JP2003228527A - Server system with redundant function and operation control method for the server system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a server system with a redundant function and an operation control method for the server system that can continuously employ redundancy in the server system irrespective of limitations on the redundancy by statuses of functions in the system. <P>SOLUTION: In a client-server system 10, operation/standby determining function parts 12a and 14a respectively in a plurality of server devices 12 and 14 recognize which server system the associated devices are by classification into an operative system and a standby system, and addresses assigned to a virtual server device 12c configured by virtual server managing function parts 12b and 14b are supplied to and stored in the operative system and standby system server devices 12 and 14 and a network device 16, which are thus managed for switching upon an anomaly in the operative system server device 12, among others, whereby the virtual server device 12c is configured and operated in the operative system server device 12. <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a server system with a redundancy function and an operation control method for the server system, and the server system with the redundancy function is, for example, TCP / IP (Transmission Contr) on Ethernet (registered trademark).
It is suitable to be applied to a server system that connects and constructs multiple client servers in a network environment that uses a layered communication protocol (such as the Internet Protocol / Internet Protocol), and the client server is redundant. It has a function. Further, the operation control method of the server system with the redundancy function is suitable for application in the case where the operation of the system is continued without affecting the operation of the application included in the system even if a failure occurs in this system. is there. In particular, the application to be applied is
It is a client / server type application that realizes a system by a server and a client via a communication protocol.

[0002]

2. Description of the Related Art In some cases, a server system is required to continue operating as a system and maintain an application service provided by the system even if a server fails. A server system for satisfying such a system operation request has a redundant function (redundant function). An example of the case described above is a client / server type application, and an example of this specific application is VoIP (Voice over Internet).
There is an IP telephone service that uses Protocol technology.

In the IP telephone service, the IP telephone is used as a client, and the connection control between the IP telephones is performed by the IP exchange which is the server before the IP telephones start communication with each other. After this, communication between IP telephones is realized.

In such applications,
If the server (IP switch) fails, the client (IP phone) cannot communicate with the connection partner, and the application cannot be used at all. Is required. When such a system down occurs, the client / server system is desired to be a server (system) having a redundant function of automatically switching a standby backup server as an active server.

As described above, making the server redundant is an important factor for improving the redundancy and reliability of the system provided. Generally, the conventional server redundancy system has two servers, an active system and a standby system, and periodically confirms normality between both servers via a network that is normally connected. In addition, the active server transfers its current system operation data to the standby server in synchronization with a certain time or when there is a change, so that the database contents of both servers are always the same. I am trying.

By confirming the normality and sharing the contents of the database, even if the standby system server cannot confirm the normality of the active system server, the standby system server automatically switches as the active system server. , It has become possible to continue the processing from the client without inviting the system down.

The operation on the client side at the time of server switching is as follows: the operating server after switching, that is, the connection server changing procedure by receiving the server switching notification message from the original standby server to each client, or the client itself. By the voluntary connection server switching change procedure due to being unable to connect to the server being used in
Switching to the standby server. Which change procedure is used for switching is defined in advance at the start of operation.

In the case of the former change procedure, it is premised that the server knows the addresses of all clients to be managed in advance, and in the latter case, the clients are in advance both active and standby. It is assumed that you know the server address.

By the way, in the latter change procedure, the current server cannot be connected because the cable or the network device on the path to which the client is connected has failed, and the client erroneously switches the server. there is a possibility. In this case, even if a message erroneously arrives from the client to the standby server, the standby server is still in the standby state and does not respond to the client's request. Therefore, this message is rejected on the standby server side. This prevents erroneous client connections. In such a case, the server system actually treats the service as unavailable to the client due to a problem on the network side.

In either change procedure, from the client's point of view, a short server disconnection time occurs during the switching process from the active system to the standby system server. Here, since the client / server system has an automated procedure for server switching, the system down time is much reduced compared to a new startup of the standby server manually, so the effect of shortening the disconnection time is Is high.

[0011]

By the way, in an actual client / server system, terminal devices and network devices of various specifications coexist. Therefore, in the conventional client / server system with redundancy function, there is a problem in operation due to the restricted function of the layer 3 network device in the server switching processing procedure. This problem is the case when applying the server switching processing procedure of the client by receiving the server switching notification message from the standby server, and the case of the voluntary server switching procedure because the client itself cannot connect to the server being used. .

The first problem described above will be further described. In this case, the information of the address value used by the client is the same between the active server and the client, because the active server has all the addresses of the managed clients in advance when executing this operation mode. To do this, the server is made to function as a server for address resolution of the clients, and the addresses for the clients that it manages are allocated when each client starts up, so that the server can grasp the address information of all the clients. In this way, the operating system server knows which address is assigned to which client, so when sending a switching notification due to a failure, it is possible to send to those client addresses. Met.

By the way, when the active server is used as an address resolution server, it is realized as follows. For example, when TCP / IP protocol is used as the communication protocol, DHCP (Dynamic Host Configuration Protocol) is used as a general terminal address resolution method.
ol) and BOOTP (B00Tstrap Protocol).

At this time, when a client exists in a network number different from the network number installed in the server, when the client issues an address resolution request message to the server, the client enters and exits the external network. The address resolution request message cannot reach the server unless the layer 3 network device such as the router transfers to the network side where the server exists.

Generally, layer 3 network equipment is DHCP
Regarding handling of broadcast packets such as / BOOTP request messages, do not forward them to other networks other than the network where the packet was generated (broadcast domain division). Therefore, the client cannot get the address from the server.

Therefore, the DHCP / BOOTP from the client
In order to transfer the request message to the external network, the DHCP from the client is connected to the layer 3 network device.
/ BOOTP relay is set up, which is a function for forwarding a / BOOTP request message to a server on another network.

As described above, the layer 3 network device is transferred to the servers of the active system and the standby system in order to cope with the redundancy of the system.
DHCP / BOOTP relay is set. This allows
When the address resolution request from the client arrives at the layer 3 network device, it is transferred to the destination to which both the active system server and the standby system server are connected. Of the two servers that have received the request, the server operating as the active system sends the address-related information to the client that sent the address resolution request. On the other hand, the standby server does not respond to the client's request at this time, and the client's address information is obtained as a copy of the operational data from the operational server, and the operational data of the system is shared between both servers. .

In the response message from the operation server to the client, incidental information such as the address of the client and the addresses of the operation system server and the standby system server are included. Once the client has this information,
The system can enter the operational state.

In the above-described normal operation when acquiring the address of the client, the layer 3 network device used for the server address to be the active system and the standby system is in standby in response to the occurrence of a failure after the start of operation. When switching to the system server, the client usually already knows the address used by itself. Therefore, the client receives the switch notification message from the server,
Even if the status prompts switching, the local address can continue to use the value up to this point, so DHCP / BOOTP will not acquire the local address again. The client updates the address value of the active server that had been held up to now from the address information of the active server after switching indicated in the received switch notification message, immediately returns to the initial standby state, and the new server address This is because it is sufficient to connect to the server under the conditions.

However, when a client whose setting has been initialized or a newly installed client is targeted, the client supplies a DHCP / BOOTP request message (address resolution request message) to the server to acquire its own address. Will be done. Layer 3 network equipment that can set two transfer destinations as the DHCP / BOOTP relay function delivers the DHCP / BOOTP request message from the client to both servers, and a response is returned from the active server (standby server) at this point. Initialization or a new client can be operated, but only one of the layer 3 network devices having the specification that the DHCP / BOOTP relay function can specify only one system, for example, only the original operational server is specified. As a result, since transfer cannot be specified for the (standby) server after switching, transfer of the DHCP / BOOTP request message issued by the client is interrupted. As a result, the client cannot start operation because the address cannot be acquired.

It should be noted that DHC is provided between the server and the client.
Instead of using a dynamic automatic address resolution function such as P / BOOTP, by manually setting various address information to each client under the same conditions as the address information of the client that the server has, the standby server Although it is possible to use the server switching processing of the client by receiving the server switching notification message, it will be managed offline so that the information between the server and the client will not be inconsistent due to the manual setting, and the system scale will be reduced. The larger the error, the higher the error occurrence rate due to inconsistency in the setting conditions and the like, which may make it difficult to operate the entire system correctly.

Next, the above-mentioned second problem will be described. When the client cannot connect to the server being used by itself, the server switching processing procedure is voluntarily performed. When carrying out this operation mode, the server addresses of the active system and the standby system are held in advance in the client. The client obtains this server address from the server through the address resolution process at the time of initial startup of the server system, or sets it by manual input. When the client determines to perform the switching, the client switches the connection destination of the server to the standby server address.

Here, when the client switches the server in this procedure, the client is in an initial standby state, but since the client already has its own address as described above, the DHCP / BOOTP processing is performed. Absent.

However, as the function of the client, the one that can define only one type of server address is
Since the server address of the standby system cannot be recognized at all, switching by the client itself could not be realized.

In such a terminal device, the switching process can be realized only by the procedure of receiving the server switching notification message from the standby server, rather than using the procedure of the switching function led by the client side.

Summarizing the above-mentioned two problems, in order to retain the server address of the client, the first requirement condition for carrying out the address resolution message transfer of these networks or the second server address management of the client is required.
There are requirements. In the first requirement, firstly, all layer 3 network devices have a function of designating an address resolution message to a plurality of transfer destinations, or second
In addition, when the address resolution transfer of the layer 3 network device is not used, it is necessary to thoroughly perform the manual setting without the mismatch of the address setting between the server and the client.

The second requirement is that all clients should recognize a plurality of server addresses in advance. In this way, the server redundancy of the application cannot be applied to the entire system in any situation, and the actual situation is limited.

The present invention solves the above-mentioned drawbacks of the prior art, and the server with a redundant function can always apply the redundancy regardless of the limitation brought about by the function of the system in the redundancy of the server system. An object is to provide an operation control method for a system and its server system.

[0029]

In order to solve the above problems, the present invention provides a client device as a terminal device connected to a network and a network in a network environment using a hierarchical communication protocol. It is equipped with a plurality of server devices that are connected to this client device through centralized management of information, cooperate with each other of these server devices, monitor mutual failures, and are involved in the operation of the active server device. A determination function block that saves the operation data to the standby server device at the same time, and determines the active server device of the plurality of server devices that is operating and the standby server device that operates when a failure occurs Network for allowing other server devices to communicate between the client device and the server device via the client device and the network A virtual server device for the network device and stores the addresses assigned to the virtual server device in the active system server, the standby system server device, and the network device, and It is characterized by including a virtual server management function block that manages including switching according to an abnormality.

The server system with redundancy function of the present invention is
It is assigned to the virtual server device constructed by the virtual server management function block by classifying it into the active system and the standby system by the judgment function block that each of the plurality of server devices has, and recognizing what system the server is Addresses are also supplied to and stored in the active and standby server devices and network devices, and the system is operated including switching according to an error in the active server device. It is possible to always recognize the destination server device to be connected as seen from the client device as one virtual server device.

In order to solve the above-mentioned problems, the present invention is supplied via a network to a client on the terminal side connected to this network in a network environment using a hierarchical communication protocol. An operation control method for a server system with a redundancy function that allocates each of a plurality of servers to operation and standby in a server system including a server that centrally manages information and switches between operation and standby of the server when the server fails. A first step of performing a predetermined setting for the server, the standby server, and the virtual server, and a second step of recognizing the operation server and the standby server based on the attributes given to each server when starting this system. Then, the address set for the operation server is changed to the address of the virtual server and the system is restarted. The process and the fourth process of transferring the operational data of this system input to the operation server to the standby server and storing it, and the connection between the client side and the virtual server as they are when an error occurs in the operation server And a fifth step of switching the standby server to the operation server.

According to the operation control method of the server system with redundancy function of the present invention, after setting the operation server, the standby server, and the virtual server, the operation server is recognized and the address set in the operation server is set. By restarting with the virtual server address and storing the virtual server address in the network layer device,
From the perspective of the client, the operating server is specified as one virtual server, and even if there is an error in the operating server, the connection between the network layer device and the virtual server is left as it is to the standby server. Can be switched.

[0033]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of a server system according to the present invention will be described in detail with reference to the accompanying drawings.

In this embodiment, the server system with redundancy function of the present invention is applied to the client / server system 10. Illustration and description of parts that are not directly related to the present invention are omitted. Here, the reference numbers of signals are represented by the reference numbers of the connecting lines in which they appear.

The client / server system has a server device that centrally manages information resources and a client device that is hardware or software connected to a network that uses the services of this server device. Client / server system 10 of the present embodiment
1 includes an active server device 12, a standby server device 14, a layer 3 network device 16 and client devices 18a, 18b, ..., 18n, as shown in FIG.

As described above, the client / server system 10 is provided with two units, that is, the actual operating server unit 12 and the standby server unit 14 in order to bring redundancy to the system. Basically, each of the active server device 12 and the standby server device 14 includes an active / standby determination function unit 12a, 14a and a virtual server management function unit 12b, 14b corresponding to each one. The operation / standby determination function units 12a and 14a make determinations when allocating attributes to be the active server device 12 and the standby server device 14, and perform setting processing of the assigned attributes.

Further, the virtual server management function units 12b and 14b have a function of constructing an active system server and a standby system server according to the determination processing result, and a switching notification from the server device side or a switching from the client side. Define the client migration procedure at the time of switching the server device according to the above, and set the IP address relationship of itself.

Here, the IP address relationship includes an IP address, a subnet mask, and a default gateway address. The subnet mask is used to identify and notify the bits assigned to the network number in the IP address, and is set to identify a plurality of physical networks with one network address. The default gateway address is the address of the router or access server used when transmitting a packet to a subnet other than the subnet to which it belongs.

The MAC (Media Access Control) of its own
Since the address is individually written in advance in the server device, there is no own MAC address in the initial setting.

By the way, the operational server device 12 receives the judgment result as described above and receives the virtual server management function unit 12b.
Thereby, the virtual server device 12c is constructed. In order to perform this construction, the virtual server management function unit 12b uses the virtual server device 1
It also manages the MAC address and IP address used by 2c. The virtual server management function unit 14b also stores the MAC address and IP address of the virtual server device 12c. Although virtual, the MAC address of the virtual server device 12c is set to be properly managed by the device manufacturer. This is because the MAC address is a device-specific address and must not be duplicated with other devices.

Three types of server devices 1 for operation, standby, and virtual 1
Separate definitions of various addresses used by 2, 14, and 12c are that each layer 2 or layer 3 network device uses the MAC address and I
This is because the P-address relationship is associated and learned so that each server device is recognized. Therefore, operation,
The standby and virtual server devices 12, 14 and 12c respectively define three sets of unique MAC addresses and IP addresses, including virtual ones, as if they exist separately.
In this way, the active server device 12 and the standby server device 14, the set of MAC address and IP address for the virtual server when the virtual server device 12c of the present embodiment operates, the active server device 12, It is stored in both standby server devices 14. The active server device 12 and the standby server device 14 are connected to the network 100.

The layer 3 network device 16 is a layer 3 that causes the network 100 on the server side to communicate with the network 200 on the client side in the network layer.
Has a switch. Layer 3 network equipment 16
Is arranged between a server device and a client device of a single system or a plurality of systems in different subnets. The layer 3 network device 16 uses DHCP (Dynami
c Host Configuration Protocol) / BOOTP (BOOTstrap Pr
It has a function to deliver broadcasts by otocol). This function is the DHCP / BOOTP relay function.

The client devices 18a, 18b, ..., 18n are
It is a personal computer that uses a server device or software that runs on this computer. The client devices 18a, 18b, ..., 18n are network 200
Is connected to the layer 3 network device 16 via.

FIG. 2 shows a client / server system 10 that performs communication between different subnets via the layer 3 network device 16. The client that executes communication is the client device 18a. The active server device 12 and the client device 18a have seven layers.

The physical layer L1 is the first layer that creates data on a physical medium using electrical current / voltage changes. The data link layer L2 is the second layer that provides a reliable transfer function via the physical link. Network layer L3
Is a third layer that provides a logical addressing function, and determines paths that can be communicated in individual systems on different logical networks. In the network layer L3, for example, a routed protocol (Routed Protocol) is placed. Transport layer L4
Is responsible for ensuring reliable data transfer between networks. Upper layer of FIG. 2, fifth
The layer L5 to the seventh layer L7 indicate a session layer, a presentation layer, and an application layer.

The session layer L5 establishes, manages, and terminates sessions between applications. The presentation layer L6 has a function for surely reading information from the application layer of one system in the application layer of another system. Finally, the application layer L7 handles applications such as email, web browsers, word processors, etc.

Since communication is performed in each layer according to a hierarchical communication protocol, when communication is attempted between a server and a client, devices are connected to each other in the second layer L2 and the third layer L3. Address is used to For example, when performing communication of TCP / IP protocol over Ethernet (registered trademark), information format 20
In the header area 22 attached at the second layer L2 level to the destination MA.
C address 22a and source MAC address 22b, via type
An IP packet 24 and an FCS (Frame Check Sequence) 26 are added. IP packet 24 has IP header 24a, TCP
It is divided into a (Transmission Control Protocol) header 24b and a data area 24c.

Further, the IP header 24a is provided with an IP address area 240 used at the third layer L3 level in an area after various information is written. A source IP address 240a and a destination IP address 240b are also written in the IP address area 240. Therefore, the message transmitted / received by the client device 18a includes the MAC address and the IP address.

As described above, FIG. 4 shows the relationship of addresses in the three server devices. The active server device 12 is
In addition to including the MAC address 22w and the IP address 240w of the own device, the virtual server device 12c is constructed in the own device in actual operation, and the operation of the operation server device 12 is performed on its behalf.
At this time, the active server device 12 is the virtual server device 12c.
It borrows MAC address 22v and IP address 240v of and communicates with the client side. In order to set the action proxy,
After confirming the active system, the active server device 12 is restarted using the MAC address 22v and the IP address 240v managed by the virtual server management function unit 12b.

The standby server device 14 determines the MAC address 22s and the IP address 24 of its own device 14 in accordance with the determination of the standby state.
The MAC address 22v supplied from the virtual server device 12c while being managed by the virtual server management function unit 14b as 0s.
And the IP address 240v is also stored here. Since the virtual server device 12c has the address in advance as described above, when the active server device 12 fails, the standby server device 14 immediately takes over the virtual address managed by the virtual server management function unit 14b and restarts. Then, act as an active system. This enables the standby server device
14 can be operated as a new active server device having a new virtual server device 12c.

Therefore, from the client side,
Only one virtual server device address can be seen in the active server device that is always connected. This makes it possible to prevent the redundancy from being limited when the address message cannot be transferred or when only a single system can be defined even if a plurality of server addresses are specified.

Next, an operation procedure at the time of starting the client / server system 10, at the time of occurrence of a failure, and at recovery from the failure will be described with reference to FIGS. 5, 6 and 7. [Initial setting] First, initial setting is performed for each of the server devices 12 and 14 (step S10). Initial settings include active and standby attribute settings, client migration procedure definitions, and IP address-related settings. The device-specific MAC address has been written for each of the active server device 12 and the standby server device 14, but no address is set for the virtual server device 12c. Therefore, the active server device 12 and the standby server device 14 are
MAC address 22v used by the virtual server device 14 as shown in FIG.
And memorize the IP address 240v. [Server device start-up process] After setting, each server device is started up and normality confirmation processing is performed (step S12). This process broadcasts (broadcast notification) a normality confirmation message that does not consider the destination address. The normality confirmation message includes an attribute indicating which is the active server.

The active system server device 12 and the standby system server device 14 confirm the operational conditions based on the contents of the normality confirmation message supplied from each (step S14). When the normality confirmation message includes, for example, the operation system attribute (YES), the supplied server device is set as the operation system server device 12 to confirm the operation condition. Active server device 12
Enables the operation with the above-mentioned settings (MAC address 22v for virtual server and IP address 240v) for initial server operation. Therefore, the active server device 12 proceeds to the restart process (step S16). When the normality confirmation message includes the standby system attribute (NO), this server device is determined to be the standby system server device 14 and the standby operation is started (step S18).

In step S16, the active server device 12
Address condition is reset once, and virtual server management function unit
Change to the address of the virtual server device 12c managed by 12b, and restart. Due to the address change, the active server device 12 has the active attribute, the MAC address 22v and the IP address 240v and continues the operation.

Thereafter, the normality confirmation messages are mutually transmitted to monitor the operations of both parties (step S20). Monitoring
This message is sent periodically. When the normal operation is performed, the system operation data is input to the active server device 12. While maintaining this system operation data by the virtual server device 12c, the operation server device 12 copies the system operation data to the standby server device 14 in order to maintain redundancy. This copy is transferred point-to-point to the address of the server that sent the normality confirmation message. The transfer is performed periodically or synchronously when a change occurs.

The operations of steps S12 to S22 as described above correspond to the restart processing for the initial operation of the server.
As a result, the application becomes available on the client devices 18a, 18b, ..., 18n, and the system is brought into the online state (step S24). That is, the client process is started.

Here, the layer 3 network equipment 16 has the DHCP / BOOTP relay function of 1 by making such a correspondence.
Even if only the system can be set, regardless of whether it is the active system or the standby system, it operates on behalf of the address of one virtual server device 12c built in either, so DHCP / BOOTP of the layer 3 network device 16 You only need to specify the virtual server address in the relay settings. Therefore, the layer 3 network device 16 may apparently register only one server device. As a result, the DHCP for address resolution issued by the client device newly and initialized in the layer 3 network device 16
Since the transfer processing of the / BOOTP request message can be realized by only a single address, the client / server system 10 can operate the system at many occasions.

In the response returned by the active server at the time of address resolution of each client, by putting only the address of the virtual server in the server address value and responding, from the client side, it is as if a single server. Device 12
Show that only c exists. Further, even in the case of manually setting the address condition to the client, the server address can be set only by setting the address of the virtual server device 12c, and the setting becomes half of the conventional one.

While performing processing for a client, a connector
In step S26 of FIG. 6, the normality confirmation processing is periodically performed mutually between the active server device 12 and the standby server device 14 via A (normality monitoring). [Fault Countermeasure Processing 1] It is determined whether or not the normality confirmation message from the active server device 12 has not reached the standby server device 14 during monitoring (step S28). If the monitoring result has not been reached (YES), it is determined that an abnormality has occurred and the process proceeds to step S30.
If the monitoring result indicates the normal operation of the operation system (NO), the system operation is continued (step S32).
Then, the system operation data is copied to the standby server device periodically or when a change occurs, and otherwise the operation is passed through to return to the normality confirmation processing (step S34).

In response to the abnormality in the active server device 12, the standby server device 14 switches to the active system (step S3).
0). This switching uses the address of the virtual server device 12c included in the virtual server management function unit 14b of the standby server device 14 to make the standby server device 14 the active server device.
At the time of this switching, the system operation data is transferred from the operation system server device 12 so far and activated using the latest conditions that have been copied.

By the way, as a client migration procedure at the time of server switching, client devices 18a, 18b, ..., 18n
If you are applying the voluntary server switching method because you cannot connect to the server device you are using by yourself, the client device 18a, 18b, ... If it is not defined, the switching cannot be realized because the address of the standby server device is not known when it is determined to switch.

However, as described above, in this system 10, the client devices 18a, 18b, ..., 18n do not have to recognize a plurality of server addresses in advance, and even if a failure occurs, a single server address is used. It suffices to always try to connect to only. This is because the standby server device 14 responds again with the address value after the elapse of a certain switching time. That is, since the address of the server device is virtually single, it is not necessary to perform the address switching process.

However, the client devices 18a, 18b, ...
18n judges the server switching while considering the connection impossible time setting with the server device for a certain time or more,
It is desirable to transit to the initial standby state. The procedure of this transition will be described later.

In addition, from the operating system server 14 (original standby system server device) after switching to the client devices 18a, 18
In the case of the switching operation by receiving the server switching notification message for b, ..., 18n, the server address switching processing on the client side is unnecessary because the address of the server device is virtually single. Also in this case, it is preferable that the client devices 18a, 18b, ..., 18n determine the switching of the server device by receiving the switching notification message from the server, and transit to the initial standby state. In this way, in the client / server system 10, the client device 18
It is not necessary to recognize multiple server addresses on the a, 18b, ..., 18n side, and the system can continue to be used. [Failure Recovery Processing] It is determined whether or not the original operating system server device 12 in the failure state has recovered (step S36). When the original active server device 12 is restored (YES), although not shown in FIG. 6, the restored original active server device 12 first sends a normality confirmation message to the network 100 at its own address.
And the normality confirmation message issued by another server device is received, and the content is confirmed (step S38). If not restored (NO), connector B
The procedure advances to the failure confirmation processing of FIG.

In this confirmation, it is determined whether or not the standby server device 14 is operating as a new active server device (step S38). After the failure occurs, if the current active server device 14 is in operation using the address of the virtual server device 12c that has been operating independently (YE
S), the current active server device 14 receives the normality confirmation message issued by the restored original active server device 12, and performs normality confirmation processing. Next, the current active server device 14
It is determined whether is normal (step S40). When this determination result is normal (YES), the appearance of a new standby server device 12 is recognized. In response to this recognition, the current operation system server device 14 sequentially copies the system operation data of its own device 14,
The data is sent to the current standby server device 12.

This recognition is based on the operation system server device 12 which has been restored.
Is in the information element of the received health confirmation message,
It will be confirmed that the active server device 14 already exists, that is, the other existing operating by using the address of the virtual server device 12c. In this case, even if the restored active server device 12 has the setting of the original active server device, the attribute of the device is set as the standby attribute to be the standby server device, and the own address (MAC address 22w and Use the IP address 240w) as it is to enter the standby state (step
S42). The current standby server device 12 is operated so as to continue the normality check and sequentially obtain a copy of the system operation data from the current active server device 14 operating as the active server device.

At this time, the client devices 18a, 18b, ...
., 18n maintain communication with the current active server device 14 that is already in operation, and there is no particular problem in using the application. Therefore, the client devices 18a, 18b, ..., 18n are unaware that the failed server device has recovered.

Also, when the original active server device 12 has been restored, but there is not yet a new active server device 14 using the address value of the virtual server device 12c (step S3
8: NO), and the restored original operating system server device 12 is set as the operating system (step S44). Then, after this, the former active server device 12 restarts its address as the virtual server device 12c, and returns to FIG. 5 via the connector C so as to accept the subsequent processing on the client side (to step S16). .

Further, the original operational server device 12 is restored,
The current active server device 14 is operating, but the normal operation confirmation message from the current active server device 14 has not been received on the original active server device 12 side (step S40: N
The original operating system server device 12 that has also returned to O) is set as the operating system (step S44). Then, the process proceeds to the so-called server start operation of FIG. 5 via the connector C (to step S16). [Fault Countermeasure Processing 2] Next, the countermeasure processing when a failure occurs in both the active system server and the standby system server will be described with reference to FIG. The standby server device 14 is activated as the current active server device in response to an error in the active system. on the other hand,
The system 10 is in a situation in which the original operational server device 12 has not been recovered yet. Under this situation, it is determined whether or not the current active server device 14 is abnormal (step S4
8). If it is determined to be abnormal (YES), the process proceeds to the determination process based on the criteria on the client device side for this (step S50). If it is determined to be normal (NO), the system operation of FIG. 6 proceeds via the connector D (to step S32).

If the current active server device 14 is abnormal,
The client cannot communicate with the address of the virtual server device 12c and cannot use the application. However, the client device side described above confirms the existence of the server device as a criterion (step S50). ). This existence confirmation is performed by using a careful server existence confirmation means based on elapse of unconnected time at sufficient intervals. The failure is judged based on this result (step S5).
2).

When it is determined that the current active server device 14 has a failure and cannot be connected at all (YES), the client devices 18a, 18b, ..., 18n initialize all conditions, that is, resolve their own address. By performing the operation from the process, the server apparatus 12, 14 prepares for the initial operation when the server apparatus 12 and 14 returns (step S54). Client device 18a, 18
When all the address information is manually set by b, ..., 18n, there is no choice but to try to connect to the server under the condition that it is registered.

After this initialization, the server devices 12 and 14 are restored.
It is determined whether or not there is an abnormality (step S56). If an abnormality is still detected even after the above processing (YES), it is determined that the system is down, the operation is stopped, and the operation is ended. Also,
When it is determined to be normal by the failure determination in step S52 (N
O) or when it is determined to be normal by the abnormality determination in step S56 (NO), the server device 12, 14 shown in FIG.
The process proceeds to the step after starting (step S12).

In the above-described embodiment, the case where the number of servers is two has been described, but even when the system is constructed by a plurality of server devices of two or more, the same virtual server device as that of the embodiment is used. Needless to say, a client / server system having a flexible redundant function can be applied by setting conditions such as addresses in each server device and performing mutual monitoring between each server device.

With the above-mentioned configuration, the server device and the client device are used via the communication protocol, the active system and the standby system are designated to this server device, and the client is carried out in the system constructed with the redundant function. -In a server type application, hold a virtual single server address in the active and standby server devices, synchronize the operational data between these server devices, and operate to communicate with the client device. Therefore, even if a failure occurs in the server device, the server address viewed from the client device side can always be handled as one server address. As a result, even if there is a layer 3 network device that cannot transfer the client's address resolution request message to a plurality of server devices in the system, the setting of each layer 3 network device is always a single virtual one. The transfer process for the server address can be completed.

Therefore, even if a server failure occurs, it is only necessary to set the transfer to a single virtual address provided by the redundancy function, so that operation is continued even in a system including layer 3 network equipment having only one relay function. Can be made.

Further, even in the case of a client device in which only a single server device address can be set, the address of the active / standby server device may always be a single virtual address, so even if a failure occurs, the active / standby It is not necessary to switch the address corresponding to the system server device, and the system operation can be continued. This means that even a terminal device that does not have a mechanism for supporting redundancy on the client device side can still benefit from the system redundancy. Therefore, it is possible to significantly improve the implementation opportunity of server redundancy in the client / server type application, which has been restricted so far.

Since the transfer processing function of the address resolution request message of the layer 3 network device used in the system can be operated only by the single setting function, a cheaper layer 3 network device can be used.
It is also effective for cost reduction in system construction.

By controlling the operation by this method, the continuity of the system operation can be enhanced, and the application of the server redundancy in the client / server type application, which has been limited in the conventional redundancy. Can be greatly improved. Further, even if the client side does not have the function of supporting the redundancy, the function can be made redundant, which can also contribute to the ease of system construction.

[0079]

As described above, according to the server system with the redundancy function of the present invention, the determination function block is classified into the active system and the standby system, and is assigned to the virtual server device formed by the virtual server management function block. Address is also supplied to and stored in the active and standby server devices and network devices, and when viewed from the client side, the destination server device of the client device can always be recognized as one virtual server device. The redundancy in the server system is not limited to the multiple designations so far, and it is possible to greatly increase and improve the opportunity to always apply the redundancy, and to improve the continuity of the system operation. Also, by setting one virtual server device,
Since only a single setting is required for the transfer processing function of the network device, an inexpensive network device can be provided, and the benefit of redundancy can be brought even if the client side does not have a redundancy support function. This is effective because it can reduce the cost.

Further, according to the operation control method of the server system with redundancy function of the present invention, after setting the operation server, the standby server, and the virtual server, the operation server is recognized and set in the operation server. Existing address is restarted with the address of the virtual server, and the address of the virtual server is stored in the device of the network layer, so that the operating server when viewed from the client is one virtual server. Even if there is an error in the operation server, the operation server can be switched to the standby server without switching the address without changing the connection between the network layer device and the virtual server. Therefore, the continuity of system operation can be improved.

As a result, it is possible to greatly improve the application of the server redundancy in the client / server type application, which has been restricted in performing the redundancy so far. Further, even if the client side does not have the function of supporting the redundancy, the function can be made redundant, which can also contribute to the ease of system construction.

[Brief description of drawings]

FIG. 1 is a block diagram showing a schematic configuration of a client / server system to which a server system with a redundancy function of the present invention is applied.

FIG. 2 is a diagram illustrating a hierarchy of a communication protocol used in the system of FIG. 1 and a data flow.

3 is a diagram illustrating a format of information in the communication protocol of FIG.

FIG. 4 is a diagram illustrating a relationship of addresses managed by each server device of the present system.

5 is a flowchart illustrating initial setting and activation of operation procedures of the client / server system of FIG. 1. FIG.

FIG. 6 is a flowchart illustrating a procedure of failure countermeasure processing 1 and failure recovery processing following the operation procedure shown in FIG.

7 is a flowchart illustrating a procedure of failure countermeasure processing 2 following the operation procedure of FIG.

[Explanation of symbols]

10 Client / server system 12 Active server device 12a, 14a Operation / standby determination function unit 12b, 14b Virtual server management function section 12c Virtual server device 14 Standby server device 16 Layer 3 network equipment 18a, 18b, ..., 18n Client device 100, 200 networks

─────────────────────────────────────────────────── ─── Continuation of front page (51) Int.Cl. ⁷ Identification code FI theme code (reference) H04L 12/28 200 H04L 12/28 200Z F term (reference) 5B034 BB03 DD02 DD05 5B045 BB28 HH01 JJ07 JJ13 JJ24 JJ26 5B089 GA11 GB02 KB04 KB06 KC21 ME02 ME04 5K033 CC01 DA01 EA04 EB06 EC01 EC03

Claims (10)

[Claims] 1. In a network environment using a hierarchical communication protocol, a client device as a terminal device connected to the network and a client device via the network are connected to manage information centrally. A plurality of server devices that operate in cooperation with each other, perform mutual failure monitoring, and save operation data related to the operation of the active server device to the standby server device at the same time, Among the plurality of server devices, a determination function block that determines an active server device that is operating and a standby server device that operates when a failure occurs, and the active server device via the client device and the network. Virtual network device for communication between the client device and the server device. A server device and stores addresses assigned to the virtual server device in the active system server, the standby system server device, and the network device, and responds to an abnormality of the active system server device. A server system with a redundancy function, which includes a virtual server management function block that manages including switching. 2. The system according to claim 1, wherein the virtual server management function block builds the virtual server device in the active server device, and the active server device is configured to operate the virtual server device. A redundancy function characterized by communicating with the server device of the standby system and the client device via the network at the address of the server device and the server device of the standby system and the client device via the network device Server system with. 3. The system according to claim 1 or 2, wherein the operational server device restarts address change for designating a server device of a communication destination viewed from the client device as the virtual server device. A server system with a redundancy function, characterized in that a virtual server management function block is operated to operate the system. 4. The system according to claim 1, wherein the standby server device manages an address managed by the virtual server management function block in the standby server device when the active server device fails. A server system with a redundancy function, which takes over and operates as a new active server device. 5. The system according to any one of claims 1 to 4, wherein the system is unique to the virtual server device together with an address and an IP (Internet Protocol) address uniquely assigned to each server device. Server system with a redundancy function, wherein the virtual server management function block of each of the server devices is provided with a unique address and an IP address. 6. A server comprising, in a network environment using a hierarchical communication protocol, a client on the terminal side connected to the network and a server for centrally managing information supplied via the network. In the operation control method of a server system with a redundancy function, which allocates each of a plurality of servers to the system for operation and standby and switches between operation and standby of the server when the server fails, the method includes the operation server and the standby server, And a first step of performing a predetermined setting for the virtual server, and a second step of recognizing the operation server and the standby server based on the attributes given to each server when the system is started up.
And a third step of restarting by using the address set in the operation server as the address of the virtual server, and transferring the operation data of the system input to the operation server to the standby server Then, a fourth step of storing and a fifth step of switching the standby server to the operation server while keeping the connection between the client side and the virtual server when an abnormality occurs in the operation server are included. An operation control method for a server system with a redundancy function, which is characterized by the following. 7. The method according to claim 6, wherein in the first step, the device of the network layer that performs communication between the client and the operation server via the network in the communication protocol is performed. An operation control method for a server system with a redundancy function, characterized by storing the address of a virtual server. 8. The method according to claim 6, wherein after the fifth step, a sixth step of determining whether or not the operation server in which the abnormality has occurred has returned, and the operation server A seventh step of determining whether the current standby server is operating using the virtual server and is operating normally when the switched standby server is set as the current standby server If only the current operation server is operating, the eighth step of determining the operating state of the current operation server, and the determination result of the seventh step are normally operating, A case where the restored operation server is made to stand by as a standby server, and when the determination result determines that the virtual server in the current operation server has stopped operating, the restored operation server is restarted as a virtual server again. If the judgment results of the ninth step and the eighth step are normal operation, the operation of the system is continued, and if the judgment result is abnormal, the client side confirms the existence of the server. 10. An operation control method for a server system with a redundancy function, comprising: a tenth step of either restoring or stopping the system depending on the presence or absence of a system failure. 9. The method according to claim 8, wherein in the ninth step, the operation data of the current operation server is stored in the restored operation server, and the operation of the server system with a redundancy function is performed. Control method. 10. The method according to claim 8, wherein when it is determined that there is the failure in the tenth step, the client side is initialized to prepare for the server side return, and further, An operation control method for a server system with a redundancy function, characterized in that the presence or absence of an abnormality in the restored server is judged, and the system is restored or stopped depending on whether it is normal or abnormal.