Here, the information processing execution unit executes information processing according to the processing request received from the client 100. The processing request number storage unit stores the number of processing requests received from the client 100. The determination value storage unit stores a determination value used for determining whether to transmit a message to the management computer 200. The index value calculation unit calculates an index value used for determining whether or not to transmit a message to the management computer 200 based on the number of processing requests stored in the processing request number storage unit. The transmission determination unit determines whether or not to transmit a message to the management computer 200 by comparing the calculated index value with the determination value stored in the determination value storage unit. The message transmission unit transmits a message to the management computer 200 according to the determination result. The load detection unit detects a value indicating the load of the web server 600.

  The web server control program 930 is a program that executes information processing according to a processing request transmitted from the client 100 and transmits a processing result for each processing request to the client 100. The web server control program 930 also provides a function for controlling the web server 600 like an operating system. In addition, in order to provide various information processing services to the client 100, a function as an application program can be provided.

  The agent program 940 is a program that collects various information for monitoring the web server 600 and transmits the information to the management computer 200. For example, the CPU usage rate and CPU usage time of the web server 600, the memory usage amount, the memory usage rate, the storage capacity of the storage device 680, the amount of data read from and written to the storage device 680, and the data transmitted to and received from other computers , The response time from when a processing request is received from the client 100 until the processing result is transmitted, the number of processing requests received from the client 100 within a predetermined time, and currently being processed according to the processing request received from the client 100 A value indicating the load of the various web servers 600 such as the number of processes, and various other information relating to the occurrence of a failure or abnormality are collected and transmitted to the management computer 200.

  Here, the agent program 940 includes a metric monitoring program 1810 and a service providing program 1820. This is shown in FIG.

  First, a processing example of the service providing program 1820 is shown in FIG. The service providing program 1820 is started together with the agent program 940 and repeats the processing from step S2210 to step S2230 until an instruction to stop the agent program 940 is issued (S2200). In step S2210, the web server 600 receives a request from the management computer 200. If there is a request from the management computer 200, the web server 600 determines the content of the request (S2220), and executes processing according to the requested content (S2230).

  If the processing according to the request content is, for example, a metric acquisition request from the management computer 200 (a transmission request for information for calculating a new determination value), it is designated from the metric history storage area 850 described later. Metric data is acquired and transmitted to the management computer 200. If the threshold (allowable value) setting request is received, a new allowable value is received from the management computer 200, and the metric monitoring threshold value (allowable value 820 for the number of processing requests and allowable value 830 for the load value) is set. Change to a new threshold.

  Next, a processing example of the metric monitoring program 1810 is shown in FIG. The web server 600 executes the metric monitoring program 1810, and based on the value indicating the load on the web server 600 and the number of processing requests received from the client 100 by the web server 600 from other programs that perform OS functions and metric measurement. The metric value calculated is collected (S1910). As a value indicating the load of the web server 600, for example, the “current CPU usage rate” can be used. Further, the value calculated based on the number of processing requests received from the client 100 by the web server 600 can be, for example, “the number of processing cases from the previous survey to the current survey”. Of course, the number of processing requests received from the client 100 may be used. Then, the web server 600 stores these metric values in the metric history storage area 850 as history data set with the collection time (S1910). The value calculated based on the number of processing requests received by the web server 600 from the client 100 is used as an index value used for determining whether or not to send a message to the management computer 200. The calculation of the index value may be performed by the web server 600 executing the metric monitoring program 1810. The CPU usage rate may be measured every time the metric acquisition request is received from the management computer 200.

  Next, the web server 600 obtains an allowable value 820 for the number of processing requests (determination value used for determining whether or not to send a message to the management computer 200) and an allowable value 830 indicating a load (S1920). . These values are transmitted when a threshold setting request is transmitted from the management computer 200 and stored in the storage device 680. Also, these values can be repeatedly transmitted from the management computer 200, for example, periodically, and updated to new values each time. By doing so, these values can be maintained in a state in which the operation state of the computer system at that time is reflected, and the load fluctuation of the web server 600 can be detected more accurately.

  Next, the web server 600 refers to the history data stored in the metric history storage area 850, calculates a value calculated based on the number of processing requests received within a predetermined time at the current time, and its allowable value 820. Are compared (S1930). If the value calculated based on the number of processing requests received within the predetermined time at the current time is greater than the allowable value 820 (if the value of the processing request allowable value 820 indicates the “lower limit value”, the processing When the request is smaller than the allowable value 820, it is determined that a message is to be transmitted to the management computer 200, and an event notification (message transmission) is performed to the management computer 200 (S1940). As an event notification method, there are various existing methods such as JMS (Java (registered trademark) Message Service), and the description thereof will be omitted.

  The above processing by the agent program 940 can be repeatedly executed, for example, periodically by an OS scheduler function (such as UNIX (registered trademark) cron) or an existing job management program. In this way, repeated or periodic metric monitoring can be performed, and an event can be quickly notified to the management computer 200 as the load increases or decreases.

  FIG. 20 shows an example of the relationship between the metric monitoring interval by the execution of the metric monitoring program 1810 and the threshold (allowable value) adjustment interval by the execution of the setting adjustment program 910.

  In the example shown in FIG. 20, the metric monitoring program 1810 is activated every 10 seconds to monitor metrics, and the setting adjustment program 910 is activated every minute to adjust the threshold based on the CPU usage rate and the number of processes. ing. That is, the load increase is detected every 10 seconds, and the fine adjustment of the threshold using the CPU usage rate and the number of processing cases is performed every minute. In the example shown in FIG. 20, the load increases from time t18, and is detected at time t19 and a server is added. Here, if the load increase is detected based on the average value of the CPU utilization for one minute, the detection is delayed until time t24. As described above, the metric monitoring program 1810 can detect an increase in load at a more appropriate timing by starting at a short interval that is different from the threshold adjustment interval using the CPU utilization rate.

  Next, as another method for realizing load fluctuation detection by the metric monitoring program 1810, it is possible to detect when the number of processing cases exceeds a threshold. FIG. 21 shows an implementation example of a metric monitoring program that can detect an increase in load when the threshold is exceeded. Note that the implementation method of the metric monitoring program 1810 does not limit the present invention.

  The web server 600 activates a request processing thread each time a request (processing request) is received from the client 100. When activated, the request processing thread first counts up (increments) the current number of processes (S2110). Then, the current processing number is compared with the threshold (S2120). If the current processing number is larger, it is determined that an event notification is to be sent to the management computer 200, and the event notification is performed (S2130).

  In this way, regardless of the load increase determination interval, it becomes possible to immediately notify the management computer 200 of an event when the current number of processing cases exceeds the threshold.

  Here, every time the step S2120 is satisfied, an event notification is performed every time, so that the management computer 200 is notified of an event more than necessary. For this reason, after the event notification is performed, it is desirable to take a measure such that the event notification is not performed again until the current number of processes is sufficiently reduced. In this case, once the event notification is made, the web server 600 stores information indicating that the event notification has been performed in the storage device 680 and the memory 620, and while this information is stored, Event notification is not performed regardless of the result of the comparison between the number of processes and the threshold. Then, when the current number of processing cases becomes smaller than the threshold by a predetermined value or more, the information is deleted. In this way, unnecessary event notification to the management computer 200 can be prevented.

  Next, the web server 600 requests the web server control program 930 to process the request (S2140). After returning a response to the request, the web server 600 counts down the current number of processes (S2150). As a result, event notification to the management computer 200 can be detected when the current number of processes exceeds the threshold.

  Here, every time a request processing thread is activated, the current processing number is counted up. However, every time a request is received from the client 100, the number of received requests is counted up by one. good. In this case, a load change is detected based on a result of comparison between the number of received requests and a threshold value, and an event is notified to the management computer 200. In this case, the event notification can be performed to the management computer 200 without activating the request processing thread, so that the event notification can be performed more quickly.

  Here, as the value indicating the load of the web server 200, various values such as the CPU usage rate, the CPU usage time, and the memory usage can be used as exemplified above. Any of these can be used as an index representing the load of the web server 200, but each has its own characteristic as shown in FIG.

  For example, the CPU utilization rate has a feature that the load on the web server 200 can be directly known. That is, it can be seen that, even if the computer is a high-performance computer or not, for example, if the CPU utilization rate is 1%, the load is still sufficient, and if it is 99%, the computer is overloaded. On the other hand, the CPU utilization rate has a feature that it fluctuates greatly when the load on the web server 200 varies even slightly. Therefore, it is necessary to carefully determine whether the load on the web server 200 is really overloaded. When the CPU usage rate is used as a load index, it is necessary to determine the state of the load by taking an average value within a predetermined time, checking whether a threshold value has been exceeded for a predetermined time, or the like. For this reason, when the CPU usage rate is used as an index of load, it takes time for the determination. Another example of an index having the same characteristics as the CPU usage rate is a memory usage rate.

  In addition, for example, the response time from the reception of the processing request from the client 100 to the transmission of the processing result can be used as an index of the load of the web server 200. In this case, from the viewpoint close to the user's standpoint It has the feature that can be evaluated. On the other hand, since it has the feature that the response time becomes longer after the load that the web server 200 cannot process is applied, it will be slow to deal with it after detecting the load increase.

  Further, a value calculated based on the number of processing requests received from the client 100 such as the number of processing requests received from the client 100 within the predetermined time and the number of processing requests currently processed is used as an index of the load of the web server 200. You can also However, in this case, the processing amount currently being processed by the web server 200 can be known, and if a threshold value (allowable value) is set, the load immediately increases when requests exceeding the threshold value are accepted. It is possible to detect this. On the other hand, even if the number of processes is the same, there is a feature that it is difficult to set an appropriate threshold because the load on the web server 200 differs depending on the processing performance of the web server 200 and the cost per process. Examples of the index having the same characteristics as the value calculated based on the number of processing requests include the number of database transactions and the number of connections.

  In the present embodiment, an allowable value (for example, an upper limit value) of the CPU usage rate is set in advance as the allowable value 830 of the value indicating the load of the web server 200, and the web server 200 is within a predetermined time from the client 100. A value obtained by multiplying the number of received processing requests by the value of the ratio of the CPU usage rate allowable value to the average value of the CPU usage rate of the web server 200 within the predetermined time is obtained from the client 100 by the web server 200. It is calculated as an allowable value (upper limit) 820 of the number of processing requests received within a predetermined time.

  In this case, when the web server 200 receives a processing request exceeding the allowable value 820 calculated in this way from the client 100, the CPU utilization rate of the web server 200 also exceeds the preset allowable value 830. It can be predicted that there will be. As described above, when it is detected that the number of processing requests received from the client 100 exceeds the allowable value (threshold value) 820, it can be immediately determined that the load has increased. By detecting that the number of processing requests that the web server 200 receives from the client 100 within a predetermined time exceeds the calculated allowable value 820, the CPU usage rate of the web server 200 is allowed. It can be immediately determined that the value 830 has been exceeded.

  In order to calculate the allowable value 820, the agent program 940 according to the present embodiment also calculates the average value of the CPU utilization rate within a predetermined time. Further, the number of processing requests received from the client 100 within a predetermined time is compared with the calculated allowable number of processing requests 820, and if it is detected that the allowable value 820 has been exceeded, the management computer A message to that effect is also sent to 200.

  Note that the calculation of the allowable value 820 of the number of processing requests received from the client 100 within a predetermined time can be performed by the management computer 200 or each web server 600.

  In the former case, each web server 600 sends the average value of the CPU utilization rate within a predetermined time (10 seconds in the example of FIG. 20) to the management computer 200 and the processing received from the client 100 within the predetermined time. Send the number of requests respectively. An allowable value 820 of the number of processing requests received within a predetermined time from the client 100 calculated by the management computer is received and stored in the memory 620 or the storage device 680. Each web server 600 compares the number of processing requests received from the client 100 within a predetermined time with the allowable value 820 of the calculated processing request number, and detects that the allowable value 820 has been exceeded. Is sent to the management computer 200 to that effect. The allowable value 820 may be updated repeatedly, for example, periodically (every minute in the example of FIG. 20). As a result, even if the content of the processing request transmitted from the client 100 changes with time, it is possible to accurately detect the load fluctuation of the web server by comparing the number of processing requests with the allowable value 820.

  On the other hand, in the latter case, each web server 600 is set in advance with a CPU usage rate allowable value (for example, 80% as the upper limit value) as the allowable value 830 indicating the load, and the CPU usage within a predetermined time period. The average value of the rate is calculated. Then, an allowable value 820 of the number of processing requests received from the client 100 within a predetermined time is calculated and stored in the memory 620 or the storage device 680. Then, a comparison is made between the number of processing requests received from the client 100 within a predetermined time and the allowable value 820 of the calculated processing request number, and if it is detected that the allowable value 820 has been exceeded, the management computer 200 is notified. Send a message to that effect. As in the former case, the allowable value 820 can be repeatedly updated, for example, periodically.

<Configuration of load balancer>
Next, the configuration of the load distribution apparatus 700 will be described. The load distribution apparatus 700 includes a CPU 710, a memory 720, a port 730, a recording medium reading device 740, an input device 750, an output device 760, and a storage device 780. The functions of these devices are the same as those of the devices included in the management computer 200 described above.

  As illustrated in FIG. 5, a load balancer control program 950, an agent program 940, and a weighting table 840 are stored in the storage device 780 included in the load balancer 700. When the CPU 710 executes the load balancer control program 950 and the agent program 940, various functions as the load balancer 700 are realized.

  The load balancer control program 950 is a program for distributing and transmitting each processing request transmitted from the client 100 to the web server 600 to each web server 600. Each processing request is distributed according to a distribution ratio stored in the weighting table 840. An example of the weighting table 840 is shown in FIG. In the example shown in FIG. 9, each processing request transmitted from the client 100 is distributed to five web servers (A to E) 200 at 10: 6: 5: 8: 5. This distribution ratio can be repeatedly updated, for example, periodically. Thereby, even if the load state of the web server 600 changes with time, the processing requests from the client 100 can be appropriately distributed to each web server 600. The distribution ratio can be updated by the management computer 200, for example. The load balancer control program 950 also provides a function for controlling the load balancer 700 like an operating system.

  The agent program 940 is a program that collects various information for monitoring the load distribution apparatus 700 and transmits the information to the management computer 200. For example, the distribution ratio stored in the weighting table 840 is periodically transmitted to the management computer 200, and the new distribution ratio transmitted from the management computer 200 is stored in the weighting table 840.

<Client configuration>
Next, the configuration of the client 100 will be described. The client 100 includes a CPU 110, a memory 120, a port 130, a recording medium reading device 140, an input device 150, an output device 160, and a storage device 180. The functions of these devices are the same as those of the devices included in the management computer 200 described above.

  As illustrated in FIG. 6, a client control program 960 is stored in the storage device 180 included in the client 100. Various functions as the client 100 are realized by the CPU 110 executing the client control program 960.

  The client control program 960 transmits a processing request to the web server 600 or receives a processing result from the web server 600 when an employee of a company uses an information processing service provided by a computer system. It is a program to do. A function for performing control as an operating system may be included.

=== Specific Example of Processing ===
Next, the content of processing when the web server 600 calculates the allowable value 820 of the number of processing requests received from the client 100 within a predetermined time, updates the distribution ratio stored in the load balancer 700, and the like. This will be specifically described with reference to FIGS.

  In the example illustrated in FIG. 10, for example, 1020 processing requests per minute are transmitted from each client 100 to the load balancer 700.

  In the load balancer 700, a distribution ratio of “10: 6: 5: 8: 5” is set for the five web servers 600. Thereby, 300 processing requests per minute (1020 × 10/34) are transmitted to the web server A600, 180 processing requests per minute are transmitted to the web server B600, and 150 processing requests per minute are transmitted to the web server C600. Processing requests are transmitted, 240 processing requests per minute are transmitted to the web server D600, and 150 processing requests per minute are transmitted to the web server E600.

  It is assumed that the average value of the CPU usage rate of each web server 600 at that time is 30%, 45%, 50%, 30%, and 75% for each of the web server A600 to the web server E600.

  In addition, it is assumed that 500 cases, 300 cases, 250 cases, 400 cases, and 250 cases are set in each web server 600 as the allowable values 820 of the processing requests received from the client 100, respectively.

  In this case, the management computer 200 follows the flowchart of the setting adjustment program 910 shown in FIG. 11, and the permissible value 820 of the processing request received from the client by each web server 600 and the load balancer 700. A value indicating the distribution ratio is calculated and transmitted to each web server 600 and the load balancer 700.

  First, the management computer 200 acquires information such as the IP addresses of the web servers 600 and the load balancer 700 from the monitoring target list table 800 (S1000). Then, the management computer 200 acquires the distribution ratio stored in the weighting table 840 of the load balancer 700 from the load balancer 700 (S1010). Next, the management computer 200 acquires the average value of CPU utilization for one minute and the number of processing requests received from the client 100 for one minute from each web server 600 (S1020).

  Then, the management computer 200 calculates a value indicating the new distribution ratio (S1030). The calculation of the value indicating the new distribution ratio is performed by calculating “the number of processing requests received from the client 100 per minute / average value of CPU utilization for one minute” for each web server 600 and calculating the ratio of the respective values. Calculation is performed by using a value indicating a new distribution ratio. That is, in the example shown in FIG. 10, the value indicating the new distribution ratio is “(300/30) :( 180/45) :( 150/50) :( 240/30) :( 150/75)” = “ 10: 4: 3: 8: 2 ".

  Next, the management computer 200 calculates a permissible value 820 for the number of processing requests received per minute from the client 100 for each web server 600 (S1040). Assuming that 80% is set in advance in the management computer 200 as the allowable value of the CPU utilization rate of each web server 600, the average number of received 1 minute processing requests and the CPU utilization rate for 1 minute at that time. From the value, the number of processing cases where the CPU usage rate is expected to be 80% is calculated. For example, in the case of the web server A600, it is calculated as 300 × (80/30) = 800 cases. In the case of the web server B600, 180 × (80/45) = 320 cases are calculated. In the case of the web server C600, 150 × (80/50) = 240 cases are calculated. In the case of the web server D600, 240 × (80/30) = 640 cases are calculated. In the case of the web server E600, 150 × (80/75) = 160 cases are calculated.

  Then, the management computer 200 transmits a value indicating the new distribution ratio to the load balancer 700 (S1050). Then, the permissible value 820 of the calculated number of processing cases is transmitted to each web server 600 (S1060).

  Thereby, a new distribution ratio and a processing request allowable value 820 can be set for the load balancer 700 and each web server 600. This setting can be repeated, for example, every minute.

  When the calculated new distribution ratio and the allowable value 820 of the number of processing requests are greatly different from the previously calculated values, the number of processing requests distributed to each web server 600 and the load of each web server 600 are different. It may become unstable.

  For this reason, in the present embodiment, a new distribution ratio can be calculated based on the average values of the calculated distribution ratio values and the previously calculated distribution ratio values. That is, the management computer 200 stores the values obtained by dividing the number of processing requests previously calculated for each web server 200 by the average value of the CPU usage rate, and newly calculated for each web server 600. The ratio of each average value between each value obtained by dividing the number of processing requests by the CPU utilization average value and each previous value may be calculated as a new distribution ratio. it can.

  In addition, the average value of the calculated allowable number of processing requests and the previously calculated value may be used as the allowable value of the number of new processing requests.

  In this way, when the allowable values of the distribution ratio and the number of processing requests are updated as needed, the change of the values can be moderated, so that more stable load distribution control can be performed. .

  Of course, the new distribution ratio is not limited to the average value of each value of the calculated distribution ratio and each value of the previously calculated distribution ratio. It can be a value between each value. Similarly, the allowable value of the number of new processing requests is not limited to the average value of the calculated allowable number of processing requests and the previously calculated value. It can be a value between the calculated values.

=== Processing when load fluctuation is detected ===
As described above, each web server 600 in which the processing request allowable value 820 received from the client 100 is set determines whether the number of processing requests from the client 100 exceeds the allowable value 820. This determination can be made by executing the agent program 940 described above. If it is detected that the allowable value 820 is exceeded, the web server 600 transmits a predetermined message (for example, the message shown in FIG. 12) to the management computer 200.

  On the other hand, the management computer 200 executes the event handling program 920 according to the flow shown in FIG. When the management computer 200 receives the message from the web server 600, the management computer 200 analyzes the received message, confirms “with which web server” and “what” has occurred (S2000 to S2020), and stores the event response list table 810. Take corrective action according to the description (S2030).

  Through the above processes, the load fluctuation of each web server 600 can be immediately detected by the fluctuation of the number of processing requests transmitted from the client 100 to the computer system. By performing the above-described autonomous control, it is possible to maintain the load on the computer system in an appropriate state without particularly obtaining an instruction from the operator. As a result, the reliability of the computer system can be improved, and a high-quality information processing service can be provided to a user who uses the computer system through the client 100.

=== Other examples ===
In this embodiment, as shown in FIG. 10, the allowable value 820 of the processing request from the client 100 is set for each web server 600, but it may be as shown in FIG. 15 and FIG. it can. That is, an allowable value 820 for a processing request from the client 100 is set for the load balancer 700.

  In this case, first, the management computer 200 acquires information such as the IP addresses of the web servers 600 and the load balancer 700 from the monitoring target list table 800 (S3000). Then, the management computer 200 acquires the distribution ratio stored in the weighting table 840 of the load balancer 700 from the load balancer 700 (S3010). Next, the management computer 200 acquires the average value of CPU utilization for one minute and the number of processing requests received from the client 100 for one minute from each web server 600 (S3020).

  Then, the management computer 200 calculates a new distribution ratio (S3030). The processing content for calculating the new distribution ratio is the same as in the case of FIG.

  Next, the management computer 200 calculates the permissible value 820 of the number of processing requests received from the client 100 per minute for each web server 600 in the same manner as in FIG. 10 (S3040). Then, the management computer calculates the total value of each allowable value (S3050).

  Then, the management computer 200 transmits the total value of the above allowable values and the new distribution ratio to the load balancer 700 (S3060).

  In the example shown in FIG. 15, the allowable value 820 of the number of processing requests set in the load distribution apparatus 700 can be the total value of the allowable values calculated for each web server 600 in FIG. Also by doing in this way, the allowable value 820 of the number of processing requests transmitted from the client 100 can be set according to the load of each web server 600 constituting the computer system. Furthermore, when the number of processing requests transmitted from the client 100 to the load balancer 700 exceeds the allowable value 820, a change in load can be detected and a response can be immediately taken.

  Further, as shown in FIG. 1, the computer system according to the present embodiment includes a management computer 200, a load balancer 700, and a web server 600, but further includes an AP server (application server) as shown in FIG. ) 300 and a DB server (database server) 400 may be provided. The configuration shown in FIG. 17 is a configuration generally known as a web three-layer system. In this case, when the web server 600 receives a processing request from the client 100, the web server 600 transmits an information processing execution request corresponding to the processing request to the AP server 300. Then, the AP server 300 performs information processing in response to the processing request and transmits the processing result to the web server 600. The web server 600 transmits the processing result to the client 100. When executing the information processing, the AP server 300 transmits a data read / write request to the database to the DB server 400 as necessary. The DB server 400 accesses the database in response to the read / write request and transmits the result to the AP server 300.

  The web server 600 stores an allowable value 820 of the number of processing requests received from the client 100 within a predetermined time, and the web server 600 receives a processing request exceeding the allowable value from the client 100. First, a message is transmitted to the management computer 200. The AP server 300 stores an allowable value of the number of processing requests received from the web server 600 within a predetermined time, and the AP server 300 receives a processing request exceeding the allowable value from the web server 600. First, a message is transmitted to the management computer 200. The DB server 400 stores an allowable value of the number of transactions transmitted from the AP server 300 within a predetermined time, and the DB server 400 receives a transaction exceeding the allowable value from the AP server 300. Transmits a message to the management computer 200. In the same manner as described above, these permissible values are a value indicating the load of each server, a value calculated based on the number of processing requests received by each server, It is calculated based on the allowable value of the value indicating the load.

  By doing so, each server can immediately transmit a load increase message to the management computer 200 when it is detected that the number of processing requests exceeds the allowable value. When the management computer 200 receives a load increase message from these servers, the management computer 200 immediately reduces the load on the server that has transmitted the message by, for example, adding a spare server to the computer system. It becomes possible. Of course, each server may transmit a load increase message to the management computer 200 when it is detected that the current number of processes exceeds the allowable value.

  Further, the computer system according to the present embodiment may be configured as shown in FIG. That is, the management computer 200 is connected to a DBMS (DataBase Management System, DataBase Management Server). The DBMS receives a transaction (processing request) from the client 100 and transmits the execution result of the transaction to the client 100. As in the above embodiment, the management computer 200 is based on a value indicating the DBMS load such as a CPU usage rate, an allowable value thereof, and a value calculated based on the number of transactions received in a predetermined time. Then, an allowable value of the number of transactions received in a predetermined time is calculated and transmitted to the DBMS. When the DBMS receives a transaction exceeding the allowable value from the client 100, the DBMS immediately transmits a message to the management computer 200.

 The computer system according to the present embodiment can be configured as shown in FIG. That is, the management computer 200 is connected to a NAS (Network Attached Storage) server. The NAS server receives a data read / write command (processing request) from the client 100 and transmits the execution result to the client 100. As in the above embodiment, the management computer 200 has a value indicating the load of the NAS server such as a CPU utilization rate, an allowable value thereof, a value calculated based on the number of read / write commands received in a predetermined time, Based on the above, an allowable value of the number of read / write commands received in a predetermined time is calculated and transmitted to the NAS server. When the NAS server receives a read / write command exceeding the allowable value from the client 100, the NAS server immediately transmits a message to the management computer 200.

  Furthermore, the computer system according to the present embodiment may be configured as shown in FIG. That is, the management computer 200 is connected to a server or PC (Personal Computer) that executes batch jobs. This server receives a batch processing request (processing request) from the client 100, and executes the batch processing at a predetermined time according to the request. As in the above embodiment, the management computer 200 calculates a value indicating the server load such as a memory usage rate, a permissible value thereof, and a value calculated based on the number of requests for batch processing (for example, a currently executing process). Based on the number of batch processes), an allowable value of the number of batch processes that can be executed simultaneously is calculated and transmitted to the server. The server immediately sends a message to the management computer 200 when the number of batch processes exceeding the allowable value is executed.

  The computer system according to the present embodiment can also be configured as shown in FIG. That is, the management computer 200 is connected to the storage controller. The storage controller receives a data read / write command (processing request) from the client 100 and transmits the execution result to the client 100. The storage controller manages the amount of data stored in the storage device. As in the above embodiment, the management computer 200 sets a value indicating the load of the storage controller, such as the rate of decrease in the free capacity of the storage device (0.1% / day), and its allowable value (25% / day). Based on a value calculated based on the number of data read / write instructions (for example, an increase amount of data stored in the storage device (1 GByte / day)) and an increase amount of data stored in the storage device An allowable value (250 Gbyte / day) is calculated and transmitted to the storage controller. When the storage controller detects an increase in the amount of data exceeding the allowable value, it immediately sends a message to the management computer 200.

  Although the present embodiment has been described above, according to the present embodiment, an allowable value indicating a load on a server such as the web server 200 is set in advance, and the number of processing requests received from the client 100 by the server is set. Based on the value calculated based on the value, the allowable value indicating the load, and the value indicating the server load, the allowable value calculated based on the number of processing requests is calculated. To do. For this reason, when the value calculated based on the number of processing requests received from the client 100 by the server exceeds the allowable value, the value indicating the load on the server also exceeds the preset allowable value. It can be predicted. When it is detected that the value calculated based on the number of processing requests received from the client 100 exceeds the allowable value, it can be immediately determined that the load has increased, so the server receives from the client 100. By detecting that the value calculated based on the number of processing requests exceeds the calculated allowable value, it is possible to immediately determine that the value indicating the server load exceeds the allowable value. It can be done. As a result, load fluctuations can be detected quickly, and measures such as adding a server in the computer system can be started immediately. In particular, in the present embodiment, even when servers having different performances are mixed in the computer system, or even when the access frequency and access pattern from the client 100 differ depending on the time zone and time, the load variation is allowed. It is possible to quickly detect a load fluctuation of each server without setting complicated values such as setting a value or a judgment value for each server or setting according to a time zone.

  Further, an increase in load is detected with a value calculated based on the number of processing requests, for example, while periodically checking a value calculated based on the number of processing requests at that time and a value indicating the load, an allowable value Tweak the. Thereby, it becomes possible to detect an increase in load when a value calculated based on the number of processing requests becomes equal to or greater than an allowable value. In other words, the fine adjustment processing of the allowable value uses a value indicating the load, and thus requires a certain time. However, the load increase detection is the allowable value of the value calculated based on the number of processing requests at that time. Can be performed more quickly. Furthermore, it is possible to deal with cases where the processing performance of each web server 600 is different, or cases where the processing content per processing request varies depending on the time zone.

  In addition, in order to update the distribution ratio of the load balancer 700 in synchronization with the allowable value, when the value calculated based on the number of processing requests of one server exceeds the allowable value, You can see that the other servers are also heavily loaded. Therefore, when the management computer 200 receives a message from a certain server, the management computer 200 can immediately start a process such as adding a server without making a determination such as “Is to deal with load leveling” or the like. is there.

  Although the best mode for carrying out the invention has been described above, the above embodiment is intended to facilitate understanding of the present invention and is not intended to limit the present invention. The present invention can be changed and improved without departing from the gist thereof, and the present invention includes equivalents thereof.

It is a block diagram which shows the whole structure of the computer system concerning this Embodiment. It is a block diagram which shows the structure of the client which concerns on this Embodiment, a management computer, a load distribution apparatus, and a web server. It is a figure which shows the memory | storage device of the management computer which concerns on this Embodiment. It is a figure which shows the memory | storage device of the web server which concerns on this Embodiment. It is a figure which shows the memory | storage device of the load distribution apparatus which concerns on this Embodiment. It is a figure which shows the memory | storage device of the client which concerns on this Embodiment. It is a figure which shows the monitoring object list table which concerns on this Embodiment. It is a figure which shows the event handling list table which concerns on this Embodiment. It is a figure which shows the weighting table which concerns on this Embodiment. It is a figure which shows the specific example of the process which concerns on this Embodiment. It is a flowchart which shows the flow of a process of the setting adjustment program which concerns on this Embodiment. It is a figure which shows the event message format which concerns on this Embodiment. It is a flowchart which shows the flow of a process of the event response program which concerns on this Embodiment. It is a figure which shows the example of the value which shows the load which concerns on this Embodiment. It is a figure which shows the other specific example of the process which concerns on this Embodiment. It is a flowchart which shows the flow of a process of the setting adjustment program which concerns on this Embodiment. It is a figure which shows the other structural example of the computer system which concerns on this Embodiment. It is a figure which shows the structure of the agent program which concerns on this Embodiment. It is a flowchart which shows the flow of a process of the metric monitoring program which concerns on this Embodiment. It is a figure which shows the threshold value adjustment interval and monitoring interval which concern on this Embodiment. It is a flowchart which shows the flow of a process of the metric monitoring program which concerns on this Embodiment. It is a flowchart which shows the flow of a process of the service provision program which concerns on this Embodiment. It is a block diagram which shows the structural example of the computer system which concerns on this Embodiment. It is a block diagram which shows the structural example of the computer system which concerns on this Embodiment. It is a block diagram which shows the structural example of the computer system which concerns on this Embodiment. It is a block diagram which shows the structural example of the computer system which concerns on this Embodiment.

Explanation of symbols

100 client 200 management computer 210 CPU
220 Memory 230 Port 240 Recording Medium Reading Device 250 Input Device 260 Output Device 280 Storage Device 300 AP Server 400 Database Server 500 First Network 510 Second Network 600 Web Server 700 Load Balancing Device 600 Monitoring Object List Table 610 Event Handling List Table 620 Load increase judgment value 630 Allowable value indicating load 840 Weighting table 850 Metric history storage area 900 Management server control program 910 Setting adjustment program 920 Event handling program 930 Web server control program 940 Agent program 950 Load balancer control program 960 Client control program 1000 Recording medium

Claims (20)

A control method for a management computer communicably connected to one or a plurality of servers for transmitting a processing result for each processing request transmitted from one or a plurality of clients to the client,
The management computer is
For each of the one or a plurality of servers, an absolute load allowable value that is a countable value indicating the absolute amount of load applied to the server is calculated, and the calculated absolute load allowable value is transmitted to the server. A value indicating the relative load according to the setting adjustment program and the processing capacity of each of the one or more servers, and the degree of the load is represented by the same value regardless of the processing capacity of each server. A storage device for storing a relative load allowable value,
A processing device for reading the setting adjustment program into a memory and executing the program;
A communication device for communicating with the one or more servers;
With
The processing device is
For each one or a plurality of servers, the absolute load of the server and the relative load are received from the server using the communication device,
For each of the one or a plurality of servers, the absolute load received from each server using the communication device, the allowable value of the relative load stored in the storage device, and each of the servers using the communication device From the relative load value received from the server, the allowable value of the absolute load of each server is calculated,
A control method for a management computer, wherein the absolute load tolerance calculated for each of the one or a plurality of servers is transmitted to each of the servers using the communication device to request a setting. . The absolute load tolerance is
Wherein the absolute load, control method for a management computer according to claim 1, characterized in that is calculated by multiplying the value of the ratio for the relative load tolerance of the relative load. The relative load is a utilization rate of a processing device provided in the server,
2. The management computer control method according to claim 1, wherein the absolute load is the number of processing requests received by the server from the client within a predetermined time. The management computer repeatedly receives the absolute load and the relative load from the server,
Each time the management computer receives the absolute load and the relative load, it calculates an allowable value of the absolute load ,
2. The management computer control method according to claim 1, wherein the management computer transmits the calculated relative load tolerance to the server each time the absolute load tolerance is calculated. The management computer is communicably connected to a plurality of servers that transmit processing results for each processing request transmitted from the client to the client, and stores an allowable value of the relative load .
The management computer receives the absolute load and the relative load from each of the servers;
The management computer calculates an allowable value of the absolute load for each server,
The management computer transmits the absolute load tolerance calculated for each server to each of the servers;
The management computer control method according to claim 1, wherein: The management computer is communicably connected to a load balancer that distributes and transmits each processing request transmitted from the client to each server,
Based on the absolute load and the relative load received from each server, the management computer determines a distribution ratio when the load distribution apparatus distributes and transmits each processing request received from the client to each server. Calculate the value shown,
The management computer transmits a value indicating the calculated distribution ratio to the load balancer;
The management computer control method according to claim 5. The value indicating the distribution ratio is:
7. The management computer control method according to claim 6, wherein the absolute load received from each server is a value indicating a ratio of each value obtained by dividing each absolute load by each relative load . The management computer repeatedly receives the absolute load and the relative load from each of the servers,
The management computer calculates a value indicating the distribution ratio every time the absolute load and the relative load are received,
7. The management computer control method according to claim 6, wherein each time the management computer calculates a value indicating the distribution ratio, the management computer transmits a value indicating the calculated distribution ratio to the load distribution apparatus. The absolute load tolerance is
It said management computer, wherein the absolute load, control method for a management computer according to claim 4, characterized in that is calculated by multiplying the value of the ratio for the relative load tolerance of the relative load. The management computer stores the previously calculated allowable absolute load value as the previously calculated allowable value,
Said management computer, when calculating the permissible value of the absolute load, and a value obtained by multiplying the value of the ratio for the relative load tolerance of the relative load on the absolute load, and the previously calculated tolerance, The management computer control method according to claim 9, wherein an allowable value of the absolute load is calculated based on the control value. The absolute load tolerance is
Administration of claim 10, wherein the value obtained by multiplying the value of the ratio for the relative load tolerance of the relative load on the absolute load, said the previously calculated tolerance, to be the average value Computer control method. The value indicating the distribution ratio is:
9. The management computer control method according to claim 8, wherein the absolute load received from each server is a value indicating a ratio of each value obtained by dividing each absolute load by each relative load . The management computer stores each value obtained by dividing the absolute load previously calculated for each server by the relative load as a previously calculated value,
The management computer calculates a value indicating the distribution ratio based on each value obtained by dividing the absolute load calculated for each server by the relative load and each previous calculated value. The control method of the management computer according to claim 12. The value indicating the distribution ratio is:
14. The value according to claim 13, wherein each of the values obtained by dividing the absolute load calculated for each server by the relative load is a value indicating a ratio of average values of the previous calculated values. Management computer control method. And one or a plurality to perform an information processing corresponding to the processing request received one or a plurality of client server, executes information processing in accordance with the message received from the one or more servers managing An absolute load that is a countable value indicating the absolute amount of load applied to the server based on the number of processing requests received from the one or more clients. An absolute load calculation unit for calculating an allowable value of the absolute load, an absolute load allowable value storage unit for storing the allowable value of the absolute load , a comparison of the calculated absolute load with the allowable value of the absolute load, and a message to the management computer a transmission determination unit for determining whether or not to transmit, a said value indicating the value of the relative load in accordance with each server's processing capacity, said each server Accepts the same values regardless of the processing capability can be indicative of the degree of load load detector for detecting a relative load, the allowable value of the transmitted said absolute load from the management computer, the absolute load tolerance the control method of the control computer in a computer system which Ru and an absolute load allowable value change unit that updates the allowable value of the stored in the storage unit the absolute load,
The management computer is
A setting for calculating an allowable value of absolute load, which is a countable value indicating the absolute amount of load applied to the server, for each of the one or a plurality of servers, and transmitting the calculated absolute load allowable value to the server A value indicating a relative load value corresponding to the adjustment program and the processing capacity of each of the one or more servers, and the degree of load can be expressed by the same value regardless of the processing capacity of the server. A storage device for storing a relative load tolerance,
A processing device for reading the setting adjustment program into a memory and executing the program;
A communication device for communicating with the one or more servers;
With
The processing device is
For each one or a plurality of servers, the absolute load of the server and the relative load are received from the server using the communication device,
For each of the one or a plurality of servers, the absolute load received from each server using the communication device, the allowable value of the relative load stored in the storage device, and each of the servers using the communication device Using the relative load value received from the server for the calculation process, calculating the absolute load allowable value of each server,
A control method for a management computer, wherein the absolute load tolerance calculated for each of the one or a plurality of servers is transmitted to each of the servers using the communication device to request a setting. . And one or a plurality to perform an information processing corresponding to the processing request received one or a plurality of client server, executes information processing in accordance with the message received from the one or more servers managing A method of controlling the server in a computer system comprising:
The management computer is
A setting for calculating an allowable value of absolute load, which is a countable value indicating the absolute amount of load applied to the server, for each of the one or a plurality of servers, and transmitting the calculated absolute load allowable value to the server A value indicating a relative load value corresponding to the adjustment program and the processing capacity of each of the one or more servers, and the degree of load can be expressed by the same value regardless of the processing capacity of the server. A storage device for storing a relative load tolerance,
A processing device for reading the setting adjustment program into a memory and executing the program;
A communication device for communicating with the one or more servers;
With
The server
And an information processing execution unit for executing processing corresponding to the processing request received from the one or plurality of client, the number of the received processing request from said one or plurality of client, of the server the relative load A metric storage unit that stores a value, an absolute load allowable value storage unit that stores the absolute load allowable value, and an absolute load allowable value that is transmitted from the management computer, and stores the absolute load allowable value storage An absolute load allowable value changing unit for updating the absolute load allowable value stored in the unit ,
Based on the number of processing requests stored in the metric storage unit, the absolute load applied to the server is calculated,
By comparing the allowable value of the calculated absolute load and the absolute load allowable value storage absolutely the stored load, a determination of whether to transmit a message to the management computer,
Depending on the result of the determination, send a message to the management computer ,
The absolute load value calculated using the absolute load value of the server, the allowable value of the relative load held by the management computer, and the relative load value of the server is the management computer. And recorded in the absolute load allowable value storage unit . The metric storage unit stores the number of processing requests received from the client within a predetermined time,
The server repeatedly updates the number of processing requests stored in the metric storage unit,
Each time the server updates the number of processing requests, it calculates a new absolute load ,
The server control method according to claim 16, wherein the server performs the determination as to whether or not to transmit a message to the management computer every time the new absolute load is calculated. The server repeatedly receives a transmission request for information for calculating a new allowable absolute load value from the management computer,
Each time the server receives the transmission request, the server transmits a value indicating the load of the server and the calculated absolute load to the management computer,
The server is based on a value indicating the load of the server transmitted by the management computer, an absolute value of the transmitted absolute load, and an allowable value indicating a load of the server stored in the management computer. Receiving the calculated new absolute load tolerance from the management computer;
Wherein the server, each time it receives the allowable value of the new absolute load, updating the allowable value of the absolute load tolerance absolute the stored load in the storage unit to the tolerance of the new the absolute load The server control method according to claim 17, wherein: Each time the server receives a processing request from the client, the server increments the number of processing requests stored in the metric storage unit,
The server calculates the absolute load allowable value every time the number of processing requests stored in the metric storage unit is incremented, and performs the determination as to whether or not to send a message to the management computer. The server control method according to claim 16. The server includes a message transmission storage unit that stores information indicating that a message has been transmitted to the management computer,
The server deletes the information stored in the message transmission storage unit when the calculated absolute load is smaller than a predetermined value by the absolute load allowable value ,
When the information indicating that a message has been transmitted to the management computer is stored in the message transmission storage unit, the server does not depend on the result of the determination as to whether or not to transmit a message to the management computer. 20. The server control method according to claim 19, wherein a message is not transmitted to the management computer.