JP2008028456A - Computer system capable of taking over service and ip address, and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To evade a contention state of IP addresses which a standby system server takes over from a principal system server even when the use of the IP addresses can not be stopped on the principal system server wherein a fault takes place. <P>SOLUTION: When detecting a fault caused in the principal system server 1-1, an intermediate driver15 of the standby system server 1-2 transmits a stop packet, for requesting stop of the use of the IP addresses used by services provided by the principal system server 1-1, to the principal system server 1-1 via a network driver 14 of the standby system server 1-2 and a network 3. When receiving the stop packet, an intermediate driver15 of the principal system server 1-1 establishes a particular state to the principal system server 1-1, wherein data transmission/reception between the network driver 14 and a TCP (Transmission Control Protocol)/IP module 16 is disabled. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention includes a computer system that includes first and second servers interconnected by a network that applies the TCP / IP protocol, and that can take over services and IP addresses between the first and second servers. Regarding the program.

  In recent client server systems, a main server and a standby server are mainly used in order to increase server (server computer) redundancy. In such a system, when a failure occurs in the primary server, the standby server can take over the function (= service: meaning that the service is provided to the client by the function).

Further, for example, in Patent Document 1, since a client (client terminal) can access a standby server without being aware of taking over the service between servers, the standby server is used by the primary server when taking over the service. It is disclosed that the IP (Internet Protocol) address that has been used is also taken over at the same time. In the following description, an IP address taken over from the primary server to the standby server is referred to as a specific IP address.
JP-A-8-212095

  In the above-described client server system, it is assumed that the primary server stops using the specific IP address when a failure occurs in the primary server. However, when the main server is slowed down under a high load condition, for example, the use of the specific IP address may not be stopped. On the other hand, the standby server detects the failure of the primary server and takes over the specific IP address. In such a case, a state (IP address conflict) occurs in which both the primary and standby servers use the same specific IP address.

When the above situation occurs, the following two problems occur.
When a client (or a router via a router) sends an ARP (Address Resolution Protocol) request to access a specific IP address, an ARP response is sent from both the primary and standby servers Is done. In this case, when the client receives the ARP response of the primary server in the failure state, the client communicates with the primary server, and thus a communication failure occurs.

  When the standby server takes over the IP address, it sets the IP address in the TCP (Transmission Control Protocol) / IP layer. At this time, the TCP / IP module of the standby server checks for an IP address conflict by a special ARP (hereinafter referred to as G-ARP) called Gravitational ARP. When another machine (primary server) already uses an IP address, an ARP response notifying IP address conflict from the other machine (primary server) is returned to G-ARP, and the standby server Fails to set the IP address.

  The present invention has been made in consideration of the above circumstances, and its purpose is to provide an IP address contention state that the standby server takes over from the primary server even when the use of the IP address cannot be stopped on the failed primary server. It is an object of the present invention to provide a computer system and a program that can avoid the problem.

  According to one aspect of the present invention, a first server and a second server are interconnected by a network to which a TCP / IP protocol is applied, and one of the first and second servers is a main server. When functioning as a system server and providing a service corresponding to a request given from a client via the network to the client, the other server of the first and second servers functions as a standby system server, When a failure of one server is detected, a computer system that takes over the service provided by the one server and the specific IP address used in the service is provided. The first and second servers of the computer system include a network driver that performs data link layer processing in TCP / IP communication, a TCP / IP module that performs processing in the TCP / IP layer in TCP / IP communication, An intermediate driver is provided between the network driver and the TCP / IP module and controls data transmission / reception between the network driver and the TCP / IP module. The intermediate driver of the second server stops using the specific IP address when the second server functions as the standby server and detects a failure of the first server. A stop packet for requesting the server is transmitted to the first server via the network driver of the second server and the network. When the intermediate driver of the first server receives the stop packet, the intermediate driver sets the first server to a specific state in which data transmission / reception between the network driver and the TCP / IP module is disabled. .

  According to the present invention, when a failure of the primary server is detected by the standby server, the standby server is addressed to the primary server before taking over the specific IP address used in the service provided by the primary server. By transmitting a stop packet, the main server can be forcibly set to a specific state in which data transmission / reception between the network driver of the main server and the TCP / IP module is disabled. This avoids an IP address race condition when the standby server uses a specific IP address even if the use of the specific IP address cannot be stopped immediately due to factors such as the main server slowing down due to a heavy load. can do.

Embodiments of the present invention will be described below with reference to the drawings.
[First Embodiment]
FIG. 1 is a diagram for explaining a basic configuration and an operation principle of a client server system including a computer system according to the first embodiment of the present invention. In FIG. 1, servers (server computers) 1-1 and 1-2 are connected by a network 3 that applies a TCP / IP protocol such as a TCP / IP network.

  A client (client terminal) 2 is also connected to the network 3. The client 2 is, for example, a personal computer that uses the servers 1-1 and 1-2 via the network 3. The client 2 transmits a specific request to an IP address associated with the service (used in the service) in order to receive provision of a desired service from a server (for example, the server 1-1) in the system. In the system of FIG. 1, for convenience of drawing, one client 2 is connected to the network 3, but generally a plurality of clients are connected to the network 3.

  The server 1-1 functions as a main server in the system state of FIG. 1, and provides a service corresponding to a request from the client (client 2) to the requesting client. Therefore, in the following description, the server 1-1 may be referred to as the primary server 1-1.

  The server 1-2 functions as a standby server in the system state of FIG. Therefore, in the following description, the server 1-2 may be referred to as a standby server 1-2. When a failure occurs in the primary server 1-1, the standby server 1-2 takes over the service provided by the primary server 1-1 and the IP address (specific IP address) used for the service. That is, the servers 1-1 and 1-2 constitute a computer system such as a cluster system.

  In the system of FIG. 1, for ease of explanation, the server 1-1 is called the primary server 1-1, and the server 1-2 is called the standby server 1-2. However, there is no essential difference between the primary server and the standby server, and the servers 1-1 and 1-2 can perform different processes independently. In other words, the server 1-1 functions as a primary server and the server 1-2 functions as a standby server for execution of a certain service, and the server 1-1 functions as a standby server and the server 1-2 for execution of another service. It can also function as a main server.

  A service process 11 is executed on the primary server 1-1 and the standby server 1-2. The service process 11 provides a service such as a Web service or a database service that can be used from the client 2 on the network 3. A plurality of service processes 11 can exist on the primary server 1-1 and the standby server 1-2, and the contents of the service may differ between the primary server 1-1 and the standby server 1-2.

  Each of the servers 1-1 and 1-2 has a cluster control unit 12. The cluster control units 12 of the servers 1-1 and 1-2 communicate with each other so that a server (hereinafter referred to as a local server) on which the cluster control unit 12 is placed and another server (hereinafter referred to as another server). (Referred to as above) and control of service process and IP address takeover. That is, when a failure occurs in the primary server 1-1, the cluster control unit 12 uses the service process 11 and the service process 11 (service provided by the service process 11) started on the primary server 1-1. The used IP address is taken over by the standby server 1-2, and the service is provided to the client 2 without being aware of the alternative control from the primary server 1-1 to the standby server 1-2. There are various types of faults, and various detection methods. In this embodiment, the communication between the cluster control unit 12 of the primary server 1-1 and the cluster control unit 12 of the standby server 1-2 is stagnated, so that the cluster control unit of the standby server 1-2. Assume that 12 detects a failure of the primary server 1-1.

  Each of the servers 1-1 and 1-2 includes a network interface (network IF) 13, a network driver 14, an intermediate driver 15, and a TCP / IP module 16.

  The network IF 13 interfaces with the network 3. The network driver 14 performs data link layer processing in TCP / IP communication. The TCP / IP module 16 is a function of the operating system (OS), and performs processing of the TCP / IP layer (processing of the network layer above the IP layer) in TCP / IP communication.

  The intermediate driver 15 is provided between the network driver 14 and the TCP / IP module 16 and controls data transmission / reception between the network driver 14 and the TCP / IP module 16. Here, the intermediate driver 15 controls data transmitted and received between the network driver 14 and the TCP / IP module 16. That is, the intermediate driver 15 logically connects the network driver 14 and the TCP / IP module 16. In the conventional system, data is directly transmitted and received between the network driver 14 and the TCP / IP module 16, and the system of FIG. 1 is different from the conventional system in this respect. The intermediate driver 15 has a MAC (Media Access Control) address as a physical address of the primary server 1-1 and the standby server 1-2, and an IP address taken over from the primary server 1-1 to the standby server 1-2. (Specific IP address) information. In this embodiment, the intermediate driver 15 of the primary server 1-1 and the standby server 1-2 receives a specific program stored in a storage device such as a disk storage device, for example. 2 is realized by reading and executing.

  Assume that a failure has occurred on the primary server 1-1 (step S1). The cluster control unit 12 of the standby server 1-2 detects a failure of the primary server 1-1 (step S2). The cluster control unit 12 of the standby server 1-2 requests the intermediate driver 15 of the server 1-2 to perform a stop process in order to cause the primary server 1-1 to stop using the IP address (step S3). Then, the intermediate driver 15 of the standby server 1-2 generates a stop packet in response to a request (stop request) from the cluster control unit 12, and transmits the stop packet to the primary server 1-1 (step). S4). Details of the generation of the stop packet will be described later.

  When receiving the stop packet, the intermediate driver 15 of the primary server 1-1 forcibly blocks all network processing on the primary server 1-1 thereafter (step S5). That is, the intermediate driver 15 of the primary server 1-1 discards all packets input to the intermediate driver 15 to be passed from the network driver 14 to the TCP / IP module 16. Similarly, the intermediate driver 15 discards all packets input to the intermediate driver 15 to be passed from the TCP / IP module 16 to the network driver 14. In this way, the primary server 1-1 becomes unable to transmit and receive data between the network driver 14 and the TCP / IP module 16 when the intermediate driver 15 of the server 1-1 receives the stop packet (inhibition). Set to a specific state.

  Now, the intermediate driver 15 of the standby server 1-2 sends a stop packet to the primary server 1-1, and then confirms an IP address (specific IP address) conflict (that is, the primary server 1-1). G-ARP (G-ARP packet) is broadcasted to the network 3 (in order to confirm whether or not the use of the IP address is stopped) (step S6). Details of the G-ARP generation will be described later.

  As described above, when the intermediate driver 15 of the primary server 1-1 receives the stop packet, it subsequently discards all packets passed from the network driver 14 to the TCP / IP module 16. For this reason, the G-ARP transmitted by the standby server 1-2 is also discarded by the intermediate driver 15 of the primary server 1-1 (step S7). Therefore, the G-ARP transmitted from the standby server 1-2 is not passed to the TCP / IP module 16 of the primary server 1-1, and the G-ARP is transferred from the primary server 1-1 to the standby server 1-2. An ARP response to the ARP is not returned. Thereby, the standby server 1-2 can take over the IP address (specific IP address) from the main server 1-1.

  As described above, in this embodiment, before the standby server 1-2 takes over the IP address, a stop packet is transmitted to the primary server 1-1, and all network processing of the server 1-1 is forcibly blocked. (Stopped), the IP address conflict state when the standby server 1-2 uses the IP address (that is, the state where the primary server 1-1 and the standby server 1-2 use the same IP address) ) Can be prevented.

  On the other hand, the cluster control unit 12 of the standby server 1-2 requests the intermediate driver 15 of the server 1-2 to stop processing, and then sets the IP address (specific IP address) of the server 1-2. This is performed for the TCP / IP module 16 (step S8).

  When the IP address is set, the TCP / IP module 16 of the standby server 1-2 broadcasts G-ARP to the network 3 to confirm the conflict of the IP address. The G-ARP transmission by the TCP / IP module 16 can be used in place of the G-ARP transmission by the intermediate driver 15 (step S6).

  Here, due to a cause such as the stop packet (step S4 in FIG. 1) transmitted by the standby server 1-2 being lost on the network 3, the intermediate driver 15 of the primary server 1-1 transmits the stop packet. The case where the packet could not be received will be described with reference to the packet flow diagram of FIG.

  First, when the intermediate driver 15 of the primary server 1-1 cannot receive the stop packet, the network processing on the primary server 1-1 is not blocked by the intermediate driver 15. When the intermediate driver 15 of the standby server 1-2 transmits a stop packet, it transmits a G-ARP (G-ARP packet) to the network 3 by broadcasting (step S11). This step S11 corresponds to step S6 in FIG.

  It is assumed that G-ARP from the standby server 1-2 is received by the network IF 13 of the primary server 1-1 and is input from the network driver 14 of the server 1-1 to the intermediate driver 15. At this time, since the intermediate driver 15 of the primary server 1-1 has not received the stop packet previously transmitted from the standby server 1-2, it does not discard the G-ARP and the server 1-1 It passes through the TCP / IP module 16.

  Then, the TCP / IP module 16 of the primary server 1-1 returns an ARP response warning the duplication of the IP address to the standby server 1-2 (step S12). When the standby server 1-2 receives the ARP response transmitted from the primary server 1-1, the standby server 1-2 determines that the intermediate driver 15 of the standby server 1-2 has failed to transmit the stop packet.

  Therefore, the intermediate driver 15 of the standby server 1-2 discards the ARP response from the primary server 1-1 without passing it to the TCP / IP module 16 of the standby server 1-2, and again transmits the stop packet to the primary server. It transmits to the server 1-1 (step S13). Next, the intermediate driver 15 of the standby server 1-2 sends G-ARP again to the primary server 1-1 to confirm whether the primary server 1-1 has stopped using the IP address. (Step S14).

  As described above, when the intermediate driver 15 of the standby server 1-2 receives the ARP response from the primary server 1-1, the intermediate driver 15 discards the ARP response and executes the processes of steps S13 and S14. Thereby, the standby server 1-2 can take over the IP address from the main server 1-1.

  When the TCP / IP module 16 of the primary server 1-1 returns an ARP response warning the duplication of the IP address (specific IP address) to the standby server 1-2 (step S12), the primary server 1-1 In order to declare that 1-1 is a valid user of a specific IP address, G-ARP is generally transmitted on the subnet. In this case, there is a possibility that the ARP table (a table in which address information including a pair of MAC address and IP address is stored) of each machine on the subnet is rewritten with respect to the specific IP address.

  However, in this embodiment, when the intermediate driver 15 of the standby server 1-2 receives the ARP response from the primary server 1-1, the intermediate driver 15 transmits the processing of steps S13 and S14, that is, the stop packet to the primary server 1-1. And an operation for transmitting the G-ARP by broadcast. The state in which the ARP table of each machine on the subnet is rewritten can be corrected by sending the G-ARP after sending the stop packet by the intermediate driver 15 of the standby server 1-2. Further, when the stop packet is received by the intermediate driver 15 of the primary server 1-1, all network processing of the server 1-1 is blocked, so that the IP address conflict state can be resolved.

Next, stop packet and G-ARP periodic transmission by the intermediate driver 15 of the standby server 1-2 will be described with reference to the sequence chart of FIG.
First, a stop packet is transmitted from the standby server 1-2 to the main server 1-1 (step S21), and then G-ARP is transmitted from the standby server 1-2 by broadcast (step S22). ). If the intermediate driver 15 of the primary server 1-1 cannot receive both the stop packet and G-ARP, the primary server 1-1 cannot stop using the IP address. However, an ARP response is not returned from the primary server 1-1 to the standby server 1-2.

  In this case, the standby server 1-2 cannot detect that the primary server 1-1 is operating. Therefore, the intermediate driver 15 of the standby server 1-2 repeats an operation including an operation of transmitting a stop packet and an operation of transmitting a G-ARP at regular intervals (steps S23, S24, S25, S26,...). That is, the intermediate driver 15 of the standby server 1-2 periodically transmits a stop packet and G-ARP.

  As described above, when the intermediate server 15 of the primary server 1-1 can receive all the stop packets and G-ARP transmitted from the standby server 1-2 as a result of the failure of the primary server 1-1, Alternatively, the standby server 1-2 avoids the IP address contention state when either the stop packet or the G-ARP stop packet cannot be received, or when both the stop packet and the G-ARP cannot be received. The IP address can be taken over from the primary server 1-1.

In the present embodiment, it is important that the intermediate driver 15 exists between the TCP / IP module 16 and the network driver 14. This will be described below.
In general, the server responds to an ARP request from a client or the like. However, when the failed primary server 1-1 responds to the ARP request, a communication failure occurs in the prior art as described in the section “Problems to be solved by the invention”. When the primary server 1-1 returns an ARP response to the G-ARP from the standby server 1-2, the standby server 1-2 fails in the IP address setting in the prior art.

  The processing related to the ARP is performed via the TCP / IP module 16 included in the OS. For this reason, the occurrence of ARP processing means that not only the TCP / IP module 16 but also the intermediate driver 15 located between the TCP / IP module 16 and the network driver 14 is in an operable state. To do. When the intermediate driver 15 is operable, even if the ARP process occurs, the standby server 1-2 avoids the IP address conflict state as described above, and the primary server 1-1 It becomes possible to take over the address.

  On the other hand, when the intermediate driver 15 does not operate, a packet transferred from the network driver 14 to the intermediate driver 15 cannot be transferred from the intermediate driver 15 to the TCP / IP module 16. Similarly, a packet transferred from the TCP / IP module 16 to the intermediate driver 15 cannot be transferred from the intermediate driver 15 to the network driver 14. When the intermediate driver 15 cannot operate as described above, the communication function on the server where the intermediate driver 15 exists stops and the above ARP problem does not occur.

  FIG. 4 is a block diagram showing a detailed configuration of the server 1-i (i = 1, 2), particularly the intermediate driver 15. The intermediate driver 15 receives a communication packet (data link layer packet including a data link layer header) from the network driver 14 of the server 1-i, performs data processing inside the intermediate driver 15, and An IP packet is sent to the TCP / IP module 16. The intermediate driver 15 receives an IP packet from the TCP / IP module 16 of the server 1-i, performs data processing inside the intermediate driver 15, and transmits a communication packet (data link layer) to the network driver 14 of the server 1-i. Packet).

  As described above, the intermediate driver 15 is provided between the network driver 14 and the TCP / IP module 16, and includes a data link input unit 151, a data link output unit 152, a stop packet processing unit 153, an IP input unit 154, an IP It includes an output unit 155, a packet passing state table TBL1, a server information table TBL2, and a remote server state table TBL3.

  Hereinafter, operations of the data link input unit 151, the data link output unit 152, the stop packet processing unit 153, the IP input unit 154, and the IP output unit 155 in the intermediate driver 15 will be sequentially described.

  First, the operation of the data link input unit 151 in the intermediate driver 15 will be described with reference to the flowchart of FIG. 5, the data format of FIG. 6, and the data structure example of the packet passing state table TBL1 of FIG.

  The data link input unit 151 receives a packet (ARP packet) 61, a packet (stop packet) 62 or a packet (data) as shown in FIG. 6A, FIG. 6B, or FIG. When the link layer packet 63) is passed as input data, the packet passing state table TBL1 is checked (step S31). The packet passage state table TBL1 holds a flag indicating whether or not a packet can pass through the intermediate driver 15, as shown in FIG. This flag is set to a state indicating that a packet can be passed (packet passable mode) when a stop packet is not received, and is set to a state indicating that a packet cannot be passed (packet pass disabled mode) when a stop packet is received. .

  The data link input unit 151 discards the input data (input packet) (step S33) when the packet passage state table TBL1 indicates that the passage is impossible (packet passage impossible mode) (step S32).

  On the other hand, when the packet passage state table TBL1 indicates that the packet can be passed (packet passable mode), the data link input unit 151 determines whether or not the input data is an ARP packet (step S34). If the input data is an ARP packet 61 as shown in FIG. 6 (a), the data link input unit 151 uses the input data (that is, the ARP packet 61) as it is as an ARP packet 64 shown in FIG. 6 (d). The packet is sent to the stop packet processor 153 (step S35), and the process is terminated.

  On the other hand, when the input data is not an ARP packet (step S34), the data link input unit 151 determines whether the input data is a stop packet (step S36). If the input data is a stop packet 62 as shown in FIG. 6B, the data link input unit 151 directly uses the input data (that is, the stop packet 62) as a stop packet 65 shown in FIG. 6E. The packet is sent to the stop packet processor 153 (step S35), and the process is terminated.

  On the other hand, when the input data is not a stop packet, that is, when the input data is a packet (data link layer packet) 63 as shown in FIG. 6C other than the ARP packet and the stop packet (step S36), the data link The input unit 151 proceeds to step S37. In step S37, the data link input unit 151 converts the input data (data link layer packet 63) from which the data link layer header is removed, that is, an IP packet including an IP header and an IP payload, as shown in FIG. ) To the IP input unit 154.

  Next, the operation of the data link output unit 152 in the intermediate driver 15 will be described with reference to the flowchart of FIG. 8 and the data format of FIG.

  The data link output unit 152 receives the destination MAC address 91 (the physical address of the destination server) and the communication data (data link layer payload) 92 shown in FIG. 9A from the stop packet processing unit 153 or the IP output unit 155. Sent. In this case, the data link output unit 152 uses the destination MAC address 91 and the communication data (data link layer payload) 92 as they are, and the destination MAC address 93 and the communication data (data link layer payload) 94 shown in FIG. To the network driver 14 (step S41), and the process ends.

Next, the operation of the stop packet processing unit 153 in the intermediate driver 15 will be described.
The stop packet processing unit 153 performs the following four processes <Process A1>
Processing to initialize packet passing state table TBL1, server information table TBL2, and other server state table TBL3 in response to a request from cluster control unit 12 <Processing A2>
Processing performed when a stop request is received from the cluster control unit 12 (processing including stop packet generation and stop packet transmission)
<Process A3>
Processing when a packet (ARP packet or stop packet) is received from the data link input unit 151.

<Process A4>
Stop packet and G-ARP periodic transmission processing.

Execute.

  Next, the processing A1 executed by the stop packet processing unit 153 will be described with reference to the flowchart of FIG. 10 and the data structure example of the server information table TBL2 of FIG. 11 and the other server status table TBL3 of FIG. To do.

  For example, the cluster control unit 12 requests initial setting of the tables TBL1, TBL2, and TBL3 in order to make the intermediate driver 15 function when the program is installed in the server 1-i. At this time, as information for initial setting, the cluster control unit 12 periodically transmits a MAC address of each of the own server and the other server, a specific IP address (takeover IP address) to be taken over, and a stop packet. Information on the transmission interval (periodic transmission interval) is given to the stop packet processing unit 153.

here,
Local server MAC address: AA: BB: CC: DD: EE: FF
MAC address of other server: FF: EE: DD: CC: BB: AA
Takeover IP address: 192.168.10.10.
Periodic transmission interval: It is assumed that 10 seconds is given.

  Then, the stop packet processing unit 153 sets, in the server information table TBL2, the MAC address, takeover IP address, and periodic transmission interval information of each of the own server and the other server given from the cluster control unit 12 ( Step S51). FIG. 11 shows an example of setting contents of the server information table TBL2 (here, the server information table TBL2 of the standby server 1-2). Here, the MAC address of the local server and the MAC address of the other server set in the server information table TBL2 of the primary server 1-1 are set in the server information table TBL2 of the standby server 1-2, respectively. Note that it matches the MAC address of the other server and the MAC address of the own server.

  Next, the stop packet processing unit 153 initializes the other server state table TBL3 (step S52). The table TBL3 holds a flag indicating whether the other system server is in a normal state or a failure state. In step S52, the flag held in the table TBL3 is initialized to a state indicating that the other system server is in a normal state. FIG. 12 shows an example of the setting contents of the other server state table TBL3 at this time.

  Finally, the stop packet processing unit 153 initializes the flag held in the packet passage state table TBL1 to a state indicating that it can pass (the state illustrated in FIG. 7) (step S53), and ends the process A1.

  Next, for the processing A2 executed by the stop packet processing unit 153 (the stop packet processing unit 153 included in the intermediate driver 15 of the standby server 1-2), refer to the flowchart of FIG. 13 and the data format of FIG. To explain.

  When the stop packet processing unit 153 receives a stop request from the cluster control unit 12, the stop packet processing unit 153 starts a process A2 according to the flowchart of FIG. First, the stop packet processing unit 153 sets the flag held in the other server state table TBL3 to a state indicating that the other server is in a failure state (step S61).

  Next, the stop packet processing unit 153 generates a stop packet (step S62). This stop packet may be in any format as long as at least data that can be identified as a stop packet is stored in the payload of the data link layer. FIG. 14 shows an example in which an IP packet 142 is used as a stop packet. The IP packet (stop packet) 142 includes an IP header and an IP payload. Here, data that can be identified as a stop packet is stored in the IP payload of the IP packet (stop packet) 142. The IP header of the IP packet (stop packet) 142 includes a source IP address and a destination IP address. In the example of FIG. 14, the takeover IP address (192.168.10.10) set in the server information table TBL2 is used for the source IP address and the destination IP address. However, as long as data that can be identified as a stop packet is stored in the IP payload as in this embodiment, other IP addresses may be used.

  The stop packet processing unit 153 acquires the MAC address of the other server set in the server information table TBL2 as the destination MAC address 141 (see FIG. 14) of the generated stop packet 142. The stop packet processing unit 153 sends the destination MAC address 141 and the stop packet 142 to the data link output unit 152 (step S63).

  Next, the stop packet processing unit 153 generates a G-ARP packet by using information set in the server information table TBL2 (step S64). Next, the stop packet processing unit 153 sends the generated G-ARP packet to the data link output unit 152 together with the broadcast address as the destination MAC address (step S65), and ends the processing A2. In this case, the G-ARP packet is transmitted on the network 3 by broadcast. A method for generating the G-ARP packet will be described in detail with reference to FIG. 18 in the process of step S79 in the flowchart of FIG.

  As described above, according to the present embodiment, the intermediate driver 15 (the intermediate driver 15 of the standby server 1-2) generates a stop packet based on the server information table TBL2, and does not pass through the TCP / IP module 16. The stop packet can be transmitted to the destination MAC address 141 of the other server (main server 1-1). Similarly, the intermediate driver 15 can generate a G-ARP packet based on the server information table TBL2, and transmit the G-ARP packet by broadcast without going through the TCP / IP module 16.

  Next, the process A3 executed by the stop packet processing unit 153 will be described with reference to the flowchart of FIG. 15 and the data formats of FIGS.

  When the stop packet processing unit 153 receives a packet (data link layer packet) from the data link input unit 151, the stop packet processing unit 153 starts processing A3 according to the flowchart of FIG. First, the stop packet processing unit 153 determines whether the input data (input packet) is an ARP request packet (step S71). If the input data is an ARP packet 161 as shown in FIG. 16A and the ARP packet 161 is an ARP packet indicating an ARP request (that is, an ARP request packet), the stop packet processing unit 153 The part obtained by removing the data link layer header from the input data (ARP request packet) is sent to the IP input unit 154 as the ARP packet 167 shown in FIG. 16E indicating the ARP request (step S72), and the process A3 is terminated.

  On the other hand, if the input data is not an ARP request packet, the stop packet processing unit 153 determines whether the input data is an ARP response packet (step S73). If the input data is an ARP packet 161 as shown in FIG. 16A and the ARP packet 161 is an ARP packet indicating an ARP response, that is, an ARP response packet 170 shown in FIG. The processing unit 153 determines whether the transmission source MAC address (physical address of the transmission source server) included in the ARP response packet 170 is the MAC address of the other server set in the server information table TBL2. (Step S74). If the source MAC address is the MAC address of the other server, the stop packet processing unit 153 determines whether the source IP address and the target IP address included in the ARP response packet 170 are the same ( Step S75). If the source IP address and the target IP address are the same, the stop packet processing unit 153 indicates in the server information table TBL2 whether the target IP address (= source IP address) is a takeover IP address (specific IP address). This is determined by comparing with the set takeover IP address (step S76). If the target IP address (= source IP address) is a takeover IP address, the stop packet processing unit 153 proceeds to step S77.

  Thus, the input data is the ARP response packet 170, the transmission source MAC address of the ARP response packet 170 is the MAC address of the other server, and the transmission source IP address and the target IP address of the ARP response packet 170 are If it is a takeover IP address, the stop packet processing unit 153 indicates that the ARP response packet 170 is an ARP response to the G-ARP of the local server (here, the standby server 1-2) (here, the ARP from the primary server 1-1). If it is a response, the process proceeds to step S77.

  In step S77, the stop packet processing unit 153 makes the stop packet 164 as shown in FIG. 16C (corresponding to the stop packet 142 shown in FIG. 14) independent from the cluster control unit 12 and the TCP / IP module 16. To generate. Next, the stop packet processing unit 153 acquires the MAC address of the other server set in the server information table TBL2 as the destination MAC address 163 (see FIG. 16C) of the generated stop packet 164. The stop packet processing unit 153 sends the destination MAC address 163 and the stop packet 164 to the data link output unit 152 (step S78).

  As described above, in the present embodiment, the stop packet addressed to the other system server (primary server 1-1) transmitted according to the flowchart of FIG. The intermediate driver 15 cannot receive the IP address and cannot stop using the IP address, and the G-ARP returns an ARP response warning the IP address conflict (IP address duplication) from the intermediate driver 15 of the other server. In this case, the stop packet can be retransmitted to the other server.

  Next, the stop packet processing unit 153 generates the G-ARP packet 166 shown in FIG. 16D using the information set in the server information table TBL2 (step S79). Here, as the source MAC address of the G-ARP packet 166, as shown in FIG. 18, the MAC address of the local server set in the server information table TBL2 is used. Further, as the target MAC address of the G-ARP packet 166, 0 address (00: 00: 00: 00: 00) is used as shown in FIG. As shown in FIG. 18, the takeover IP address set in the server information table TBL2 is used for the transmission source IP address and the target IP address of the G-ARP packet 166.

  Next, the stop packet processing unit 153 acquires the broadcast address (FF: FF: FF: FF: FF: FF) as the destination MAC address 165 shown in FIGS. The stop packet processing unit 153 sends the destination MAC address 165 and the generated G-ARP packet 166 to the data link output unit 152 (step S80), and ends the processing A3.

  On the other hand, even if the input data is the ARP response packet 170 (step S73), the transmission source MAC address of the ARP response packet 170 is not the MAC address of the other server (step S74), or the transmission of the ARP response packet 170 is performed. If the original IP address and the target IP address are not the same (step S75), or if they are the same but not the takeover IP address (step S76), the stop packet processing unit 153 sends the ARP response packet 170 to the G- It is determined that the response is not an ARP response to the ARP. In this case, the stop packet processing unit 153 sends the part obtained by removing the data link layer header from the ARP response packet to the IP input unit 154 as an ARP packet (ARP response packet) 167 shown in FIG. 16E (step S72). The process is terminated.

  If the input data is not an ARP response packet (step S73), that is, if the input data is neither an ARP request packet nor an ARP response packet, the stop packet processing unit 153 determines whether the input data is a stop packet. (Step S81). If the input data is a stop packet 162 as shown in FIG. 16B, the stop packet processor 153 indicates that the source MAC address included in the stop packet 162 is server information as shown in FIG. It is determined whether or not it matches the MAC address of the other server set in the table TBL2 (step S82). If the source MAC address matches the MAC address of the other server, the stop packet processing unit 153 sends the source IP address and destination included in the stop packet 162 (its IP header) as shown in FIG. It is determined whether the IP address matches the takeover address set in the server information table TBL2 (steps S83 and S84). If the source IP address and the destination IP address match the takeover address, the stop packet processing unit 153 proceeds to step S85.

  Thus, the input data is the stop packet 162, the source MAC address of the stop packet 162 is the MAC address of the other server, and the source IP address and the destination IP address of the stop packet 162 are the takeover IP. If it is an address, the stop packet processing unit 153 determines that the stop packet 162 is for the local server and proceeds to step S85.

  In step S85, the stop packet processing unit 153 changes the flag held in the packet passage state table TBL1 to a state indicating that the packet cannot be passed (a state opposite to the state shown in FIG. 7) (step S85). A3 is terminated.

  On the other hand, whether the input data is not the stop packet 162 (step S81), or even if it is the stop packet 162, the data set in the server information table TBL2 at least one of the source MAC address, source IP address, and destination IP address (Step SS82, S83, S84), the stop packet processing unit 153 determines that the input data is not for the local server. In this case, the stop packet processing unit 153 discards the input data (step S86) and ends the process A3.

  Next, the process A4 executed by the stop packet processing unit 153 will be described with reference to the flowchart of FIG. 20 and the data format of FIG. This process A4 is executed at regular transmission intervals set in the server information table TBL2.

  The stop packet processing unit 153 checks the status of the other server with reference to the other server status table TBL3 (step S91). If the other server is in a normal state (step S92), the stop packet processing unit 153 does nothing and ends the process A4.

  On the other hand, if the other server is in a failure state (step S92), that is, if a failure has occurred in the other server, the stop packet processing unit 153 (corresponds to the stop packet 142 shown in FIG. 14). A stop packet 212 as shown in FIG. 21A is generated (step S93). Next, the stop packet processing unit 153 acquires the MAC address of the other server set in the server information table TBL2 as the destination MAC address 211 (see FIG. 21A) of the generated stop packet 212. The stop packet processing unit 153 sends the destination MAC address 211 and the stop packet 212 to the data link output unit 152 (step S94).

  Next, the stop packet processing unit 153 generates a G-ARP (G-ARP packet) as shown in FIG. 21B (corresponding to the G-ARP 166 shown in FIG. 18) based on the information in the server information table TBL2. ) 214 is generated (step S95). The stop packet processing unit 153 acquires the broadcast address (FF: FF: FF: FF: FF: FF) as the destination MAC address 213 illustrated in FIG. The stop packet processing unit 153 sends the destination MAC address 213 and the generated G-ARP packet 214 to the data link output unit 152 (step S96), and ends the processing A4.

  Next, the operation of the IP input unit 154 will be described with reference to the flowchart of FIG. 22 and the data format of FIG.

  An IP packet 231 or an ARP packet 232 as shown in FIG. 23A or FIG. 23B is sent to the IP input unit 154 by the data link input unit 151 or the stop packet processing unit 153. When receiving the IP packet 231 or the ARP packet 232, the IP input unit 154 converts the received packet as an IP packet 233 or an ARP packet 234 as shown in FIG. 23C or FIG. 16 (step S101), and the process is terminated.

  Next, the operation of the IP output unit 155 will be described with reference to the flowchart of FIG. 24 and the data format of FIG.

  The IP output unit 155 receives the IP packet 252 and the destination MAC address 251 as shown in FIG. 25A or the ARP packet 254 and the destination MAC as shown in FIG. 25B by the TCP / IP module 16. Address 253 is sent. When receiving such data (destination MAC address 251 + IP packet 252 or destination MAC address 253 + ARP packet 254), the IP output unit 155 checks the packet passage state table TBL1 (step S111).

  If the packet passage state table TBL1 indicates that packet passage is impossible (step S112), the IP output unit 155 discards the input data (step S113). On the other hand, when the packet passage state table TBL1 indicates that the packet can be passed (step S112), the IP output unit 155 sends the input data to the data link output unit 152 as it is (step S114), and the process is terminated. Here, when the input data is the destination MAC address 251 + IP packet 252 shown in FIG. 25A, the input data is sent to the data link output unit 152 as the destination MAC address 255 + IP packet 256 shown in FIG. . Similarly, when the input data is the destination MAC address 253 + ARP packet 254 shown in FIG. 25B, the input data is sent to the data link output unit 152 as the destination MAC address 257 + ARP packet 258 shown in FIG. .

[Second Embodiment]
FIG. 26 is a diagram for explaining the basic configuration and operation principle of a client server system including a computer system according to the second embodiment of the present invention. In FIG. 26, elements similar to those in FIG.

  The system of FIG. 26 applied in the second embodiment is different from the system of FIG. 1 applied in the first embodiment, and a shared disk (storage) shared by the servers 1-1 and 1-2 ( Shared disk device) 4. In such a system, it is assumed that the use of the IP address cannot be stopped due to, for example, the main server 1-1 being slowed down in a high load state. In this state, since the service process 11 is operating on the primary server 1-1 that is slowing down, the primary server 1-1 and the standby server 1-2 share the shared disk 4. Then, the data area of the shared disk 4 may be destroyed.

  Considering this point, the second embodiment uses an emergency stop function provided by the OS when the intermediate driver 15 of the primary server 1-1 receives a stop packet from the standby server 1-2. The intermediate driver 15 immediately stops the entire primary server 1-1.

Next, the operation of the system of FIG. 26 will be described with a focus on differences from the system of FIG.
First, it is assumed that a failure has occurred on the primary server 1-1 (step S121). When the cluster control unit 12 of the standby server 1-2 detects a failure of the primary server 1-1 (step S122), the cluster control unit 12 stops the use of the IP address of the primary server 1-1. The intermediate driver 15 is requested to stop processing (step S123). The intermediate driver 15 of the standby server 1-2 generates a stop packet and transmits the stop packet to the main server 1-1 (step S124). The steps so far are the same as steps S1 to S4 in FIG.

  When the intermediate driver 15 of the primary server 1-1 receives the stop packet, it immediately forcibly stops the entire primary server 1-1 using the emergency stop function provided by the OS (step S125). In this respect, the system of FIG. 26 is different from the system of FIG. 1 that blocks all network processing on the primary server 1-1.

  As described above, in the present embodiment, when the intermediate driver 15 of the primary server 1-1 receives the stop packet, the entire primary server 1-1 is immediately stopped to provide the primary server 1-1. It is possible to prevent the shared disk 4 (shared data area) from being destroyed when the standby server 1-2 takes over the service.

  The intermediate driver 15 of the standby server 1-2 sends a stop packet to the primary server 1-1 and then checks the IP address (specific IP address) contention in the same manner as in step S6 of FIG. Then, G-ARP (G-ARP packet) is broadcast and transmitted to the network 3 (step S126).

  On the other hand, after requesting the stop process to the intermediate driver 15 of the server 1-2, the cluster control unit 12 of the standby server 1-2 sets the IP address as in step S8 of FIG. This is performed for the TCP / IP module 16 of 1-2 (step S127).

In the second embodiment, the configuration of the intermediate driver 15 is the same as that of the first embodiment. For this reason, FIG. 4 is used as appropriate in the following description.
In the second embodiment, the functions (operations) of the data link input unit 151 and the IP output unit 155 in the intermediate driver 15 are partially different from those in the first embodiment. On the other hand, the functions (operations) of the data link output unit 152, the stop packet processing unit 153, and the IP input unit 154 in the intermediate driver 15 are the same as those in the first embodiment. Therefore, operations of the data link input unit 151 and the IP output unit 155 in the second embodiment will be sequentially described.

First, the operation of the data link input unit 151 will be described with reference to the flowchart of FIG. 27 and the data format of FIG. 28, focusing on the differences from the first embodiment.
The data link input unit 151 receives a packet (ARP packet) 281, a packet (stop packet) 282 or a packet (data) as shown in FIG. 28A, FIG. 28B or FIG. When the link layer packet) 283 is passed as input data, the packet passing state table TBL1 is checked (step S131).

  The data link input unit 151 discards the input data (input packet) when the packet passage state table TBL1 indicates that the packet cannot be passed (step S132). Unlike the first embodiment, the data link input unit 151 stops the entire local server using the emergency stop function provided by the OS (step S134).

  On the other hand, when the packet passage state table TBL1 indicates that packet passage is possible (step S132), the data link input unit 151 performs the same operation as in the first embodiment. That is, the data link input unit 151 executes processing (steps S135 to S138) corresponding to steps S34 to S37 in the flowchart of FIG. Thus, when the input data is the ARP packet 281 shown in FIG. 28A or the stop packet 282 shown in FIG. 28B, the ARP packet 281 or the stop packet 282 is directly shown in FIG. 28C. The packet is sent to the stop packet processing unit 153 as the ARP packet 284 or the stop packet 285 shown in FIG. When the input data is a packet (data link layer packet) 283 as shown in FIG. 28C other than the ARP packet and the stop packet, the IP packet excluding the data link layer header is shown in FIG. Is sent to the IP input unit 154 as the IP packet 286 shown in FIG.

  Next, the operation of the IP output unit 155 will be described with reference to the flowchart of FIG. 29, focusing on the differences from the first embodiment. Here, the data format of FIG. 25 is used.

  Data (destination MAC address 251 + IP packet 252 or destination MAC address 253 + ARP packet 254) as shown in FIG. 25 (a) or FIG. 25 (b) is sent to the IP output unit 155 by the TCP / IP module 16. When receiving such data, the IP output unit 155 checks the packet passage state table TBL1 (step S141).

  If the packet passage state table TBL1 indicates that packet passage is impossible (step S142), the IP output unit 155 discards the input data (step S143). Then, unlike the first embodiment, the IP output unit 155 stops the entire local server using the emergency stop function provided by the OS (step S144).

  On the other hand, when the packet passage state table TBL1 indicates that the packet can be passed (step S142), the IP output unit 155 performs the same operation as in the first embodiment. That is, the IP output unit 155 executes a process (step S145) corresponding to step S114 in the flowchart of FIG.

[Third Embodiment]
FIG. 30 is a diagram for explaining the basic configuration and operation principle of a client server system including a computer system according to the third embodiment of the present invention. In FIG. 30, the same elements as those in FIG.

  The basic configuration of the system of FIG. 30 applied in the third embodiment is the same as the system of FIG. 1 applied in the first embodiment. The feature of the third embodiment is that G-- sent by the TCP / IP module 16 (the TCP / IP module 16 in the standby server 1-2) instead of the stop packet applied in the first embodiment. The ARP is used as a stop packet.

Next, the operation of the system of FIG. 30 will be described focusing on the differences from the system of FIG.
First, it is assumed that a failure has occurred on the primary server 1-1 (step S151). When the cluster control unit 12 of the standby server 1-2 detects a failure of the primary server 1-1 (step S152), it sets an IP address for the TCP / IP module 16 of the server 1-2 (step S152). Step S153).

  When the IP address is set, the TCP / IP module 16 of the standby server 1-2 confirms whether the primary server 1-1 has stopped using the IP address, so that the G-ARP (G- ARP packet) is transmitted by broadcast to the network 3 (step S154).

  When the intermediate driver 15 of the primary server 1-1 receives G-ARP from the standby server 1-2, it is assumed that the use of the IP address has been stopped, and the standby server 1-2 in the first embodiment is requested. As in the case where the stop packet is received from, all network processing on the primary server 1-1 is blocked thereafter (step S155). That is, the intermediate driver 15 of the primary server 1-1 discards all packets input to the intermediate driver 15 to be passed from the network driver 14 to the TCP / IP module 16. Similarly, the intermediate driver 15 discards all packets input to the intermediate driver 15 to be passed from the TCP / IP module 16 to the network driver 14. Therefore, the G-ARP transmitted from the standby server 1-2 is not passed to the TCP / IP module 16 of the primary server 1-1, and the ARP response is sent from the primary server 1-1 to the standby server 1-2. Will not be replied. Thereby, the standby server 1-2 can prevent an IP address conflict state when taking over the IP address from the primary server 1-1.

Next, periodic transmission of G-ARP by the intermediate driver 15 of the standby server 1-2 will be described with reference to the sequence chart of FIG.
First, it is assumed that G-ARP is broadcasted from the TCP / IP module 16 of the standby server 1-2 (step S161). If the intermediate driver 15 of the primary server 1-1 cannot receive the G-ARP, the primary server 1-1 cannot stop using the IP address. However, an ARP response is not returned from the primary server 1-1 to the standby server 1-2.

  In this case, the intermediate driver 15 of the standby server 1-2 cannot detect that the primary server 1-1 is operating. Therefore, the standby server 1-2 intermediate driver 15 transmits the G-ARP by broadcast, for example, at regular intervals (that is, periodically) independently of the TCP / IP module 16 (steps S162, S163,...).

  As described above, as a result of the failure in the primary server 1-1, the G-ARP transmitted from the standby server 1-2 may or may not be received by the intermediate driver 15 of the primary server 1-1. However, the standby server 1-2 can take over the IP address from the main server 1-1 by avoiding the IP address conflict state.

In the third embodiment, the configuration of the intermediate driver 15 is the same as that of the first embodiment. For this reason, FIG. 4 is used as appropriate in the following description.
In the third embodiment, the functions (operations) of the data link input unit 151 and the stop packet processing unit 153 in the intermediate driver 15 are partially different from those in the first embodiment. On the other hand, the functions of the data link output unit 152, the IP input unit 154, and the IP output unit 155 in the intermediate driver 15 are the same as those in the first embodiment. Therefore, operations of the data link input unit 151 and the stop packet processing unit 153 according to the third embodiment will be sequentially described.

First, the operation of the data link input unit 151 will be described with reference to the flowchart of FIG. 32 and the data format of FIG. 33, focusing on the differences from the first embodiment.
When the data link input unit 151 receives a packet (ARP packet) 331 or a packet (data link layer packet) 332 as shown in FIG. 33A or 33B from the network driver 14 as input data Then, the packet passing state table TBL1 is checked (step S171).

  If the packet passage state table TBL1 indicates that the packet can be passed (step S172), the data link input unit 151 determines whether the input data is an ARP packet (step S174).

  If the input data is not an ARP packet, the data link input unit 151 unconditionally goes to step S37 in FIG. 5 without determining whether or not the input data is a stop packet unlike the first embodiment. Corresponding processing (step S176) is performed. That is, when the input data is a packet (data link layer packet) 332 as shown in FIG. 33B other than the ARP packet, the data link input unit 151 starts from the input data (data link layer packet 332) to the data link layer header. The part excluding, that is, the IP packet, is sent to the IP input unit 154 as the IP packet 334 shown in FIG. When the input data is an ARP packet 331 as shown in FIG. 33A, the data link input unit 151 sends the ARP packet 331 as it is to the stop packet processing unit 153 as the ARP packet 333 shown in FIG. Send (step S175).

Next, the operation of the stop packet processing unit 153 in the intermediate driver 15 will be described.
The stop packet processing unit 153 performs the following three processes <Process B1>
Processing to initialize packet passing state table TBL1, server information table TBL2, and other server state table TBL3 in response to a request from cluster control unit 12 <Processing B2>
Process when ARP packet is received from data link input unit 151 <Process B3>
G-ARP periodic transmission processing.

Execute.
Since the process B1 is the same as the process A1 in the first embodiment, a description thereof will be omitted.

Next, the process B2 executed by the stop packet processing unit 153 will be described with reference to the flowchart of FIG. 34 and the data formats of FIGS.
When the stop packet processing unit 153 receives the ARP packet 351 as illustrated in FIG. 35A from the data link input unit 151, as illustrated in FIG. 36, the stop packet processing unit 153 includes the source MAC address included in the ARP packet 351. Is the MAC address of the other server set in the server information table TBL2 (step S181).

  If the source MAC address is the MAC address of the other server, the stop packet processing unit 153 determines whether the source IP address and the target IP address included in the ARP packet are the same (step S182). ). If the source IP address and the target IP address are the same, the stop packet processing unit 153 determines whether the target IP address (= source IP address) is the takeover IP address, as shown in FIG. This is determined by comparing with the takeover IP address set in TBL2 (step S183). If the target IP address (= source IP address) is a takeover IP address, the stop packet processing unit 153 proceeds to step S184.

  Thus, the transmission source MAC address of the ARP packet 351 received from the data link input unit 151 is the MAC address of the other server, and the transmission source IP address and the target IP address of the ARP packet 351 are the takeover IP address. In this case, the stop packet processing unit 153 determines that the ARP packet 351 is a G-ARP (G-ARP packet) from another server, and proceeds to step S184. In step S184, the stop packet processing unit 153 changes the flag held in the packet passing state table TBL1 to a state indicating that the packet cannot pass (a state opposite to the state shown in FIG. 7), and ends the process B2. To do.

  On the other hand, the source MAC address of the ARP packet 351 is not the MAC address of the other server (step S181), or the source IP address and the target IP address of the ARP packet 351 are not the same (step S182), or the same. If it is not the takeover IP address (step S183), the stop packet processing unit 153 proceeds to step S185. In step S185, the stop packet processing unit 153 sends the portion (IP packet) obtained by removing the data link layer header from the ARP packet 351 to the IP input unit 154 as the ARP packet 352 shown in FIG. Exit.

  Next, the process B3 executed by the stop packet processing unit 153 will be described with reference to the flowchart of FIG. 37 and the data format of FIG. This process B3 is executed at regular transmission intervals set in the server information table TBL2.

  The stop packet processing unit 153 checks the status of the other server with reference to the other server status table TBL3 (step S191). If the other server is in a normal state (step S192), the stop packet processing unit 153 ends the process without doing anything.

  On the other hand, when the other system server is in a failure state (step S192), the stop packet processing unit 153 does not generate and transmit a stop packet as shown in FIG. 38, unlike the first embodiment. A G-ARP (G-ARP packet) 382 is generated (step S193). The G-ARP packet 382 is generated as follows using the information set in the server information table TBL2. First, as the source MAC address of the G-ARP packet 382, as shown in FIG. 38, the MAC address of the local server set in the server information table TBL2 is used. As the target MAC address of the G-ARP packet 382, 0 address is used as shown in FIG. As shown in FIG. 38, the takeover IP address set in the server information table TBL2 is used for the source IP address and the target IP address of the G-ARP packet 382.

  Next, the stop packet processing unit 153 acquires the broadcast address (FF: FF: FF: FF: FF: FF) as the destination MAC address 381 shown in FIG. The stop packet processing unit 153 sends the destination MAC address 381 and the generated G-ARP packet 381 to the data link output unit 152 (step S194), and ends the process.

[Fourth Embodiment]
FIG. 39 is a diagram for explaining the basic configuration and operation principle of a client server system including a computer system according to the fourth embodiment of the present invention. In FIG. 39, elements similar to those in FIG. 1 or FIG.

  The basic configuration of the system of FIG. 39 applied in the fourth embodiment is the same as the system of FIG. 26 applied in the second embodiment. A feature of the fourth embodiment is that G-ARP is used as a stop packet in the same manner as in the third embodiment, instead of the stop packet applied in the second embodiment.

Next, the operation of the system of FIG. 39 will be described focusing on the differences from the system of FIG.
First, it is assumed that a failure has occurred on the primary server 1-1 (step S201). When the cluster control unit 12 of the standby server 1-2 detects a failure of the primary server 1-1 (step S202), it sets an IP address for the TCP / IP module 16 of the server 1-2 (step S202). Step S203).

  When the IP address is set, the TCP / IP module 16 of the standby server 1-2 confirms whether the primary server 1-1 has stopped using the IP address, so that the G-ARP (G- ARP packet) is transmitted by broadcast to the network 3 (step S204).

  When receiving the G-ARP from the standby server 1-2, the intermediate driver 15 of the primary server 1-1 receives the stop packet from the standby server 1-2 in the second embodiment. The entire primary server 1-1 is immediately stopped using the emergency stop function provided by (step S205).

  As described above, in this embodiment, when the intermediate driver 15 of the primary server 1-1 receives the G-ARP, the entire primary server 1-1 is immediately stopped to provide the primary server 1-1. It is possible to prevent the shared disk 4 (shared data area) from being destroyed when the standby server 1-2 takes over the service that has been provided.

In the fourth embodiment, the configuration of the intermediate driver 15 is the same as that of the first embodiment. For this reason, FIG. 4 is used as appropriate in the following description.
In the fourth embodiment, the function (operation) of the data link input unit 151 in the intermediate driver 15 is partly different from the first to third embodiments. On the other hand, the function of the data link output unit 152 in the intermediate driver 15 is the same as that of the first (or third) embodiment, and the function of the stop packet processing unit 153 is the same as that of the third embodiment. The function of the IP input unit 154 in the intermediate driver 15 is the same as that of the first (or third) embodiment, and the function of the IP output unit 155 is the same as that of the third embodiment. Therefore, the operation of the data link input unit 151 of the fourth embodiment will be described with reference to the flowchart of FIG. 40 and the data format of FIG. 41, focusing on the differences from the second embodiment.

  When the data link input unit 151 receives a packet (ARP packet) 411 or a packet (data link layer packet) 412 as shown in FIG. 41A or 41B as input data from the network driver 14 Then, the packet passage state table TBL1 is checked (step S211).

  When the packet passage state table TBL1 indicates that the packet can pass (step S212), the data link input unit 151 determines whether the input data is an ARP packet (step S215).

  If the input data is not an ARP packet, the data link input unit 151 does not determine whether the input data is a stop packet unlike the second embodiment, and unconditionally goes to step S138 in FIG. Corresponding processing (step S217) is performed. That is, when the input data is a packet (data link layer packet) 412 as shown in FIG. 41B other than the ARP packet, the data link input unit 151 starts from the input data (data link layer packet 412) to the data link layer header. The part excluding, that is, the IP packet, is sent to the IP input unit 154 as the IP packet 414 shown in FIG.

  On the other hand, if the packet passage state table TBL1 indicates that packet passage is impossible (step S212), the data link input unit 151 discards the input data (input packet) and is provided by the OS, as in the second embodiment. The entire local server is stopped using the emergency stop function being performed (steps 213 and S214).

  Note that the present invention is not limited to the above-described embodiments as they are, and can be embodied by modifying constituent elements without departing from the scope of the invention in the implementation stage. Moreover, various inventions can be formed by appropriately combining a plurality of constituent elements disclosed in the above embodiments. For example, some components may be deleted from all the components shown in each embodiment.

The figure for demonstrating the basic composition and operation | movement principle of a client server system containing the computer system which concerns on the 1st Embodiment of this invention. The figure which shows the flow of a packet when the intermediate | middle driver of a main system server cannot receive the stop packet transmitted by the standby system server in the said 1st Embodiment. The sequence chart for demonstrating the stop packet and the periodic transmission of G-ARP in the said 1st Embodiment. The block diagram which shows the detailed structure of the intermediate | middle driver especially of the server shown by FIG. The flowchart which shows the operation | movement procedure of the data link input part in the 1st embodiment. The figure which shows the format of the data input into a data link input part in the same 1st Embodiment, and the data output from the said data link input part. The figure which shows the data structure example of a packet passage state table. The flowchart which shows the operation | movement procedure of the data link output part in the 1st embodiment. The figure which shows the format of the data input from the data link output part in the said 1st Embodiment, and the data output from the said data link output part. The flowchart which shows the procedure of the process A1 performed by the stop packet process part in the said 1st Embodiment. The figure which shows the data structure example of a server information table. The figure which shows the data structure example of another system server state table. The flowchart which shows the procedure of the process A2 performed by the stop packet process part in the said 1st Embodiment. The figure which shows the format of the data output to a data link output part by the stop packet process part in process A2. The flowchart which shows the procedure of the process A3 performed by the stop packet process part in the said 1st Embodiment. The figure which shows the format of the data input into a stop packet process part in the process A3, and the data output from the said stop packet process part. The figure which shows the format of an ARP response packet in association with a server information table. The figure which matches the format of a G-ARP packet with a server information table, and shows it. The figure which shows the format of a stop packet matched with a server information table. The flowchart which shows the procedure of the process A4 performed by the stop packet process part in the said 1st Embodiment. The figure which shows the format of the data output from a stop packet process part in process A4. The flowchart which shows the operation | movement procedure of the IP input part in the said 1st Embodiment. The figure which shows the format of the data input into an IP input part in the 1st Embodiment, and the data output from the said IP input part. The flowchart which shows the operation | movement procedure of the IP output part in the said 1st Embodiment. The figure which shows the format of the data input into the IP output part in the same 1st Embodiment, and the data output from the said IP output part. The figure for demonstrating the basic composition and operation | movement principle of a client server system containing the computer system which concerns on the 2nd Embodiment of this invention. The flowchart which shows the operation | movement procedure of the data link input part in the said 2nd Embodiment. The figure which shows the format of the data input into a data link input part in the same 2nd Embodiment, and the data output from the said data link input part. The flowchart which shows the operation | movement procedure of the IP output part in the said 2nd Embodiment. The figure for demonstrating the basic composition and operation | movement principle of a client server system containing the computer system which concerns on the 3rd Embodiment of this invention. The sequence chart for demonstrating the periodic transmission of G-ARP in the said 3rd Embodiment. The flowchart which shows the operation | movement procedure of the data link input part in the 3rd Embodiment. The figure which shows the format of the data input into the data link input part in the same 3rd Embodiment, and the data output from the said data link input part. The flowchart which shows the procedure of the process B2 performed by the stop packet process part in the said 3rd Embodiment. The figure which shows the format of the data input into a stop packet process part in the process B2, and the data output from the said stop packet process part. FIG. 36 is a diagram showing the format of the G-ARP packet shown in FIG. 35 in association with a server information table. The flowchart which shows the procedure of process B3 performed by the stop packet process part in the said 3rd Embodiment. The figure which matches the format of the G-ARP packet output from a stop packet process part in process B3 with a server information table. The figure for demonstrating the basic composition and operation | movement principle of a client server system containing the computer system which concerns on the 4th Embodiment of this invention. The flowchart which shows the operation | movement procedure of the data link input part in the same 4th Embodiment. The figure which shows the format of the data input into the data link input part in the 4th Embodiment, and the data output from the said data link input part.

Explanation of symbols

  1-1 ... server (primary server, first server computer), 1-2 ... server (standby server, second server computer), 1-i ... server, 2 ... client (client terminal), 3 ... Network, 4 ... Shared disk, 11 ... Service process, 12 ... Cluster control unit, 13 ... Network IF, 14 ... Network driver, 15 ... Intermediate driver, 16 ... TCP / IP module, 151 ... Data link input unit, 152 ... Data Link output unit, 153... Stop packet processing unit, 154... IP input unit, 155... IP output unit, TBL1... Packet passing state table, TBL2... Server information table, TBL3.

Claims (14)

The first server and the second server are interconnected by a network to which the TCP / IP protocol is applied, and one of the first and second servers functions as a main server via the network. When providing a service corresponding to a request given by a client to the client, the other server of the first and second servers functions as a standby server, and when a failure of the one server is detected, In the computer system that takes over the service provided by the one server and the specific IP address used in the service,
The first and second servers are
A network driver that performs data link layer processing in TCP / IP communication;
A TCP / IP module that performs processing of the TCP / IP layer in TCP / IP communication;
An intermediate driver located between the network driver and the TCP / IP module and controlling data transmission / reception between the network driver and the TCP / IP module;
The intermediate driver of the second server stops using the specific IP address when the second server functions as the standby server and detects a failure of the first server. Sending a stop packet for requesting the server to the first server via the network driver of the second server and the network;
The intermediate driver of the first server sets the first server to a specific state in which data transmission / reception between the network driver and the TCP / IP module is disabled when the stop packet is received. A computer system characterized by that. The intermediate driver of the second server includes a server information table storing a physical address of the first server;
The computer system according to claim 1, wherein the intermediate driver of the second server uses the physical address of the first server stored in the server information table as a destination physical address of the stop packet. .   When the intermediate driver of the first server receives the stop packet, the data to be transmitted / received between the network driver and the TCP / IP module is passed through the intermediate driver. By switching from the passable state that allows the data to be transmitted / received between the network driver and the TCP / IP module to the non-passable state that inhibits data from passing through the intermediate driver, the first 2. The computer system according to claim 1, wherein the server is set to the specific state. The intermediate driver of the second server transmits the stop packet, and then broadcasts a special address resolution protocol request for confirming the conflict of the specific IP address to the network,
When the intermediate driver of the first server receives the special address resolution protocol request in the passable state, the intermediate driver transmits the special address resolution protocol request to the TCP / IP module of the first server;
When the TCP / IP module of the first server receives the special address resolution protocol request, the TCP / IP module sends an address resolution protocol response warning the duplication of the specific IP address, which is a response to the special address resolution protocol request. Return to the second server,
When the intermediate driver of the second server receives the address resolution protocol response to the special address resolution protocol request, the intermediate driver transmits the stop packet again to the first server, and then the special address. The computer system according to claim 3, wherein a resolution protocol request is broadcast to the network. The intermediate driver of the first and second servers includes a server information table storing physical addresses of the first and second servers and the specific IP address,
The intermediate driver of the second server uses the physical address of the first server stored in the server information table as the destination physical address of the stop packet, and stores the second driver stored in the server information table. 5. The computer system according to claim 4, wherein a physical address of the server and the specific IP address are a source physical address and a target IP address of the special address resolution protocol request, respectively.   The intermediate driver of the second server broadcasts the first transmission operation for transmitting the stop packet to the first server and the special address resolution protocol request after the first transmission operation. 5. The computer system according to claim 4, wherein the operation including the second transmission operation is repeated. The intermediate driver of the first and second servers includes a server information table storing physical addresses of the first and second servers and the specific IP address,
The intermediate driver of the second server uses the physical address of the first server stored in the server information table as the destination physical address of the stop packet, and stores the second driver stored in the server information table. The computer system according to claim 6, wherein a physical address of the server and the specific IP address are a source physical address and a target IP address of the special address resolution protocol request, respectively.   When the intermediate driver of the first server receives the stop packet, the intermediate driver stops the first server by using a system emergency stop function provided by an operating system. The computer system according to claim 1, wherein the specific state is set. The first server and the second server are interconnected by a network to which the TCP / IP protocol is applied, and one of the first and second servers functions as a main server via the network. When providing a service corresponding to a request given by a client to the client, the other server of the first and second servers functions as a standby server, and when a failure of the one server is detected, In the computer system that takes over the service provided by the one server and the specific IP address used in the service,
The first and second servers are
A network driver that performs data link layer processing in TCP / IP communication;
A TCP / IP module that performs TCP / IP layer processing in TCP / IP communication;
An intermediate driver for controlling data transmitted and received between the network driver and the TCP / IP module;
The TCP / IP module of the second server confirms the contention of the specific IP address when the second server functions as the standby server and detects a failure of the first server. Broadcast a special address resolution protocol request to the network,
When the intermediate driver of the first server receives the special address resolution protocol request, the intermediate driver is requested to stop using the specific IP address used in the service provided by the first server. The computer system is characterized in that the first server is set in a specific state in which data transmission / reception between the network driver and the TCP / IP module is disabled.   When the intermediate driver of the first server receives the special address resolution protocol request, the status of the intermediate driver indicates that the data to be transmitted / received between the network driver and the TCP / IP module By switching from a passable state that allows passage through an intermediate driver to a nonpassable state that prevents data to be transmitted and received between the network driver and the TCP / IP module from passing through the intermediate driver. The computer system according to claim 9, wherein the first server is set to the specific state.   The intermediate driver of the second server transmits the special address resolution protocol request to the network by broadcast again after the TCP / IP module of the second server transmits the special address resolution protocol request. The computer system according to claim 10, wherein the operation is repeated.   When the intermediate driver of the first server receives the special address resolution protocol request, the intermediate driver stops the first server by using a system emergency stop function provided by an operating system. 10. The computer system according to claim 9, wherein one server is set in the specific state. A network driver connected to a network to which the TCP / IP protocol is applied and capable of functioning as either a primary server or a standby server, and performing data link layer processing in TCP / IP communication; and TCP A server computer including a TCP / IP module that performs TCP / IP layer processing in IP / IP communication, and functions as the main server in response to a request given from a client via the network When providing a service to the client and functioning as the standby server, the other server computer provides the service by detecting a failure of the other server computer connected to the network and functioning as the primary server Services and special features used in those services A program executed by a server computer to take over the IP address,
In the server computer,
The server computer detects that a failure has occurred in the other server computer while the server computer functions as the standby server and the other server computer functions as the primary server. A stop packet for requesting the other server computer to stop using the specific IP address is sent to the other computer via the network driver and the network;
In the state where the server computer functions as the primary server and the other server computer functions as the standby server, a stop packet for requesting the use stop of the specific IP address is sent to the other server. Setting the server computer to a specific state in which data transmission / reception between the network driver and the TCP / IP module is disabled when transmitted from the computer to the server computer and received by the server computer; Determining whether the server computer is set to the specific state when data transmission / reception between the network driver and the TCP / IP module is required;
When the server computer is set in the specific state, a program for executing a step of suppressing data transmission / reception between the network driver and the TCP / IP module. A network driver connected to a network to which the TCP / IP protocol is applied and capable of functioning as either a primary server or a standby server, and performing data link layer processing in TCP / IP communication; and TCP A server computer including a TCP / IP module that performs TCP / IP layer processing in IP / IP communication, and functions as the main server in response to a request given from a client via the network When providing a service to the client and functioning as the standby server, the other server computer provides the service by detecting a failure of the other server computer connected to the network and functioning as the primary server Services and special features used in those services A program executed by a server computer to take over the IP address,
In the server computer,
The identification that the other server computer takes over as a result of a failure in the server computer in a state where the server computer functions as the primary server and the other server computer functions as the standby server When the server computer receives a special address resolution protocol request broadcast to the network by the other server computer in order to confirm an IP address conflict, the use of the specific IP address is requested. For example, setting the server computer to a specific state in which data transmission / reception between the network driver and the TCP / IP module is disabled;
Determining whether the server computer is set to the specific state when data transmission / reception between the network driver and the TCP / IP module is required;
When the server computer is set in the specific state, a program for executing a step of suppressing data transmission / reception between the network driver and the TCP / IP module.

Images