WO2017047065A1

WO2017047065A1 - Cluster system, information processing device, synchronization method for cluster system, and storage medium for storing program

Info

Publication number: WO2017047065A1
Application number: PCT/JP2016/004136
Authority: WO
Inventors: 文隆江部
Original assignee: 日本電気株式会社
Priority date: 2015-09-17
Filing date: 2016-09-12
Publication date: 2017-03-23
Also published as: JPWO2017047065A1; JP6555353B2

Abstract

Provided is a cluster system which is capable of restricting node operating costs. The cluster system 100 is provided with an active node 110, a standby node 120, a control device 130 and a storage unit 140. The active node 110 has: a sync data generating unit 112 which generates sync data for synchronizing storage content of a storage unit 111 of the active node 110 with storage content of a storage unit 121 of the standby node 120; a sync data transmitting unit 113 which stores the sync data in the storage unit 140; and a start instruction transmitting unit 114 which transmits a start instruction to the control device 130. The control device 130 starts the standby node 120 when the start instruction is received. The standby node 120 has a sync data acquisition unit 122 which acquires the sync data stored in the storage unit 140, and an update unit 123 which updates the storage content in the storage unit 121.

Description

Cluster system, information processing apparatus, cluster system synchronization method, and storage medium for storing program

The present invention relates to a cluster system, an information processing apparatus, a cluster system synchronization method, and a storage medium for storing a program.

As a method for improving the availability of IT (Information Technology) service, there is a method of constructing a HA (High Availability) cluster system. The HA cluster system includes an active node that actually provides a service and a standby node that takes over the service when a failure occurs in the active node. In order for the standby node to take over the service from the active node and provide it, it is necessary to take over not only the program used by the active node but also the data used by the program.

As a method of transferring data from the active node to the standby node, there are a method of using a shared disk (shared disk type), a method of mirroring data (data mirror type), and the like. In the case of the data mirror type, since data is synchronized by transferring data from the active node to the standby node using a network, the active node and the standby node must always be activated.

Further, in Patent Document 1, in a clustering system composed of a primary server and a slave server, updates to the storage means connected to the primary server are updated in the storage means connected to the slave server via a common disk. It is disclosed that it is reflected.

Also, in recent years, cloud services have become widespread, and cases where IaaS (Infrastructure as a Service) is used instead of purchasing physical machines to provide IT services are increasing.

Here, when constructing an HA cluster system using IaaS, it is difficult to use a shared disk, so the data mirror type is used as a method of taking over data. Further, when an HA cluster system is constructed between an on-premises environment and a cloud environment by using an existing node and a cloud service, a data mirror type is used.

JP 2001-337856 A

In the synchronization method by mirroring, both the active node and the standby node need to be activated at the same time when synchronizing. This is because data written to the active node is always transferred to the standby node. In addition, if the active node fails when the standby node is in the stopped state, the standby node does not take over the data written to the active node even if it tries to provide services on the standby node. Can not provide services. Further, when the active node is in a stopped state, the standby node cannot acquire the latest data from the active node.

For this reason, it is necessary to keep both the active node and the standby node in the active state. For this reason, it becomes difficult to suppress the operation cost of the node.

Even in the technique described in Patent Document 1, no consideration is given to this point, and the operation cost of the node cannot be suppressed.

SUMMARY OF THE INVENTION An object of the present invention is to solve such a problem, and a cluster system, an information processing apparatus, a cluster system synchronization method, and a memory for storing a program that can suppress the operation cost of the node. To provide a medium.

A cluster system according to one aspect of the present invention includes a first node that includes a first storage unit and operates as an active system, and a second node that includes a second storage unit and operates as a standby system. A first control device that controls activation of the second node, and a third storage unit provided separately from the first node and the second node, wherein the first node includes: Synchronization data generating means for generating synchronization data for synchronizing the storage content of the first storage unit and the storage content of the second storage unit when the storage content of the first storage unit is updated; Transmitting the synchronization data generated by the synchronization data generation unit to the third storage unit and storing the synchronization data in the third storage unit; and transmitting the synchronization data by the synchronization data transmission unit If the second node A first activation instruction transmitting means for transmitting an activation instruction for instructing a transition from a stopped state to an activated state to the first control device, wherein the first control device receives the activation instruction; The second node is activated, and the second node acquires the synchronization data stored in the third storage unit, and the synchronization data acquisition unit acquires the synchronization data. Update means for updating the storage content of the second storage unit by reflecting the update content indicated by the synchronization data in the storage content of the second storage unit.

In addition, in the information processing device according to one embodiment of the present invention, when the storage content of the first storage unit and the first storage unit is updated, the storage content and other information of the first storage unit are updated. Synchronous data generating means for generating synchronous data for synchronizing the storage contents of the second storage unit included in the processing apparatus, and the synchronous data generated by the synchronous data generating means for the own apparatus and the other information processing A synchronization data transmission means for transmitting to a third storage section provided separately from the apparatus and storing the same in the third storage section, and the other information when the synchronization data transmission means transmits the synchronization data. There is provided an activation instruction transmission means for transmitting an activation instruction for instructing the processing apparatus to shift from the stopped state to the activated state to an activation control unit that controls activation of the other information processing apparatus.

In addition, the cluster system synchronization method according to one aspect of the present invention is configured such that when the storage content of the first storage unit included in the first node operating as the active system is updated, the storage of the first storage unit is performed. Generating synchronization data for synchronizing the contents and the contents stored in the second storage unit of the second node operating as the standby system, and generating the generated synchronization data from the first node and the second The third storage unit is transmitted to the third storage unit provided separately from the node and stored in the third storage unit, the second node is shifted from the stop state to the start state, and the third storage unit The synchronization data stored in the storage unit is acquired, and the update content indicated by the acquired synchronization data is reflected in the storage content of the second storage unit to update the storage content of the second storage unit It is.

In addition, the cluster system synchronization method according to one aspect of the present invention is configured such that when the storage content of the first storage unit is updated, the storage content of the first storage unit and the second information processing apparatus have Generating synchronization data for synchronizing the storage contents of the storage unit, and transmitting the generated synchronization data to a third storage unit provided separately from the own apparatus and the other information processing apparatus, Control activation of the other information processing device with a start instruction for instructing the other information processing device to shift from the stopped state to the activated state when stored in the third storage unit and transmitting the synchronization data It is the method of transmitting to the starting control part.

In addition, the program stored in the storage medium according to one aspect of the present invention is stored when the storage content of the first storage unit is updated in the computer of the information processing apparatus having the first storage unit. Synchronizing data for synchronizing the storage content of one storage unit and the storage content of a second storage unit included in another information processing apparatus is generated, and the generated synchronization data is included in the first storage unit. In the case of transmitting to the third storage unit provided separately from the information processing device and the other information processing device, storing the third storage unit, and transmitting the synchronization data, the other information processing This is a program for executing transmission of an activation instruction for instructing to shift the apparatus from a stopped state to an activated state to an activation control unit that controls activation of the other information processing apparatus.

According to the present invention, it is possible to provide a cluster system, an information processing apparatus, a cluster system synchronization method, and a storage medium for storing a program that can suppress the operation cost of the node.

It is a block diagram which shows the outline | summary of a structure of the cluster system concerning embodiment. 1 is a block diagram showing a configuration of a cluster system according to a first exemplary embodiment. 1 is a block diagram showing a configuration of an active node according to a first exemplary embodiment. FIG. 3 is a block diagram showing a configuration of a standby node according to the first exemplary embodiment. 3 is a sequence chart illustrating an example of a synchronization operation of the cluster system according to the first exemplary embodiment; 3 is a sequence chart illustrating an example of a synchronization operation of the cluster system according to the first exemplary embodiment; FIG. 3 is a block diagram showing a configuration of a cluster system according to a second exemplary embodiment. 10 is a sequence chart showing an example of a failover operation in the cluster system according to the second exemplary embodiment;

<Outline of Embodiment of the Present Invention>
Prior to the description of the embodiment, an outline of the embodiment according to the present invention will be described. FIG. 1 is a block diagram illustrating an outline of a configuration of a cluster system 100 according to the embodiment. As shown in FIG. 1, the cluster system 100 includes an active node 110 that is a first node that operates as an active system, a standby node 120 that is a second node that operates as a standby system, and activation of the standby node 120. A control device 130 (also referred to as a first control device) and a storage unit 140.

Here, the “node” corresponds to hardware that executes software, a virtual machine, or the like, and typically corresponds to an operation unit of an OS (Operating System). For example, the active node 110 and the standby node 120 may each be realized in an on-premises environment, or may be realized by a cloud service that is an environment using a cloud computing infrastructure using a virtual machine. As an example, the active node 110 is realized in an on-premises environment, and the standby node 120 is realized by a cloud service. Note that the active node 110 and the standby node 120 may be referred to as information processing apparatuses.

The cluster system 100 normally provides a predetermined service by the active node 110, and performs a failover when an abnormality occurs in the active node 110. The standby node 120 provides the predetermined service instead of the active node 110. HA cluster system.

Here, the active node 110 has a storage unit 111 (also referred to as a first storage unit), and stores data in the storage unit 111. The data stored in the storage unit 111 includes, for example, data written by application software that provides the predetermined service, information on a file system managed by an OS (Operating System), and the like. The storage unit 111 is not limited to these, and various other data may be stored.

The standby node 120 includes a storage unit 121 (also referred to as a second storage unit), and stores the same data as the data stored in the storage unit 111 of the active node 110. The storage unit 111 of the active node 110 and the storage unit 121 of the standby node 120 are synchronized via a storage unit 140 (also referred to as a third storage unit) provided separately from the active node 110 and the standby node 120. Take. Specifically, synchronization is achieved by the following configurations of the active node 110 and the standby node 120.

The active node 110 includes a synchronization data generation unit 112, a synchronization data transmission unit 113, and an activation instruction transmission unit 114.

The synchronization data generation unit 112 is data for synchronizing the storage content of the storage unit 111 and the storage content of the storage unit 121 when the storage content of the storage unit 111 is updated. Generate.

The synchronization data transmission unit 113 transmits the synchronization data generated by the synchronization data generation unit 112 to the storage unit 140 and causes the storage unit 140 to store the synchronization data.

When the synchronization data transmission unit 113 transmits the synchronization data, the activation instruction transmission unit 114 transmits an activation instruction that instructs the standby node 120 to shift from the stop state to the activation state. On the other hand, when receiving the activation instruction, the control device 130 controls the standby node 120 to be activated. As a result, the standby node 120 changes from the stopped state to the activated state.

On the other hand, the standby node 120 includes a synchronization data acquisition unit 122 and an update unit 123. The synchronization data acquisition unit 122 acquires the synchronization data stored in the storage unit 140. Further, the update unit 123 updates the storage content of the storage unit 121 by reflecting the update content indicated by the synchronization data acquired by the synchronization data acquisition unit 122 in the storage content of the storage unit 121.

When the standby node 120 is activated by the control of the control device 130, the synchronization data acquisition unit 122 acquires the synchronization data, and the update unit 123 updates the storage contents of the storage unit 121 according to the acquired synchronization data. As a result, the cluster system 100 synchronizes the storage unit 111 and the storage unit 121. When the update process of the storage unit 121 is completed, the update unit 123 executes a shutdown process and shifts the standby node 120 to the stopped state again.

According to the cluster system 100, synchronization can be realized even if both the active node 110 and the standby node 120 are not activated simultaneously. For this reason, even if the active node 110 is not activated when the standby node 120 acquires the synchronization data, the standby node 120 can acquire the synchronization data. In addition, the activation of the standby node 120 can be limited when acquiring synchronous data.

For this reason, the operation cost can be suppressed in the following points. Since it is not necessary to always operate the standby node 120, the operation cost of the standby node 120 can be suppressed. In particular, when the standby node 120 is realized in an environment where the usage fee is charged, the charged usage fee can be suppressed. Here, the operation cost is not necessarily limited to the usage fee, but includes costs such as labor related to operation. Further, when the standby node 120 acquires the synchronization data, the synchronization is completed even if the active node 110 is not activated. Therefore, when the standby node 120 acquires the synchronization data, the active node 110 is not necessarily in the activated state. Also good. For this reason, the operation cost of the active node 110 is also suppressed.
<Embodiment 1>
The first embodiment will be described below.

FIG. 2 is a block diagram of a configuration of the cluster system 200 according to the first embodiment. As illustrated in FIG. 2, the cluster system 200 includes an active node 210, an external storage unit 240, and the cloud service 20. The active node 210, the external storage unit 240, and the cloud service 20 are connected to each other via the network 25 so that they can communicate with each other. The network 25 may be a wired network or a wireless network.

In the present embodiment, the active node 210 is an information processing apparatus realized in an on-premises environment. The detailed configuration of the active node 210 will be described later with reference to FIG. The active node 210 corresponds to the active node 110 described above.

The cloud service 20 is an environment that provides a service using a cloud computing infrastructure using a virtual machine, and is typically an information processing system including a server computer. The cloud service 20 includes a service control unit 230 and a standby node 220. Therefore, in the present embodiment, the service control unit 230 and the standby node 220 are information processing apparatuses realized using a cloud computing infrastructure.
The service control unit 230 and the standby node 220 are connected to the network 25. Note that the service control unit 230 may communicate with the active node 210 via a network different from the network 25. The detailed configuration of the standby node 220 will be described later with reference to FIG. The service control unit 230 corresponds to the control device 130 described above, and the standby node 220 corresponds to the standby node 120 described above.

The cluster system 200 realizes an HA cluster system with such a configuration. That is, the cluster system 200 normally provides a service in accordance with an application 213 described later by the active node 210 and performs a failover when an abnormality occurs in the active node 210, and replaces the active node 210 with a standby node. 220 continues to provide this service according to an application 223 described below. Each of the active node 210, the service control unit 230, and the standby node 220 includes a CPU (Central Processing Unit) and a memory (not shown) and functions as a computer.

The external storage unit 240 is a storage device provided separately from the active node 210 and the standby node 220, as shown in FIG. The external storage unit 240 corresponds to the storage unit 140 described above. The external storage unit 240 may be an object storage provided by the cloud service 20. Since the storage service is generally available at all times, it is considered suitable as a storage destination for synchronous data. As will be described later, the external storage unit 240 stores synchronization data for synchronizing the storage contents between the active node 210 and the standby node 220.

The service control unit 230 controls each function in the cloud service 20. In particular, in the present embodiment, the service control unit 230 controls activation of the standby node 220. The service control unit 230 controls to activate the standby node 220 when receiving an activation instruction to be described later. Thereby, the standby node 220 shifts from the stopped state to the activated state. The control by the service control unit 230 can be realized by executing a program under the control of the CPU, for example. More specifically, it is realized by executing a program stored in a storage device (not shown) of the service control unit 230 under the control of the CPU. In addition, the present invention is not limited to software by a program, and may be realized by any combination of hardware, firmware, and software.

Next, specific configurations of the active node 210 and the standby node 220 will be described. First, the active node 210 will be described.

FIG. 3 is a block diagram showing the configuration of the active node 210. As shown in FIG. 3, a storage unit 211, a file system 212, an application 213, clusterware 214, a data control unit 215, a data transfer unit 216, and a standby node activation control unit 217 are included.

The file system 212, the application 213, the clusterware 214, the data control unit 215, the data transfer unit 216, and the standby node activation control unit 217 can be realized, for example, by executing a program under the control of the CPU. More specifically, for example, a program stored in a storage device such as the storage unit 211 is executed under the control of the CPU. In addition, each component is not limited to being realized by software by a program, but may be realized by any combination of hardware, firmware, and software.

The storage unit 211 is a storage device corresponding to the storage unit 111 described above, and stores data.
The file system 212 is a data management function provided by the OS or the like, and manages data and management information (data size, access right to data, etc.) of the data. The application 213 is application software that provides a predetermined service such as a business application.

The clusterware 214 cooperates with the clusterware of other nodes (clusterware 224: see FIG. 4), and adjusts so that the coordinated operation is possible between the nodes. Specifically, for example, the clusterware 214 monitors the presence or absence of an abnormality such as a failure of the own node and other nodes.
The clusterware 214 adjusts the operation of each node by communicating between the nodes using the network 25 or another network, for example. For example, the clusterware 214 adjusts which node is activated when an application is activated, and checks whether another node is activated by heartbeat communication.

When an abnormality such as a failure occurs in the active node 210, the clusterware 214 cooperates with the clusterware 224 to start the application 223 by the standby node 220 from the activation of the application 213 by the active node 210 (see FIG. 4). Switch to startup. Specific examples of such switching include the following. After the failure of the active node 210 is detected by the clusterware 214, the execution of the application 213 on the active node 210 is stopped by the communication between the active node 210 and the standby node 220, and the execution of the application 223 on the standby node 220 is stopped. Be started. Further, the interruption of the heartbeat communication from the active node 210 is detected by the clusterware 224 of the standby node 220, and the application 223 of the standby node 220 is activated assuming that the active node 210 is down.

The data control unit 215 controls input / output with respect to the storage unit 211. In general, input / output of the file system 212 is directly performed on the storage unit 211, but in this embodiment, the processing of the data control unit 215 is hooked to the input / output processing between the file system 212 and the storage unit 211. is doing. The data control unit 215 corresponds to the synchronous data generation unit 112 described above, and when the storage content of the storage unit 211 is updated, the storage content of the storage unit 211 and the storage of the storage unit 221 described later of the standby node 220 are stored. Generate synchronization data to synchronize the contents.

The write request from the application 213 passes through the file system 212 to the data control unit 215. As a result, data is written in the storage unit 211.

The data control unit 215 creates synchronous data based on the data stored in the storage unit 211. In the present embodiment, the synchronization data is difference data between the first storage content of the storage unit 211 and the second storage content of the storage unit 211 after the update with respect to the first storage content. That is, the synchronization data is difference data between two stored contents based on, for example, time or the order of data written in the storage unit 211.

The data transfer unit 216 corresponds to the synchronous data transmission unit 113 described above, and stores the synchronous data generated by the data control unit 215 in the external storage unit 240 via the network 25. In the present embodiment, since difference data is used, the amount of data stored in the external storage unit 240 can be suppressed.

The standby node activation control unit 217 corresponds to the activation instruction transmission unit 114 described above, and requests the service control unit 230 to activate the standby node 220. When the data transfer unit 216 transmits and stores the synchronization data in the external storage unit 240, the standby node activation control unit 217 issues an activation instruction that instructs the standby node 220 to shift from the stopped state to the activated state. 230.

Next, the standby node 220 will be described. In the example illustrated in FIG. 2, the number of standby nodes 220 is one, but may be plural. In this embodiment, the standby node 220 is configured as an IaaS node provided by the cloud service 20 and is charged according to usage time. In the case of pay-as-you-go, operating costs can be reduced by operating the node for the required time when necessary.

FIG. 4 is a block diagram showing the configuration of the standby node 220. As shown in FIG. As illustrated in FIG. 4, the standby node 220 includes a storage unit 221, a file system 222, an application 223, clusterware 224, a data transfer unit 225, and a data control unit 226.

The file system 222, the application 223, the clusterware 224, the data transfer unit 225, and the data control unit 226 can be realized, for example, by executing a program under the control of the CPU. More specifically, for example, it is realized by executing a program stored in a storage device such as the storage unit 221 under the control of the CPU. In addition, each component is not limited to being realized by software by a program, but may be realized by any combination of hardware, firmware, and software.

The storage unit 221 is a storage device corresponding to the above-described storage unit 121, and stores the same data as the data stored in the storage unit 211 of the active node 210. The file system 222 has the same data management function as the file system 212. The application 223 is application software similar to the application 213, and provides the same service as the service provided by the application 213. The clusterware 224 is similar to the clusterware 214, and cooperates with the clusterware 214 to make adjustments so that a coordinated operation between nodes is possible.

The data transfer unit 225 corresponds to the above-described synchronization data acquisition unit 122, and acquires the synchronization data stored in the external storage unit 240 via the network 25. Further, the data transfer unit 225 determines whether or not the synchronization data to be acquired is synchronization data that may be applied to the storage unit 221 when acquiring the synchronization data. That is, it is determined whether or not the synchronization data to be acquired matches the stored contents of the storage unit 211 of the synchronization partner. The configuration for performing this determination process in the data transfer unit 225 may be referred to as determination means. Such a determination can prevent inconsistency between the storage contents of the storage unit 211 of the synchronization partner node and the storage contents of the storage unit 221 of the own node.

Whether or not the synchronization data to be acquired may be applied to the storage unit 221 is determined, for example, based on whether or not the synchronization data to be acquired is newer than the data currently stored in the storage unit 221. Is done. In this case, whether or not the synchronization data to be acquired is newer than the data currently held in the storage unit 221 depending on the date and time when the data was written in the storage unit 211 or the writing order in the storage unit 211. It may be determined. Note that the writing order is recorded by the data control

units

215 and 226, for example.

Further, whether or not the synchronization data to be acquired may be applied to the storage unit 221 may be determined based on, for example, whether or not the data is consistent. Under the situation where business applications are started on a plurality of different nodes, if a write occurs on a plurality of nodes at the same timing, a branch occurs in the data flow. In this case, consistency cannot be determined only by whether or not the data is new. Therefore, the determination may be made by confirming whether the node that has performed the writing is the synchronization partner node based on the history information that identifies the node that has performed the writing.

The data control unit 226 controls input / output with respect to the storage unit 221, similarly to the data control unit 215. However, while the data control unit 215 generates synchronization data, the data control unit 226 corresponds to the update unit 123 described above, and the update content indicated by the synchronization data acquired by the data transfer unit 225 is updated. The storage content of the storage unit 221 is updated by reflecting it in the storage content of the storage unit 221. In the present embodiment, the data control unit 226 applies the difference data acquired by the data transfer unit 225 to the storage content of the storage unit 221 and updates the storage content. Note that the data control unit 226 performs an update process on the synchronization data determined to be consistent with the storage content of the storage unit 221. When the update process of the storage unit 221 is completed, the data control unit 226 executes a shutdown process and shifts the standby node 220 to a stopped state.

Next, the synchronization operation in the cluster system 200 will be described. FIG. 5 is a sequence chart showing an example of the synchronization operation of the cluster system 200. In the sequence chart shown in FIG. 5, the standby node 220 is in a stopped state until step 103.

In step 100 (S100), the data control unit 215 of the active node 210 creates synchronization data from the data stored in the storage unit 211.

Next, in step 101 (S101), the data transfer unit 216 of the active node 210 transmits the synchronization data generated in step 100 to the external storage unit 240 using the network 25 and stores it in the external storage unit 240. To do.

Next, in step 102 (S102), the standby node activation control unit 217 of the active node 210 requests the service control unit 230 to activate the standby node 220. That is, the standby node activation control unit 217 transmits an activation instruction to the service control unit 230.

Next, in step 103 (S103), the service control unit 230 controls the standby node 220 to be activated.

In step 104 (S104), after the start-up of the standby node 220 is completed, the data transfer unit 225 of the standby node 220 determines whether or not the synchronization data stored in the external storage unit 240 can be applied to the storage unit 221 of the standby node 220. judge. If applicable, the data transfer unit 225 acquires the synchronization data from the external storage unit 240 using the network 25.

Subsequently, in step 105 (S105), the data control unit 226 of the standby node 220 updates the storage content of the storage unit 221 using the synchronization data acquired in step 104.

In step 106 (S106), when the synchronization is completed, the standby node 220 shuts down itself and stops the node.

As described above, the cluster system 200 transmits and receives synchronous data via the external storage unit 240. For this reason, synchronization is possible even if the active node 210 and the standby node 220 are not simultaneously activated. That is, since the active node 210 and the standby node 220 access the external storage unit 240, the synchronization data is stored by the active node 210 (step 101), and the synchronization data is acquired by the standby node 220 (step 104). The synchronization process can be executed regardless of the activation state of the counterpart node. Further, since the synchronization data is in the external storage unit 240, even when data synchronization is performed when there are a plurality of standby nodes 220, the load on the active node 210 can be suppressed and synchronization can be realized.

Further, when operating with the active node 210, the standby node 220 can be brought into a stopped state except when the synchronization process is performed. For this reason, it becomes possible to hold down operation cost. In particular, the sum of the operation cost of the standby node 220 when the activation of the standby node 220 is limited to the synchronization process and the operation cost of the external storage unit 240 when the external storage unit 240 is always activated is When it is lower than the operation cost when it is always activated, the operation cost is lower than when the active node 210 and the standby node 220 are always activated and directly synchronized between them.

Note that when the external storage unit 240 becomes unavailable due to a failure or the like, the active node 210 and the standby node 220 in the cluster system 200 perform direct synchronization processing. Specifically, if the data control unit 226 of the standby node 220 fails to acquire the synchronization data stored in the external storage unit 240, the data control unit 215 of the active node 210 requests the synchronization data from the active node 210. Transmits synchronization data to the standby node 220 in response to this request. The operation in this case will be described below. FIG. 6 is a sequence chart showing an example of the synchronization operation of the cluster system 200.

In step 200 (S200), an abnormality such as a failure occurs in the external storage unit 240.

In step 201 (S201), the data transfer unit 225 of the standby node 220 tries to acquire the synchronization data from the external storage unit 240, but the acquisition fails because a failure has occurred.

In step 202 (S202), the data transfer unit 225 of the standby node 220 requests synchronization data from the active node 210. Specifically, the data control unit 226 acquires information on data currently stored in the storage unit 221 from the data control unit 226 and requests the active node 210 to generate synchronization data.

In step 203 (S203), the data control unit 215 of the active node 210 generates synchronization data in accordance with the request in step 202.

In step 204 (S204), the data transfer unit 216 of the active node 210 transmits the generated synchronization data to the standby node 220.

In step 205 (S205), the data transfer unit 225 of the standby node 220 acquires the synchronization data transmitted from the active node 210. And the data control part 226 updates the memory content of the memory | storage part 221 using the acquired synchronous data.

In step 206 (S206), when the synchronization is completed, the standby node 220 shuts down itself and stops the node.
<Embodiment 2>
Next, a second embodiment will be described. When the standby node 220 is activated only when necessary as in the cluster system 200 according to the first embodiment, if the active node 210 goes down when the standby node 220 is stopped, the cluster system is There is a possibility that business cannot be continued. Therefore, in the present embodiment, an external control unit 260 that monitors the activation state of the active node 210 is provided.

FIG. 7 is a block diagram of a configuration of the cluster system 300 according to the second embodiment. As shown in FIG. 7, the cluster system 300 is the same as the cluster system 200 except that an external control unit 260 (second control device) is added. That is, the cluster system 300 includes an active node 210, an external storage unit 240, the cloud service 20, and an external control unit 260. Hereinafter, the cluster system 300 will be described with respect to differences from the cluster system 200, and redundant description will be omitted.

The external control unit 260 is an information processing apparatus having an active node activation monitoring unit 261 and a standby node activation control unit 262. The external control unit 260 may be realized in an on-premises environment, or may be realized in an environment using a cloud computing infrastructure using a virtual machine. For example, the external control unit 260 may be realized by the cloud service 20. In this case, the node with the lowest cost of the cloud service 20 may be used. The external control unit 260 is always activated while the cluster system 300 is activated, but the overall cost can be suppressed by operating the external control unit 260 at a lower cost than the standby node 220.

The active node activation monitoring unit 261 and the standby node activation control unit 262 can be realized, for example, by executing a program under the control of the CPU. More specifically, for example, a program stored in a storage device (not shown) is executed under the control of the CPU. In addition, each component is not limited to being realized by software by a program, but may be realized by any combination of hardware, firmware, and software.

The external control unit 260 can communicate with the active node 210 and the service control unit 230 via the network 25. The external control unit 260 may communicate with these via a network different from the network 25. The external control unit 260 has a CPU (Central Processing Unit) and a memory (not shown) and functions as a computer.

The active node activation monitoring unit 261 monitors the operating state of the active node 210. That is, the active node activation monitoring unit 261 (sometimes referred to as a monitoring unit) monitors whether the active node 210 is normally activated. The active node activation monitoring unit 261 detects this abnormality when an abnormality such as a failure occurs in the active node 210. The monitoring by the active node activation monitoring unit 261 is performed by, for example, heartbeat communication, but various known monitoring methods may be used.

The standby node activation control unit 262 requests the service control unit 230 to activate the standby node 220 in the same manner as the standby node activation control unit 217. Specifically, the standby node activation control unit 262 transmits an activation instruction to the service control unit 230 when the active node activation monitoring unit 261 detects an abnormality in the active node 210. The standby node activation control unit 262 may be referred to as a second activation instruction transmission unit.

Next, the operation of the cluster system 300 will be described. FIG. 8 is a sequence chart showing an example of failover operation in the cluster system 300 when the active node 210 goes down while the standby node 220 is stopped.

In step 300 (S300), an abnormality such as a failure occurs in the active node 210, and the active node 210 goes down.

In step 301 (S301), the active node activation monitoring unit 261 of the external control unit 260 detects that the active node 210 is down.

Next, in step 302 (S302), the standby node activation control unit 262 of the external control unit 260 requests the service control unit 230 to activate the standby node 220.
That is, the standby node activation control unit 262 transmits an activation instruction to the service control unit 230.

Next, in step 303 (S303), the service control unit 230 controls the standby node 220 to be activated.

Next, in step 304 (S304), when the activation of the standby node 220 is completed, the clusterware 224 of the standby node 220 confirms that the active node 210 is down.

Next, in step 305 (S305), the data transfer unit 225 of the standby node 220 determines whether or not the synchronization data stored in the external storage unit 240 is applicable to the storage unit 221 of the standby node 220. If applicable, the data transfer unit 225 acquires the synchronization data from the external storage unit 240 using the network 25.

Next, in step 306 (S306), the data control unit 226 of the standby node 220 updates the storage contents of the storage unit 221 using the synchronization data acquired in step 305.

Next, in step 307 (S307), when the synchronization is completed, the clusterware 224 activates the application 223.

The operation has been described above, but when the synchronization has already been completed, the update process of the storage unit 221 may be omitted.

As described above, according to the cluster system 300, even if the active node 210 goes down when the standby node 220 is stopped, the standby node 220 is activated, and the business can be continued as the cluster system.

Here, the operation cost of the standby node 220 when the activation of the standby node 220 is limited to the synchronization processing, the operation cost of the external storage unit 240 when the external storage unit 240 is always activated, and the external control unit 260 are always set. When the sum of the operation cost of the external control unit 260 when activated is lower than the operation cost when the standby node 220 is always activated, the active node 210 and the standby node 220 are always activated to Operational costs are lower than with direct synchronization.

Note that the performance of the standby node 220 when activated based on the activation instruction by the standby node activation control unit 217 of the active node 210 is the same as that when activated based on the activation instruction by the standby node activation control unit 262 of the external control unit 260. The performance may be lower than that of the standby node 220. For example, when realizing the standby node 220 in the cloud service 20, the standby node 220 may be constructed while suppressing performance such as the number of CPUs or the amount of memory. Thus, when the standby node 220 is activated not for providing a service but for synchronization, the amount of resources to be used can be suppressed. It is possible to reduce operational costs.

Note that the present invention is not limited to the above-described embodiment, and can be appropriately changed without departing from the spirit of the present invention. For example, each node, the service control unit 230, or the external control unit 260 may be realized in a physical environment. Note that activation of the standby node 220 in the physical environment can be realized by using a Wake-on-LAN function or the like.

Also, the program can be stored using various types of non-transitory computer readable media and supplied to a computer. Non-transitory computer readable media include various types of tangible storage media (tangible storage medium). Examples of non-transitory computer-readable media include magnetic recording media (eg flexible disks, magnetic tapes, hard disk drives), magneto-optical recording media (eg magneto-optical discs), CD-ROMs (Read Only Memory), CD-Rs, CD-R / W, semiconductor memory (for example, mask ROM, PROM (Programmable ROM), EPROM (Erasable ROM), flash ROM, RAM (random access memory)) are included. The program may also be supplied to the computer by various types of temporary computer-readable media. Examples of transitory computer readable media include electrical signals, optical signals, and electromagnetic waves. The temporary computer-readable medium can supply the program to the computer via a wired communication path such as an electric wire and an optical fiber, or a wireless communication path.

The present invention has been described above using the above-described embodiment as an exemplary example. However, the present invention is not limited to the above-described embodiment. That is, the present invention can apply various modes that can be understood by those skilled in the art within the scope of the present invention.

This application claims priority based on Japanese Patent Application No. 2015-183641 filed on September 17, 2015, the entire disclosure of which is incorporated herein.

20 Cloud service 25

Network

100, 200, 300

Cluster system

110, 210

Active node

111, 121, 140, 211, 221 Storage unit 112 Synchronization data generation unit 113 Synchronization data transmission unit 114 Activation

instruction transmission unit

120, 220 Standby node 122 Synchronization Data acquisition unit 123 Update unit 130

Controller

212, 222

File system

213, 223

Application

214, 224

Clusterware

215, 226

Data control unit

216, 225

Data transfer unit

217, 262 Standby node activation control unit 230 Service control unit 240 External storage Unit 260 external control unit 261 active node activation monitoring unit

Claims

A first node having a first storage and operating as an active system;
A second node having a second storage unit and operating as a standby system;
A first control device that controls activation of the second node;
A third storage unit provided separately from the first node and the second node;
The first node is:
Synchronization data generating means for generating synchronization data for synchronizing the storage content of the first storage unit and the storage content of the second storage unit when the storage content of the first storage unit is updated; ,
Synchronization data transmission means for transmitting the synchronization data generated by the synchronization data generation means to the third storage unit and storing it in the third storage unit;
A first activation instruction for transmitting, to the first control device, an activation instruction for instructing the second node to shift from a stopped state to an activated state when the synchronous data transmitting means transmits the synchronous data; Having a transmission means,
When the first control device receives the activation instruction, the first control device activates the second node;
The second node is
Synchronization data acquisition means for acquiring the synchronization data stored in the third storage unit;
Updating means for reflecting the update content indicated by the synchronization data acquired by the synchronization data acquisition means in the storage content of the second storage unit to update the storage content of the second storage unit;
Cluster system.
Monitoring means for monitoring an operating state of the first node;
A second activation instruction transmission unit configured to transmit the activation instruction to the first control device when an abnormality of the first node is detected by the monitoring unit;
A second control device;
The cluster system according to claim 1.
The performance of the second node when activated based on the activation instruction transmitted by the first activation instruction transmitting means is activated based on the activation instruction transmitted by the second activation instruction transmitting means. Lower than the performance of the second node when
The cluster system according to claim 2.
When the synchronization data acquisition unit fails to acquire the synchronization data stored in the third storage unit, the synchronization data acquisition unit requests the synchronization data from the first node,
The synchronous data transmission means transmits the synchronous data to the second node in response to a request from the synchronous data acquisition means;
The cluster system according to any one of claims 1 to 3.
A determination unit that determines whether or not the synchronization data to be acquired by the synchronization data acquisition unit is consistent with the storage content of the first storage unit;
The update unit performs update using the synchronization data determined to be consistent with the storage content of the first storage unit by the determination unit.
The cluster system according to any one of claims 1 to 4.
The synchronous data is difference data between the first storage content of the first storage unit and the second storage content of the first storage unit after the update to the first storage content.
The cluster system according to any one of claims 1 to 5.
A first storage unit;
When the storage content of the first storage unit is updated, synchronization data for synchronizing the storage content of the first storage unit with the storage content of the second storage unit included in another information processing apparatus is generated. Synchronization data generating means for
Synchronous data transmission for transmitting the synchronous data generated by the synchronous data generating means to a third storage unit provided separately from the own apparatus and the other information processing apparatus and storing the same in the third storage unit Means,
Activation control for controlling activation of the other information processing apparatus with an activation instruction for instructing the other information processing apparatus to shift from a stopped state to an activated state when the synchronization data is transmitted by the synchronization data transmitting unit Having an activation instruction transmitting means for transmitting to the unit,
Information processing device.
When the storage content of the first storage unit included in the first node operating as the active system is updated, the storage content of the first storage unit and the second node included in the second node operating as the standby system are updated. Generate synchronization data to synchronize the storage contents of
The generated synchronization data is transmitted to a third storage unit provided separately from the first node and the second node, and stored in the third storage unit,
Moving the second node from a stopped state to an activated state;
Obtaining the synchronization data stored in the third storage unit;
Updating the storage content of the second storage unit by reflecting the update content indicated by the acquired synchronous data in the storage content of the second storage unit;
Cluster system synchronization method.
When the storage content of the first storage unit is updated, synchronization data for synchronizing the storage content of the first storage unit and the storage content of the second storage unit included in another information processing device is generated. ,
The generated synchronization data is transmitted to a third storage unit provided separately from its own device and the other information processing device, and stored in the third storage unit,
When transmitting the synchronization data, a start instruction for instructing the other information processing apparatus to shift from the stop state to the start state is transmitted to the start control unit that controls the start of the other information processing apparatus.
Cluster system synchronization method.
In the computer of the information processing apparatus having the first storage unit,
When the storage content of the first storage unit is updated, synchronization data for synchronizing the storage content of the first storage unit with the storage content of the second storage unit included in another information processing apparatus is generated. And
The generated synchronization data is transmitted to a third storage unit provided separately from the information processing apparatus having the first storage unit and the other information processing apparatus, and stored in the third storage unit Let
When transmitting the synchronization data, transmitting a start instruction for instructing the other information processing apparatus to shift from a stopped state to a start state to a start control unit that controls the start of the other information processing apparatus; A storage medium for storing a program for executing the program.