JPH07146851A - Constructing method of distributed system - Google Patents

Abstract

PURPOSE:To provide the constructing method of a distributed system where the necessity of the inputting of set data by a system controller is eliminated every time a terminal is introduced and few failure is generated in an initialization processing. CONSTITUTION:When a client machine 300 is connected with a network 350, an initialization program loader 302 is started. Based on the set data 342, the procedure data 343 and procedure control data 344 obtained from an introduction exclusive server 341, the initialization processing of a client program 301 is executed. As a result, the labor and time for inputting the set data which is required for the initialization processing of the client program 301 when the client machine 300 is introduced is saved for a system controller.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method of constructing a distributed system composed of a plurality of distributed software, and particularly to constructing a distributed system suitable for initializing a client program of the distributed software. Regarding the method.

[0002]

2. Description of the Related Art In the conventional method of constructing a distributed system, the following two methods are mainly used to initialize the client program for accessing the server program.

(1) The system administrator manually inputs the setting data such as the position of the terminal where the server program exists to the client program of each terminal according to the procedure described in the manual, and the initialization process is performed. To do.

(2) As seen in Japanese Patent Application Laid-Open No. 4-302347, for client programs of a plurality of terminals,
The server program sends the setting data and executes the initialization process remotely online.

[0005]

In the method (1), it is necessary to manually perform the initialization process every time the system administrator installs the terminal. Also, the system administrator needs to store the setting data such as the position of the terminal where the server program exists.

In the method (2), the system administrator does not have to manually enter the setting data, but on the other hand, the remote control by the server program is proportional to the increase in the number of client programs of the target terminal. The load of execution increases. Also, if there are multiple server programs and multiple initialization procedures corresponding to each server program are required for the initialization processing of the client program, the execution order of the multiple initialization procedures is planned and the initialization processing is performed. No means are provided to control the. Therefore, there is a problem that the initialization process is defective depending on the order of execution of the initialization procedure.

An object of the present invention is to provide a method of constructing a distributed system in which the system administrator does not need to input setting data every time a terminal is installed.

Still another object of the present invention is to prevent the occurrence of an initialization failure due to an erroneous execution sequence when a plurality of initialization procedures corresponding to a plurality of server programs are required for initialization processing of a client program. To provide a method of constructing the distributed system.

[0009]

In order to achieve the above object, in the method of constructing a distributed system according to the present invention, data necessary for initialization of a client program, that is, setting data and procedure data, and a procedure are provided on a communication network. There is a setting management program dedicated to the introduction of terminals that manages control data. When the terminal to be installed is connected to the communication network, the initial program loader prepared for that terminal queries the setting management program and downloads multiple setting data and procedure data and procedure control data all at once. Execute the conversion process.

Particularly in a preferred mode of the present invention, when the initialization processing is sequentially performed based on the setting data and the procedure data of the initialization processing for a plurality of server programs,
Accumulates the log that executes the initialization process corresponding to each procedure data, inspects the log when the procedure described in each procedure data ends, and when the initialization process failure is detected, the procedure in which the failure occurred To try again.

[0011]

When the terminal to be installed is connected to the communication network, the initial program loader prepared for the terminal issues a request to the setting management program dedicated to the terminal installation, and the data necessary for initializing the client program, that is, the setting. Since the data, the procedure data, and the procedure control data are downloaded and the initialization process is executed, the system administrator does not need to input the setting data and the like for each installed terminal.

Particularly, since the initial program loader sequentially performs the initialization processing based on the setting data and the procedure data of the initialization processing for the plurality of server programs, the initialization of the client program is executed in the wrong order. Never.

Further, by accumulating a log of initialization processing corresponding to each procedure data and inspecting the log when the procedure described in each procedure data is completed, the progress of the initialization processing can be confirmed. Yes, you can identify failed steps and steps that should be retried. Further, when a defect in the initialization process is detected, if a process for retrying the procedure in which the defect has occurred is provided, a failure (for example, stop of the server program) causing the defect is recovered after a certain time. Since the possibility is high, the initialization process can be continuously executed by retrying.

[0014]

Embodiments of the present invention will be described in detail with reference to the drawings. In the following, for simplification of description, the terminal where the client program exists is the client machine, the terminal where the server program that accesses to execute the client program exists is the server machine, and the data referenced when installing the client machine is managed. The terminal in which the setting management program that exists is called the installation server machine.

FIG. 1 is a flow chart showing a processing procedure in a client machine and an installation server machine of a distributed system according to an embodiment of the present invention. In this embodiment, the client machine has steps 108 to 11 in FIG.
The feature is that the first process is provided, data is requested from the introduction server machine, and the initial process is performed by itself. FIG. 2 is a flowchart showing an embodiment of a procedure for retrying the initialization process based on a defect detected during the initialization process of the client program in step 111 of the processing procedure in the client machine of FIG.

FIG. 3 is a block diagram showing an outline of the distributed system of the embodiment. The client machine 300, the server machines 310, 320 and 330, and the introduction server machine 340 are interconnected via a communication network 350. The client program 301 existing in the client machine 300 accesses any one of the server programs 311, 321, and 331 in order to perform its processing. Server program 31
1, 321 and 331 are server machines 310 and 320
And 330 are distributed. As an example, the client program 301 is the server programs 311, 3
21 and 331 are all accessed, and when a client machine 300 equipped with such a client program is newly introduced into the system, the client program 301 includes the server programs 311, 321 and 331 in which these client programs are distributed. In order to make each of them accessible, each initialization process is required for the client program 301. In order to execute this multiple types of initialization processing without conflict,
The client machine 300 is further equipped with an initial program loader 302. Further, the installation server machine 340 is loaded with a setting management program 341. The setting management program 341 includes setting data 342 for a client program corresponding to each server program, procedure data 343 indicating a setting procedure of each setting data, and mutual dependency between each procedure data when performing a plurality of initialization processes. It is a program for managing the procedure control data 344 for controlling the procedure of the initialization processing based on the above. That is, the setting management program 341 connects the database 342 and the database 343 to each other in response to a request, and provides them to the initial program loader 302. Particularly, when the initial program loader 302 performs a plurality of initialization processes,
The procedure control data 344 is transferred to the initial program loader 302.
To provide. The initial program loader 302 requests the setting management program 341 for necessary data and downloads it, and uses the data to download the client program 3
01 initialization is performed. 2 of the same client machine
In preparation for re-introduction after the first time, the execution progress of the initialization processing is accumulated in the setting log 303.

FIG. 4 is a block diagram showing an outline of a conventional distributed system for comparison, and each server machine has an initialization processing function. That is, the server machines 410, 420, 430 respectively mounting the server programs 411, 421, 431 are provided with the initial program loaders 412, 422, 432. These initial program loaders use the client program 40 of the client machine 400 based on the setting data and procedure data of each.
1 is initialized. Therefore, the feature of the embodiment of FIG. 3 is that the client machine has the initial program loader for initializing its own client program, and the database for the initialization processing is centrally arranged on the installation server machine. In point.

Next, referring to the flow chart of FIG.
The operation of each section of will be described in more detail.

First, the operation of the client machine 300 will be described. First, the client machine 300 is connected to the communication network 350 (step 100). Next, when the power of the client machine 300 is turned on, the initial program loader 302 is activated (step 101). The initial program loader 302 creates a list of the client programs 301 existing in the client machine 300 by searching a file tree, for example (step 102). Here, the problem is the position of the installation server machine 340 where the setting management program 341 (hereinafter abbreviated as an installation server) that manages the data necessary for the initialization of the client program 301 exists. For example, when re-introducing a client machine that has been initialized in the past into the distributed system, the installation-dedicated server machine 340 can be identified based on the past connection progress by referring to the setting log 303. Therefore, it is determined whether the initialization processing has been performed in the past (step 10
3), in the case of Yes, the client program 301 is specified by designating the terminal showing the connection progress left in the setting log 303.
And the introduction server 341 are communicatively connected (step 10).
5). Further, as in the case where the system administrator explicitly defines it, it is determined whether the position of the installation server machine 340 can be specified in advance (step 104). If Yes, the client program 301 and the installation server 341 are determined according to the determination. And are connected for communication (step 105). For example, on UNIX workstations, the file / etc / s
ervices and / etc / hosts installed server machine 340
The location information of can be set. If both Step 103 and Step 104 are No, the introduction server 341 is searched by using broadcast communication, and then communication connection is established (Step 106). The initial program loader 302 connected to the introduction server 341 requests data required for initialization by the client program 301 to the introduction server (step 108) and has a reply (step 1).
09). When there is a response from the installation server, setting data, procedure data, and procedure control data 344 required for initialization.
Are collectively downloaded (step 110), and initialization processing is executed (step 111).

Next, the operation of the introduction server machine will be described. The system administrator turns on the power of the installation server machine 340 (step 151). Next, the system administrator activates the introduction server (setting management program) 341 that processes the request of the initial program loader 302. The introduction server 341 waits until there is an initialization request from the initial program loader 302 of each client machine (step 154, step 153). Introduction server 341
When the initialization request is received (step 155), the databases search the databases 342 and 343 and extract the setting data and procedure data according to the request (step 156). Especially when initialization of the client program is requested for a plurality of server programs (step 15).
7) There is setting data and procedure data 143 corresponding to each server program, and it is necessary to control the procedure of the initialization process based on the interdependency between the procedure data. Therefore, the procedure control data 344 indicating the interdependence between the procedure data is searched, and the procedure control data according to the request is extracted (step 158). The extracted setting data and procedure data, or further procedure control data is downloaded to the client machine 300 (step (15
9).

Table 1 is a table for explaining the configurations of the setting data database 342 and the procedure data database 343 handled in step 156 of FIG.

[0022]

[Table 1]

The setting data is various data to be set necessary for initializing the client program.
As shown in the figure, information such as the starting position of the server program, the starting option of the client program, and the position of the duplicate server is included. Further, the procedure data is data that describes an initial procedure required for a client program that accesses each server program. For example, it includes information for instructing inspection such as activation confirmation of the server program and creation of a database unique to the client program. The procedure data and the setting data are set for each server program and form a database 342 and a database 343. That is, in FIG. 3, the respective elements of the procedure data database 343 are 413, 423 of FIG.
433, and each element of the setting data database 342 corresponds to 414, 424, 434 in FIG. These are realized by variable declarations and control statements in declaration statements that can be described in UNIX shell language, for example.

Table 2 is a table for explaining the structure of the procedure control data handled in step 158 of FIG.

[0025]

[Table 2]

As described above, the procedure control data 344 describes the interdependence relationship between the initialization processes corresponding to the plurality of servers (how sequentially the initialization processes corresponding to the plurality of servers should be performed). It was done. For example,
It includes information such as that the initialization of the client program for the server program 311 must be completed before the initialization of the client program for the server programs 321 and 331 (example of Table 2). Based on this information, the execution order of the procedure data is controlled.

Within the range described above, the present embodiment can easily and easily save the time and effort for the system administrator to manually set various data for initializing the client program every time the client machine is installed in the system. The distributed system can be expanded, and even if the client program accesses multiple server programs, the multiple types of initialization required to access each server program are automatically executed without conflict with each other. It can be understood that it is possible to prevent the initialization failure due to the wrong execution order from occurring. Furthermore, in the embodiment, since the history of the initialization process is saved in the setting log, the progress status of the process can be confirmed during the initialization process,
Further, when the client machine once separated from the system is re-introduced into the system, the initialization process is quickly performed using the history.

In the present embodiment, if a defect occurs in the initialization process, the initialization process can be continued by retrying. This will be described with reference to FIG.

FIG. 2 shows step 111 of FIG. 1 in detail. First, the initialization process is started based on the procedure control data 344 (step 200). If there is no procedure data or setting data for instructing initialization of the client program 301 (step 201), the initialization success is reported to the system administrator (step 206), and the process ends. The initial program loader 302 searches the setting log 303 (step 202), proceeds to step 210 if it is confirmed that it has not been executed, and confirms the result if initialization processing is being performed (step 203). ), If it is necessary to set again, the process proceeds to step 210, and if not, the process proceeds to step 205 (step 204). If there is the setting data or procedure of the initialization process to be executed, the process returns to step 202, and if not, step 2
The procedure proceeds to 06 (step 205).

The process of actually performing the initialization process is steps 210 to 217. First, the prerequisites such as the activation of the server program are confirmed (step 210). Based on the setting data and the procedure data, initialization processing such as registering the attribute of the client program in the server program is executed. The execution result is stored in the setting log 303 (step 211). Each time the initialization process based on one piece of procedure data is completed, the setting log 303 is searched, and if a defect is found, 1 is added to the number of trials 304 (step 214) and the process proceeds to step 216. Otherwise, the trial count 304 is cleared to zero (step 21
5), go to step 205. It is compared whether the cumulative number of trials exceeds the specified value, for example, 3 times (step 21).
6) If it exceeds, the failure of the initialization process is reported to the system administrator, for example, by displaying a message indicating the occurrence of a defect on the screen, and the process ends (step 218).
If it does not exceed the specified value, the retry of the procedure in which the defect has occurred is planned and the process returns to step 211 (step 21).
7).

[0031]

As described above, according to the present invention, since the setting management program for managing the setting data and the procedure data and distributing the setting data upon request is provided, the system administrator does not need to manually perform the initialization processing of the client program. . Further, since the initialization process is automatically performed by the initial program loader of the client machine after the client machine is connected to the communication network, the server machine that executes the initialization process of the client program of the entire communication network as in the conventional case. Does not exist, and the load does not concentrate on a specific terminal. Furthermore, when executing the initialization process based on multiple procedures corresponding to multiple server programs, the setting management program distributes the procedure control data according to the setting data and procedure data, which may cause an incorrect execution order. It is possible to prevent the occurrence of the initialization failure.

[Brief description of drawings]

FIG. 1 is a flowchart collectively showing a processing procedure in a client machine and an installation-dedicated server machine according to an embodiment of the present invention.

FIG. 2 is a flowchart showing a procedure of retrying the initialization process based on a defect detected at the initialization process in the client machine in the embodiment.

FIG. 3 is a configuration diagram showing an outline of a distributed system in one embodiment.

FIG. 4 is a configuration diagram showing an outline of a conventional distributed system.

[Explanation of symbols]

340 ... Installation server machine, 341 ... Setting management program, 301 ... Client program, 302
...... Initial program loader

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification number Internal reference number FI Technical display location G06F 15/16 370 N 7429-5L (72) Inventor Naofumi Shoji 1099 Ozenji, Aso-ku, Kawasaki-shi, Kanagawa Hitachi, Ltd. System Development Laboratory

Claims (5)

[Claims] 1. A first terminal equipped with a client program and a second terminal equipped with a server program that is accessed when the client program is executed, wherein the first and second terminals form a communication network. In the method of constructing a distributed system connected via the above, the third terminal in which a setting management program for managing data for initialization processing necessary for initialization processing of the client program using the server program exists is previously described above. When connecting to the communication network and introducing the first terminal to the distributed system, the initial program loader included in the first terminal issues a request to the setting management program to transfer the data for the initialization processing. The data is downloaded to the first terminal, and the first initialization data using the downloaded initialization processing data is used. Method for constructing a distributed system and performing initialization period of its own client program in the terminal. 2. Communication between a plurality of server machines in which a plurality of server programs are installed in a distributed manner and a client machine in which a client program that accesses any of the plurality of server programs and executes its own processing is communicated. In a method of constructing a distributed system connected via a network, for each of the plurality of server programs, setting management for managing data for initialization processing necessary for initialization processing of a client program that uses the server program Connect the installation server machine equipped with the program to the communication network in advance,
When the client machine is installed in the distributed system, an initial program loader included in the client machine issues a request to the setting management program to download necessary data out of the initialization processing data to the client machine. Then, the method for constructing a distributed system is characterized in that the client computer performs an initialization process of its own client program using the downloaded data. 3. The initialization data includes setting data and setting procedure data unique to each of the plurality of server programs, and an interdependence relationship between procedures of the initialization processing of each of the plurality of server programs. The method for constructing a distributed system according to claim 2, further comprising procedure control data indicating 4. A first terminal, in which a plurality of terminals are connected via a communication network, and which has a client program for issuing a request and an initial program loader for making the client program available, the client program of the first terminal A plurality of second terminals in which a server program for processing the above request exists, and setting data necessary for the client program of the first terminal to access the server program of the second terminal, using the setting data A third terminal having a procedure management data describing a method for initializing the client program of the first terminal and a setting management program having a process of downloading the setting data and the procedure data to another terminal. In a method of constructing a distributed system including: the first terminal in the communication network. When continuing, the step of the initial program loader of the first terminal requesting the setting data and the procedure data to the setting management program of the third terminal; and the setting management program of the third terminal, A step of collectively downloading to the initial program loader of the first terminal, the plurality of setting data and the procedure data, and procedure control data describing an execution order of the plurality of procedure data; Using the procedure data and the procedure control data, the initial program loader of the first terminal can access the server programs of the plurality of second terminals.
A method for constructing a distributed system, comprising the step of executing an initialization process of a client program of the terminal. 5. The method for constructing a distributed system according to claim 4, wherein the initial program loader of the first terminal can access the server programs of a plurality of the second terminals. In the step of executing the initialization process of the client program, the step of sequentially executing the initialization process of the client program of the first terminal according to the execution control data and the plurality of procedure data, and corresponding to each procedure data The step of accumulating the log of the above initialization processing, the step of inspecting the log of the above initialization processing when the procedure described in each procedure data is completed, and the initialization processing when a defect of the initialization processing is detected. The procedure in which the processing failure has occurred is retried, and cumulative information corresponding to the cumulative number of retries and the procedure data is accumulated and accumulated. When the accumulated information exceeds a predetermined value, method for constructing a distributed system, characterized in that it comprises a step of stopping the initializing process of the client program by the initial program loader.