JP3782429B2 - Load balancing system and computer management program - Google Patents

Description

  The present invention relates to a load distribution system for distributing loads by distributing and processing requests received from a plurality of client computers, and particularly suitable for dealing with a sudden spike state, and It relates to a computer management program.

  In order to efficiently process execution requests from a large number of client computers, a load distribution system that distributes these execution requests to a plurality of computers is known (for example, see Non-Patent Documents 1 and 2). This type of load distribution system, for example, a WEB (web) server load distribution system, includes a plurality of WEB server computers and a load distribution apparatus.

  In a load distribution system of a WEB server, a request (execution request) from a WEB browser or the like executed on a client computer is sent to a load distribution apparatus via an external network. When receiving the request sent from the client computer, the load balancer selects a server computer that should process the request from a plurality of WEB server computers. The selection of the server computer is performed in consideration of not applying a great load only to the specific server computer. Therefore, as a server computer selection method, for example, (1) round robin method, (2) weighted round robin method, or (3) a method of selecting a server computer with the smallest number of sessions at that time is applied. The round robin method is a method of selecting each server computer equally in a certain order. The weighted round robin method is based on the round robin method and changes the frequency with which the computer is selected according to the capability of each server computer.

  The load balancer sends the received request to the selected server computer via the internal network. A WEB server is running on the server computer. The server computer performs processing by the WEB server based on the sent request. Then, the server computer sends a reply to the request to the load balancer via the internal network. The load balancer returns the sent reply to the client computer that sent the request via the external network.

On the other hand, when a failure occurs in an operating computer, the service (business) executed on the computer where the failure has occurred is in another computer in the system, for example, in a standby state (hot standby state). A computer system called a cluster system that can be taken over (failed over) to another computer is known (see, for example, Non-Patent Document 3).
R. Buyya, “High Performance Cluster Computing: Architecture and Systems (Volume 1 & 2)”, 1999, Prentice Hall Tony Bourke, “Server Load Balancing Technology”, O'Reilly Japan, Inc., December 13, 2001 Tetsuo Kaneko, Yoshiya Mori, "Cluster Software", Toshiba Review, Vol.54 No.12 (1999), p.18-21

In the above WEB server load distribution system, the number of WEB server computers that process requests from client computers is generally determined according to the amount of requests and the like. However, in recent years, the number of users of the Internet has increased to a huge number, and the number of requests far exceeding the expectation of the operator of the WEB site tends to concentrate on the load distribution system. in this case,
(A) Response time for requests becomes extremely poor. (B) Connection to a WEB server computer becomes impossible. (C) Troubles caused by the so-called “spike phenomenon” that the WEB server computer goes down often occur.

  If an increase in requests (access) is expected to some extent, the operator of the WEB site can deal with the spike phenomenon by, for example, renting an additional WEB server computer. However, in many cases, it is impossible to anticipate an increase in access and it is difficult to cope with it.

  The present invention has been made in view of the above circumstances, and an object thereof is to provide a load distribution system and a computer management program that can easily cope with a sudden spike phenomenon.

  According to one aspect of the present invention, a plurality of server computers that process requests from a plurality of client computers, and requests from a plurality of client computers are distributed to a plurality of load balancing target computers including the plurality of server computers. A load distribution system including a load distribution apparatus is provided. This load balancing system is a spare computer that can be used as the load balancing target computer, and is a spare computer that is activated when used as the load balancing target computer, and all loads including the plurality of server computers. Server configuration management means for determining the number of spare computers to be newly used as load balancing target computers or to be terminated as load balancing target computers based on the load amount of the balancing target computers, Preliminary computer management means for managing the use status of computers, and the number of units determined by the server configuration management means should be newly used as a load distribution target computer or terminated as a load distribution target computer Preliminary computer management means for determining a spare computer to be used, and a preliminary meter determined by the preliminary computer management means A standby computer start / stop unit for starting or stopping a computer and a backup computer determined by the backup computer management unit are added as load balancing target computers to be load balanced by the load balancing device, or the load balancing Setting change means for excluding from load balancing by the device.

  In a load balancing system having such a configuration, based on the load amount of all load balancing target computers including a plurality of server computers, the use as a load balancing target computer should be newly started or as a load balancing target computer A spare computer to be terminated is determined, and the determined spare computer is started or stopped. Further, the determined spare computer is added as a load distribution target computer to be a load distribution target by the load distribution device or excluded from the load distribution target by the load distribution device. In this way, in the load balancing system configured as described above, even if a "spike state" load occurs, the required number of spare computers are activated and the spare computer is included in the load balancing target. If the `` spike state '' is resolved, the spare computer that was the target of load distribution is stopped and excluded from the load distribution target, so it is possible to predict without always starting many server computers. Can easily cope with difficult spikes.

  Here, a plurality of server sites comprising a set of the plurality of server computers, the load balancer, the server configuration management means and the load balancer setting change means, the spare computer, the spare computer management means and the spare Prepare a spare site with a set of computer start / stop means and configure these sites to be connected to the network, and the server configuration management means in the server site newly distributes the load within the server site. When the number of spare computers to be used as target computers or to be terminated as load balancing target computers is determined, the start or end of use of spare computers in the server site is networked to the spare site. The spare computer management means in the spare site requests the spare computer from the server configuration management means. In accordance with a request for start of use or end of use, a configuration may be adopted in which a spare computer that should start using as a load distribution target computer or end use as a load distribution target computer in the requesting server site is determined. .

  In such a configuration, each server site can use a spare computer in the spare site according to the load state of the server site, that is, only when a “spike state” load occurs.

  In addition, when boot images including operating systems executed on the plurality of server computers are not of the same type, the plurality of server computers are grouped into a plurality of server computer groups according to the type of boot image executed on the computers. It is also possible to adopt a configuration in which management is performed separately. In this case, the server configuration management unit should obtain the load amount of the entire group in the group unit, and based on the load amount, the server configuration management unit should newly start use as a load distribution target computer in the group unit, or load distribution It is preferable to determine the number of spare computers to be used as the target computer and to request the start or end of use of the spare computers in the group. Further, the spare computer management means may determine a spare computer to be used as a load distribution target computer in a group requested by the server configuration management means or to end use as a load distribution target computer. In addition, when the spare computer to be used is determined by the spare computer management means, the spare computer is executed as a boot image to be booted by the server computer in the server computer group in which the use of the spare computer is started. By adding a boot image selection means for selecting a boot image of the type that has been set, that is, a boot image specific to the group, the spare computer may be booted with a boot image specific to the group.

  With this configuration, even if a “spike state” load occurs in a certain server computer group, a spare computer necessary to distribute the load is determined, and the determined spare computer is Then, it is possible to boot with the same type of boot image as that being executed on the server computers in the group so as to be a load balancing target computer belonging to the group. In other words, regardless of the “spike state” load that occurs in any server computer group in the system, a spare computer can be assigned regardless of the type of boot image being executed on the server computer in the server computer group. And a limited number of spare computers can be used effectively. Here, a plurality of server sites comprising a set of a plurality of server computers managed separately into the plurality of server computer groups, the load distribution device, the server configuration management means, and the load distribution device setting change means, Each of the servers is prepared with a spare computer, a spare computer management means, a spare computer start / stop means, a boot image setting means, and a spare site having a set of boot image storage means provided for each type of boot image. It is also possible to adopt a configuration in which the spare computers in the spare site can be commonly used from the site.

  According to the present invention, even if a sudden spike phenomenon occurs, a spare computer is activated and newly used as a load balancing target computer. When the spike phenomenon is settled, the activated spare computer is Since it is excluded from load balancing and stopped, it is possible to easily cope with a sudden spike phenomenon that is difficult to predict without always starting many server computers.

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

[First Embodiment]
FIG. 1 is a block diagram showing the configuration of the load distribution system according to the first embodiment of the present invention. The system shown in FIG. 1 includes a plurality of, for example, six WEB server computers 11-1 to 11-6 and a load balancer 12. The WEB server computers 11-1 to 11-6 and the load balancer 12 are interconnected by a network (internal network) 13 inside the system. The load balancer 12 is connected to a plurality of client computers (not shown) via a network (external network) 14 outside the system.

  The WEB server computers 11-1 to 11-6 process requests from client computers. When receiving a request from a client computer via the external network 14, the load balancer 12 selects a server computer that should process the request from the WEB server computers 11-1 to 11-6. For this selection, for example, any of the selection methods (1) to (3) described in the background art can be used. The load balancer 12 also sends a request from the client computer to the selected server computer. In other words, the load balancer 12 distributes requests from a large number of client computers to the WEB server computers 11-1 to 11-6 for processing, thereby distributing the load.

  The load balancer 12 includes a computer load information acquisition unit 120. The computer load information acquisition unit 120 acquires information (computer load information) indicating the load state of each computer that is the target of load distribution by the load distribution device 12. Here, as the load information of each load distribution target computer acquired by the load distribution device 12, the number of sessions set between the load distribution target computer and the load distribution target computer is used. The load distribution target computer is a computer activated as a load distribution target among the WEB server computers 11-1 to 11-6 and the provisioning node computers 15-1 to 15-4 described below.

  The provisioning node computers 15-1 to 15-4 are standby computers that are normally in a power-off state and are not activated. The provisioning node computers 15-1 to 15-4 are positioned as load distribution target candidates for eliminating the state when a sudden “spike state” occurs. However, provisioning node computers that are already subject to load distribution are excluded from this candidate. In the present embodiment, the “spike state” is regarded as a state in which the load amount of the entire computer that is the load distribution target exceeds a predetermined threshold. The provisioning node computers 15-1 to 15-4 are connected to the internal network 13.

  A management server computer 15 is also connected to the internal network 13. On the management server computer 15, the management server 150 operates. The management server 150 is realized by a CPU (not shown) in the management server computer 15 reading and executing a predetermined computer management program. This computer management program is provided by a storage medium such as a CD-ROM, a flexible disk, or a memory card. The computer management program provided by this storage medium is used by being stored in, for example, a disk drive of the management server computer 15. It is also possible to transfer the computer management program to the management server computer 15 via the external network 14 and store it in the disk drive of the computer 15.

  The management server 150 includes functional elements of a server configuration management unit 151, a provisioning node management unit 152, a provisioning node start / stop unit 153, and a setting change unit 154.

  The server configuration management unit 151 has a function of managing a computer (server computer) that operates as a WEB server. Further, the server configuration management unit 151 has a function of determining the number of provisioning node computers that should start using (should be newly targeted for load distribution) or should end use (should be excluded from load balancing). This determination is made based on the load state of the computer that is the load distribution target by the load distribution device 12. Therefore, the server configuration management unit 151 holds a load threshold / plan usage provisioning node computer number table 151a. The table 151a associates the threshold value of the load amount of the entire load distribution target computer (load threshold value) with the number of provisioning node computers that are to be load distribution targets corresponding to the threshold value (the number of planned use provisioning node computers). Remember.

An example of the load threshold / plan usage provisioning node computer number correspondence table 151a is shown in FIG. In the example of FIG. 5, the table 151a stores each of the three types of load threshold values N _S1 , N _S2, and N _S3 and the number of planned use provisioning node computers in association with each other. The threshold values N _S1 , N _S2 and N _S3 correspond to the first, second and third spike states, respectively. The table 151a is preset in the computer management program, and is loaded into the main memory (not shown) of the computer 15 when the management server computer 15 is activated.

  The provisioning node management unit 152 has a function of managing the usage status of the provisioning node computers 15-1 to 15-4. Further, the provisioning node management unit 152, when the server configuration management unit 151 determines the number of provisioning node computers that should start using or end use, includes the provisioning node computers 15-1 to 15-4. The provisioning node computer that should actually start use or end use is determined (selected).

  The provisioning node start / stop unit 153 has a function of starting or stopping the provisioning node computer determined by the provisioning node management unit 152. The provisioning node start / stop unit 153 starts / stops the provisioning node computer as follows. The provisioning node activation / deactivation unit 153 activates the provisioning node computer using the Wake On LAN technology. The provisioning node activation / deactivation unit 153 sends a special packet specifying the provisioning node computer to be activated to the internal network 13 in order to activate the provisioning node computer. In the special packet, the MAC address of the LAN card (network card) installed in the provisioning node computer is used to specify the provisioning node computer to be activated. The LAN card installed in each provisioning node computer operates as a control circuit for connecting to a network (in this case, the internal network 13). This LAN card stands by for monitoring special packets flowing through the internal network 13. Therefore, a standby current is always supplied to the LAN card in a state where the power of the provisioning node computer is turned off. The LAN card has a function for starting (booting) a provisioning node computer in which the LAN card is mounted when receiving a special packet in which its own MAC address is specified (that is, a special packet addressed to itself). The provisioning node computer having such a LAN card is set in a state where it can always be started (standby state). In addition, the provisioning node start / stop unit 153 stops the provisioning node computer as follows. The provisioning node start / stop unit 153 sends, via the internal network 13, a shutdown command that designates the power-off as an option to the provisioning node computer to be stopped. By sending this command, the provisioning node start / stop unit 153 stops the provisioning node computer. Specifically, for example, in the operating system UNIX (registered trademark), “init” can be used as a shutdown command. The “init” command has an option for specifying power-off. Therefore, when a shutdown command specifying power-off is executed, the power-supply driver is called up to the computer power-off. In order to execute a shutdown command from a remote computer, a remote shell (rss) command may be used.

  The setting change unit 154 has a function of controlling the load distribution device 12 so that the provisioning node computer activated or stopped by the provisioning node activation / deactivation unit 153 is a target of load distribution by the load distribution device 12 or not. Have. That is, the setting change unit 154 permits or stops the distribution of the request from the load distribution device 12 to the provisioning node computer activated or stopped by the provisioning node activation / deactivation unit 153 to the load distribution device 12. Has the function to give.

Next, the operation of the load distribution system of FIG. 1 will be described with reference to the flowchart of FIG. First, the server configuration management unit 151 in the management server computer 15 receives from the computer load information acquisition unit 120 of the load distribution device 12 the WEB server computers 11-1 to 11-1 that are the targets of load distribution by the load distribution device 12. Load information of each computer (load balancing target computer) including 11-5 is acquired (step S1). Therefore, when a provisioning node computer used as a load distribution target computer exists in the provisioning node computers 15-1 to 15-4, load information of the computer is also acquired. Here, as the computer load information of each load distribution target computer, the number of sessions set between each client computer and the load distribution target computer by the load distribution device 12 is acquired. In step S1, the server configuration management unit 151, the computer load information for each load balancing target computer, obtains the load N _S of the whole load balancing target computer. Here by adding the amount of load indicated by the computer load information for each load balancing target computer (visits), the total load N _S is obtained. Incidentally, weights the load by the weight corresponding (e.g. reverse- proportional) to the processing capacity of each load balancing target computer, the sum of the weighted load may be the overall load N _S.

Next, the server configuration management unit 151 refers to the load threshold / planned use provisioning node computer number correspondence table 151a, based on the contents of the table 151a, the current load N plan using provisioning node computer corresponding to _S The number is determined (step S2). In the present embodiment, the contents of the table 151a are as shown in FIG. In this case, the server configuration management unit 151
If N _S <N _S1 , the planned number of provisioning node computers = 0
If N _S1 ≦ N _S <N _S2 , the number of planned use provisioning node computers = 2
If N _S2 ≦ N _S <N _S3 , the planned number of provisioning node computers = 4
If N _S ≧ N _S3 , the planned number of provisioning node computers = 6
Decide like this.

  Next, the server configuration management unit 151 compares the determined planned use provisioning node computer number with the provisioning node computer number currently in use managed by the provisioning node management unit 152 (step S3). If the number of computers is equal, the server configuration management unit 151 determines that it is not necessary to newly allocate a provisioning node computer as a load balancing target computer (that is, a computer to start using). In this case, after waiting for a certain time (step S4), the server configuration management unit 151 repeats the processing from step S1.

  On the other hand, if the number of planned use provisioning node computers is larger (step S3), the server configuration management unit 151 determines that a provisioning node computer needs to be newly allocated as a load balancing target computer. In this case, the server configuration management unit 151 requests the provisioning node management unit 152 to newly start using the provisioning node computer. In response to this, the provisioning node management unit 152 checks whether there is a usable provisioning node computer among the provisioning node computers 15-1 to 15-4 (step S5). If there is no provisioning node computer that can be used, that is, if all the provisioning node computers 15-1 to 15-4 are already in use as load balancing target computers, the provisioning node management unit 152 to the server configuration management unit 151 is notified. Then, after waiting for a fixed time (step S4), the server configuration management unit 151 repeats the processing from step S1.

  On the other hand, if there is a provisioning node computer that can be used, that is, a provisioning node computer that is not currently used (activated), the provisioning node management unit 152 uses the provisioning node as a computer (load balancing target computer) that starts use. A process of assigning computers (step S6) is performed. This process is performed as follows based on the magnitude relationship between the number N of provisioning node computers that can be used and the difference Nd between the numbers of both the calculations. First, if N ≧ Nd, the provisioning node management unit 152 selects a number of provisioning node computers that match the difference Nd from the available provisioning node computers, and assigns them as computers that start use. . If N <Nd, the server configuration management unit 151 selects all the N provisioning node computers that can be used, but assigns them as computers to start using, although the number of computers is insufficient. This computer allocation information is notified from the provisioning node management unit 152 to the server configuration management unit 151, the provisioning node activation / deactivation unit 153, and the setting change unit 154, respectively.

  Then, the provisioning node activation / deactivation unit 153 activates the provisioning node computer allocated (determined) as a computer to be used by the server configuration management unit 151 among the provisioning node computers 15-1 to 15-4 ( Step S7). At this time, the setting change unit 154 controls the load balancer 12 to replace the provisioning node computer activated by the provisioning node activation / deactivation unit 153 (determined to start use by the provisioning node management unit 152) with the load distribution device. 12 is added as a load distribution target computer to be subjected to load distribution by step 12 (step S8). Here, the order of steps S7 and S8 may be reversed. When the target of load distribution by the load distribution device 12 is changed under the control of the setting change unit 154, the server configuration management unit 151 waits for a certain time (step S4) and then repeats the processing from step S1.

  On the other hand, if the planned use provisioning node computer number is smaller than the number of provisioning node computers currently in use (step S3), the server configuration management unit 151 uses the provisioning node computer having the difference Nd between the two computer numbers. It is determined that it is necessary to terminate. In this case, the server configuration management unit 151 requests the provisioning node management unit 152 to end the use of Nd provisioning node computers. In response, the provisioning node management unit 152 determines the number of provisioning node computers equal to the difference Nd between the numbers of both computers to be used out of the currently used provisioning node computers (step S9). The provisioning node computer information to be used, determined by the provisioning node management unit 152, is notified from the provisioning node management unit 152 to the server configuration management unit 151, the provisioning node activation / deactivation unit 153, and the setting change unit 154, respectively. .

  Then, the setting change unit 154 controls the load balancer 12 so that the provisioning node computer whose use has been determined to be excluded is not subject to load balancing by the load balancer 12 (step S10). On the other hand, the provisioning node start / stop unit 153 stops the provisioning node computer whose use is determined to be ended by the provisioning node management unit 152 (step S11). Here, the order of steps S10 and S11 may be reversed. When the number of provisioning node computers equal to the difference Nd is stopped, the server configuration management unit 151 waits for a certain time (step S4) and then repeats the processing from step S1.

As apparent from the above description, if the table 151a shown in FIG. 5 in the Figure 1 applied system load N _S becomes the load threshold N _S1 or (where N _S <N _S2), Two provisioning node computers are used as load balancing target computers. In this state, if the load amount N _S becomes equal to or greater than the load threshold N _S2 (where N _S <N _S3 ), two more provisioning node computers are added as load balancing target computers. In this case, a total of four provisioning node computers, that is, all provisioning node computers 15-1 to 15-4 in the system are used. In this state, loading N _S is if becomes load threshold N _S3 above, must be further added the two provisioning node computer as load balancing target computer. However, at this point, all the provisioning node computers 15-1 to 15-4 in the system are already in use, and there is no provisioning node computer that can be newly used. Therefore, there is no use of more than four provisioning node computers.

Further, in the state of the provisioning computing nodes 15-1 through 15-4 is in use, if the load amount N _S becomes lower than the load threshold N _S2 (where N _S ≧ N _S1), 2 units of provisioning node computer Not subject to load balancing. In this case, two provisioning node computers are used. In this state, if the load amount N _S becomes lower than the load threshold N _S1, further two provisioning node computer are load balanced excluded. In this case, the use of the provisioning node computer is stopped. From this, it is understood that the table 151a is a table in which the use start / use end condition of the provisioning node computer corresponding to the load is defined.

  Thus, according to the load distribution system according to the first embodiment of the present invention, when a spiked load occurs, the provisioning node computer (spare computer) is activated, and the activated provisioning node computer is also In addition, load distribution is performed by the load distribution device 12. In such a state, when the load decreases, the provisioning node computer is excluded from the load distribution target, and the provisioning node computer is stopped. Thereby, it is possible to cope with a sudden spike state without always activating many server computers. In other words, in this embodiment, even if the number of requests far exceeding the expectation of the operator of the system (the system including the WEB server computers 11-1 to 11-6) is concentrated, it is listed in the background art section. The occurrence of troubles (A) to (C) due to the “spike phenomenon” can be avoided.

  In the present embodiment, a configuration in which the management server 150 operates on the independent management server computer 15 is applied. However, the management server 150 may be moved on any of the WEB server computers 11-1 to 11-6, for example. In this case, it is preferable to apply a configuration in which the function provided by the management server 150 does not stop even if a failure occurs in the WEB server computer on which the management server 150 is operating. Specifically, at least the WEB server computers 11-1 to 11-6 may constitute the cluster system described in the prior art document. By applying this cluster system, even if a failure occurs in the WEB server computer in which the management server 150 is operating among the WEB server computers 11-1 to 11-6, the function of the management server 150 (that is, server configuration management). The operation of the unit 151, the provisioning node management unit 152, the provisioning node activation / deactivation unit 153, and the setting change unit 154) can be taken over by another WEB server computer. Further, the number of WEB server computers and the number of provisioning node computers are not limited to the above-described embodiment, and there may be at least two (that is, a plurality of) WEB server computers and at least one provisioning node computer. The computer load information acquisition unit 120 in the load balancer 12 may acquire the load NS.

[Second Embodiment]
FIG. 3 is a block diagram showing the configuration of the load distribution system according to the second embodiment of the present invention. In addition, the same code | symbol is attached | subjected to the component equivalent to FIG. The system of FIG. 3 is different from the system of FIG. 1 in that the load balancer 22 corresponding to the load balancer 12 in FIG. 1 does not have a computer load information acquisition unit. Instead, WEB server computers 21-1 to 2-6 corresponding to the WEB server computers 11-1 to 11-6 in FIG. 1 have computer load information acquisition units F1 to F6, respectively. Unlike the computer load information acquisition unit 120 in FIG. 1, the computer load information acquisition units F1 to F6 acquire only the load information of the corresponding WEB server computers 21-1 to 21-6. Further, the computer load information acquired by the computer load information acquisition units F1 to F6 is not the number of sessions but, for example, the usage rate of the CPU included in each of the WEB server computers 21-1 to 21-6. If one WEB server computer is equipped with a plurality of CPUs, a value obtained by adding the usage rates of the plurality of CPUs is used as load information (computer load) of the WEB server computer. As long as the load reflects the request processing in the WEB server computer, not only the CPU usage rate, but also the memory usage rate, the usage rate of the input / output device, etc. may be used alone or in combination with the CPU usage rate. Is possible.

In the system of FIG. 3, a management server computer 25 is used instead of the management server computer 15 in FIG. On this management server computer 25, the management server 250 operates. The management server 250 includes functional elements of a server configuration management unit 251, a provisioning node management unit 152, a provisioning node start / stop unit 153, and a setting change unit 154. The server configuration management unit 251 is different from the server configuration management unit 151 in FIG. 1 on the basis of the computer load information acquired from the computer load information acquisition unit in each load distribution target computer. Find the load. This total load is called N _C. Therefore, in the present embodiment, the provisioning node computers 15-1 to 15-4 positioned as spare load distribution target computers are also connected to the computer load information acquisition units F1 to F6 in the WEB server computers 21-1 to 21-6. It is assumed that a computer load information acquisition unit (not shown) corresponding to is provided.

The server configuration management unit 251 holds a load threshold / plan usage provisioning node computer number correspondence table 251a. The table 251a associates the threshold value of the load amount of the entire load distribution target computer (load threshold value) with the number of provisioning node computers that should be load distribution targets corresponding to the threshold value (the number of planned use provisioning node computers). Remember. An example of this table 251a is shown in FIG. In the example of FIG. 6, the table 251a stores each of the three types of load threshold values N _C1 , N _C2, and N _C3 in association with the planned use provisioning node computer number. The threshold values N _C1 , N _C2 and N _C3 correspond to the load threshold values N _S1 , N _S2 and N _S3 stored in the table 151a shown in FIG.

Next, the operation of the load distribution system of FIG. 3 will be described with reference to the flowchart of FIG. 4 with a focus on differences from the system of FIG. In the load distribution system of FIG. 3, steps S21 to S31 corresponding to steps S1 to S11 in the flowchart of FIG. 2 are executed. The main difference from the system of FIG. 1 is how to determine the load amount in step S21 (corresponding to step S1 in FIG. 2). That is, in step S21, the server configuration management unit 251 includes the computer (load distribution target computer) that is currently subject to load distribution, including the computer load information acquisition units F1 to F6 in the WEB server computers 21-1 to 21-6. The load information of the load balancing target computer is acquired from the computer load information acquisition unit in (). Then, the server configuration management unit 251 obtains the entire load amount N _C from the acquired load information (computer load information) of each load distribution target computer. Here, an added value of the load amount (CPU usage rate) indicated by the load information of each load distribution target computer is obtained as the total load amount N _C. Incidentally, weights the load by the weight corresponding (e.g. proportional) to the processing capacity of each load balancing target computer (e.g., the processing speed of CPU), also the sum of the weighted load as a whole of load N _C Good. Operation after the server configuration management unit 251 to determine the load N _C is the N _C load from N _S, the load threshold is changed from the N _S1, N _S2 and N _S3 in N _C1, N _C2 and N _C3 Except for these points, the second embodiment is the same as the first embodiment.

[Third Embodiment]
FIG. 7 is a block diagram showing a configuration of a load distribution system according to the third exemplary embodiment of the present invention. In addition, the same code | symbol is attached | subjected to the component equivalent to FIG. 1 or FIG. In the system of FIG.
The load balancer 12 shown in FIG. 1 is used as the load balancer, and the WEB server computers 21-1 to 21-6 shown in FIG. 3 are used as the WEB server computers. The system of FIG. 7 is characterized in that each load distribution target computer includes a computer load information acquisition unit 120 in the load distribution device 12 and computer load information acquisition units F1 to F6 in the WEB server computers 21-1 to 21-6. The computer load information acquisition unit is used. That is, in the system of FIG. 7, the load information (information indicating the number of sessions) of each load distribution target computer acquired by the computer load information acquisition unit 120 and the load distribution target computer inside are determined in determining the number of planned use provisioning node computers. Both the load information (information indicating the CPU usage rate) acquired by the computer load information acquisition unit is used.

Therefore, in the system of FIG. 7, the management server computer 35 is used instead of the management server computer 15 in FIG. 1 or the management server computer 25 in FIG. On this management server computer 35, a management server 350 operates. The management server 350 includes functional elements of a server configuration management unit 351, a provisioning node management unit 152, a provisioning node start / stop unit 153, and a setting change unit 154. The server configuration management unit 351 has the functions of both the server configuration management unit 151 in FIG. 1 and the server configuration management unit 251 in FIG. That server configuration management unit 351 acquires the load N _S based on the obtained computer load information by computer load information acquisition unit 120 of the load balancer 12. The server configuration management unit 351 also acquires the load amount N _C based on the computer load information acquired by the computer load information acquisition unit in each load distribution target computer.

The server configuration management unit 351 holds a load threshold / planned provisioning node computer number correspondence table 351a. An example of this table 351a is shown in FIG. Features of the table 351a, the number of planned use provisioning computing nodes, load N _S to the corresponding load threshold (N _S1 here, N _S2 and N _S3), and load N _C to the corresponding load threshold (N in this case _C1 , N _C2 and N _C3 ) are stored in association with each other.

  Next, the operation of the load distribution system of FIG. 7 will be described with reference to the flowchart of FIG. 8 with a focus on differences from the system of FIG. In the load distribution system of FIG. 7, step S41 corresponding to step S1 in the flowchart of FIG. 2 and step S42 corresponding to step S21 in the flowchart of FIG. 4 are executed. In the load distribution system of FIG. 7, step S43 corresponding to step S2 in the flowchart of FIG. 2 and steps S44 to S52 corresponding to steps S3 to S11 in the flowchart of FIG. 2 are executed.

That is, in the load distribution system of FIG. 7, the server configuration management unit 351 is based on the load information (number of sessions) of each load distribution target computer acquired by the computer load information acquisition unit 120 in the load distribution apparatus 12. Request load N _S of the whole load balancing target computer (step S41). The server configuration management unit 351 also loads the load amount N _{C of the} entire load distribution target computer based on the load information (CPU usage rate) of the computer acquired by the computer load information acquisition unit of each load distribution target computer. Is obtained (step S42).

Server configuration management unit 351 when determining the load N _S and N _C, the process proceeds to step S43. Server configuration management unit 351 in this step S43, with reference to the load threshold / planned use provisioning node computer number correspondence table 351a, to determine the plan used provisioning node calculator number n1 corresponding to the current load N _S. Similarly, the server configuration management unit 351 also determines the planned use provisioning node computer number n2 corresponding to the current load amount N _C based on the contents of the table 351a. Next, the server configuration management unit 351, selects compares the load N _S and N respectively _C corresponding negative planning using provisioning node computer speeds n1 and n2, whichever is greater, as planned use provisioning computing nodes the number of actually used To do. Subsequent operations are the same as those in the first embodiment.

The number of the thus in the present embodiment, on the basis of different types load N _S and N _C, provisioning computing nodes required to reduce the spike state represented by the load N _S and N _C n1 and n2 are determined individually. The larger one of the computer numbers n1 and n2 is selected as the number of planned use provisioning node computers actually used. Thus, as compared with the first and second embodiments use a number of provisioning computing nodes is determined from only one of loadings N _S and N _C, can be more effectively address a spike condition.

[Fourth Embodiment]
FIG. 9 is a block diagram showing a configuration of a load distribution system according to the fourth exemplary embodiment of the present invention. In addition, the same code | symbol is attached | subjected to the component equivalent to FIG. First, the difference between the system of FIG. 9 and the system of FIG. 7 will be described. In the system of FIG. 7, although not specifically stated, it is assumed that each WEB server computer (21-1 to 21-6) executes the same type of boot image. The same applies to the first and second embodiments. The boot image includes at least an operating system (OS). In addition to the OS, the boot image may include an application program that operates under the OS. As described above, when the boot image including the OS executed on each WEB server computer is of the same type, the boot for booting (starting up) the computer with respect to the provisioning node computer used when the spiked load occurs. Images can be associated with each other.

  On the other hand, in the system of FIG. 9, the boot image to be executed differs depending on the WEB server computer. Here, the WEB server computer is managed by being grouped for each type of boot image executed on the computer. Further, the load used for determining the number of provisioning node computers to be used is also obtained for each group. In this case, the type of boot image for booting the provisioning node computer used at that time differs depending on which group has a spiked load. Therefore, a boot image cannot be associated with a provisioning node computer in advance. If a provisioning node computer is prepared for each group, it is possible to associate a boot image with the computer in advance. However, for that purpose, a large number of provisioning node computers must be prepared, which is not practical. For example, when provisioning node computers are divided into three groups depending on the type of boot image to be applied, if four provisioning node computers are prepared for each group, a total of 3 × 4 = 12 provisioning node computers are required. .

  Therefore, in the system of FIG. 9, the configuration described below is applied so that the provisioning node computer can boot with any boot image. First, there are three types of boot images applied in the system of FIG. 9, boot images # 1, # 2, and # 3. For this reason, the group of WEB server computers is also divided into three groups of WEB server computer groups 40-1 (# 1) to 40-3 (# 3) according to the three types of boot images # 1 to # 3 applied. It is managed separately. The WEB server computer group 40-1 (# 1) includes p WEB server computers 41-11 to 41-1p booted with the boot image # 1. The WEB server computer group 40-2 (# 2) includes q WEB server computers 41-21 to 41-2q booted with the boot image # 2. The WEB server computer group 40-3 (# 3) includes r WEB server computers 41-31 to 41-3r booted with the boot image # 3.

In the system of FIG. 9, a management server computer 45 corresponding to the management server computer 35 in FIG. 7 is used. On the management server computer 45, the management server 450 operates. The management server 450 includes functional elements of a server configuration management unit 451, a provisioning node management unit 152, a provisioning node start / stop unit 153, a setting change unit 154, and a provisioning node boot image selection unit 455. Server configuration management unit 451, WEB server computer group #i (i = 1, 2, 3) determine the load N _S and N _C in the unit, the group #i units in accordance with the load amount N _S and N _C It has a function to determine the number of planned use provisioning node computers.

The server configuration management unit 451 holds a table 451a shown in FIG. Features of the table 451a is the number of planned use provisioning computing nodes, each WEB server computer group # 1 to # 3, load N _S to the corresponding load threshold (N _S11 here, N _S12, N _S13 to N _S31 , N _S32, N _S33), and that are stored in association with both N _C11, N _C12, N _C13 to _{N C31, N C32, N C33} ) is a load N _C to the corresponding load threshold (here It is in.

  The provisioning node boot image selection unit 455 sends the provisioning node computer determined to be used by the provisioning node management unit 152 as a result of the spiked load in the WEB server computer group 40-i (#i) to the group 40- It has a function of setting a type of boot image #i specific to i (#i). The boot image #i is a boot image used in the WEB server computer in the group 40-i (#i). The setting of the boot image #i for the provisioning node computer by the provisioning node boot image selection unit 455 includes the provisioning node computer and a disk (disk drive) in which the boot image #i used for booting the computer is stored. ) Is associated with a table 470 described later. That is, the provisioning node boot image selection unit 455 has a function of selecting a disk in which the boot image #i used for booting the provisioning node computer is stored.

  When the load of the WEB server computer group 40-1 (# 1) is high, the provisioning node boot image selection unit 455 determines the boot image # used by the WEB server computers 41-11 to 41-1p belonging to the group. The disk in which 1 is stored (disk D1 described later) is associated with a provisioning node computer newly incorporated in the group. Similarly, when the load of the WEB server computer group 40-2 (# 2) is high, the provisioning node boot image selection unit 455 boots used by the WEB server computers 41-21 to 41-2q belonging to the group. The disk storing the image # 2 (disk D2 described later) is associated with a provisioning node computer newly incorporated in the group. Similarly, when the load of the WEB server computer group 40-3 (# 3) is high, the provisioning node boot image selection unit 455 boots used by the WEB server computers 41-31 to 41-3r belonging to the group. The disk in which image # 3 is stored (disc D3 described later) is associated with a provisioning node computer that is newly incorporated into the group.

  In this embodiment, WEB server computers 41-11 to 41-1p, WEB server computers 41-21 to 41-2q, WEB server computers 41-31 to 41-3r, a management server computer 45, and a provisioning node computer 15-1 to 15-4 are connected to a storage area network (hereinafter referred to as SAN) 46. A storage device 47 is also connected to the SAN 46. The storage device 47 includes disks (disk drives) D1 to D3. That is, the above computers and the disks D1 to D3 in the storage apparatus 47 are connected by the SAN 46. Boot images # 1 to # 3 used in the WEB server computers in the WEB server computer groups 40-1 to 40-3 are stored in advance in the disks D1, D2 and D3, respectively.

  The storage device 47 has a database holding a table 470. The table 470 stores information in which the disks D1 to D3 in the storage device 47 are associated with computers using the disks D1 to D3 as boot disks. The computer using the disk Di (i = 1, 2, 3) as a boot disk is a boot computer that is booted (started up) by the boot image #i stored in the disk Di.

  An example of the table 470 is shown in FIG. In the example of FIG. 13, the table 470 is stored in the disks (boot disks) D1 to D3 (for example, disk identifiers) storing the boot images # 1 to # 3 and the disks D1 to D3. Information (for example, computer identifiers) of computers booted by the boot images # 1 to # 3 (that is, computers using the disks D1 to D3 as boot disks) is stored in association with each other. In this embodiment, the identifiers of the WEB server computers 41-11 to 41-1p are WS11 to WS1p, the identifiers of the WEB server computers 41-21 to 41-2q are WS21 to WS2q, and the WEB server computers 41-31 to 41-31 to The identifier of 41-3r is assumed to be WS31 to W3r. The identifiers of the provisioning node computers 15-1 to 15-4 are P1 to P4.

  In the system of FIG. 9, the following procedure is used to boot the provisioning node computer 15-j (j is any one of 1 to 4) with the boot image #i. First, the provisioning node boot image selection unit 455 searches the table Di 470 stored in the database in the storage apparatus 47 for the identifier of the disk Di storing the boot image #i. Then, the provisioning node boot image selection unit 455 operates the table 470 to additionally set the identifier (Pj) of the provisioning node computer 15-j in association with the identifier (Di) of the disk Di. By the table operation of the provisioning node boot image selection unit 455, the disk Di is logically connected as the boot disk of the provisioning node computer 15-j. That is, the disk Di storing the boot image #i used for booting the provisioning node computer 15-j is selected.

  In this case, when the provisioning node computer 15-j is activated by the provisioning node activation / deactivation unit 153, the provisioning node computer 15-j retrieves its own computer identifier (Pj) from the table 470 held in the storage device 47 via the SAN 46. The provisioning node computer 15-j then uses the disk identifier (Di) associated with its own computer identifier (Pj) in the table 470 to use the boot image (in this case, the boot image) to boot itself. Identify the disk Di storing #i). Accordingly, the provisioning node computer 15-j reads the boot image #i from the identified disk Dj (that is, the disk Dj associated with itself) and boots itself with the boot image #i.

  Therefore, on the table 470, for example, as shown in FIG. 14, when the disk D1 is newly associated with two provisioning node computers 15-1 and 15-2, the computers 15-1 and 15-2 D1 is shared as a boot disk. In this case, the provisioning node computers 15-1 and 15-2 are booted with the boot image # 1 stored in the disk D1.

Next, the operation of the load distribution system of FIG. 9 will be described with reference to the flowchart of FIG. 10 with a focus on differences from the system of FIG. First, the server configuration management unit 451 sets a variable i for designating a WEB server computer group to an initial value 1 (step S61). Next, the server configuration management unit 451 becomes a target of current load distribution including the WEB server computer belonging to the WEB server computer group 40-i (#i) from the computer load information acquisition unit 120 in the load balancer 12. and that acquires load information of each computer (load balancing target computer) (computer load information), obtains the load N _S of the whole group 40-i (#i) (step S62). Further, the server configuration management unit 451 includes, from the computer load information acquisition unit in each load distribution target computer, including the computer load information acquisition unit in the WEB server computer belonging to the WEB server computer group 40-i (#i). acquires load information of the load balancing target computer, obtains the load N _C of the entire group 40-i (#i) (step S63). Processing in steps S62 and S63 are points for obtaining the load N _S and N _C of the entire group #i specified by the variable i, the step of FIG. 8 S41 to determine the load N _S and N _C of the entire system And S42.

When the server configuration management unit 451 obtains the load N _S and N _C for the entire group 40-i (#i), the process proceeds to step S64. Server configuration management unit 451 in this step S64 refers to the entry of a group of tables 451a 40-i (#i), plans used corresponding to the current load N _S of the group 40-i (#i) The number of provisioning node computers n1 is determined. Server configuration management unit 451 similarly determines, based on the entry contents of the group tables 451a 40-i (#i), the group 40-i (#i) current plans use provisioning node corresponding to the load N _C of The number of computers n2 is determined. Next, the server configuration management unit 451 compares the group 40-i loading N _S and N planned use provisioning node _C correspond to computer speeds n1 and n2 of (#i), whichever is greater, it is actually used Select as planned use provisioning node number of computers.

  The operation after step S64 is substantially the same as the operation after step S43 in FIG. However, the following operations are different in this embodiment. First, it is assumed that a spike load of the WEB server computer group 40-i (#i) has occurred. In this case, in order to reduce the load, at least one provisioning node computer is assigned to the group 40-i (#i) by the provisioning node management unit 152 as a provisioning node computer to start using (step S69). . Then, in the present embodiment, control is transferred from the provisioning node management unit 152 to the provisioning node boot image selection unit 455. In response to this, the provisioning node boot image selection unit 455 stores the provisioning node computer assigned to the group 40-i (#i) and the boot image #i unique to the group 40-i (#i). Corresponding to the disk Di that is present. That is, the provisioning node boot image selection unit 455 selects the boot image #i (disk Di on which the boot image #i is stored) to boot the provisioning node computer assigned to the group 40-i (#i) ( Step S70).

  When the provisioning node boot image selection unit 455 selects the boot image #i to boot the provisioning node computer to be used, the provisioning node start / stop unit 153 operates. The provisioning node activation / deactivation unit 153 activates the provisioning node computer assigned by the provisioning node management unit 152 among the provisioning node computers 15-1 to 15-4 (step S71). The activated provisioning node computer is associated with the disk Di storing the boot image #i to boot the computer by the table operation of the provisioning node boot image selection unit 455. That is, the boot image #i to boot the provisioning node computer that has been activated is selected by the provisioning node boot image selection unit 455. Therefore, the provisioning node computer that has been activated reads the selected boot image #i from the disk Di and executes booting.

  When the provisioning node computer is activated by the provisioning node activation / deactivation unit 153, the setting change unit 154 indicates that the activated provisioning node computer belongs to the WEB server computer group 40-i (#i) and the group 40- The load balancer 12 is controlled so as to become a load balancing target computer together with other computers in i (#i) (step S72). Here, the order of steps S71 and S72 may be reversed.

  When the load distribution target by the load balancer 12 is changed under the control of the setting change unit 154, the server configuration management unit 451 waits for a certain period of time (step S66), and then switches the WEB server computer group to be processed ( Step S67). In the present embodiment in which the WEB server computer group is a group of three groups # 1 to # 3, if the current group is group # 1, it is group # 2, if it is group # 2, it is group # 3, and group # 3. If so, it is switched to group # 1. Then, the process is repeated from step S62 for the switched WEB server computer group.

  As described above, in this embodiment, depending on the type of boot image executed by each WEB server computer in the system, the WEB server computer can be a WEB server computer group 40-1 (# 1) to 40-3 (# 3). ). When a spike load occurs in any of the groups 40-i (#i) among the groups 40-1 (# 1) to 40-3 (# 3), the provisioning node computers 15-1 to The provisioning node computer selected from 15-4 is associated with the boot image #i (disk D1 storing the boot image #i) unique to the group 40-i (#i). As a result, the selected provisioning node computer can be activated as a load distribution target computer belonging to the group 40-i (#i). For this reason, the provisioning node computers 15-1 to 15-4 can be used effectively.

[Modification of Fourth Embodiment]
In the system of FIG. 9, a configuration is used in which the provisioning node computer 15-j is booted with the boot image #i using the disk Di in the storage device 47. However, the disk Di in the storage device 47 is not necessarily required to boot the provisioning node computer 15-j with the boot image #i. Accordingly, FIG. 15 is a block diagram of a modification of the fourth embodiment that enables the provisioning node computer 15-j to boot with the boot image #i without using the disk Di in the storage apparatus 47. This will be described with reference to the configuration diagram. In addition, the same code | symbol is attached | subjected to the component equivalent to FIG.

  In FIG. 15, a WEB server computer 41-11 and a provisioning node computer 15-1 are connected to an input / output bus (external input / output bus) 51. The WEB server computer 41-11 belongs to the WEB server computer group 40-1 (see FIG. 9). Although omitted in FIG. 15, a server computer other than the WEB server computer 41-11 belonging to the WEB server computer group 40-1, each server computer belonging to the other WEB server computer groups 40-2 and 40-3, and provisioning Provisioning node computers other than the node computer 15-1 are also connected to the input / output bus 51.

  The computers 41-11 and 15-1 include CPUs 52a and 52b, local disks (local disk drives) 53a and 53b, and input / output buses (internal input / output buses) 54a and 54b, respectively. The local disk 53a in the computer 41-11 stores a boot image # 1 unique to the WEB server computer group 40-1 (# 1) used for booting the computer 41-11. The CPU 52a and the local disk 53a are interconnected by an input / output bus 54a, and the CPU 52b and the local disk 53b are interconnected by an input / output bus 54b. The local disks 53a and 53b are connected to the input / output bus 51. The input / output buses 54 a and 54 b are also connected to the input / output bus 51.

  Assume that the provisioning node computer 15-1 in FIG. 15 is assigned to the WEB server computer group 40-1 shown in FIG. In this case, the provisioning node boot image selection unit 455 shown in FIG. 9 has one of the WEB server computers 41-11 to 41-1p belonging to the group 40-1 such as the WEB server computer 41-. 11 is requested to copy the boot image to the provisioning node computer 15-1. This means that the boot image # 1 stored in the local disk 53a in the WEB server computer 41-11 belonging to the group 40-1 is selected as a boot image for booting the provisioning node computer 15-1. Represents. At this time, the provisioning node computer 15-1 has not been activated yet.

  The CPU 52a in the WEB server computer 41-11 reads the boot image # 1 stored in the local disk 53a in the computer 41-11 in accordance with a copy request from the provisioning node boot image selection unit 455, and inputs / outputs the input / output bus. 51 to the local disk 53b of the provisioning node computer 15-1. By the copy operation of the CPU 52a in the WEB server computer 41-11, the boot image # 1 can be copied to the local disk 53b of the provisioning node computer 15-1 even when the provisioning node computer 15-1 is not activated. Many cluster computers and blade server computers have such a function.

  When the copy of the boot image # 1 is completed, the provisioning node activation / deactivation unit 153 shown in FIG. 9 activates the provisioning node computer 15-1. Then, the provisioning node computer 15-1 boots itself using the boot image # 1 copied to its own local disk 53b. More specifically, the CPU 52b in the provisioning node computer 15-1 reads the boot image # 1 copied from the local disk 53b to the disk 53b via the input / output bus 54b and boots. In FIG. 15, input / output buses 54a and 54b are provided inside the computers 41-11 and 15-1, respectively, and an input / output bus 51 is provided outside the computers 41-11 and 15-1. However, such a bus configuration is an example. In short, the computer 41-11 having the local disk 53a storing the boot image directly or indirectly copies the boot image (data) to the local disk 53b included in the computer 15-1 that is not in the activated state. All you need to do is

[Fifth Embodiment]
FIG. 16 is a block diagram showing a configuration of a load distribution system according to the fifth exemplary embodiment of the present invention. In addition, the same code | symbol is attached | subjected to the component equivalent to FIG. The system of FIG. 16 includes a plurality of, for example, four different operation matrixes (for example, different companies), WEB server sites 50-1 to 50-4, and the WEB server sites 50-1 to 50-4. This is in that a spare site (hereinafter referred to as a shared provisioning node computer site) 60 that can be used in common is provided.

  The WEB server site 50-1 includes the same elements as the components excluding the provisioning node computers 15-1 to 15-4, the SAN 46, and the storage apparatuses 47-1 to 47-3 from the system of FIG. That is, the WEB server site 50-1 includes the load balancer 12, for example, three WEB server computer groups 40-1 to 40-3, and a management server computer 55 corresponding to the management server computer 45 in FIG. Yes. A management server 550 operates on the management server computer 55. The management server 550 includes a server configuration management unit 551 and a setting change unit 154. The remaining WEB server sites 50-2 to 50-4 have the same configuration (not shown) as the WEB server site 50-1. Note that the number of WEB server computer groups, the number of WEB server computers in the group, and the like in each WEB server site 50-1 to 50-4 are not necessarily equal. However, in this embodiment, in order to simplify the description, it is assumed that each WEB server computer group in the WEB server sites 50-1 to 50-4 corresponds to one of the boot images # 1 to # 3.

Server configuration management unit 551 in the management server 550, like the server configuration management unit 451 in FIG. 9, has a function of obtaining the load N _S and N _C of the entire WEB server computer group 40-i (#i) . The server configuration management unit 551, like the server configuration management unit 451 in FIG. 9, based on the table 451a of the data structure shown in FIG. 12 and load N _S and N _C, the start or Using It has a function of determining the number of provisioning node computers to be stopped. The server configuration management unit 551 further has a function of requesting (requesting) a later-described management server 650 of the shared provisioning node computer site 60 to start or stop using the number of provisioning node computers determined by itself.

  On the other hand, the shared provisioning node computer site 60 includes provisioning node computers 15-1 to 15-4 and a storage device 47. The storage device 47 includes disks D1 to D3 in which boot images # 1 to # 3 are stored and a database that holds a table 470, as in the case of being used in the system of FIG. The provisioning node computers 15-1 to 15-4 and the storage apparatus 47 are interconnected by a SAN 46. With this configuration, any provisioning node computer among the provisioning node computers 15-1 to 15-4 can be booted with any one of the boot images # 1 to # 3.

  The provisioning node computers 15-1 to 15-4 are connected via the network 70 to the load balancer 12, management server computer 55, and WEB server computer groups 40-1 to 40-3 (groups) provided in the WEB server site 50-1. WEB server computers 40-1 to 40-3). In addition, the provisioning node computers 15-1 to 15-4 are connected to the load balancer, management server computer, and WEB server computer group (all of which are shown in FIG. (Not shown). Thereby, an arbitrary provisioning node computer can be assigned as a load distribution target computer belonging to an arbitrary WEB server computer group in an arbitrary WEB server site.

  Accordingly, the shared provisioning node computer site 60 processes the provisioning node use start request (use start request) and the use end request (use end request) from the WEB server sites 50-1 to 50-4 of the respective operation hosts. The management server computer 65 is provided. This management server computer 65 is also connected to the network 70 and the SAN 46. On the management server computer 65, a management server 650 having a function of processing a provisioning node use start request and use end request operates. The management server 650 includes functional elements of a provisioning node management unit 652, a provisioning node start / stop unit 153, and a provisioning node boot image selection unit 455.

  The provisioning node management unit 652 has a function of managing the usage status of the provisioning node computers 15-1 to 15-4. The provisioning node management unit 652 also receives provisioning node use start requests or use end requests from the WEB server sites 50-1 to 50-4, and processes the request by the provisioning node start / stop unit 153, and further provisioning. The node boot image selection unit 455 has a function of controlling.

  The provisioning node management unit 652 holds a provisioning node computer use record file 652a having the data structure shown in FIG. This record file 652a stores a provisioning node computer use record. In each record, information indicating the usage status of the provisioning node computer activated in response to a request from the WEB server sites 50-1 to 50-4 is recorded. Specifically, the provisioning node computer usage record includes items (fields) 81 to 84. The item 81 is used to record the time when the provisioning node computer is activated (that is, the activation start time). The item 82 is used to record the time when the provisioning node computer is stopped (that is, the start end time). The activation start time and activation end time recorded in the items 81 and 82 respectively indicate the time (activation time) when the provisioning node computer is activated. The item 83 is used for recording information for identifying a provisioning node computer, for example, a provisioning node identifier. The item 84 is used to record information on the WEB server site using the provisioning node indicated by the identifier recorded in the item 83. In addition to the format shown in FIG. 19, various formats such as a log format can be used for each record stored in the record file 652a. In addition, information about the start time and start time can be recorded as separate records. Moreover, it is good also as a structure by which the information showing the time (stop time) when the provisioning node computer has stopped is recorded on a record instead of the time when the provisioning node computer is starting.

Next, the operation of the load distribution system of FIG. 16 will be described with reference to the flowcharts of FIGS. 17 and 18, focusing on the differences from the system of FIG. The server configuration management unit 551 operating on the management server computer 55 of the WEB server site 50-1 first executes the same processing (steps S81 to S85) as steps S61 to S65 in FIG. That server configuration management unit 551, after setting a variable i to an initial value 1 (step S81), the load of the whole WEB server computer group 40-i (#i) specified with the variable i N _S and N _C Is obtained (steps S82 and S83). Next, the server configuration management unit 551 refers to the table 451a, to determine the group 40-i current load N _S and N planned use provisioning node _C correspond to computer speeds n1 and n2 of (#i) (Step S83). In this step S83, the server configuration management unit 551, load N of _S and N the number plan used provisioning computing nodes _C correspond to n1 and n2, planned use provisioning node towards number calculator is large, actually used Select as the number of computers. Next, the server configuration management unit 551 compares the selected planned use provisioning node computer number with the provisioning node computer number currently in use (step S85). The number of provisioning node computers currently in use is managed by the provisioning node management unit 652 of the shared provisioning node computer site 60.

  If the number of computers is equal, the server configuration management unit 551 waits for a certain time (step S86), and then operates the variable i to switch the WEB server computer group (step S87). Thereby, the process is repeated from step S82 for the WEB server computer group after switching.

  On the other hand, if the number of planned use provisioning node computers is larger (step S85), the server configuration management unit 551 determines that a provisioning node computer needs to be newly assigned as a load balancing target computer. In this case, the server configuration management unit 551 provisions the management server 650 on the management server computer 65 for managing the provisioning node computer provided in the shared provisioning node computer site 60 with the number of differences Nd between the two computers. The start of use of the node computer is requested via the network 70 (step S88).

  Also, if the number of provisioned use provisioning node computers is smaller than the number of provisioning node computers currently in use (step S85), the server configuration management unit 551 uses the provisioning node computers of the number of differences Nd between the numbers of both computers. It is determined that it is necessary to terminate. In this case, the server configuration management unit 551 requests the management server 650 on the management server computer 65 to end the use of the Nd number of provisioning node computers via the network 70 (step S91).

  The provisioning node management unit 652 in the management server 650 is waiting for a request from the WEB server sites 50-1 to 50-4 (request for starting or ending use of the provisioning node computer) (step S101). In this state, it is assumed that the provisioning node management unit 652 has received a request from the management server 550 on the management server computer 55 provided in the WEB server site 50-1, for example. In this case, the provisioning node management unit 652 determines whether the type of the received request is a provisioning node computer use start request or a use end request (step S102).

  If it is a request to start using a provisioning node computer, the provisioning node management unit 652 checks whether there is a usable provisioning node computer among the provisioning node computers 15-1 to 15-4 (Step S1). S103). If there is a provisioning node computer that can be used, the provisioning node management unit 652 allocates the provisioning node computer as a computer to start using (step S104).

  The provisioning node boot image selection unit 455 in the management server 650 associates the provisioning node computer assigned by the provisioning node management unit 652 with the disk Di in which the boot image #i is stored. That is, the provisioning node boot image selection unit 455 selects the boot image #i (disk Di storing the boot image #i) to boot the provisioning node computer assigned by the provisioning node management unit 652 (step S105). . Then, the provisioning node start / stop unit 153 in the management server 650 operates. The provisioning node activation / deactivation unit 153 activates the provisioning node computer assigned by the provisioning node management unit 652 among the provisioning node computers 15-1 to 15-4 (step S106). The provisioning node computer that has been started up is associated with the disk Di by which the boot image #i to boot the computer is stored by the provisioning node boot image selection unit 455. Therefore, the provisioning node computer that has been activated reads the boot image #i stored in the disk Di associated with the computer and executes booting.

  When a provisioning node computer assigned as a computer to start use is activated, the provisioning node management unit 652 displays the activation start time, which is the time at that time, in the provisioning node computer use record file 652a, and the provisioning node computer activated. And a record in which information indicating a WEB server site (here, the WEB server site 50-1) using the provisioning node computer is stored (step S107). Here, as the information indicating the WEB server site, the domain name of the site is used. The domain name and the IP (Internet Protocol) address are associated with each other by a DNS (Domain Name System) (not shown).

  When the provisioning node management unit 652 executes step S107, the processing result for the request from the management server 550 of the WEB server site 50-1 (provisioning node computer use start request) is sent to the management server 550 via the network 70. Notification is made (step S108).

  On the other hand, if there is no provisioning node computer that can be used (step S103), the provisioning node management unit 652 informs the request to start using the provisioning node computer (management server 550 of the WEB server site 50-1). (Step S108).

  If the request from the management server 550 of the WEB server site 50-1 is a provisioning node computer use end request (step S102), the provisioning node management unit 652 selects the provisioning node computer currently in use. The provisioning node computer whose use is to be ended is determined (step S109). The provisioning node start / stop unit 153 stops the provisioning node computer whose use is determined to be ended by the provisioning node management unit 652 (step S110). Then, the provisioning node management unit 652 searches the record file 652a for a record in which the identifier of the provisioning node computer that has been stopped is recorded, and the item 82 of the activation end time is blank, and the item in the record The current time is recorded in 82 as the activation end time (step S111). Then, the provisioning node management unit 652 notifies the management server 550 of the processing result in response to the request from the management server 550 of the WEB server site 50-1 (the provisioning node computer use end request) via the network 70 (step S1). S108).

  When the server configuration management unit 551 in the management server 550 of the WEB server site 50-1 requests the management server 650 of the shared provisioning node computer site 60 to start using the provisioning node computer (step S88), the management server 650 It waits for the result of the processing to be notified from the provisioning node management unit 652 therein. When the result of the provisioning node computer use start request is notified, the server configuration management unit 551 determines from the result whether the request is successful (step S89). If successful, the provisioning node management unit 652 passes control to the setting change unit 154. Then, the setting change unit 154 adds the provisioning node computer activated in response to the provisioning node computer use start request as a load distribution target computer to be a load distribution target in the WEB server computer group 40-i (#i). Thus, the load balancer 12 is controlled (step S90).

  As described above, in this embodiment, the provisioning node computers of the shared provisioning node computer site (preliminary site) 60 can be used for the WEB server sites 50-1 to 50-4 of the four different operation hosts. In addition, the record file 652a can be applied to information such as billing for an operation base using a provisioning node computer.

  In the present embodiment, it is assumed that a plurality of WEB server computers in each of the WEB server sites 50-1 to 50-4 are managed in groups for each type of boot image executed on the computer. However, in each WEB server site 50-1 to 50-4, all the WEB server computers in the server site may be configured to execute the same type of boot image. In this case, it is not necessary to group the WEB server computers in the WEB server sites 50-1 to 50-4. Further, if the WEB server computers of any of the WEB server sites 50-1 to 50-4 are configured to execute the same type of boot image, the storage device 47 and the provisioning node boot image selection in the shared provisioning node computer site 60 are selected. The part 455 is also unnecessary. Further, it is possible to use a DNS (Domain Name System) round robin function as a load balancer instead of the dedicated load balancer 12. The same applies to the load balancer applied in the first to fourth embodiments. The DNS round robin function is a function that distributes the load of a server with many accesses by applying DNS that associates a domain name with an IP address. Specifically, it is a function that prevents a concentration of access to one computer by assigning a plurality of IP addresses to one domain name and answering in turn each time an inquiry is made.

  Note that the present invention is not limited to the above-described embodiment as it is, and can be embodied by modifying the constituent elements without departing from the scope of the invention in the implementation stage. Further, various inventions can be formed by appropriately combining a plurality of constituent elements disclosed in the embodiment. For example, some components may be deleted from all the components shown in the embodiment. Furthermore, you may combine suitably the component covering different embodiment.

1 is a block diagram showing a configuration of a load distribution system according to a first embodiment of the present invention. The flowchart for demonstrating operation | movement of the said 1st Embodiment. The block diagram which shows the structure of the load distribution system which concerns on the 2nd Embodiment of this invention. The flowchart for demonstrating operation | movement of the said 2nd Embodiment. The figure which shows the example of a data structure of the table 151a in FIG. The figure which shows the example of a data structure of the table 251a in FIG. The block diagram which shows the structure of the load distribution system which concerns on the 3rd Embodiment of this invention. The flowchart for demonstrating operation | movement of the said 3rd Embodiment. The block diagram which shows the structure of the load distribution system which concerns on the 4th Embodiment of this invention. The flowchart for demonstrating operation | movement of the said 4th Embodiment. The figure which shows the example of a data structure of the table 351a in FIG. The figure which shows the example of a data structure of the table 451a in FIG. FIG. 10 is a diagram showing an example data structure of a table 470 in FIG. 9. FIG. 14 is a diagram showing a state in which identifiers (P1 and P2) of provisioning node computers 15-1 and 15-2 are added to the table 470 in the state shown in FIG. 13 in association with the identifier (D1) of the disk D1. The block block diagram which shows the modification of the 4th Embodiment of this invention. The block diagram which shows the structure of the load distribution system which concerns on the 5th Embodiment of this invention. The flowchart for demonstrating operation | movement of the WEB server site in the said 5th Embodiment. The flowchart for demonstrating operation | movement of the shared provisioning node computer site in the said 5th Embodiment. The figure which shows the example of a data structure of the record file 652a in FIG.

Explanation of symbols

  11-1 to 11-6, 21-1 to 21-6, 41-11 to 41-1p, 41-21 to 41-2q, 41-31 to 41-3r ... WEB server computer, 12, 22 ... load distribution Devices, 15, 25, 35, 45, 55 ... management server computers, 15-1 to 15-4 ... provisioning node computers (standby computers), 40-1 to 40-3 ... WEB server computer groups, 47 ... storage devices, 50-1 to 50-4 ... WEB server site, 53a, 53b ... local disk (boot image storage means), 60 ... shared provisioning node computer site (spare site), 70 ... network, 120 ... computer load information acquisition unit, 150 , 250, 350, 450, 550... Management server, 151, 251, 351, 451, 551... Server configuration management unit, 151 a, 251 a, 351 a, 451 a, 470. 152,652 ... provisioning node management unit (preliminary computer management means), 153 ... provisioning node start / stop unit (preliminary computer start / stop means), 154 ... setting change unit, 455 ... boot image selection unit, 652a ... provisioning node computer Usage record files, F1 to F6, F11 to F1p, F21 to F2q, F31 to F3r ... computer load information acquisition unit, D1 to D3 ... disk (disk drive, boot image storage means).

Claims (13)

A plurality of server computers that process requests from a plurality of client computers, and are managed by being divided into a plurality of server computer groups depending on the type of boot image including the operating system executed by the server computer A calculator,
A load distribution device that distributes requests from the plurality of client computers to a plurality of load distribution target computers including the plurality of server computers;
A spare computer that can be used as a load balancing target computer, which is not normally started , and is started when used as the load balancing target computer; and
Using each of the plurality of server computer groups as a unit, based on the load amount of all load balancing target computers including the server computers in the group among the plurality of server computers, newly used as load balancing target computers Server configuration management means for determining the number of spare computers to be started or to end use as load balancing target computers;
It is a spare computer management means for managing the usage status of the spare computer, and the number of units determined by the server configuration management means in units of the group should be newly used as a load balancing target computer in the group, Or a spare computer management means for determining a spare computer to be terminated as a load balancing target computer;
Standby computer start / stop means for starting or stopping the standby computer determined by the standby computer management means;
A setting change unit that adds the spare computer determined by the spare computer management unit as a load balancing target computer to be a load balancing target by the load balancing device, or excludes it from a load balancing target by the load balancing device;
A server computer group in which the determined spare computer is started to be used as the load balancing target computer as a boot image to boot the spare computer determined by the spare computer management means from among the plurality of types of boot images A boot image selecting means for selecting a boot image specific to
When the backup computer is started by the backup computer start / stop unit, the backup computer is booted with the boot image selected by the boot image selection unit. A plurality of boot image storage means for individually storing the plurality of types of boot images;
Correspondence information storage means for storing information associating each of the plurality of boot image storage means with a computer booted with the boot image stored in the boot image storage means,
The boot image selection means stores a boot image specific to a server computer group in which the spare computer determined by the spare computer management means among the plurality of boot image storage means is used as the load balancing target computer. The boot image storage means selects the boot image to boot the spare computer by associating the spare computer determined by the spare computer management means with the correspondence information storage means,
The spare computer is a boot image selected by the boot image selection unit when booted by the spare computer start / stop unit, and is associated with the spare computer by the correspondence information storage unit The load distribution system according to claim 1, wherein the boot image is read from the means and boots itself. Each of the plurality of server computers and the spare computer has boot image storage means for storing a boot image used to boot itself.
The boot image selection unit is configured to provide the boot computer included in the arbitrary computer to an arbitrary server computer in a server computer group in which the standby computer determined by the standby computer management unit starts to be used as the load balancing target computer. Selecting a boot image to boot the spare computer by requesting a copy of the boot image from the image storage means to the spare computer determined by the spare computer management means;
Each of the plurality of server computers is a boot image selected by the boot image selection unit when the copy is requested by the boot image selection unit, and is stored in the boot image storage unit included in itself. The load distribution system according to claim 1, wherein the image is copied to the boot image storage unit included in the spare computer determined by the spare computer management unit. A plurality of server sites each comprising a set of the plurality of server computers, the load distribution device, the server configuration management means, and the setting change means managed separately in the plurality of server computer groups;
A spare site comprising a set of the spare computer, the spare computer management means, the spare computer start / stop means, the boot image selection means, and the plurality of boot image storage means;
A network connecting the plurality of server sites and the spare site;
The server configuration management means included in each of the plurality of server sites should start to be used as a new load distribution target computer or be used as a load distribution target computer in units of the server computer group. When determining the number of spare computers to be terminated, request the start or end of use of spare computers in the group to the spare site via the network,
The spare computer management means in the spare site has the number of spare computers determined by the server configuration management means in response to a request for starting or ending use of the spare computer from the server configuration management means in the server site. A computer that is to be used as a load balancing target computer in a group requested by the server configuration management unit in the server site where the server configuration management unit exists, or used as a load balancing target computer The load balancing system according to claim 1, wherein a spare computer to be terminated is determined.   The spare computer management means in the spare site is activated when the number of spare computers determined by the server configuration management means included in each of the plurality of server sites is activated or stopped. 5. The load distribution system according to claim 4, wherein a set of information representing time or stop time, information representing the spare computer, and information representing the server site is recorded. A plurality of server computers that process requests from a plurality of client computers;
A load distribution device that distributes requests from the plurality of client computers to a plurality of load distribution target computers including the plurality of server computers;
A spare computer that can be used as a load distribution target computer, and is activated when used as the load distribution target computer;
Based on the load amount of all load balancing target computers including the plurality of server computers, the number of spare computers to be newly used as load balancing target computers or to be terminated as load balancing target computers is determined. A server configuration management means for determining;
A spare computer management means for managing the usage status of the spare computer, wherein the number determined by the server configuration management means should be newly used as a load distribution target computer or used as a load distribution target computer A spare computer management means for determining a spare computer to be terminated,
Standby computer start / stop means for starting or stopping the standby computer determined by the standby computer management means;
A setting change unit that adds the spare computer determined by the spare computer management unit as a load balancing target computer to be a load balancing target by the load balancing device, or excludes the spare computer from a load balancing target by the load balancing device;
Comprising
The load distribution device is a first computer load indicating a load amount corresponding to the number of requests being processed by a load distribution target computer including the plurality of server computers that are targets of load distribution by the load distribution device. Including first load information acquisition means for acquiring information,
Each of the plurality of server computers and the spare computer includes second load information acquisition means for acquiring second computer load information indicating a load amount corresponding to processing of a request executed by itself,
The server configuration management means obtains the load amount of the entire load distribution target computer from the first computer load information acquired by the first load information acquisition means included in the load distribution apparatus, and newly A first number of spare computers to be used as load balancing target computers or to be terminated as load balancing target computers is determined, and load balancing targets among the plurality of server computers and the spare computers are determined. The load amount of the entire load distribution target computer is obtained from the second computer load information acquired by the second load information acquisition means included in the computer that is the load distribution target by the load distribution device as a computer. The second computer of the spare computer that should be newly used as a load distribution target computer or that should be terminated as a load distribution target computer Determines, either one of the first and second number is selected as the number of spare computer to be finished using a load balancing target computer should start using as, or load balancing target computer
A load balancing system characterized by that . A plurality of servers sites with multiple server computers, load balancer, the servers configuration manager and configuration changing means set respectively,
Preliminary computer, and the backup site with a set of spare computer management means and spare computers start / stop means,
A network connecting the plurality of server sites and the spare site;
Comprising
The plurality of server computers process requests from a plurality of client computers,
The load distribution apparatus distributes requests from a plurality of client computers to a plurality of load distribution target computers including the plurality of server computers in the server site where the load distribution apparatus exists,
The spare computer can be used as a load balancing target computer and is activated when used as the load balancing target computer,
Before SL server configuration management means, based on the load of all load balancing target computer comprising a plurality of server computers in the server site to the server configuration manager is present, newly load balancing within the server site Determine the number of spare computers that should start using as target computers or end use as load balancing target computers,
Before Symbol preliminary computer management unit, the managing usage of the spare computers, said a preliminary calculator number determined by the server configuration management unit newly loaded in the server site to the server configuration manager is present Decide which spare computer should start to be used as a distribution target computer or should end use as a load distribution target computer .
The spare computer start / stop means starts or stops the spare computer determined by the spare computer management means,
The setting change unit adds the spare computer determined by the spare computer management unit as a load balancing target computer to be a load balancing target by the load balancing device, or excludes it from a load balancing target by the load balancing device. Do
A load balancing system characterized by that . The spare computer management means in the spare site is activated when the number of spare computers determined by the server configuration management means included in each of the plurality of server sites is activated or stopped. The load distribution system according to claim 7 , wherein a set of information indicating time or stop time, information indicating the spare computer, and information indicating the server site is recorded. A plurality of server computers that process requests from a plurality of client computers, and a plurality of sets of load distribution devices that distribute requests from a plurality of client computers to a plurality of load distribution target computers including the plurality of server computers. A server site and a spare computer that can be used as the load balancing target computer, the spare site having a spare computer that is activated when used as the load balancing target computer connected by a network In a distributed system, when a spiked load occurs in each of the plurality of server sites, a spare computer is assigned from the spare site to the server site as a new load balancing target computer. A computer management program for
In the computer located at the server site,
Based on the load amount of all load balancing target computers including the plurality of server computers in the server site, the server site should start to be used as a new load balancing target computer, or as a load balancing target computer Determining the number of spare computers to be terminated, and requesting the spare site to start or terminate the use of the spare computers in the server site via the network;
In response to a request to start using the spare site, a spare computer to be used as a load distribution target computer is determined in the server site of the request source at the spare site, and the determined spare computer is activated. If the backup site is notified to the server site, the notified backup computer is added as a load distribution target computer to be a load distribution target by the load distribution device in the server site. When,
In response to the use termination request for the spare site, a spare computer to be used as a load distribution target computer in the server site of the request source is determined at the spare site, and the determined spare computer is When the standby site is notified of the suspension from the spare site, the notified spare computer is excluded from the load balancing targets by the load balancer in the server site. Computer management program for. A plurality of server computers that process requests from a plurality of client computers, and a plurality of sets of load distribution devices that distribute requests from a plurality of client computers to a plurality of load distribution target computers including the plurality of server computers. A server site and a spare computer that can be used as the load balancing target computer, the spare site having a spare computer that is activated when used as the load balancing target computer connected by a network In a distributed system, when a spiked load occurs in the server site and is applied to the spare site, a spare computer is added as a new load balancing target computer from the spare site in response to a request from the server site. Computer management program to assign to server sites There,
In the computers located at the spare site,
In each of the plurality of server sites, based on the load amount of all load balancing target computers including the plurality of server computers in the server site, the server site starts to be used as a new load balancing target computer. As a result of determining the number of spare computers to be used as load balancing target computers, the start of use or termination of use of spare computers within the server site is made from the server site to the spare site. Determining whether a use start or a use end is requested when requested via the network; and
When the start of use is requested, determining a spare computer to start using as a load balancing target computer in the requesting server site;
When a spare computer to be used is determined, starting the determined spare computer;
When the end of use is requested, determining a spare computer to end use as a load balancing target computer in the requesting server site;
When a spare computer to be discontinued is determined, stopping the determined spare computer;
A computer management program for causing the spare site to notify the requesting server site of a processing result for the use start or use end request from the server site. A plurality of server computers that process requests from a plurality of client computers, and are managed by being divided into a plurality of server computer groups depending on the type of boot image including the operating system executed by the server computer A computer, a load distribution device that distributes requests from the plurality of client computers to a plurality of load distribution target computers including the plurality of server computers, and a spare computer that can be used as the load distribution target computer, This is applied to a load balancing system that is not normally started but includes a spare computer that is started when used as the load balancing target computer. When a spiked load occurs in the server computer group, a new load is generated. Total for assigning spare computer as load balancing target computer A machine management program,
In the calculator,
For each group of the plurality of server computer groups, the load amount of the entire group is obtained based on the load amounts of all load balancing target computers including the server computers belonging to the group among the plurality of server computers. Steps,
When the load amount of the entire group is obtained in units of each of the plurality of server computer groups, based on the load amount, the group should start to be used as a load balancing target computer or load Determining the number of spare computers that should be discontinued as distributed computers;
When the number of spare computers to be used in the server computer group is determined, determining a spare computer to be newly used as a load distribution target computer in the group for the determined number; ,
When a spare computer to be used in the computer group is determined, the determined spare computer is executed by a server computer belonging to the group among the plurality of server computers as a boot image to be booted. Selecting a boot image specific to that group of types,
When a spare computer to be used in the computer group is determined and a boot image unique to the group is selected, the determined spare computer is booted so as to boot with the selected boot image. And steps to
When a spare computer to be used in the computer group is determined, adding the determined spare computer as a load balancing target computer to be a load balancing target in the group by the load balancer When,
When the number of spare computers to be used in the server computer group is determined, determining a spare computer to be used in the group of the determined number;
When a spare computer to be terminated in the computer group is determined, stopping the determined spare computer;
When a spare computer to be used in the computer group is determined, a step of excluding the determined spare computer from a load distribution target in the group by the load balancer is executed. Computer management program. A plurality of server computers that process requests from a plurality of client computers, and are managed by being divided into a plurality of server computer groups depending on the type of boot image including the operating system executed by the server computer A plurality of server sites each including a computer and a set of load distribution devices that distribute requests from the plurality of client computers to a plurality of load distribution target computers including the plurality of server computers, and used as the load distribution target computer In a load balancing system in which a spare computer including a spare computer activated when used as a load balancing target computer is connected by a network, the spare computer is applied to each of the plurality of server sites. The server in the server site If the load spike condition occurs in calculation machine group, a computer management program for causing assigned to the group preliminary computer from the preliminary site as a new load balancing target computer,
In the computer located at the server site,
In the server site, based on the load amount of all load balancing target computers including the server computers belonging to the group among the plurality of server computers in units of each of the plurality of server computer groups, the entire group Determining the load amount of
In the server site, when the load amount of the entire group is obtained for each group of the plurality of server computer groups, a load distribution target is newly added to the group in the server site based on the load amount. Determining the number of spare computers to be used as computers or to be terminated as load balancing target computers;
In the server site, when the number of spare computers to be used in one group among the plurality of server computer groups belonging to the server site is determined, Requesting the start of use of the determined number of spare computers in the one group in the server site via the network;
In response to a request to start using the spare site, a spare computer to be used as a load distribution target computer in the requested group in the requesting server site is determined at the spare site, and the determined spare When the spare site is notified to the requesting server site that the computer has been booted with the type of boot image being executed on the server computer belonging to the group, the notified spare computer is Adding as a load balancing target computer to be a load balancing target in the group by the load balancing device;
In response to a request for termination of use for the spare site, a spare computer to be terminated as a load distribution target computer in the requested group in the requesting server site is determined at the spare site. When the spare computer is notified from the spare site to the requesting server site, the notified spare computer is excluded from the load balancing target within the group by the load balancer. And a computer management program for executing the steps. A plurality of server computers that process requests from a plurality of client computers, and are managed by being divided into a plurality of server computer groups depending on the type of boot image including the operating system executed by the server computer A plurality of server sites each including a computer and a set of load distribution devices that distribute requests from the plurality of client computers to a plurality of load distribution target computers including the plurality of server computers, and used as the load distribution target computer In a load balancing system in which a spare computer including a spare computer that is activated when used as the load balancing target computer is connected by a network, the spare computer is applied to the spare site, and the server site The server computer group in If the load spike condition occurs, a computer management program for allocating a spare computer from the preliminary site to the group as a new load balancing target computer in response to a request from the server site,
In the computer located at the spare site,
In each of the plurality of server sites, based on the load amount of all load balancing target computers including the server computers belonging to the group among the plurality of server computers, in units of each of the plurality of server computer groups. As a result of determining the number of spare computers that should be newly used as load balancing target computers in the group in the server site or to end use as load balancing target computers, When a site is requested to start or end use of the determined number of spare computers in the group in the server site via the network, either start of use or end of use is requested. A step of determining whether or not
When the start of use is requested, the number of spare computers to be used as load balancing target computers in the requested group in the request source server site is determined for the number determined in the request source server site. Steps,
When a spare computer to be used as a load distribution target computer is determined in the requested group in the request source server site, the request source server is used as a boot image to boot the determined spare computer. Selecting a boot image specific to the group of the type being executed on the server computer belonging to the group in the site;
When a standby computer to be used as a load balancing target computer is determined in the requested group in the requesting server site and a boot image unique to the group is selected, boot with the selected boot image Activating the determined spare computer, and
When the use termination is requested, the number of spare computers to be terminated as load balancing target computers in the requested group in the request source server site is determined for the number determined in the request source server site. And steps to
When a standby computer to be terminated as a load balancing target computer in the requested group in the requesting server site is determined, stopping the determined standby computer;
And a step of notifying the requesting server site of the processing result for the use start or use end request from the spare site.

Images