JP2000132441A - Decentralized data base system, processor, and data base access method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide the decentralized data base system, processor, and data base access method which can improve the turn-around time by shortening the wait time for access to a data base even in the case of fault occurrence. SOLUTION: If a fault occurs to the data base 2A and access is disabled, a DB arrangement converting mechanism 1A selects an active node 3 so that a data base 2B at a mirror destination is accessed. Consequently, the active node 3 performs all access processing such as retrieval, updating, etc., by using the data base 2B at the mirror destination instead of the fault data base 2A until the data base 2A recovers. A standby node 4 changes from a standby state to an active state to perform processings for access to the data bases 2A and 3B.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to a distributed database system, a processing device, and a database access method for executing processing such as access to a database distributed on a network.

[0002]

2. Description of the Related Art Conventionally, in a distributed database system of this type, each CPU is provided with a processing unit such as a central processing unit (hereinafter referred to as a CPU) for accessing a database in order to cope with a failure. Then, the database is disk mirrored or the database is duplicated.

[0003] The processing operation of a conventional distributed database system that performs this disk mirroring will be described with reference to the configuration diagram shown in FIG. First, in a state where the active node 2 has failed, the terminal 6
Sends a database 2A search request to
The access request node 1 is a database layout conversion mechanism 1
A determines the standby node 4 instead of the active node 2 as the access request destination and sends an access request to the standby node 4.

The standby node 4 receiving this access request
Starts the process of accessing the database by itself, accesses the database 2A (retrieval processing) and transfers the access result (retrieval result) to the DB access request node 1 when it becomes active. I do. If a failure has occurred in the database 2A, the mirrored database 2B is referred to. In addition, under the above-mentioned conditions, when the processing request from the terminal 6 is an update process, the standby node 4 updates the database 2A and performs mirroring of the updated database 2A on the database 2B.

Next, FIG. 8 shows a configuration diagram of a conventional distributed database system in which the database is duplicated.

The processing operation of the conventional distributed database system will be described. When a search request for the database 2 is sent from the terminal 6 in a state where the active node 2 has failed, the DB access request node 1 uses the database arrangement conversion mechanism 1A as an access request destination, and And determines the active node 7 and sends an access request to the active node 7. The active node 7 receiving the access request accesses (searches) the database 2A and transfers the access result (search result) to the DB access request node 1. If the processing request from the terminal 6 is an update process of the database 2A, after the database 2A is updated first, the database 2A is updated.
B is mirrored, so active node 2
, The update process to the database 2A by the active node 2 and the mirroring to the database 2B by the active node 7 are not executed until the active node 2 recovers.

Further, a distributed database system in which the database is duplicated is disclosed in Japanese Patent Laid-Open No. Hei 8-2.
No. 78911. In the system disclosed in this publication, one disk device storing a database is associated with one database site for performing access processing to the disk device, and a plurality of sets of such a relationship are configured. Multiplexed. For example, in the case where the original database and the duplicate database are provided and duplexed, when a failure occurs in one processing system, the processing is performed in the other processing system without stopping the system. I try to continue.

[0008]

[0007] However, in the conventional distributed database system shown in FIG.
When there is a search request to A, the standby node 4 can access the database 2A instead of the active node 2, but the standby node 4 activates a process for accessing the database and enters the active state. Until the access processing to the database 2A cannot be executed. That is, when the standby node 4 is switched to the active node, the above process must be started in the standby node 4. For this reason, there is a problem that the access to the database is stopped until the preparation for accessing the database is completed.

In the conventional distributed database system shown in FIG. 8, the active node and the database in the database layer are duplicated in order to maintain the same access processing capacity after the failure as before the failure. In the database layer, twice as many active nodes are required as compared to the case without duplication. That is, in the case of no duplication, for example, three active nodes are required respectively corresponding to three databases, whereas in the case of duplication, in addition to the three active nodes, a database in which each database is mirrored , Three additional active nodes are required. That is, six active nodes are required.

At the time of updating the database, the database 2A at the mirror source is first updated, and then the database 2B at the mirror destination is mirrored. If a failure has occurred in the node 2, the active node 2 is restored until the active node 2 recovers.
Is not executed by the active node 7 and the update process to the database 2A by the active node 7 is not executed. That is, there is a problem in that access to the database is suspended until the active node that has failed recovers.

In the system disclosed in the above publication, as in the case of the conventional distributed database system shown in FIG. 4, in order to maintain the same access processing capacity after a failure as before the failure, the database site is duplicated. It is twice as much as when it is not used. For example, when duplicating two different databases, a total of four database sites are required, two database sites corresponding to each original database and two database sites corresponding to each duplicate database. It is. Of course, when three different databases are duplicated, a total of six database sites are required.

An object of the present invention is to provide a distributed database system capable of improving a turnaround time by shortening a database access waiting time even when a failure occurs.

[0013]

To achieve the above object, a distributed database system according to a first aspect of the present invention comprises a plurality of nodes each accessing a database via a logically connected communication path. , An access requesting node that transmits an access request to the plurality of nodes, wherein each of the plurality of nodes accesses a different one of the plurality of databases, and Equipped with access means to access the backup database corresponding to the database accessed by the node,
The access request node transmits a request for transmitting an access request to any one of the plurality of nodes, and determines whether a node corresponding to the access request from the transmission means can access the database. When the determination means determines that the access is not possible, the access request is transmitted to the node that accesses the backup database corresponding to the database to which the inaccessible node should access. And processing means.

In such a distributed database system, the access requesting node sends an access request to the node accessing the backup database corresponding to the predetermined database to be accessed by the inaccessible node by the processing means. Send Therefore, even when a failure occurs in a predetermined node to which an access request is to be transmitted, the backup database corresponding to the database to be accessed by the predetermined node is accessed, so that the predetermined node does not The request for access to the power database should not be stopped.

Further, in the distributed database system according to the first aspect, the plurality of databases connected to the plurality of nodes and the plurality of backup databases are physically connected to the plurality of backup databases. A standby node that receives a request, wherein the access request node further includes a unit that transmits an access request to the standby node when the access unit determines that the access is not possible; In response to the request, the inaccessible node may have access means for accessing the database and accessing the backup database to be accessed.

In such a distributed database system, the standby node accesses a predetermined database to be accessed by the inaccessible node in response to an access request from the access request by an access means. And access the backup database. Therefore, even when a failure occurs in a predetermined node to which an access request is to be transmitted, the backup database corresponding to the database to be accessed by the predetermined node is accessed, so that the predetermined node does not The request for access to the power database should not be stopped.

Incidentally, a plurality of the standby nodes may be provided. In this case, the access request node transmits an access request to any one of the plurality of standby nodes. For this reason, any of the standby nodes can access a predetermined database to which the inaccessible node should access and also access a backup database.

Further, in the distributed database system according to the first aspect, when the access request node is determined to be inaccessible by the determining means, any one of the plurality of standby nodes may be used as the access node. Means for transmitting an access request to the access request; means for determining whether or not there is a response to the access request transmitted from the transmitting means; and when the determination means determines that there is no response, the access request And a means for transmitting an access request to a standby node different from the standby node that has not responded to the request.

In such a distributed database system, the access request node transmits an access request to a standby node in which a failure has not occurred among a plurality of standby nodes. For this reason, even if a failure has occurred in a predetermined standby node, another normal standby node accesses a predetermined database to be accessed by the inaccessible node and a backup database. Can be accessed.

The access request node transmits information indicating a state transition of the plurality of nodes and the standby node, information indicating an access processing type, and transmission of an access request determined based on the information. It is also possible to provide a table storing information indicating the destination. In this case, the transmission destination of the access request is determined according to the state transition of the node and the type of access processing.

A processing device according to a second aspect of the present invention has a processing device having an access request node for transmitting an access request to a plurality of nodes that access a database via logically connected communication paths. And a first transmitting unit that transmits an access request to any one of the plurality of nodes, and whether a node corresponding to the access request from the transmitting unit can access the database. An access request is transmitted to the determining means for determining and, if the determining means determines that the access is not possible, to the node accessing the backup database corresponding to the database to be accessed by the inaccessible node. And processing means for performing the processing.

In such a processing device, the access requesting node sends an access request to the node accessing the backup database corresponding to the predetermined database to be accessed by the inaccessible node by the processing means. Send. Therefore, even when a failure occurs in a predetermined node to which an access request is to be transmitted, the backup database corresponding to the database to be accessed by the predetermined node is accessed, so that the predetermined node does not The request for access to the power database should not be stopped.

In the processing apparatus according to the second aspect, when the determination means determines that the access is not possible, the plurality of databases connected to the plurality of nodes and the plurality of backup databases are provided. May be provided with a second transmitting means for transmitting an access request to a standby node physically connected to the communication node.

In such a processing device, the access request node transmits an access request to a standby node in which a failure has not occurred among a plurality of standby nodes.
For this reason, even if a failure has occurred in a predetermined standby node, another normal standby node accesses a predetermined database to be accessed by the inaccessible node and a backup database. Can be accessed.

A database access method according to a third aspect of the present invention accesses a database different from a plurality of databases through a logically connected communication path, and a database accessed by another node. In a database access method in a distributed database system having a plurality of nodes that access a backup database corresponding to a plurality of nodes, and an access request node that transmits an access request to the plurality of nodes, the access request node includes the plurality of nodes. A transmitting step of transmitting an access request to any one of the nodes; a determining step of determining whether a node corresponding to the access request transmitted by the transmitting step can access a database; According to the decision step Transmitting an access request to a node accessing a backup database corresponding to the database to be accessed by the inaccessible node when the access is determined to be inaccessible. And

In such a database access method, the access requesting node, by the processing step, sends an access request to a node accessing a backup database corresponding to a predetermined database to be accessed by an inaccessible node. Send Therefore, even when a failure occurs in a predetermined node to which an access request is to be transmitted, the backup database corresponding to the database to be accessed by the predetermined node is accessed, so that the predetermined node does not The request for access to the power database should not be stopped.

[0027]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing one embodiment of a distributed database system according to the present invention.

As shown in FIG. 1, a database access request node (hereinafter referred to as a DB access request node) 1 which is a functional layer, and active nodes 2 and 3 and a standby node 4 which are database layers are connected to a network 5 respectively. It is connected.

The DB access request node 1 and each node in the database layer exchange an access request, an access result, and data indicating that a failure has occurred via the network 5.

The DB access request node 1 is connected to a terminal 6 which is a presentation layer, and is provided with a database layout conversion mechanism 1A. From the terminal 6, processing requests such as creation, retrieval, and updating of a database are transmitted.

The DB layout conversion mechanism 1A includes a database 2
When a failure occurs in the database A or the database 3A and access becomes impossible, the active node is used in the case of an access request to the database 2A so that the database access processing is performed on the mirror destination database. In the case of an access request to the database 3A, the active node 2 is selected. As a result, until the failure database is restored, all access processing such as searching and updating is performed on the mirror destination database in place of the failure database.

The DB allocation conversion mechanism 1A stores request destination data (for example, a table format) indicating a request destination of an access process determined according to a relationship between a state transition of a node in a database layer and a type of an access process. Data). This request destination data will be described later.

The active node 2 has a database 2
A and 3B are accessed, and the active node 3 accesses the databases 3A and 2B. That is, each active node accesses the original database (mirror source) corresponding to its own node and the mirroring database of the original database corresponding to the other active nodes. In other words, for example, database 2
Active node 2 accesses A (mirror source)
However, the active node 3 accesses the database 2B that mirrors the database 2A.
It is. Therefore, if the processing request (transaction processing) from the terminal 6 is a processing other than the update processing such as the search of the database 2A, the database access request node (hereinafter referred to as DB access request node) 1 In the case of the updating process, the access request is sent to the active nodes 2 and 3. Thereby, the contents of the database 2A and the database 2B can be kept the same. The processing described above is also performed for the access processing to the database 3A.

The standby node 4 is normally in a standby state, and starts a process to be in an active state (ie, a process for performing database access) in response to an access request from the DB access request node 1. As a result, access to each of the above databases becomes possible.

Next, an example of the request destination data is shown in FIG.
FIG. 2A shows an example of a request destination of an access process to the database 2A. In FIG. 2A, reference numeral 2
The row indicated by 0 indicates that when the active node 2 is normal, the active nodes 2 and 3 are the request destinations in the update processing, and that the active node 2 is the request destination in the processing other than the update processing. Have been. In the row indicated by reference numeral 30, when the active node 2 has failed and is switching to the standby node 4, the active node 3 is the request destination for both the update processing and the processing other than the update processing. It is shown. In the row indicated by reference numeral 40, when the switching is completed and the standby node 4 changes to the active node, in the updating process,
It is shown that the active nodes 4 and 3 are request destinations, and in processes other than the update process, the active node 4 is the request destination.

FIG. 2B shows an example of a request destination of an access process to the database 3A. In FIG. 2B, in the row indicated by reference numeral 50, when the active node 3 is normal, the active nodes 2 and 3 are request destinations in the update processing, and in the processing other than the update processing,
It is shown that the active node 3 is the request destination. In the row indicated by reference numeral 60, during a period in which a failure has occurred in the active node 3 and switching to the standby node 4 is being performed, the active node 2 is the request destination for both the update processing and processing other than the update processing. It is shown. In the line indicated by reference numeral 70, the switching is completed,
When the standby node 4 changes to the active node,
In the update processing, the active nodes 4 and 2 are the request destinations, and in the processing other than the update processing, the active node 4 is the request destination.

In this embodiment, the DB access request node 1, the DB layout conversion mechanism 1A, the active nodes 2, 3 and the standby node 4 store programs for realizing the functions of the respective constituent elements. A hard disk and a memory, and a central processing unit, and the central processing unit reads the program from the hard disk or the memory and executes the program to realize the program.

Next, the operation of accessing the database will be described with reference to the flowchart shown in FIG. In this embodiment, a processing operation in the case of an access request to the database 2A will be described.

First, when the DB access request node 1 receives a processing request from the requesting terminal 6 to the database 2A as a transaction process (step 101), the active node 2 receives a request based on a predetermined protocol. It is determined whether or not it is normal (step 10).
2). If the active node 2A is normal, the DB access request node 1 determines whether the processing request is an update process (step 103). If the update request is an update process, the DB allocation conversion mechanism 1A accesses the active node 2 A request is sent and an access request is sent to the active node 3 (step 104). The active node 2 that has received the access request executes an access process (update process) on the database 2A (step 10).
5). On the other hand, in the active node 3, the database 2
B is mirrored so as to be the same as the storage content of the database 2A (step 106).

In the above step 103, in the case of a process other than the update process (for example, a process of search, reference, etc.), the DB arrangement conversion mechanism 1A
An access request is sent (step 107). The active node 2 performs processing other than updating on the database 2A (step 108).

If a failure has occurred in the active node 2 in step 102 and access to the database 2A is impossible, the DB layout conversion mechanism 1A determines a node in the database layer that performs access processing to the databases 2A and 3B. Then, the active node 2 is switched to the standby node 4 (step 109), and the active node 3 is selected and the access request is transmitted so that the mirror destination database 2B is accessed (step 11).
0).

The switched standby node 4 activates a process for accessing the databases 2A and 3B.

On the other hand, the active node 3 performs all access processing such as reference, search, and update on the database 2B (step 111). That is, during the node switching period, the DB arrangement conversion mechanism 1A
Assume that all access request destinations to the database 2A are to the active node 3. Thus, after a failure occurs in the active node 2 and access to the database 2A becomes impossible, the standby node 4
Until the access to A becomes possible, the active node 3 performs the access process to the mirror destination database 2B in place of the failed active node 2, thereby stopping the access request to the database 2A. It will not be done.

When the standby node 4 changes from the standby state to the active state (step 112) and the access processing to the databases 2A and 3B becomes possible, the standby node 4 functions as an active node. . Here, the standby node 4 performing this function is referred to as an active node 4. Then, the DB arrangement conversion mechanism 1A
Instructs the active node 3 to read the contents of the database 2B and transfer it to the active node 4. The active node 4 reflects the transferred contents of the database 2B on the database 2A (step 113). As a result, the contents of the database 2A and the database 2B become the same.

For subsequent access requests to the database 2A, the DB layout conversion mechanism 1A selects the active node 4 (step 114). On the other hand, only when the access request is an update process, the active node 3
An access request to the database 2B is sent (step 115). As a result, the database 2B
Performs mirroring of the database 2A. Of course, when there is an access request to the database 3A, the active node 3 is selected, and only in the case of the update process, the active node 2 is selected.

The active node 2 becomes a standby node after the recovery from the failure, and performs the same function as the standby node 4 described above. That is, the roles of the active node 2 and the active node 4 are switched from the initial state.

Thereafter, each time a failure occurs in the database layer node, the above-described switching processing is performed.

FIG. 4 is a configuration diagram showing an aspect of the system after the occurrence of a failure in the active node 2 in the above-described processing operation.

As shown in FIG. 4, when there is an access request to the database 2A and a failure occurs in the active node 2, the active node 2 switches to the standby node 4, and the standby node 4 executes the above process. to start. At the same time, the active node 3 performs access processing to the database 2B, which is the mirror destination of the database 2A, while the standby node 4 is in the active state. When the standby node 4 becomes active, the contents of the database 2B are reflected on the database 2A as described above. The standby node 4 becomes the active node 4 and performs access processing to the databases 2A and 3B. On the other hand, the active node 3 performs access processing to the databases 3A and 2B.

As described above, according to this embodiment, the mirroring destination database of the database to be accessed by the failed active node in the database layer is accessed from another active node in the database layer. Even after the active node fails, the failure occurs even during the period of switching the standby node to the active node (that is, during the period from when the standby state is changed to the active state until the database can be accessed). The processing request to the database accessed by the active node that caused the error does not stop. This means that the waiting time for access to the database is significantly reduced. That is, when a failure occurs in the active node, the time from when the slave terminal 6 issues a processing request to when it receives a response can be significantly reduced as compared with the related art. Therefore, the turnaround time can be improved.

According to the present embodiment, the means for accessing the database is different from that of the conventional distributed database system shown in FIG. 8 and the system disclosed in Japanese Patent Application Laid-Open No. 8-278911. Can be suppressed to the minimum necessary, and the database can be duplicated.

In the embodiment described above, one standby node is provided. However, as shown in FIG. 5, the standby node may be duplicated. FIG. 5 shows a configuration in which a standby node 50 is added to the configuration shown in FIG.
With this configuration, when a failure occurs in the standby node 4 for some reason, the active node that has failed is switched to the standby node 50 and the processing operation illustrated in FIG. 3 is performed. Can be. Of course,
By multiplexing the standby nodes in triple, quadruple,..., It is possible to further cope with the failure of the standby nodes.

In the above embodiment, two active nodes, two mirror source databases, and two mirror destination databases are provided.
One or more active nodes, a mirror source database, and a mirror destination database may be provided. FIG. 6 shows a configuration diagram in the case where four components are provided. In the configuration shown in FIG. 1, active nodes 61 and 62, a mirror source database 61A,
62A, a mirror database 61B of the database 61A, and a mirror database 62B of the database 62A are added. With this configuration, even when the same load as the load in the embodiment shown in FIG. 1 is processed, the load can be distributed among the active nodes, so that the data processing amount of one active node is reduced. Is done. Further, when the data processing amount is the same, more data can be handled as compared with the embodiment shown in FIG.

In the above-described embodiment, the database of the mirror source and the database of the mirror destination are respectively constituted by separate hard disks. However, the present invention is not limited to this. Different types of databases can be stored. For example, database 2A and database 3B
Are stored in one hard disk.

Further, in the above-described embodiment, when a failure occurs in the active node, switching to the standby node is performed. However, the present invention is not limited to this, and the following can be performed. That is, when a failure occurs in a communication path between a predetermined active node and a predetermined database, and the predetermined active node cannot access the database, switching to the standby node is performed. Can also. Of course, in the case where a plurality of standby nodes are provided, if a failure occurs in a communication path between one standby node and the database, it is possible to switch to another standby node.

Further, in the above embodiment, the DB access request node 1 is connected to each of the active nodes and one or a plurality of standby nodes via a bus-type network. In this case, the active node is switched to the standby node instead of the active node, but this may be performed as follows. That is, the DB access request node 1 is connected to each of the active nodes and the single or plural standby nodes by a dedicated line, and the plural active nodes and the single or plural standby nodes are normal. When a failure occurs in such a dedicated line, the above-described node switching processing may be performed.

[0057]

As described above, according to the present invention, even when a failure has occurred in a predetermined node to which an access request is to be transmitted, the predetermined node can handle a database to be accessed. Since the backup database is accessed, the access request for the database to be accessed by the predetermined node does not stop.

Therefore, even when a failure occurs, the turnaround time can be improved by shortening the waiting time for access to the database.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an embodiment of a distributed database system according to the present invention.

FIG. 2 is an example of request destination data held by a database layout conversion mechanism.

FIG. 3 is a flowchart illustrating a processing operation according to the embodiment.

FIG. 4 is a configuration diagram illustrating an aspect of a system after a failure of an active node occurs.

FIG. 5 is a block diagram showing another embodiment of the distributed database system according to the present invention.

FIG. 6 is a block diagram showing another embodiment of the distributed database system according to the present invention.

FIG. 7 is a configuration diagram showing a configuration of a conventional distributed database system.

FIG. 8 is a configuration diagram showing a configuration of a conventional distributed database system.

[Explanation of symbols]

Reference Signs List 1 database access request node 1A database layout conversion mechanism 2, 3, 61, 62 active node 2A, 3A, 61A, 62A mirror source database 2B, 3B, 61B, 62B mirror destination database 4, 50 standby node 5 network 6 terminal

Claims (9)

[Claims] 1. A distributed database system comprising: a plurality of nodes each accessing a database via logically connected communication paths; and an access request node transmitting an access request to the plurality of nodes. Each of the plurality of nodes includes access means for accessing a different one of the plurality of databases and accessing a backup database corresponding to a database accessed by another node; Transmitting means for transmitting an access request to any one of the plurality of nodes; determining means for determining whether a node corresponding to the access request from the transmitting means can access a database; If access is determined to be inaccessible by the determination means Is distributed database system, characterized in that the inaccessible node to a node to access the database for backup corresponds to the database to be accessed, and a processing means for transmitting an access request. 2. The system according to claim 1, further comprising: a standby node physically connected to the plurality of databases and the plurality of backup databases connected to the plurality of nodes, and receiving a request from the access request node. The requesting node includes means for transmitting an access request to the standby node when it is determined that the access is inaccessible by the determining means, wherein the standby node responds to the access request and the inaccessible node 2. The distributed database system according to claim 1, further comprising access means for accessing a database to be accessed and a database for backup. 3. A plurality of said standby nodes, wherein said access request node transmits an access request to any one of said plurality of standby nodes when said access means determines that access is not possible. 3. The distributed database system according to claim 2, further comprising: 4. The access request node, when the determination unit determines that access is not possible, a unit that transmits an access request to any one of the plurality of standby nodes; Means for determining whether there is a response to the access request transmitted from the means; and a standby node different from the standby node having no response to the access request if the determination means determines that there is no response. 4. The distributed database system according to claim 3, further comprising means for transmitting an access request to the distributed database. 5. An access request node, comprising: information indicating a state transition of the plurality of nodes and a standby node; information indicating an access processing type; and transmission of an access request determined based on the information. It has a table that stores information indicating the destination, and according to the transition of the state of the node and the type of access processing,
5. The distributed database system according to claim 1, wherein a destination of the access request is determined with reference to the table. 6. The standby node, corresponding to each of the plurality of nodes, responds to an access request from the access request node, substitutes the inaccessible node, and accesses a predetermined database. The distributed database system according to claim 1, wherein: 7. A processing apparatus for transmitting an access request to a plurality of nodes accessing a database via communication paths logically connected to each other, comprising: First transmitting means for transmitting an access request to the database, determining means for determining whether a node corresponding to the access request from the transmitting means can access the database, and determining that the access is not possible by the determining means Processing means for transmitting an access request to a node accessing a backup database corresponding to a database to be accessed by the inaccessible node. 8. When the determining means determines that access is not possible, access is made to the plurality of databases connected to the plurality of nodes and the standby node physically connected to the plurality of backup databases. Second to send request
8. The processing apparatus according to claim 7, further comprising a transmitting unit. 9. A plurality of databases that access different ones of the plurality of databases via communication paths that are logically connected to each other and access a backup database corresponding to a database accessed by another node. A database access method in a distributed database, comprising: a node; and an access request node that transmits an access request to the plurality of nodes, wherein the access request node is for any one of the plurality of nodes. A transmitting step of transmitting an access request through the transmitting step; a determining step of determining whether a node corresponding to the access request transmitted by the transmitting step is capable of accessing the database; and a case in which the determining step determines that access is not possible. Inaccessible Database access method node to the node to access the database for backup corresponds to the database to be accessed, characterized in that it comprises a transmission step of transmitting an access request.