JP2011253371A - Scale-out system, and method and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a scale-out technique of a server capable of suppressing unnecessary consumption of computer resources and effectively performing load distribution.SOLUTION: A scale-out system comprises: a host server 3 having a virtual server group consisting of a plurality of virtual servers; a monitoring server 4 having a function to monitor the virtual server group of the host server 3; a dependency relation management server 5 having a function to manage dependency relation of software of the virtual server group of the host server 3; and a configuration change server 6 having a function to make a request for configuration change to virtualization software for managing the virtual server group of the host server 3. The monitoring server 4 detects a virtual server that is the cause of heavy load, and the dependency relation management server 5 obtains a system depending on a process of the detected virtual server. The configuration change server 6 moves only the obtained system to a virtual server of a scale-out destination.

Description

  The present invention relates to a server scale-out technique, and more particularly to a system, method, and program for performing server scale-out using software dependency.

  In general, in a server system composed of multiple servers, a load balancing device is placed in the front-end part so that load can be distributed under high load, and transactions are given priority to servers with low load. Requests are being made. For example, Patent Document 1 (Japanese Patent Application Laid-Open No. 2005-141441) describes a method of statistically processing an access log of access to a server group that is a target of load distribution and distributing transactions based on the processing result. Yes.

  However, with the method disclosed in Patent Document 1, it is difficult to improve the processing capacity of the entire system according to the amount of transactions, and it is not possible to deal with transactions that exceed the total processing capacity of the servers constituting the system. There is. A server scale-out is known as a mechanism for solving this problem.

JP-A-2005-141441

  A general scale-out method is realized by creating a copy of a server whose performance is to be improved and distributing transactions for the server with a load balancer. Server replication performed in this scale-out method is a complete copy of the target server. For example, a physical server can be added by installing all the software installed on the replication source server, and a virtual server can be realized by creating a clone. In order to distribute the transaction by the load balancer, the host name and the IP address are changed. However, regarding the installed software, the replication source and the replication destination have the same configuration.

  That is, when the number of servers is increased by the conventional scale-out method, all software is subject to processing distribution. For example, if a server running two systems is scaled out when only one system is loaded, not only the loaded system but also the unloaded system is 2 Computer resources will be used. The system here refers to a plurality of software groups operating with a purpose.

  The problem with the conventional scale-out method described above is that a scale-out server can be prepared for each system so that scale-out can be performed for each required function. There is a problem that the efficiency is low.

  Therefore, in the present invention, a server that solves the problems of the conventional scale-out method, and its typical purpose is to suppress unnecessary consumption of computer resources and to perform load distribution efficiently. Is to provide scale-out technology.

  The above and other objects and novel features of the present invention will be apparent from the description of this specification and the accompanying drawings.

  Of the inventions disclosed in the present application, the outline of typical ones will be briefly described as follows.

  That is, in order to achieve the above-described object, the present invention realizes a mechanism for automatically scaling out a server using software dependency.

  The present invention relates to a host server having a virtual server group composed of a plurality of virtual servers, a monitoring server having a function of monitoring the virtual server group of the host server, and a function of managing the software dependency of the virtual server group of the host server. And a configuration change server having a function of requesting a configuration change to the virtualization software that manages the virtual server group of the host server.

  In this scale-out system, the monitoring server detects the virtual server that caused the high load, and the dependency management server acquires the system that depends on the detected virtual server process. The above-mentioned problem is solved by moving only the system to the virtual server of the scale-out destination by the configuration change server.

  The present invention can also be applied to a method using a system for performing the scale-out and a program for causing a computer to function as the system for performing the scale-out.

  Among the inventions disclosed in the present application, effects obtained by typical ones will be briefly described as follows.

  That is, the scale-out of the present invention makes it possible to distribute the load in units of systems, so that unnecessary computer resource consumption can be suppressed and load distribution can be performed efficiently. Moreover, since all the operations related to scale-out are automatically performed by the system of the present invention, the burden on the server operator is reduced.

It is a system configuration figure showing an example of the composition of the scale-out system in one embodiment of the present invention. In the scale-out system of FIG. 1, it is explanatory drawing which shows an example of the information item of a dependence relationship management database. 2 is a flowchart illustrating an example of the operation of the system from the occurrence of a high load to the completion of server scale-out in the scale-out system of FIG. 1. It is explanatory drawing which shows an example of implementation | achievement of a scale out in the scale out system of FIG.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. Note that components having the same function are denoted by the same reference symbols throughout the drawings for describing the embodiment, and the repetitive description thereof will be omitted.

<Configuration of scale-out system>
A system configuration of a system for performing scale-out in the present invention (herein referred to as a scale-out system) will be described with reference to FIG. FIG. 1 is a system configuration diagram illustrating an example of a configuration of a scale-out system according to an embodiment of the present invention.

  The scale-out system in the present embodiment includes a load distribution device 2 connected to the Internet 1, a host server 3 connected to the load distribution device 2, a monitoring server 4 connected to the host server 3 and connected to the same network, a dependency relationship It consists of a management server 5 and a configuration change server 6. As a path for performing scale-out, the monitoring server 4 is connected to the dependency management server 5, and the dependency management server 5 is connected to the configuration change server 6.

  The host server 3 has a virtualization function, and includes a virtual server group 9 including a virtual server A7 and a virtual server B8 that provide services, and a virtual server for copying the virtual server A7 and the virtual server B8. A virtual server creation template 10 including an A template 11 and a virtual server B template 12 is provided.

  The monitoring server 4 has a function of monitoring the virtual server group 9 and detects a high load state of each virtual server A7 and virtual server B8 of the virtual server group 9.

  The dependency relationship management server 5 has a function of managing the software dependency relationship of the virtual server group 9 and includes a dependency relationship management database 13 for managing the software dependency relationship of the virtual server group 9.

  The configuration change server 6 has a function of requesting a configuration change to the virtualization software that manages the virtual server group 9, and includes a virtual server replication script 14 that performs replication of the virtual server, and an operating process A process stop script 15 for executing stop.

  This scale-out system is constructed using, for example, a computer system including a CPU and hardware such as various memories and peripheral devices. In this computer system, various types of software (scale-out programs) stored in a program storage memory are executed under the control of the CPU, whereby the load distribution device 2 and the host server 3 having the respective functions are executed. The monitoring server 4, the dependency relationship management server 5, and the configuration change server 6 can be realized. The virtual server creation template 10 and the dependency relationship management database 13 can be realized on a data storage memory.

  In this scale-out system, a client terminal (not shown) is connected to the Internet 1 to which the load distribution apparatus 2 is connected, and the load distribution apparatus 2 handles the virtual server group 9 of the host server 3 with respect to access from this client terminal. Scale-out to distribute transactions is possible. In this scale-out, when creating a replica of a virtual server for which performance is to be improved, it can be realized by creating a clone of the virtual server.

<Information item of dependency management database>
The dependency management database 13 arranged in the above-described dependency management server 5 will be described with reference to FIG. FIG. 2 is an explanatory diagram showing an example of information items in the dependency management database 13.

  The dependency management database 13 in the present embodiment holds various types of information related to the software of the virtual server group 9. The dependency relationship management database 13 stores server name 13 (1), system name 13 (2), software name 13 (3), and process name 13 (4) as information items.

  The server name 13 (1) stores the name of the virtual server. In this example, “virtual server A” and “virtual server B” are stored corresponding to FIG.

  The system name 13 (2) stores the name of the system. In this example, “System 01”, “System 02”, and “System 03” are stored. The system mentioned here refers to a plurality of software groups that operate with purpose on a virtual server.

  The software name 13 (3) stores the name of the software. In this example, “Soft w”, “Soft x”, “Soft y”, and “Soft z” are stored. The software mentioned here refers to software in a system (a plurality of software groups).

  The process name 13 (4) stores the name of the process. In this example, “Proc α”, “Proc β”, “Proc γ”, and “Proc Ω” are stored. The process here refers to an operation by software.

<Operation of scale-out system>
With reference to FIGS. 1 and 2, the operation performed in the scale-out of the present invention will be described with reference to FIG. FIG. 3 is a flowchart showing an example of the operation of the system from the occurrence of a high load to the completion of server scale-out. FIG. 4 is an explanatory diagram showing an example of realizing scale-out.

  The operations performed in the scale-out according to the present embodiment are started when a high load state occurs in one virtual server, for example, virtual server A7.

  First, in step ST1, the monitoring server 4 detects the high load state of the monitored server, here the virtual server A7.

  In step ST2, the monitoring server 4 sets the process of the virtual server A7 causing the high load, for example, the process name “Proc α” together with the server name, here “virtual server A”, to SNMP (simple network management protocol) is input to the trap message and transmitted to the dependency management server 5.

  Subsequently, in step ST <b> 3, the dependency relationship management server 5 receives the SNMP trap transmitted from the monitoring server 4.

  Then, in step ST4, the dependency management server 5 that has received the SNMP trap uses the name of the process that caused the high load, which is one of the information input to the SNMP trap, here “Proc α”. The system name on which this process operates is retrieved and acquired from the dependency management database 13, and a list of processes depending on the target system is acquired. In this example, the process name “Proc α” (13 (4) first line) of “virtual server A” (13 (1) first line) is the system name “System 01” (13 (2) first line). The software name “Soft w” (13 (3), the first line) that operates in the first), and the system name “System 01” (13 (2), the second line) includes the software name “Soft x”. "(13 (3) 2nd line)", the process name "Proc α" (13 (4) 1st line) and the process name "Proc β" (13 (4) 2nd line) are listed. Get as.

  Furthermore, in step ST5, the dependency management server 5 retrieves and acquires from the dependency management database 13 a process which has caused a high load, in this case, a process which does not depend on the process name “Proc α”. Get a list of independent processes. In this example, in “virtual server A” (13 (1) 3rd line), the software name “Soft y” (13 (3) 3rd line) in the system name “System 02” (13 (2) 3rd line). ) Is operating, the process name “Proc γ” (13 (4) 3rd line) is acquired as a list.

  Then, in step ST6, the dependency management server 5 uses the server name, which is another information input to the SNMP trap, here “virtual server A”, and deploys a virtual server template, here a virtual server. The server A template 11 is selected and a configuration change request is transmitted to the configuration change server 6. At this time, the process change request (“Proc α” and “Proc β”) acquired in step ST4 and the process list acquired in step ST5 (“Proc γ”) are input to the configuration change request as data.

  Subsequently, in step ST <b> 7, the configuration change server 6 receives the configuration change request transmitted from the dependency management server 5.

  In step ST8, the configuration change server 6 executes the virtual server replication script 14 and creates a replica of the virtual server A7 using the virtual server A template 11.

  Furthermore, in step ST9, the configuration change server 6 determines whether or not the process list (“Proc γ”) acquired in step ST5 is input to the received configuration change request data, that is, the process that caused the high load. Determine whether there is a process that does not depend on. If there is a process that does not depend on (Yes), the process proceeds to step ST10, and if there is no process that does not depend on (No), the process proceeds to a scale-out completion state. In this example, since there is a process (“Proc γ”) that does not depend on the process (“Proc α”) that caused the high load, the process proceeds to step ST10.

  In step ST10, the configuration change server 6 executes the process stop script 15 on the virtual server A7, and stops all processes that do not depend on the process list acquired in step ST5, that is, the process causing the high load. Let In this example, the process (“Proc γ”) is stopped.

  As a result, in this example, the scale-out as shown in FIG. 4 is realized for the virtual server A7. That is, as shown in FIG. 4, the virtual server A (replication source) 7 can operate a system (“System 01”) in which a process (“Proc α”) and a process (“Proc β”) operate, The system (“System 02”) on which the process (“Proc γ”) operates is stopped. The virtual server A (replication destination) 7 ′ includes a system (“System 01”) in which a process (“Proc α”) and a process (“Proc β”) operate, and a process (“Proc γ”). A working system ("System 02") is replicated.

<Effects of the present embodiment>
According to the scale-out system in the present embodiment described above, the monitoring server 4, the dependency management server 5, and the configuration change server 6 connected to the host server 3 including the virtual server group 9 are included. By moving only the system that depends on the virtual server process to the virtual server that is the scale-out destination, it becomes possible to distribute the load on a system-by-system basis. It becomes possible to perform load balancing.

  In addition, since all operations related to scale-out are automatically performed by the scale-out system according to the present embodiment, the burden on the server operator is reduced.

  As mentioned above, the invention made by the present inventor has been specifically described based on the embodiment. However, the present invention is not limited to the embodiment, and various modifications can be made without departing from the scope of the invention. Needless to say.

  The server scale-out technique of the present invention can be used in a system, method, and program for performing server scale-out using software dependency.

DESCRIPTION OF SYMBOLS 1 Internet 2 Load balancer 3 Host server 4 Monitoring server 5 Dependency management server 6 Configuration change server 7 Virtual server A
8 Virtual server B
9 Virtual server group 10 Virtual server creation template 11 Virtual server A template 12 Virtual server B template 13 Dependency management database 14 Virtual server replication script 15 Process stop script

Claims (5)

A host server having a virtual server group composed of a plurality of virtual servers;
A monitoring server having a function of monitoring the virtual server group of the host server;
A dependency management server having a function of managing software dependency of the virtual server group of the host server;
A configuration change server having a function of requesting a configuration change to virtualization software that manages the virtual server group of the host server;
The monitoring server detects a virtual server that has caused a high load, acquires a system that depends on the detected virtual server process with the dependency management server, and uses the configuration change server to acquire only the acquired system. A scale-out system, wherein the scale-out system is moved to a scale-out destination virtual server. The scale-out system according to claim 1, wherein
The dependency management server includes a virtual server name of the virtual server and a plurality of servers operating on the virtual server in order to associate the virtual server causing the high load with a system dependent on the process of the virtual server. A dependency management database for storing information consisting of the system name of the system consisting of the software group and the process name of the process by the software operating in the system,
The configuration change server has a virtual server replication script for executing replication of the virtual server, and a process stop script for executing process stop during operation.
The scale-out system, wherein the host server has a template for creating a virtual server for duplicating each virtual server of the virtual server group. The scale-out system according to claim 2,
The monitoring server detects a high load state of the virtual server, inputs the process name of the virtual server causing the high load into an SNMP trap message together with the server name, and sends it to the dependency management server. Send to
The dependency management server receives the SNMP trap transmitted from the monitoring server, and uses the process name that caused the high load input to the SNMP trap to determine the name of the system in which this process operates. Retrieve from the dependency management database, obtain a list of processes that depend on this system, and further retrieve from the dependency management database a process that does not depend on the process that caused the high load, The independent process list is acquired, and the template of the virtual server to be deployed is selected using the server name input to the SNMP trap, and the dependent process list and the independent process list are configured. Enter the change request and send it to the configuration change server,
The configuration change server receives the configuration change request transmitted from the dependency management server, executes the virtual server replication script to create a replica of the virtual server using the template of the virtual server, Further, it is determined whether or not there is an independent process in the received configuration change request data. If there is an independent process, the process stop script is executed on the virtual server, and the independent process list is executed. The scale-out system is characterized in that the scale-out is completed by stopping all of the processes, and the scale-out is completed when there is no independent process. A host server having a virtual server group composed of a plurality of virtual servers;
A monitoring server having a function of monitoring the virtual server group of the host server;
A dependency management server having a function of managing software dependency of the virtual server group of the host server;
A scale-out method by a scale-out system having a configuration change server having a function of requesting a configuration change to virtualization software that manages the virtual server group of the host server,
The dependency management server associates a virtual server that causes a high load with a system that depends on a process of the virtual server, and a virtual server name of the virtual server and a plurality of servers that operate on the virtual server. A dependency management database that stores information including a system name of a system composed of software groups and a process name of a process by software operating in the system;
The configuration change server has a virtual server replication script for executing replication of the virtual server, and a process stop script for executing process stop during operation.
The host server has a template for creating a virtual server for duplicating each virtual server of the virtual server group,
The monitoring server detects a high load state of the virtual server, and inputs the process name of the virtual server that has caused the high load in an SNMP trap message together with the server name to the dependency management server. Sending to the
The dependency management server receives the SNMP trap transmitted from the monitoring server, and uses the process name that caused the high load input in the SNMP trap to determine the system name on which the process operates. Retrieve from the dependency management database, obtain a list of processes that depend on this system, and further retrieve from the dependency management database a process that does not depend on the process that caused the high load, The independent process list is acquired, and the template of the virtual server to be deployed is selected using the server name input to the SNMP trap, and the dependent process list and the independent process list are configured. Inputting into a change request and transmitting to the configuration change server;
The configuration change server receives a configuration change request transmitted from the dependency management server, executes the virtual server replication script to create a replica of the virtual server using the template of the virtual server, Further, it is determined whether or not there is an independent process in the received configuration change request data. If there is an independent process, the process stop script is executed on the virtual server, and the independent process list is executed. A scale-out method, comprising the steps of: stopping all the processes of the process to complete the scale-out, and completing the scale-out when there is no independent process. A host server having a virtual server group composed of a plurality of virtual servers;
A monitoring server having a function of monitoring the virtual server group of the host server;
A dependency management server having a function of managing software dependency of the virtual server group of the host server;
A scale-out program for causing a computer to function as a scale-out system having a configuration change server having a function of requesting a configuration change to virtualization software that manages the virtual server group of the host server,
The dependency management server associates a virtual server that causes a high load with a system that depends on a process of the virtual server, and a virtual server name of the virtual server and a plurality of servers that operate on the virtual server. A dependency management database that stores information including a system name of a system composed of software groups and a process name of a process by software operating in the system;
The configuration change server has a virtual server replication script for executing replication of the virtual server, and a process stop script for executing process stop during operation.
The host server has a template for creating a virtual server for duplicating each virtual server of the virtual server group,
The monitoring server detects a high load state of the virtual server, and inputs the process name of the virtual server that has caused the high load in an SNMP trap message together with the server name to the dependency management server. Sending to the
The dependency management server receives the SNMP trap transmitted from the monitoring server, and uses the process name that caused the high load input in the SNMP trap to determine the system name on which the process operates. Retrieve from the dependency management database, obtain a list of processes that depend on this system, and further retrieve from the dependency management database a process that does not depend on the process that caused the high load, The independent process list is acquired, and the template of the virtual server to be deployed is selected using the server name input to the SNMP trap, and the dependent process list and the independent process list are configured. Inputting into a change request and transmitting to the configuration change server;
The configuration change server receives a configuration change request transmitted from the dependency management server, executes the virtual server replication script to create a replica of the virtual server using the template of the virtual server, Further, it is determined whether or not there is an independent process in the received configuration change request data. If there is an independent process, the process stop script is executed on the virtual server, and the independent process list is executed. A scale-out program, wherein the process is stopped to complete the scale-out, and if there is no independent process, the step of completing the scale-out is executed.

Images