JP2010198060A - Server system and server system control method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To efficiently operate a server system having redundant configurations including active servers and standby servers. <P>SOLUTION: The server system includes: a plurality of redundant configurations 10, 20 respectively having the active servers 1A, 2A and the standby servers 1B, 2B; and a management server 3 for managing the redundant configurations 10, 20. The management server 3 partially changes-over the standby servers, based on the resource states of the active servers. When the resource states of the active servers are high-load states, for example, the management server 3 reconfigures the redundant configurations, so as to allow the number of the active servers to be larger than the number of the active servers in a normal state. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a server system and a server system control method, and more particularly to a server system having a redundant configuration having an active server and a standby server and a server system control method.

  In a server system, since it is necessary to provide services continuously even if a failure occurs in a server, a redundant configuration including an active server and a standby server is generally used. In a server system having a redundant configuration, when a failure occurs in the active server, the service can be continuously provided by immediately switching to the standby server.

  Patent Documents 1 and 2 disclose a server system having a redundant configuration including such an active server and a standby server. The server system disclosed in Patent Document 1 is shown in FIG. A server system 110 illustrated in FIG. 5 has a redundant configuration including a server 121 and a server 122, and the server 121 and the server 122 are connected via a network 170. When the server 121 is operating as the active system and the server 122 is the standby system, when a failure occurs in the server 121, the files in the HDD (magnetic storage device) 141 are stored in the file replication control unit 161 of the server 121 and the server 122. The data is transferred to the HDD 142 of the server 122 by the file replication control unit 162.

  A server system disclosed in Patent Document 2 is shown in FIG. The server system shown in FIG. 6 includes an active server 210 and a standby server 211. The active server 210 includes a process 220, a shared memory 230, a CPU monitoring function unit 240, a replication function unit 250, and a management table 260. The process 220 is a program for performing data processing, and is a unit of processing executed by the CPU. The shared memory 230 writes and reads data. The CPU monitoring function unit 240 monitors the CPU usage rate. The replication function unit 250 performs replication (memory synchronization). The management table 260 stores management information of data that has not been replicated. The standby server 211 has the same configuration as that of the active server 210. When the service of the active server 210 is stopped due to a failure or the like, the active server 210 and the standby server 211 are switched, and the standby server 211 starts the service as the active system.

  FIG. 4 is a diagram showing a server system having a plurality of redundant configurations 100 including an active server 100A and a standby server 100B. In the server system shown in FIG. 4, the servers 100A, 100B, 200A, and 200B constituting the redundant configurations 100 and 200 are managed by the management server.

JP 2007-328595 A JP 2007-286952 A

  However, in the redundant server system shown in FIGS. 4 to 6, the system is configured by combining two servers, an active server and a standby server. In such a system configuration, the standby server normally remains in a standby state and is not used. As shown in FIG. 4, in a server system having a plurality of redundant configurations in which two servers, an active server and a standby server, are combined, the standby server is used when the system is composed of n active servers. Including 2n servers in total is required. In such a system configuration, only half the number of servers in the entire system can be exhibited, which is not efficient.

  Accordingly, an object of the present invention is to efficiently operate a server system having a redundant configuration including an active server and a standby server.

  The server system according to the present invention includes a plurality of redundant configurations having an active server and a standby server, and a management server that manages the plurality of redundant configurations, and the management server includes a resource state of the active server. Based on the above, a part of the standby server is switched to the active server.

  A server system control method according to the present invention is a server system control method including a plurality of redundant configurations having an active server and a standby server, and includes the following steps. Monitoring the resource status of the active server. A step of switching a part of the standby server to the active server based on the monitoring result of the resource status of the active server.

  According to the present invention, it is possible to efficiently operate a server system having a redundant configuration including an active server and a standby server.

It is a figure which shows the server system concerning embodiment. (A) is a figure which shows the structure of the server system at the time of a low load state, (b) is a figure which shows the structure of the server system at the time of a high load state. It is a figure which shows the operation | movement sequence in case the server system concerning embodiment reconfigure | reconfigures a system (from a low load state to a high load state). It is a figure which shows the operation | movement sequence in case the server system concerning embodiment reconfigure | reconfigures a system (from a high load state to a low load state). It is a figure for demonstrating the server system concerning background art. It is a figure for demonstrating the server system concerning background art. It is a figure for demonstrating the server system concerning background art.

Embodiments of the present invention will be described below with reference to the drawings.
FIG. 1A is a diagram showing a server system according to the present embodiment in a low load state (normal state). In FIG. 1A, the redundant configuration 10 includes an active server 1A and a standby server 1B. The redundant configuration 20 includes an active server 2A and a standby server 2B. The normal server system according to the present embodiment has a plurality of redundant configurations each including one active server and one standby server. Also, the operation of these redundant configurations is managed by the management server 3.

  The management server 3 has a function of monitoring the resource state (for example, CPU usage rate, memory usage rate, storage usage rate) of each active server. Then, it is determined whether the resource state of the active server is a high load state (overload state), and the redundant configuration of the active server and the standby server is reconfigured. That is, when the management server 3 determines that the active servers 1A and 2A are in a high load state by monitoring the resource state of the management server 3, the management server 3 is one or more servers operating as standby servers. Switch to the active system and reconfigure the redundant configuration. Conversely, if the management server 3 determines that the resource state of the active server has changed from a high load state to a low load state, the management server is redundant so that it returns to the configuration (normal state) of two units. Reconfigure the configuration. That is, the reconfiguration of the redundant configuration can be a 1: 1 configuration in a low load state and an n: 1 configuration in a high load state.

  FIG. 1B is a diagram illustrating a configuration example of a server system in a high load state. In FIG. 1B, the server 1B that was a standby system in a low load state is used as an active server, and a 3: 1 redundant configuration 30 of three active servers and one standby server is reconfigured. .

  In FIGS. 1A and 1B, the server system having two redundant configurations (10, 20) has been described as an example. However, the server system according to the present embodiment has three or more redundant configurations. The same can be applied to a server system having the same. That is, in the server system according to the present embodiment, the management server sets the ratio of the active server to the standby server to (2n-1): 1 when the resource state of the active server is high. be able to. The operation of the server system according to this embodiment will be specifically described below.

  FIG. 2 is a diagram for explaining an operation of reconfiguring the system when the server system according to the embodiment changes from a low load state to a high load state. Also in FIG. 2, in a low load state (normal state), the server 1 </ b> A operates as an active server and the server 1 </ b> B operates as a standby server, and these constitute a redundant configuration 10. Further, the server 2A operates as an active server, and the server 2B operates as a standby server, and these constitute a redundant configuration 20. The management server monitors the resource status of the active servers 1A and 2A. In addition, the active server 1A and the standby server 1B perform replication processing with each other. The active server 2A and the standby server 2B also perform replication processing with each other.

  The management server monitors the resource statuses of the active servers 1A and 2A. For example, if the resource usage rate of the active servers exceeds a threshold for a certain time, it is determined that the load is in a high load state (S1). When determining that the management server is in a high load state, the management server instructs the standby server (server 1B) to transition to the active system (S2). At this time, the management server stores information on the standby server 2B, which becomes a new standby server of the active server 1A, and information on the standby server 2B, which becomes the standby server when the server 1B transitions to the active server. Notify 1B.

  In addition, the standby server 1B that has received an instruction to transition to the active system takes over the replication data of the active server 1A to the standby server 2B (S3). Then, the server 1B notifies the active server 1A paired with the server 1B that the standby server has been changed from the server 1B to the server 2B (S4).

  Next, the standby server 1B notifies the management server that the transition from the standby server to the active server has been completed (S5). Then, the server 2B is a standby server, and the server 1B starts a service as an active server. The management server that has received the notification of completion of the state transition from the server 1B updates the operation state of the system and starts operation with the system configuration after reconfiguration (S6). In addition, the server 1B transitioned to the active system notifies the standby system server 2B notified from the management server of an active system addition request so as to perform replication processing with the server 1B (S7). Note that the order of notification to each server may be changed as long as consistency can be obtained.

  The standby server 2B takes over the replication data of the active server 1A from the standby server 1B, and then performs a replication process for the active server 1A. The standby server 2B also performs replication processing for the active server 1B in response to the active addition request received from the active server 1B.

  Next, the operation of the server system when the server system according to the present embodiment changes from a high load state to a low load state will be described with reference to FIG. Since the initial state in FIG. 3 is a high load state, the servers 1A, 1B, and 2A operate as an active server, and the server 2B operates as a standby server, which constitutes a redundant configuration 30. The management server monitors the resource status of the active servers 1A, 1B, 2A. In addition, the active server 1A and the standby server 2B perform replication processing with each other. The active server 1B and the standby server 2B also perform replication processing with each other. The active server 2A and the standby server 2B also perform replication processing with each other.

  The management server monitors the resource status of the active servers 1A, 1B, 2A. For example, when the resource usage rate of the active server falls below the threshold value for a certain time, it is determined that the load is not in a high load state (S11). . When determining that the management server is not in a high load state, the management server instructs the active server (server 1B) to transition to the standby system (S12). At this time, the management server notifies the server 1B of the information of the active server 1A that replaces the active server 1B and the information of the active server 1A that is paired when the server 1B transitions to the standby system. .

  The active server 1B that has received the instruction to transition to the standby system takes over the data being executed on the active server 1B to the active server 1A and performs replication processing on the active server 1A. (S13). The active server 1A takes over the data processed by the active server 1B and continues the processing instead of the active server 1B.

  After taking over all processing, the active server 1B notifies the standby server 2B of the active server 1A that the server 1B has transitioned to the standby server of the active server 1A (S14). Next, the active server 1B notifies the standby server 2B of the active server 1B of an active system deletion request so as to delete the server 1B from the active system (S15). Next, the active server 1B notifies the management server that the transition from the active server to the standby server has been completed (S16).

  After all notifications are completed, the active server 1B transitions to the standby server and starts the operation of the active server 1A as the standby server. Here, the order of notification to each server may be changed within a range where consistency can be obtained. The standby server 2B that has received notification from the active server 1B that the active server 1B has changed to the standby server of the active server 1A ends the replication process with the server 1A. The standby server 2B that has received the active deletion request from the active server 1B ends the replication process with the server 1B.

  The management server that has received the notification of completion of the state transition from the active server 1B updates the operating state of the system (S17), and starts operation with the system configuration after reconfiguration. The server system according to the present embodiment provides an application service while repeating the processes of FIGS. 2 and 3 according to the resource state of the server.

  In the server system according to the present embodiment described above, in the low load state (normal state), there are two active servers (1A, 2A) and two standby servers (1B, 2B), In a high load state, there are three active servers (1A, 1B, 2A) and one standby server (2B). However, the server system according to the present embodiment is not limited to such a combination. For example, the server system may be configured by three active servers and three standby servers. In this case, it can be reconfigured as four active servers and two standby servers in a high load state. Further, it can be reconfigured as five active servers and one standby server.

  Further, for example, the server system may be configured by four active servers and four standby servers. In this case, it can be reconfigured as five active servers and three standby servers in a high load state. It can also be reconfigured as 6 active servers and 2 standby servers. Furthermore, it can be reconfigured as seven active servers and one standby server. As described above, the management server can appropriately set the ratio of the active server to the standby server in consideration of the load on the active server.

  In the above example, the case where the configuration of the active server and the standby server in the low load state is a 1: 1 configuration has been described. However, in the server system according to the present embodiment, the configuration ratio of the active server and the standby server in a low load state can be other than 1: 1. For example, in a low load state (normal state), there can be four active servers and two standby servers, and in a high load state, there can be five active servers and one standby server.

  That is, in the server system according to the present embodiment, when the resource state of the active server is in a high load state, a part of the standby server is transitioned to the active server. As a result, the number of active servers can be increased in the high load state than in the low load state, and the server system can be operated efficiently.

  In addition, a control method for a server system having a plurality of redundant configurations having an active server and a standby server according to the present embodiment includes the following steps. The process of monitoring the resource status of the active server. A step of switching a part of the standby server to the active server based on the monitoring result of the resource status of the active server.

  In the server system control method according to the present embodiment, when the resource state of the active server is in a high load state, the number of active servers may be larger than the number of standby servers. In the server system control method according to the present embodiment, when the resource state of the active server is a high load state, the ratio of the active server to the standby server is set to (2n−1): 1. May be. In the server system control method according to the present embodiment, when the resource state of the active server is the normal state, the ratio of the active server to the standby server may be 1: 1.

  Also in the server system control method according to the present embodiment, when the resource state of the active server is in a high load state, a part of the standby server is changed to the active server. As a result, the number of active servers can be increased in the high load state than in the low load state, and the server system can be operated efficiently.

  Although the present invention has been described with reference to the above-described embodiment, the present invention is not limited to the configuration of the above-described embodiment, and those skilled in the art within the scope of the invention of the claims of the present application claims. It goes without saying that various modifications, corrections, and combinations that can be achieved are included.

1A Active server 1B Standby server (low load), active server (high load)
2A Active server 2B Standby server 3 Management servers 10, 20, 30 Redundant configuration

Claims (12)

A plurality of redundant configurations having an active server and a standby server;
A management server for managing the plurality of redundant configurations,
The server system, wherein the management server switches a part of the standby server to the active server based on the resource state of the active server.   The management server reconfigures the redundant configuration so that the number of active servers is larger than the number of active servers in a normal state when the resource state of the active servers is a high load state. Item 4. The server system according to Item 1.   2. The server system according to claim 1, wherein the management server sets a ratio of the active server to the standby server to (2n−1): 1 when a resource state of the active server is a high load state. .   The management server according to any one of claims 1 to 3, wherein a ratio of the active server to the standby server is 1: 1 when the resource state of the active server is a normal state. Server system. The plurality of redundant configurations include a first redundant configuration having a first active server and a first standby server, and a second redundancy having a second active server and a second standby server. And having a configuration,
2. The management server according to claim 1, wherein when the resource state of the first or second active server is in a high load state, one of the first and second standby servers is set as an active server. The server system described.   The server system according to claim 1, wherein the management server monitors a resource state of the active server based on a CPU usage rate of the active server.   The server system according to any one of claims 1 to 6, wherein the management server monitors a resource state of the active server based on a memory usage rate of the active server.   The server system according to any one of claims 1 to 7, wherein the management server monitors a resource state of the active server based on a storage usage rate of the active server. A server system control method comprising a plurality of redundant configurations having an active server and a standby server,
Monitor the resource status of the active server,
A control method of a server system, wherein a part of the standby server is switched to an active server based on a monitoring result of a resource state of the active server.   The server system according to claim 9, wherein when the resource state of the active server is in a high load state, the redundant configuration is reconfigured so that the number of the active servers is larger than the number of the standby servers. Control method.   The server system control method according to claim 9, wherein when the resource state of the active server is a high load state, a ratio of the active server to the standby server is set to (2n−1): 1.   12. The server system control method according to claim 9, wherein when the resource state of the active server is a normal state, the ratio of the active server to the standby server is set to 1: 1.

Images