JP2005056347A - Server function takeover method and server function takeover program - Google Patents

Abstract

An object of the present invention is to provide a server function takeover method for quickly taking over the function of a host in which a failure has occurred in order to avoid a system failure when a failure occurs in a server.
In a network environment in which an Internet data center, a system administrator, and a general user are connected via the Internet, each server activates a program for executing the function of the server. The startup process, the server load information acquisition process in which each server acquires the load information of its own server, the failed server detection process for detecting the failed server, and the function of the failed server can be taken over. A takeover server detection step of detecting a server and a server function takeover step of taking over the function of the server detected in the takeover server detection step are sequentially executed.
[Selection] Figure 1

Description

  In an e-mail delivery system that consists of multiple servers via the Internet, when a failure occurs in some servers, the server that takes over the functions of the failed servers in other servers and prevents the system from going down The present invention relates to a function takeover method and a server function takeover program for causing a computer to execute the method.

  Generally, in an e-mail delivery system via the Internet, a system is configured by connecting a plurality of servers in order to improve e-mail delivery efficiency. In particular, measures such as redundancy of hardware and server programs are taken to cope with hardware failures. For example, in the case of a system configuration that includes a plurality of e-mail distribution servers and a management server that manages them, when a failure occurs in one of the e-mail distribution servers, another e-mail distribution server takes over unsent mail To do. This takeover may be performed automatically, or only unsent e-mails may be dealt with manually (see Patent Document 1 below). On the other hand, as a countermeasure for the failure of the management server, prepare two systems with the same configuration as the management server itself, operate one (main system / production system), and perform the same operation on the other (standby system / standby system). While waiting (so-called hot standby), or operating one side (main system / production system) and leaving the other side (standby system / standby system) not operating (so-called cold standby) ) Is taken.

JP 2001-43105 A

  In the technology described in Patent Document 1, when a failure of any server is detected in a system in which a plurality of servers are connected via a network, the function in which the failure has occurred is assigned a predefined priority. Based on the new server. For this reason, when the load of the server as the takeover destination is high, there is a problem that the takeover process itself of the server function that should be required to be quick is delayed.

  In addition, in the management server failure handling method using hot standby as described above, hardware redundancy is required, and thus an increase in cost is inevitable. In addition, the man-hours required for setting / testing become large. On the other hand, with the management method for failure of the management server by cold standby, the work man-hours until recovery are increased and it takes time. For this reason, there arises a problem that the administrator cannot quickly manage the system (status grasp, system control).

  The present invention has been made in view of the above, and a server function takeover method for quickly taking over the function of a server in which a failure has occurred to another server in order to avoid a system failure when a failure occurs in the server. The purpose is to provide. In particular, an object of the present invention is to provide a server function takeover method and a server function takeover program that can take over the function of a server in which a failure has occurred quickly without making the hardware configuration of the server redundant.

  In order to solve the above-described problems and achieve the object, the server function takeover method of the present invention is an electronic mail distribution system using a plurality of servers and using the Internet, and the control information of each server in each server Referring to the information table in which the server is defined, a program starting step for starting a program for executing the function of the own server, a server load information acquiring step for acquiring load information of the own server in each server, and the information table The failure server detection process for detecting the failed server and the weight table in which the weight for each information function and the server function is defined are referred to and the function of the failed server is taken over. A takeover server detecting step of detecting a server capable of performing the above operation, and referring to the information table, A server function takeover process take over functions issued a server, characterized in that it contains.

  The server function takeover method and server function takeover program according to the present invention include a server in which a failure occurs when a failure occurs in any of the servers in an e-mail distribution system in which a plurality of servers are connected via a network. It is possible to take over the functions that the server has taken over to a server with a low load, so that it is possible to complete the takeover of functions of a server in which a failure has occurred quickly without causing a system down.

  Embodiments of a server function takeover method and a server function takeover program according to the present invention will be described below in detail with reference to the drawings. In addition, this invention is not limited by this embodiment.

  FIG. 1 is a schematic diagram of a network configuration for realizing the server function takeover method of the present invention. The method of the present invention is realized in a network environment in which an Internet data center (IDC) 100, a system administrator 110, and a general user 120 are connected via the Internet 130. The Internet data center 100 includes a management server, a plurality of electronic mail distribution servers, a file server, a firewall, and a load balancer.

  FIG. 2 is a diagram showing an example of the configuration of an Internet data center for realizing the server function takeover method of the present invention. The Internet data center 100 includes a router 201, a firewall 202, a load balancer 203, switching hubs (SW) 204 and 205, a management server 206, e-mail distribution servers 207, 208, 209 and 210, and a file server 211. The

  The router 201 performs routing of a communication path in the Internet data center 100. The firewall 202 controls the flow of traffic on the communication path in the Internet data center 100 to ensure the security of the Internet data center 100. The load balancer 203 reduces the load on the management server 206 and the e-mail distribution servers 207, 208, 209, and 210. The administrator's browser is connected to the management server 206 provided with a service by the load balancer 203 via the Internet 130. The switching hub 204 connects the load balancer 203 to the management server 206 and the e-mail distribution servers 207, 208, 209, and 210. The switching hub 205 connects the management server 206 and the electronic mail distribution servers 207, 208, 209, 210 and the file server 211.

  The management server 206 stores the destination address of the email and the content of the email, and sequentially sends the destination address of the email and the content of the email to the email delivery servers 207, 208, 209, and 210 as necessary. Send. The electronic mail distribution servers 207, 208, 209, and 210 transmit the content of the electronic mail sent from the management server 206 to the corresponding mail address. The file server 211 stores a shared file for the management server 206 and the e-mail distribution servers 207, 208, 209, and 210 to execute a predetermined service. The number of management servers 206 and electronic mail distribution servers 207, 208, 209, and 210 can be further increased.

  FIG. 3 is a chart showing an example of an information table during normal operation in the server function takeover method of the present invention. FIG. 4 is a chart showing an example of the service weight table at this time. These information table and service weight table are set in advance by the system administrator 110, and the method of the present invention is executed based on information defined in each area of the information table and service weight table. These information table and service weight table are stored in the storage means of the file server 211 or the management server 206 and the e-mail distribution servers 207, 208, 209, and 210.

  The following information is described in each area of the information table. In the “host name” area, the name of each server defined in advance is described. Here, the management server 206 is set to host1. In addition, the e-mail distribution servers 207 to 210 are set as hosts 2 to 5, respectively. In the “service in charge” area, the type of service assigned to the host is described. When another service is newly assigned during execution of a certain service, the type of the service is also described. The "takeover priority" area is a reference for determining which host will take over the service (function) of a failed host when a failed host (hereinafter referred to as a failed host) is detected. A numerical value is described.

  The larger this value is, the higher the priority of service takeover is, and the value 0 indicates that the host is stopped or has failed. Therefore, a host having a “takeover priority” area value larger than that of any other host will take over the service of the failed host. In this case, the value of the “takeover priority” area of the host that has taken over the service is decremented by a certain value. The initial value of the “takeover priority” area is set based on the specifications of the host, and a higher value is set for a host with higher specifications. If the load of the host that has taken over the service becomes very high, it will be adjusted up or down as needed. In the “check time” area, the current time is recorded at regular intervals. While referring to the time, it is possible to detect the failure of the local host by monitoring the statuses of the local host and other hosts (details will be described later). In each area of the service weight table, the type of service provided by each host and the weight value corresponding to the service are described.

  Next, an execution procedure of the server function takeover method of the present invention will be described. FIG. 5 is a flowchart showing the execution procedure of this server function takeover method. First, each host refers to the “service in charge” area of the information table created by the system administrator 110 corresponding to the host, and starts a service program for the host (step S501). In the present invention, it is assumed that each host executes a service as the management server 206 and a service as an e-mail distribution server, and a program for executing these services is stored in the storage means of each host in advance. Stored. Further, a program for that purpose may be stored in the storage means of the file server 211 and downloaded from the file server 211 when necessary. Next, each host writes the current time (the time when the program was started in step S501) in the “check time” area of the information table corresponding to the host (step S502).

  Subsequently, each host detects its current CPU load (step S503). Here, it is detected whether the CPU idle is 20% or less. When the CPU idle is 20% or less (step S503: Yes), the process proceeds to step S504. If the CPU idle is not less than 20% (step S503: No), the process proceeds to step S506.

  If the CPU idle is 20% or less in step S503 (step S503: Yes), each host detects whether the “takeover priority” area of the information table corresponding to the host is 2 or more (step S503). S504). If the "takeover priority" area is 2 or more (step S504: Yes), the information table is updated by subtracting 1 from the "takeover priority" area corresponding to the own host (step S505), and the process proceeds to step S509. On the other hand, if the “takeover priority” area is not 2 or more (step S504: No), the process proceeds to step S509.

  If the CPU idle is not less than 20% in step S503, each host detects the current CPU load of its own host again (step S506). Here, it is detected whether or not the CPU idle is 50% or more. If the CPU idle is 50% or more (step S506: Yes), the process proceeds to step S507. If the CPU idle is not 50% or more (step S506: No), the process proceeds to step S509.

  If the CPU idle is 50% or more in step S506 (step S506: Yes), each host detects whether the “takeover priority” area of the information table corresponding to the host is 1 or more and less than 10. (Step S507). If the "takeover priority" area is 1 or more and less than 10 (step S507: Yes), the information table is updated by adding 1 to the "takeover priority" area of the information table corresponding to the own host (step S508), The process proceeds to step S509. On the other hand, if the “takeover priority” area is not 1 or more and less than 10 (step S507: No), the process proceeds to step S509 as it is.

  Next, each host detects whether or not 20 seconds or more have elapsed since the last time (in the process of step S502) written in the “check time” area of the information table corresponding to the host (step S509). If it is detected that 20 seconds or more have elapsed since the previous writing time (step S509: Yes), the process proceeds to step S510. If it is not detected that 20 seconds or more have passed since the previous writing time (step S509: No), the process returns to step S503 again to continue the processing.

  If it is detected in step S509 that 20 seconds or more have elapsed from the previous write time (step S509: Yes), it is detected whether there is a failed host. Specifically, compare the "check time" area of the local host with the "check time" area of the other host whose "takeover priority" area has a value of 1 or more, and if the difference is ± 30 seconds or more Whether or not is detected (step S510). In this step, it is possible to detect whether or not a failure has occurred in the own host by comparing each host with another host. When the difference is ± 30 seconds or more (step S510: Yes), it is detected that a failure has occurred in the local host, and the process proceeds to step S511. When the difference is not more than ± 30 seconds (step S510: No), the process returns to step S503 again and the processing is continued.

  Subsequently, except for the failed host, each host detects whether or not the value of the “takeover priority” area of the information table corresponding to the own host is the maximum compared to the values of other hosts (step S511). If the value of the “takeover priority” area of the own host is the maximum (step S511: Yes), the process proceeds to step S512. When the value of the “takeover priority” area of the own host is not the maximum (step S511: No), the value of the “takeover priority” area of the own host is the same and the same as that of the other host, and It is detected whether or not the own host is described in the highest column of the information table except for the failed host (step S513).

  When the value of the “takeover priority” area of the local host is the same as that of the other host and the maximum, and the local host is described in the topmost column of the information table except for the failed host (step S513: Yes) Advances to step S512. When the value of the “takeover priority” area of the local host is the same as that of the other host and is not the maximum, or when the local host is not described in the highest column of the information table except for the failed host (step S513: No) Returns to step S503 again and continues processing. In the processing of step S511 and step S513, a host candidate (hereinafter referred to as a candidate host) that takes over the service that was handled by the failed host is selected.

  The candidate host refers to the service weight table (see FIG. 4) and obtains a weight value corresponding to the “service in charge” of the faulty host (step S512). Subsequently, the candidate host sequentially evaluates the value of the “takeover priority” area of the own host and the weight value of the failed host (step S514). Specifically, the value of (the “takeover priority” value of the own host−the weight value of the failed host) is calculated.

  Next, the candidate host determines whether there is a service that was handled by the failure host that can be taken over by the server (step S515). Specifically, it is performed by detecting whether or not the evaluation result (calculated value) in step S514 is 0 or more. If there is a service that is handled by the failed host that can be taken over by the local server (step S515: Yes), the process proceeds to step S516. If there is no service that was handled by the failed host that can be taken over by the local server (step S515: No), the process returns to step S503 again to continue the processing.

  If the candidate host can take over the service in charge of the failed host, the value of the “takeover priority” area of the information table corresponding to the failed host is set to 0 (step S516). Next, the host that takes over the service that was handled by the failed host refers to the information table and starts the “service in charge” program that was handled by the failed host (step S517).

  Next, the host that has taken over the service updates the information table by subtracting the weight value of the newly activated service from the value in the “takeover priority” area of the information table corresponding to the host (step S518). Subsequently, the host that has taken over the service additionally registers the newly started responsible service in the “responsible service” area of the information table corresponding to the host (step S519).

  In the server function takeover method of the present invention, when a failure occurs in any of the servers, the services (functions) that the failure server has handled are immediately transferred to the server with a low load. Since it is possible to take over, it is possible to complete the takeover of the server function quickly without causing a system down. The host that has taken over the service of the failed host will provide the newly taken over service together with the service that has been provided by the local host. For example, when the electronic mail distribution server takes over the service of the management server, the service as the original electronic mail distribution server and the service as the management server are executed together. Further, it is not necessary to make the hardware configuration redundant in order to take over the server function, and there is no risk of cost.

  Here, for example, it is assumed that a failure has occurred in the host 1 executing the service as the management server. In this case, since a failure has occurred in host1, the “check time” area of the information table corresponding to host1 is not updated. Therefore, the other hosts determine that a failure has occurred in host 1 and each compares the value of the current “takeover priority” area. As a result, the service of host1 takes over host2, which has the highest value in the "takeover priority" area except for host1 (see FIG. 3). Then, host2 activates the function of the management server (Web server program such as Apache) in its own host, and adds the service name (management server) to the “service in charge” area. Further, the value in the “takeover priority” area is decremented by a certain value. That is, the weight value 1 corresponding to the newly added management server service is reduced. FIG. 6 shows the information table updated in this case. After this, if the load of host2 further increases, the value is further decremented by a certain value. The next time a faulty host is detected, a more generous (low load) host will take over the service.

  Next, a processing procedure when the host in which the failure has occurred is recovered will be described. FIG. 7 is a flowchart showing a processing procedure in this case. Here, it is assumed that a failure occurred in the management server is dealt with. First, the restored host initializes (NULL clear) the “service in charge” area of the information table corresponding to the host (step S701). Next, the restored host writes the current time (recovered time) in the “check time” area of the information table corresponding to the own host (step S702). Subsequently, the restored host writes 1 in the “takeover priority” area of the information table corresponding to the own host (step S703). Then, the procedure of the flowchart shown in FIG. 5 is followed (step S704).

  The server function takeover method described in this embodiment can be realized by executing a program prepared in advance on a computer such as a personal computer or a workstation. This program is recorded on a computer-readable recording medium such as a hard disk, a flexible disk, a CD-ROM, an MO, and a DVD, and is executed by being read from the recording medium by the computer. The program may be a transmission medium that can be distributed via a network such as the Internet.

(Supplementary note 1) An electronic mail distribution system including a plurality of servers and using the Internet,
A program starting step of starting a program for executing the function of the own server with reference to an information table in which control information of each server is defined in each server,
A server load information acquisition step of acquiring load information of the own server in each server;
Referring to the information table, a faulty server detection step of detecting a faulty server;
A takeover server detecting step of detecting a server capable of taking over the function of a server in which a failure has occurred with reference to the information table and a weight table in which a weight for each server function is defined;
With reference to the information table, a server function takeover step of taking over the function to the server detected in the takeover server detection step;
A server function takeover method characterized by comprising:

(Additional remark 2) The said program starting process starts the program for making each server perform the function as a system management server or an electronic mail delivery server, The server function takeover of Additional remark 1 characterized by the above-mentioned Method.

(Additional remark 3) The program for performing the function as a management server or an electronic mail delivery server of the said system is stored in the memory | storage means in each said server, or the other apparatus in the said system, It is characterized by the above-mentioned. The server function takeover method according to attachment 2.

(Supplementary Note 4) The information table includes at least a name area in which the name of each server is described, a function area in which the function of each server is described, and a time area in which the time at which each server confirms the status of the server is described. The server function takeover according to any one of appendices 1 to 3, characterized in that a takeover priority area is provided in which a numerical value indicating a priority in the case of taking over another server function by the local server is described. Method.

(Additional remark 5) The said information table is produced by the system administrator, and is stored in the memory | storage means in each said server, or the other apparatus in the said system, Any one of Additional remark 1-4 characterized by the above-mentioned. The server function takeover method described in 1.

(Supplementary note 6) The weight table is created by a system administrator and stored in storage means in each server or other device in the system. The server function takeover method described in 1.

(Supplementary note 7) The server function takeover method according to any one of supplementary notes 1 to 6, wherein the server load information acquisition step detects a load of the CPU of the server itself.

(Supplementary note 8) The failure server detection step includes the update time of the time area corresponding to the own server and the takeover after a predetermined time has elapsed since the time described in the time area of the information table serving as a reference of the own server. Compared to the update time of the time area corresponding to another server whose numerical value greater than or equal to a predetermined value is described in the priority area, and if there is an opening that exceeds the predetermined value, the local server has failed The server function takeover method according to any one of appendices 1 to 7, wherein the server function is detected.

(Supplementary Note 9) The takeover server detecting step includes a takeover priority maximum value server detecting step of detecting a server having a maximum takeover priority area value of the information table except for a server in which a failure has occurred. The server function takeover method according to any one of Supplementary notes 1 to 8, which is characterized by the following.

(Additional remark 10) In the said takeover priority maximum value server detection process, when a some server is detected, the server described in the highest column of the said information table is given priority excluding the server which failed. The server function takeover method according to appendix 9, further comprising a priority server detection step of detecting.

(Additional remark 11) Further, the takeover server detection step refers to the weight table, and determines whether or not the server detected in the takeover priority maximum value server detection step can take over the function. The server function takeover method according to supplementary note 9, including a process.

(Supplementary Note 12) Further, the takeover server detection step includes a function takeover availability determination step of referring to the weight table and determining whether or not the server detected in the priority server detection step can take over the function. The server function takeover method according to attachment 10, wherein the server function is taken over.

(Supplementary note 13) In the server function takeover step, the name of the function taken over in the function area of the information table of the server that took over the function of the server in which the fault has occurred is described together with the current function name, and the fault The server function takeover method according to any one of appendices 1 to 12, wherein a program for executing a function taken over by a server taking over the function of the generated server is started.

(Supplementary note 14) The server according to any one of supplementary notes 1 to 13, further comprising a recovery step of starting the restored server and incorporating it into the system when the failed server is restored. Function takeover method.

(Supplementary Note 15) A server function takeover program for a plurality of servers used in an electronic mail delivery system using the Internet,
A program starting step of starting a program for executing the function of the own server with reference to an information table in which control information of each server is defined in each server,
A server load information acquisition step for acquiring the load information of the own server in each server;
Referring to the information table, a faulty server detection step of detecting a faulty server;
A takeover server detection step of detecting a server capable of taking over the function of the server in which the failure has occurred with reference to the information table and a weight table in which a weight for each server function is defined;
With reference to the information table, a server function takeover step of taking over a function to the server detected in the takeover server detection step;
Server function takeover program, which causes a computer to execute

  As described above, the server function takeover method according to the present invention is useful for an electronic mail delivery system using the Internet, and particularly for avoiding a system down when a failure occurs in a server installed in the Internet data center. Are suitable.

It is the schematic of the network structure for implement | achieving the server function takeover method of this invention. It is a figure which shows an example of a structure of the internet data center for implement | achieving the server function takeover method of this invention. It is a chart which shows an example of the information table at the time of normal operation in the server function takeover method of the present invention. It is a graph which shows an example of the service weight table in the server function takeover method of this invention. It is a flowchart which shows the execution procedure of the server function takeover method of this invention. It is a graph which shows the information table updated by the server function takeover method of this invention. It is a flowchart which shows the process sequence when the host in which the failure generate | occur | produced is recovered.

Explanation of symbols

100 Internet Data Center (IDC)
110 System administrator 120 General user 130 Internet 201 Router 202 Firewall 203 Load balancer 204, 205 Switching hub 206 Management server 207, 208, 209, 210 E-mail distribution server 211 File server

Claims (5)

An e-mail delivery system using a plurality of servers and using the Internet,
A program starting step of starting a program for executing the function of the own server with reference to the information table in which the control information of each server is defined in each server;
A server load information acquisition step of acquiring load information of the own server in each server;
Referring to the information table, a faulty server detection step of detecting a faulty server;
A takeover server detection step of detecting a server capable of taking over the function of a server in which a failure has occurred with reference to the information table and a weight table in which a weight for each server function is defined;
With reference to the information table, a server function takeover step of taking over a function to the server detected in the takeover server detection step;
A server function takeover method characterized by comprising:   The server function takeover method according to claim 1, wherein the server load information acquisition step detects a load of a CPU of the own server.   The failure server detection step includes a predetermined time in an update time of the time area corresponding to the own server and a takeover priority area after a predetermined time has elapsed since the time described in the time area of the information table serving as a reference of the own server. Compare with the update time of the time domain corresponding to the other servers described in the above numerical values, and detect that a failure has occurred in the local server when there is an opening that exceeds a predetermined value. The server function takeover method according to claim 1 or 2.   The takeover server detecting step includes a takeover priority maximum value server detecting step of detecting a server having a maximum takeover priority area value of the information table except for a server in which a failure has occurred. Item 4. The server function takeover method according to any one of Items 1 to 3. A server function takeover program for a plurality of servers used in an electronic mail delivery system using the Internet,
A program starting step of starting a program for executing the function of the own server with reference to an information table in which control information of each server is defined in each server,
A server load information acquisition step for acquiring the load information of the own server in each server;
Referring to the information table, a faulty server detection step of detecting a faulty server;
A takeover server detection step of detecting a server capable of taking over the function of the server in which the failure has occurred with reference to the information table and a weight table in which a weight for each server function is defined;
With reference to the information table, a server function takeover step of taking over a function to the server detected in the takeover server detection step;
Server function takeover program, which causes a computer to execute

Images