JP2010212807A - Fail-over method, system thereof, node, and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a fail-over method, a system thereof, a node, and a program such that information used to decide whether to shift from a standby system node to an in-use system node is made consistent between standby system nodes. <P>SOLUTION: In the fail-over method, when each standby system node detects a fault occurring to an in-use system node, information on the standby system node itself at a certain point of time after the fault detection of the in-use system node is transmitted to other standby system nodes, information on the standby system node itself and information on the other standby system nodes are acquired, and it is decided whether to shift from the standby system node to an in-use system node by using the acquired information on the standby system nodes. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a failover method, a system thereof, a node, and a program.

  In a distributed computer system composed of a plurality of nodes, a manager for managing a node to be managed exists on the system. For example, when a failure occurs in a node, the processing performed on the node can be taken over by another node, so that high availability can be achieved. For example, based on information of a plurality of nodes In addition, it is possible to improve the efficiency of the processing of the entire system by shifting the processing performed at a certain node to another node.

  However, if a failure occurs in the active manager that is in operation, the agent cannot be managed, and the processing of the entire system stops. At that time, it is necessary to continue the system by a specific method.

  One way to achieve this is to prepare a standby manager in advance, and when a failure occurs in the active manager, the standby manager operates as a replacement for the failed active manager. There is a way.

  For example, when there are multiple standby managers that can be transferred to the active manager, the standby managers are discussed among all the standby managers as to which standby manager is to be used as the active manager. Start operation as an active system. As a result, when a failure occurs in the active manager, it is possible to continue processing as the entire system.

  Patent Document 1 describes an invention in which a standby server that automatically monitors a plurality of redundant servers and detects a failure of the active server automatically performs server switching. . In the present invention, the first server monitors the second server, the second server monitors the third server, and the like. This is the active server.

JP 2006-229512 A

  However, in the technique of Patent Document 1, the standby server monitors only one server. Therefore, when a failure occurs in the active server, the standby server that takes over the processing of the active server is the standby server that has been monitoring the active server. Therefore, the standby server that has taken over the processing of the active system may not always be suitable in terms of capabilities such as the free capacity of resources.

  Also, when negotiating between multiple standby managers and determining the standby server that will take over the active process, the standby server that has taken over the active process is often suitable in terms of capability. Consultations between standby managers were not easy. This is because the information used to determine whether or not to shift from the standby manager to the active manager when discussing among a plurality of standby managers is different for each standby manager.

  Therefore, the present invention has been invented in view of the above-described problems, and its purpose is to perform a failover that matches information used for determining whether to move from a standby node to an active node between standby nodes. To provide a method, its system, a node and a program.

  The present invention for solving the above-described problems is a method for controlling a failover system including an active node and at least two standby nodes. Each of the standby nodes functions as an active node. If it is detected that the active node is not transmitted, the information of the standby node at a certain time after the detection of the active node is transmitted to the other standby node, the information of the other standby node is obtained, This is a failover method for determining whether to move from the standby node to the active node using information on the standby node and the acquired information on the other standby node.

  The present invention for solving the above-described problems has at least one or more active nodes and at least two or more standby nodes, and each of the standby nodes includes the at least one or more active nodes. Detection means for detecting that it is not functioning as a standby node information transmission means for transmitting information of its standby node at a certain time after detection of the active node of the detection means to another standby node; Determining whether to transition from the standby node to the active node using the acquired standby node information by acquiring the information of the own standby node and the information of the other standby node; And a failover system having means.

  The present invention for solving the above problems is a node in a failover system having at least one active node and at least two standby nodes, wherein at least one active node functions as an active node. Detecting means for detecting that there is no standby node information transmitting means for transmitting information of its standby node at a certain time after detection of the active node of the detecting means to another standby node; Determining means for determining whether to move from the standby node to the active node using the acquired information of the standby node using the acquired information of the standby node; It is a node that has.

  The present invention for solving the above-described problems is a node program in a failover system having at least one active node and at least two standby nodes, wherein at least one active node is used as an active node. A detection process for detecting non-functioning, a standby node information transmission process for transmitting information of its standby node at a certain time after detection of the active node to another standby node, and the own node A node that acquires information on the standby node and information on the other standby node, and uses the acquired information on the standby node to determine whether to move from the standby node to the active node; This is a program to be executed.

  According to the present invention, the information used for determining whether to move from the standby node to the active node can be matched between the standby nodes.

FIG. 1 is a diagram for explaining an outline of an embodiment of the present invention. FIG. 2 is a block diagram of the first embodiment. FIG. 3 is a sequence diagram of the first embodiment. FIG. 4 is a block diagram of the first embodiment. FIG. 5 is a block diagram of a system according to the second embodiment. FIG. 6 is a block diagram of a system according to the third embodiment. FIG. 7 is a sequence diagram of the third embodiment. FIG. 8 is a block diagram of the second embodiment. FIG. 9 is a sequence diagram of the second embodiment.

  An outline of an embodiment of the present invention will be described.

  FIG. 1 is a diagram for explaining an outline of an embodiment of the present invention.

  In FIG. 1, three standby nodes 1, 2, 3 are shown, and each standby node 1, 2, 3 has a determination algorithm 4. This determination algorithm 4 is an algorithm for determining whether or not the node itself shifts from the standby node to the active node based on the information of the standby nodes of the three standby nodes 1, 2, and 3.

  When each of the standby nodes 1, 2, and 3 detects a failure of the active node, it transmits information of its own standby node at a certain point in time after detecting the failure of the active node to the other standby nodes. Then, each standby node 1, 2, 3 is based on the information of its own standby node transmitted to the other standby node, the information of the other standby node, these information, and the determination algorithm 4. , It determines whether it will shift from the standby system to the active system. The failure of the active node includes not only the failure of the active node itself but also the case where it is not functioning as the active node due to a failure of the line other than the active node, etc. Collectively, it is described as a failure of the active node.

  Here, it is a certain point in time after detecting the failure of the active node when transmitting information of its own standby node, but it differs in time from the same standby node for the failure detection of one active node This is intended to exclude the transmission of information of a plurality of standby nodes. This is because when information on a plurality of standby nodes that are different in time is transmitted, all the standby nodes cannot perform the migration determination with the same information. For example, if standby node 1 transmits information about its own standby node at time t1 and information about its own standby node at time t2, standby node 2 becomes standby node 1 at time t1. The standby node 3 may use the information of the standby node 1 at time t2. In this case, determination cannot be made using the same information in all standby nodes.

  As a specific method for preventing such a state, when a failure of the active node is detected, the information of its own standby node is transmitted to other standby nodes only once after the detection. In this way, for the failure detection of one active node, it becomes the information of the standby node at a certain point after the failure detection of the active node, and all the standby nodes make the migration determination with the same information. Can do. However, the same standby node information at a certain time may be transmitted a plurality of times. This is because the same information is used regardless of which information is used. In addition, if identification information for identifying the active node in which a failure is detected is added to the information of the standby node and transmitted, even if multiple active node failures occur, which of the active nodes It is possible to identify whether it is standby node information for a failure.

  In addition, the standby node information includes CPU operation rate, resource free capacity, failure rate, and information on bandwidth used for data transmission or reception, but is not limited thereto.

  Furthermore, it is preferable that the determination algorithm 4 used for the determination is the same in each of the standby nodes 1, 2, and 3. This is because if the determination algorithm 4 is the same and the standby node information used by the determination algorithm 4 is common, the determination results of the standby nodes 1, 2, and 3 are the same.

  With such a configuration, the information for determining whether to move from each standby node to the active node can be made consistent among all the standby nodes. Since the determination result is the same among all the standby nodes, each of the standby nodes does not need to transmit the determination result to other standby nodes. However, the present invention does not exclude that each of the standby nodes transmits the determination result to another standby node, and may be transmitted.

Hereinafter, specific embodiments will be described in detail with reference to the drawings.
<First Embodiment>
Referring to FIG. 2, in the first embodiment of the present invention, a system in which a failover method is realized has two standby managers (nodes) 100 and 101.

  The manager 100 includes a standby list storage unit 110, an information acquisition unit 120, a distribution unit 130, an aggregation unit 140, and a failure detection unit 150.

  The manager 101 includes a standby system list storage unit 111, an information acquisition unit 121, a distribution unit 131, an aggregation unit 141, and a failure detection unit 151.

  Each of these devices generally operates as follows.

  When the failure detection unit 150 detects a failure of the active manager, the failure detection unit 150 notifies the distribution unit 130 of the failure. When the distribution unit 130 is notified of the failure of the active manager, the distribution unit 130 acquires a list of all standby managers from the standby list storage unit 110, acquires information about its own standby manager from the information acquisition unit 120, and This information is transmitted only once to the aggregating unit 140 and the aggregating unit 141 in each standby manager.

  Similarly, when the failure detection unit 151 detects a failure of the active manager, the failure detection unit 151 notifies the distribution unit 131 of the failure. When the distribution unit 131 is notified of the failure of the active manager, the distribution unit 131 acquires a list of all standby managers from the standby list storage unit 111, acquires information about its own standby manager from the information acquisition unit 121, and This information is transmitted only once to the aggregating unit 140 and the aggregating unit 141 in each standby manager.

  Here, the standby node information includes the CPU operation rate, the resource free capacity, the failure rate, information on the bandwidth used for data transmission or reception, and the like, but is not limited thereto.

  The aggregation unit 140 has an algorithm 1401 for determining whether the own standby system manager shifts from the standby system to the active system. The algorithm 1400 is the same determination algorithm in the standby managers 100 and 101, but the type of the determination algorithm is not particularly limited. Then, the aggregating unit 140 aggregates the information of each standby system manager received from the distributing units 130 and 131 in each standby system manager, and moves itself to the active system manager by the determination algorithm 1400 using the aggregated information as data. It is determined whether or not.

  Similarly, the aggregation unit 141 includes an algorithm 1401 that determines whether the own standby system manager shifts from the standby system to the active system. The algorithm 1401 is the same determination algorithm in the standby managers 100 and 101, but the type of the determination algorithm is not particularly limited. Then, the aggregating unit 141 aggregates the information of each standby system manager received from the distribution units 130 and 131 in each standby system manager, and shifts itself to the active system manager by the determination algorithm 1401 using the aggregated information as data. It is determined whether or not.

  Next, the overall operation of the present embodiment will be described in detail with reference to the sequence diagrams of FIGS.

  First, when a failure occurs in the active manager, the failure detection unit 150 of the manager 100 detects the failure (S210), and the failure detection unit 150 notifies the distribution unit 130. Similarly, when a failure occurs in the active manager, the failure detection unit 151 of the manager 101 detects the failure (S211), and the failure detection unit 151 notifies the distribution unit 131.

  Next, the distribution unit 130 acquires a list of all standby system managers (in this example, managers 100 and 101) from the standby system list storage unit 110 (S220), and acquires information on the manager 100 from the information acquisition unit 120. (S230). Then, the distribution unit 130 transmits its own information to the aggregation units (in this example, the aggregation units 140 and 141) of all managers (S240). Similarly, the distribution unit 131 acquires a list of all standby system managers (in this example, managers 100 and 101) from the standby system list storage unit 111 (S221), and acquires information on the manager 101 from the information acquisition unit 121. (S231). Then, the distributing unit 131 transmits its information to the aggregating units of all managers (in this example, the aggregating units 140 and 141) (S241).

  Finally, the aggregating unit 140 aggregates information from the distributing units of all managers (in this example, the distributing units 130 and 131) (S250), and determines whether or not it itself shifts to the active system (S260). . Similarly, the aggregation unit 141 aggregates information from the distribution units (distribution units 130 and 131 in this example) of all managers (S251), and determines whether or not the manager itself shifts to the active system (S261). .

  From the standby system list storage units 110 and 111, a method for acquiring all standby system lists (S220 and S221), the specific contents of information held by the information acquisition units 120 and 121 (S230 and S231), and the distribution units 130 and 131 The communication method to the aggregation units 140 and 141 (S240 and S241), and the specific method (S260 and 261) for determining whether to move to the active system performed by the aggregation units 140 and 141 are changed in a timely manner within the scope of the present invention. Can do.

  As described above, in this embodiment, the information used for the determination of the transition to the active system is the same in all the aggregation units. Accordingly, it is possible to autonomously select a manager to be transferred to the active system from a plurality of standby system managers.

  Next, Example 1 corresponding to the first embodiment will be described.

  As shown in FIG. 4, the system according to the first embodiment includes a computer network 400, nodes 410 to 412 and 419 that are components thereof, four managers 420 to 422 and 429 that operate on each node, The manager holds four standby system list storage units 430 to 432, 439, four information acquisition units 440 to 442, 449, four distribution units 450 to 452, 459, and four Aggregating units 460 to 462 and 469.

  In the system as described above, when the manager 429 on the node 419 operates as the active system and the managers 420 to 422 on the other nodes 410 to 412 operate as the standby system, the node 419 is disconnected from the computer network 400. The case where it is cut will be described.

  First, each distribution unit 450 to 452 acquires managers 420 to 422 that are lists of all standby system managers from each standby system list storage unit 430 to 432, and each distribution unit 450 to 452 owns from each information acquisition unit 440 to 442. Information is acquired and the information is transmitted to the aggregation units 460 to 462.

  Next, each aggregation unit 460 to 462 aggregates the information of the managers 420 to 422 received from all of the standby system list storage units 430 to 432, and determines whether or not it shifts to the active manager.

  Assuming that the manager to be transferred to the active manager as a result of aggregating the information of the managers 420 to 422 is, for example, the manager 421, each of the aggregating units 460 to 462 determines that the manager 421 shifts to the active system. 421 shifts to the active system, and the managers 420 and 422 continue to operate in the standby system without shifting to the active system.

  In the conventional method, the information of the managers 420 to 422 held for determining whether or not to shift to the active system could not be matched in all managers. Therefore, it is possible to determine whether or not each manager shifts to the active system.

  In the first embodiment described above, the distribution units 450 to 452 acquire the managers 420 to 422 that are lists of all standby system managers from the standby system list storage units 430 to 432, and send information to the aggregation units 460 to 462. Although it is transmitted, it may also be transmitted to the aggregation unit 469 of the manager 429 that is the active manager. Further, the manager 429 that is the active manager may perform the same operation as the other standby managers 420 to 422.

  Another example of the first embodiment will be described.

  In the above-described embodiment, assuming that the managers 421 and 422 are the managers to be transferred to the active manager as a result of aggregating the information of the managers 420 to 422, for example, the aggregating means 460 to 462 have the managers 421 and 422 respectively. The managers 421 and 422 are determined to shift to the active system, the managers 421 and 422 shift to the active system, and the manager 420 continues to operate without being transferred to the active system.

  Still another example of the first embodiment will be described.

In the embodiment described above, as a result of aggregating the information of the managers 420 to 422, if there is no manager to be transferred to the active manager, for example, each of the aggregating means 460 to 462 has no manager to shift to the active system. The managers 420 to 422 continue to operate while remaining in the standby system without shifting to the active system.
<Second Embodiment>
A second embodiment of the present invention will be described in detail with reference to the drawings. FIG. 5 is a block diagram of a system according to the second embodiment.

  The system according to the second embodiment includes two standby managers 200 and 201 and a standby list storage unit 310.

  The manager 200 includes an information acquisition unit 220, a distribution unit 230, an aggregation unit 240, and a failure detection unit 250.

  The manager 201 includes an information acquisition unit 221, a distribution unit 231, an aggregation unit 241, and a failure detection unit 251.

  The second embodiment is different from the first embodiment in that each manager does not have a standby list storage unit for storing a list of all standby managers, but a standby list storage unit shared by each manager. 210. Since other parts are the same as those in the first embodiment, detailed description thereof is omitted.

  Each of these devices generally operates as follows.

  When the failure detection unit 250 of the manager 200 detects a failure of the active manager, the failure detection unit 250 notifies the distribution unit 230 of the failure. The distribution unit 230 acquires a list of all standby system managers from the standby system list storage unit 210, acquires its own information from the information acquisition unit 220, and acquires the own information acquired by the aggregation units 240 and 241 in each standby system manager. Send information.

  Similarly, when the failure detection unit 251 of the manager 201 detects a failure of the active manager, the failure detection unit 251 notifies the distribution unit 231 to that effect. The distribution unit 231 acquires a list of all standby system managers from the standby system list storage unit 210, acquires its own information from the information acquisition unit 221, and acquires its own information acquired by the aggregation units 240 and 241 in each standby system manager. Send information.

  Next, the aggregating unit 240 aggregates the information of each standby system manager received from the distribution units 230 and 231 in each standby system manager, and determines whether or not it shifts to the active system manager by a determination algorithm. .

  Similarly, the aggregating unit 241 aggregates the information of each standby system manager received from the distribution units 230 and 231 in each standby system manager, and determines whether or not it shifts to the active system manager by a determination algorithm. .

In the second embodiment, each manager does not have a standby system list storage unit, and is configured to acquire a list of all standby system managers from the standby system list storage unit 210 shared by each manager. Accordingly, it is only necessary to update the manager list or the like only for the standby system list storage unit 210.
<Third Embodiment>
A third embodiment of the present invention will be described in detail with reference to the drawings. FIG. 6 is a block diagram of a system according to the third embodiment.

  Referring to FIG. 5, the failover system according to the third embodiment of the present invention has two standby managers 300 and 301.

  The manager 300 includes a standby list storage unit 310, an information acquisition unit 320, a distribution unit 330, an aggregation unit 340, and a failure detection unit 350.

  The manager 301 includes a standby list storage unit 311, an information acquisition unit 321, a distribution unit 331, an aggregation unit 341, and a failure detection unit 351.

  The third embodiment differs from the first embodiment in that the failure detection units 350 and 351 not only detect the failure of the active manager, but also identify which active manager has failed, Information is transmitted to the distribution units 330 and 331 of each standby manager. The distribution units 330 and 331 add identification information for identifying the failed active manager to the information of the standby manager and transmit the identification information to the aggregation units 340 and 341. Furthermore, the aggregation units 340 and 341 of the respective standby managers are different in that migration determination is performed using information on standby systems having the same identification information of the failed active manager.

  With such a configuration, even when a plurality of active managers fail almost simultaneously, determination is not made using information on the standby manager at the time of failure of different active managers.

  Since other parts are the same as those in the first embodiment, detailed description thereof is omitted.

  Next, the overall operation of the present embodiment will be described in detail with reference to the sequence diagrams of FIGS.

  First, when a failure occurs in the active manager, the failure detection unit 350 of the manager 300 detects the failure (S310), and notifies the distribution unit 330 which failure of the active manager is detected. Similarly, when a failure occurs in the active manager, the failure detection unit 351 of the manager 301 detects the failure (S311), and notifies the distribution unit 331 which failure of the active manager has been detected.

  The distribution unit 330 acquires a list of all standby system managers (managers 300 and 301) from the standby system list storage unit 310, acquires the information of the manager 300 from the information acquisition unit 320, and the current working that failed in the information of the manager 300 The identification information of the system manager is added and transmitted to the aggregating units (aggregating units 340 and 341 in this example) of all managers (S340).

Similarly, the distribution unit 331 acquires a list of all standby managers (managers 300 and 301) from the standby list storage unit 311, acquires information on the manager 301 from the information acquisition unit 321, and stores the information on the manager 301. The identification information of the failed active manager is added and transmitted to the aggregation units (in this example, the aggregation units 340 and 341) of all managers (S341).
These steps are performed each time a failure of the active manager is detected.

  Finally, the aggregating unit 340 aggregates manager information from the distributing units of all managers (distributing units 330 and 331 in this example) (S350), and the identification information of the active manager in which the failure has occurred is the same manager. Using the information, it is determined whether or not the device itself shifts to the active system (S360). Similarly, the aggregating unit 341 aggregates manager information from the distributing units of all managers (in this example, distributing units 330 and 331) (S350), and the identification information of the active manager in which the failure has occurred is the same manager. Using the information, it is determined whether or not the device itself shifts to the active system (S360).

  In this way, when multiple active managers fail almost simultaneously, the information acquisition unit of each standby manager acquires and distributes information on its own standby manager as many as the number of active managers that have failed. The unit adds the identification information for identifying the active manager in which the failure has occurred to the information and transmits the information. The aggregation unit of each standby manager performs the migration determination for each identification information, that is, the number of active managers in which a failure has occurred. By adopting such a configuration, it is possible to determine which information about the standby manager information at the time of the failure of the active manager, and in each migration determination performed multiple times in all aggregation units However, the information used for determination is the same.

  The second embodiment will be described with reference to the block diagram of FIG. 8 and the sequence diagram of FIG. This example shows an example of the operation when a plurality of active managers fail almost simultaneously.

  As shown in FIG. 5, a computer network 500, five nodes 510 to 512, 518, and 519 that are components thereof, and five managers 520 to 522, 528, and 529 operating on each of the nodes, Five standby list storage units 530 to 532, 538, and 539 held by each manager, five information acquisition units 540 to 542, 548, and 549, and five distribution units 550 to 552 and 558 , 559, five aggregation units 560-562, 568, 569 and five fault detection units 570-572, 578, 579, managers 528 and 529 on nodes 518 and 519 are active. When the managers 520 to 522 on the nodes 510 to 512 are operating as standby systems, the node 51 And node 519 will be described in the form of operation when it is disconnected from the computer network 500.

  First, each of the failure detection units 570 to 572 detects a failure of the node 518 and notifies the distribution units 550 to 552 (S610 to S612).

  In response to the notification from the failure detection units 570 to 572, the distribution units 550 to 552 acquire managers 520 to 522 that are lists of all standby system managers from the respective standby system list storage units 530 to 532 (S620 to S620). S622), own information is acquired from each information acquisition part 540-542 (S630-S632). Then, the identification information of the node 518 is added, and the own information is transmitted to all the aggregation units 560 to 562 (S640 to S642).

  In addition, each of the failure detection units 570 to 572 detects a failure of the node 519 and notifies the distribution units 550 to 552 (S613 to S615).

  In response to the notification from the failure detection units 570 to 572, the distribution units 550 to 552 acquire managers 520 to 522 that are lists of all standby system managers from the respective standby system list storage units 530 to 532 (S623 to S623). S625), own information is acquired from each information acquisition part 540-542 (S633-S635). Then, identification information of the node 519 is added and the information is transmitted to all the aggregation units 560 to 562 (S643 to S645).

  Next, each aggregation part 560-562 aggregates the information received from the distribution part 550-552 (S650-S652).

  Subsequently, the aggregating units 560 to 562 perform determination using information having the same identification information of the failed active manager (S660 to S662). For example, the aggregating unit 560 receives information from each of the distributing units 550 to 552 twice, and the six pieces of information are related to the failure of the active manager 528 and the failure of the active manager 529. . Therefore, the information regarding the failure of the three active managers 528 is used to determine the migration regarding the active manager 528, and the information regarding the failure of the three active managers 529 is used to determine the transition of the migration regarding the active manager 529. Make a decision. For example, when each of the aggregation units 560 to 562 determines that the standby managers to be transferred to the active system are 521 and 522, respectively, as a result of the failure of the active managers 528 and 529, the standby managers 521 and 522 Both of them shift to the active system, and the standby manager 520 continues to operate without changing to the active system.

  The present invention can be applied to a distributed computer system composed of a plurality of nodes, in which those nodes are managed by a manager existing in the system.

100, 101 Manager 110, 111 Standby system list storage unit 120, 121 Information acquisition unit 130, 131 Distribution unit 140, 141 Aggregation unit 150, 151 Fault detection unit

Claims (31)

A method for controlling a failover system comprising an active node and at least two standby nodes,
Each of the standby nodes
When detecting that the active node is not functioning as the active node, the information of its own standby node at a certain time after the detection of the active node is transmitted to other standby nodes,
Obtain information of the other standby node,
A failover method for determining whether to move from a standby node to an active node using information on its own standby node and the acquired information on another standby node.   The failover method according to claim 1, wherein each of the standby nodes does not transmit the determined result to another standby node. Each standby node has a determination algorithm for determining a transition from the standby node to the active node using the acquired standby node information, and the determination algorithm is the same for each standby node. The failover method according to claim 1 or 2. 4. The failover method according to claim 1, wherein each of the standby nodes transmits the same information as information of its standby node to another standby node. 5.   5. Each standby node, after detecting that it is not functioning as an active node, transmits the information of its own standby node once to another standby node. Failover method. Each standby system
Adding identification information identifying the active node that has detected that it is not functioning as an active node to the information of its own standby node,
6. The failover method according to claim 1, wherein it is determined whether to move from the standby node to the active node using information of the standby node having the same identification information.   7. The information of the standby node is at least one of CPU operation rate, resource free capacity, failure rate, and band information used for data transmission or reception. 7. Failover method. Having at least one active node and at least two standby nodes;
Each of the standby nodes is
Detecting means for detecting that the at least one active node is not functioning as an active node;
Standby node information transmission means for transmitting information of its own standby node at a certain time after detection of the active node of the detection means to other standby nodes;
Determination means for acquiring information of the own standby node and information of the other standby node and determining whether to move from the standby node to the active node using the acquired standby node information And a failover system.   The failover system according to claim 8, wherein each of the standby nodes does not transmit the determination result to another standby node. The determination means has a determination algorithm for determining a transition from a standby node to an active node using the acquired standby node information, and the determination algorithm is the same for each standby node. The failover system according to claim 8 or 9.   The failover system according to any one of claims 8 to 10, wherein the standby node information transmitting unit transmits the same information as information of its standby node to another standby node.   The standby node information transmitting unit transmits information on its own standby node to another standby node once after detecting the active node of the detection unit. Failover system. The standby node information transmission means adds identification information for identifying the active node detected as not functioning as the active node to the information of its own standby node and transmits it to other standby nodes. And
The failover system according to any one of claims 8 to 12, wherein the determination unit determines whether to move from the standby node to the active node using information of the standby node having the same identification information.   14. The information of the standby node is at least one of CPU operation rate, resource free capacity, failure rate, and band information used for data transmission or reception. 14. Failover system.   15. The failover system according to claim 8, wherein each standby node has standby node information storage means for storing information of a standby node that transmits information of the standby node. Standby node information storage means for storing information of the standby node that transmits the information of the standby node;
The failover system according to claim 8, wherein each of the standby nodes shares the standby node information storage unit. A node in a failover system having at least one active node and at least two standby nodes,
Detecting means for detecting that at least one active node is not functioning as an active node;
Standby node information transmission means for transmitting information of its standby node at a certain time after detection of the active node of the detection means to other standby nodes;
Determination means for acquiring information of the own standby node and information of the other standby node and determining whether to move from the standby node to the active node using the acquired standby node information A node having   The node according to claim 17, wherein the node does not transmit the determination result to another standby node. The determination unit has a determination algorithm for determining a transition from the standby node to the active node using the acquired standby node information.
The node according to claim 17 or 18, wherein the determination algorithm is the same in each standby node.   The node according to any one of claims 17 to 19, wherein the standby node information transmission unit transmits the same information as information of its standby node to another standby node.   21. The standby node information transmission unit according to claim 17, wherein after detecting the active node of the detection unit, the standby node information transmission unit transmits information of its own standby node once to another standby node. node. The standby node information transmitting means adds identification information for identifying the active node that has detected that the active node is not functioning as the active node to the information of its own standby node, so that other standby nodes To the node,
The node according to any one of claims 17 to 21, wherein the determination unit determines whether to move from the standby node to the active node using information of a standby node having the same identification information.   The information of the standby node is at least one of CPU operation rate, resource free capacity, failure rate, and band information used for data transmission or reception. Nodes.   The node according to any one of claims 13 to 17, further comprising standby node information storage means for storing information of a standby node that transmits information of the standby node. A node program in a failover system having at least one active node and at least two standby nodes,
A detection process for detecting that at least one active node is not functioning as an active node;
Standby node information transmission processing for transmitting information of its own standby node at a certain time after detection of the active node to other standby nodes;
Determination processing for acquiring information of the own standby node and information of the other standby node, and determining whether to move from the standby node to the active node using the acquired standby node information A program that causes nodes to execute.   The program according to claim 25, wherein, in the determination process, the determination result is not transmitted to another standby node. 27. The program according to claim 25 or claim 26, wherein the determination processing determines whether to move from the standby node to the active node using a determination algorithm that is the same in each standby node.   The program according to any one of claims 25 to 27, wherein the standby node information transmission process transmits the same information to another standby node as information of its own standby node.   The program according to any one of claims 25 to 28, wherein the standby node information transmission processing transmits the information of its own standby node once to another standby node after detecting the active node. In the standby node information transmission process, identification information for identifying the active node that is detected as not functioning as an active node is added to the information of its own standby node and transmitted to another standby node. And
30. The program according to claim 25, wherein the determination processing determines whether to move from the standby node to the active node using information of a standby node having the same identification information.   31. The information of the standby node is at least one of CPU operation rate, resource free capacity, failure rate, and bandwidth information used for data transmission or reception. Program.

Images