JPH11203257A - Computer system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a computer system of high reliability that does not stop supplying services to a client. SOLUTION: A server of a computer system to perform large scale process management adopts a duplex system of CPUs 11 and 21, a duplexed disk and a duplexed network. Also applications activated in the server are roughly divided into communication control applications 13 and 23 for communicating with a client group 41, data base applications 12 and 22 for accessing a data base and other data working applications 14 and 24, wherein plural applications stay in each of the CPUs 11 and 12, and perform at least confirmation of the presence of applications, communication-enable/disable detection of the applications and database access enable/disable detection of the applications.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a client-server type computer system, and more particularly to a system for detecting a failure of a server application and a communication control system after recovery from the failure.

[0002]

2. Description of the Related Art In recent computer systems, a client-server type has become widespread, and the server is roughly divided into an application for accessing a database, an application for controlling communication, and other data processing applications. .

[0003] In a computer system for performing a large-scale process control, it is natural that hardware is duplicated.
It is common to group applications having the same function in preparation for application downtime and communicate between the groups. In addition, the server receives requests with different processing contents from the client, which occur asynchronously, and controls these communications.

[0004]

In order to cope with a failure of a computer system, a CP in an operating system is used.
U (Central Processing Unit) and hardware such as disks were duplicated. In response to software failures, the mechanism to change the communication destination between applications was such that if a failure that exists on the operating system but did not function occurred, the transmitted message was lost. .

Conventionally, there has been a function of automatically detecting a failure of an application and restarting the node at its own node. However, when an environmental factor, for example, when an occupied memory allocated to each application is exceeded, the function is restarted. Had a problem that the message was not processed after repeated restart. In addition, since the communication destination between the applications is changed and the message is transmitted through a plurality of paths, it can be detected when processing a request from a client in order to detect that it functions correctly for each path. Therefore, there was a problem that the response to the problem was delayed.

On the other hand, in order to speed up a response in response to a request from a client, a method of sequentially transmitting data to a plurality of destinations has been used. However, after recovery from a failure, the number of messages before processing and the weight of each message are reduced. In some cases, the sum of the data becomes unbalanced for each transmission path, and a message transmitted earlier may be overtaken by a message transmitted later. SUMMARY OF THE INVENTION It is an object of the present invention to eliminate the above-mentioned problems and to provide a highly reliable computer system that does not stop services to clients.

[0007]

In order to achieve the above object, the present invention provides: [1] In a client-server type computer system, as a means for detecting a failure of an application existing on a server, the existence of the application is confirmed. Means, means for detecting the communication availability of the application, and means for detecting the database access availability of the application.

[2] In the computer system according to the above [1], a restart means for an abnormal application and a restart means for an application in a group are provided as means for responding when a fault is detected by the fault detecting means for the application. It is provided with. [3] In the computer system according to the above [2], a means for searching for a path with a small communication wait after the recovery from the failure and changing the communication destination as a means for searching for an optimal communication path after the recovery from the failure; A means for determining a communication destination based on overtaking conditions, and after a failure recovery,
A means for shifting the application to an initial start node is provided.

[0009]

Embodiments of the present invention will be described below in detail with reference to the drawings. FIG. 1 is a schematic system configuration diagram of a computer system showing an embodiment of the present invention. In this figure, 1 and 2 are databases (D
B), 11 is a first CPU (central processing unit), 12 is a first DB application (group), 13 is a first communication control application (group), 14 is a first data processing application (group), 15 is the first
Is an abnormality detection application.

Reference numeral 21 denotes a second CPU, 22 denotes a second DB application (group), 23 denotes a second communication control application (group), 24 denotes a second data processing application (group), and 25 denotes a second data processing application (group). 2 is an abnormality detection application. Further, the first abnormality detection application 15 and the second abnormality detection application 25 are connected to an abnormality detection monitor 31.

Also, the first communication control application 1
A client (for example, a personal computer terminal) group 41 is connected to the third communication control application 23 and the second communication control application 23. As described above, in the server of the computer system that performs the large-scale process management, the CPUs 11 and 21 are duplicated, the disks are duplicated, and the network is duplicated.

The applications started by the server are roughly divided into communication control applications 13 and 23 for communicating with the client group 41, DB applications 12 and 22 for accessing the database, and other data processing applications 14 and 24. Each of the CPUs 11 and 21 resides in a plurality. Anomaly detection applications 15, 25 for detecting these anomalies
Are resident in each of the CPUs 11 and 21 and display the detection result on the abnormality detection monitor 31.

FIG. 2 is an operation flowchart of a computer system showing an embodiment of the present invention. Hereinafter, the operation of the computer system will be described with reference to FIG. There are roughly two types of application failures. One is a failure that leaves the computer and the other is a failure that exists on the computer but does not function.

In order to make these determinations, the abnormality detection application checks the existence of each application. That is, a search for an application existing on the OS (operating system) is performed (Step S).
1) Check whether all applications exist (step S2). Next, in order to detect an application that exists on the computer but does not function, the abnormality detection application transmits a message with a response request to all the applications to determine whether communication is possible (step S3) and checks whether there is a response. (Step S4).

When each application receives a message with a response request, it returns a response to the transmission source. The abnormality detection application waits for a certain period of time, but if it exceeds a certain period, it is determined that there is an abnormality (step S
7). In addition, if a database is accessed with a security setting error or a resource that exceeds the resources allocated to each application, a failure may occur such that the database can be referenced and cannot be updated.

In order to detect this failure, the abnormality detection application transmits a message with a DB check response request to the application accessing the DB (step S5). DB for each application
DB when receiving message with check response request
, Reference, addition, update, and deletion of the predetermined table, and reports the access result to the abnormality detection application.

In the abnormality detection application, reception is waited for a certain period of time. If the period exceeds a certain period, it is detected as abnormal, and it is further determined whether the report is abnormal or normal based on the contents of the report (step S6). Next, a table that defines the correspondence of the application determined to be abnormal by the abnormality detection application is searched to determine the correspondence.

In the abnormality detection correspondence table, for example,
As shown in FIG. 3, a corresponding method is defined for each application that should exist. The response method includes a restart at the start node and a restart method at another node. When it is necessary to access data at high speed, it is common to share data on a memory, and applications having the same function are managed as the same group.

The operation flow when there is a faulty application will be described below. FIG. 4 is an operation flowchart when there is a failure application according to the embodiment of the present invention. First, a failed application is searched (step S11), and it is checked whether the failed application is stopped (step S12). When restarting the failed application, first, if the failed application has not been stopped, stop processing is performed (step S13).

Next, when a failure occurs in one application in the group, it is necessary to move to another node for each group. Therefore, an abnormality detection correspondence table is searched to determine whether there is an application in the same group (step S14), and if there is an application in the same group, stop processing is performed (step S15).

Next, the restarting node is searched for the stopped application by referring to the abnormality detection correspondence table (step S16), and the application is started (step S17). Finally, abnormalities of these applications frequently occur in the CPU 11, for example.
When the application is migrated to 22 nodes, the load balance will be lost and the response will decrease.
After the system administrator removes the abnormality, the system administrator requests the abnormality detection application to shift to the initial startup node.

Hereinafter, the operation of shifting to the initial activation node will be described. FIG. 5 is a flowchart showing the operation of shifting to the initial startup node according to the embodiment of the present invention. First, a request for transition processing to the initial activation node is received (step S2).
1). Next, an application existing on the OS is searched (step S22).

Next, a search is made for an initial startup node of the application (step S23). The abnormality detection application compares the current start node of each application with the initial start node (step S24), and if they do not match, the initial start node transmits a restart request (steps S25 and S26).

On the other hand, the weight of processing is defined for each type of message in order to search for a path with a light load according to the type and amount of request of the client at the time of failure recovery. Further, a database is generally composed of a plurality of tables, and the database is normalized by each transaction. For example, if the tables for managing the device status and the device history are managed separately,
When the overtaking of the transaction occurs, inconsistency may occur between the state and the history. However, by overtaking the message in the search request of the single table, the fast processing can be performed.

For this reason, for example, as shown in FIG. 6, it is previously defined whether each message can be overtaken. After recovery from the failure, the communication control application adds the processing weight of each message waiting for the message,
Manages the communication load table. FIG. 7 is a flow chart of an operation of searching for a route with little communication waiting and changing a communication destination after recovery from a failure of the computer system according to the embodiment of the present invention.

First, failure recovery is performed (step S3).
1). Next, a message is received from the client (step S32). Next, a quantity search of the pre-processing message is performed (step S33). Next, a weight search of the pre-process message is performed (step S34). Next, the sum of the weights of the pre-processing message is calculated (step S3).
5).

Next, it is checked whether or not overtaking is possible (step S36). Next, when overtaking is possible, the message is transmitted to an application having a small sum of weights of the pre-processing message (step S).
37). On the other hand, if overtaking is not possible, the message is transmitted to an application having a large sum of the weights of the pre-processing message (step S38).

As described above, first, the pre-processing message is searched for each application, and then the weight of each pre-processing message is searched and the sum of the pre-processing quantity is calculated. If the message received from the client can be overtaken, the message is transmitted to a route with a small sum. If the message received from the client cannot be overtaken, the message is transmitted to a route having a large sum.

As described above, according to this embodiment, in a client-server type computer system, a failure detecting means for an application existing on a server,
By providing the corresponding means, it is possible to prevent the service to the client from being stopped. In addition, by providing a restart determination unit for abnormal applications, it is possible to prevent the application from being repeatedly stopped and restarted due to environmental factors, and it is not necessary to consider abnormalities during abnormal processing. It is possible to reduce the development cost by reducing the cost.

Further, the provision of the means for searching for a route with little communication waiting after the recovery from the failure and changing the communication destination makes it possible to reduce the stop time due to the failure. In addition,
The present invention can be applied to all client-server type computers. Further, the present invention is not limited to the above-described embodiments, and various modifications are possible based on the gist of the present invention, and these are not excluded from the scope of the present invention.

[0031]

As described above, according to the present invention, the following effects can be obtained. (1) According to the first aspect of the present invention, in a client-server type computer system, a service for a client is not stopped by providing a failure detection unit for an application existing on a server and a corresponding unit. Can be

(2) According to the second aspect of the present invention, the provision of the restart means for the abnormal application can prevent the stop and restart of the application from being repeated due to environmental factors. It is not necessary to consider abnormalities during processing, and development costs can be reduced by reducing development man-hours. (3) According to the third aspect of the present invention, the provision of the means for searching for a route with less communication waiting after the recovery from the failure and determining the communication destination can reduce the stop time due to the failure.

[Brief description of the drawings]

FIG. 1 is a schematic system configuration diagram of a computer system showing an embodiment of the present invention.

FIG. 2 is an operation flowchart of a computer system showing an embodiment of the present invention.

FIG. 3 is a diagram illustrating an abnormality detection correspondence table of the computer system according to the embodiment of the present invention.

FIG. 4 is an operation flowchart in the case where there is a failure application according to the embodiment of the present invention.

FIG. 5 is a flowchart illustrating a transition operation to an initial activation node according to the embodiment of the present invention.

FIG. 6 is a diagram illustrating a message, a processing weight, and an overtaking availability correspondence table of the computer system according to the embodiment of the present invention.

FIG. 7 is a flowchart of an operation of searching for a route having a small communication wait and changing a communication destination after recovery from a failure of the computer system according to the embodiment of the present invention.

[Explanation of symbols]

1, 2 Database (DB) 11 1st CPU (central processing unit) 12 1st DB application (group) 13 1st communication control application (group) 14 1st data processing application (group) 15 1st Anomaly detection application 21 Second CPU 22 Second DB application (group) 23 Second communication control application (group) 24 Second data processing application (group) 25 Second anomaly detection application 31 Anomaly detection monitor 41 Client group

Claims (3)

[Claims] In a client-server type computer system, (a) existence check means of the application, (b) communication availability detection means of the application, and (b) means for detecting a failure of an application existing on a server; c) a computer system comprising: a database access permission / non-permission detection unit for the application. 2. The computer system according to claim 1, wherein: (a) a restart means for an abnormal application; and (b) a restart for an application in a group, when the fault is detected by the fault detecting means for the application. A computer system comprising: starting means. 3. The computer system according to claim 2, wherein, as means for searching for an optimal communication path after recovery from the failure, (a) means for searching for a path with less communication waiting after recovery from the failure and changing the communication destination; A computer system comprising: b) means for deciding a communication destination based on a message overtaking availability condition after restoration of a failure; and (c) means for shifting the application to an initial start node after restoration from the failure.