JP2001134545A - Server managing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To solve that abnormality detection as to whether or not all tasks needed for a server application to start are in operation can not be performed although server machine main body and network trouble, etc., can be detected. SOLUTION: A status file 18 is provided which can record and manage whether or not not only a server machine main body and a network function, but also a server application and OS modules are functioning normally. When the self-diagnosis result of the machine main body, the monitor result of the network function by the server application, the operation state detection result of the server application, and the operation state detection result of OS modules that the server application use are all 'OK (normal)', it is judged that the system is in normal operation and actuation is carried out. When one or more results are '(NG)' different from the normal operation, an 'error' is decided and abnormality of the server application which appears to operate normally is speedily detected.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a computer network system in which a plurality of computers are connected via a network, and more particularly, to a LAN (Local Area Network).
The present invention relates to a server management method suitable for efficiently managing a failure that occurs in a server in a client-server system including a network and a WAN (Wide Area Network).

[0002]

2. Description of the Related Art Servers in a client server system in which a computer is connected to a network such as a LAN or a WAN include a groupware server, a database server, a network server, a printer server, a file server, and the like. Is used in a wide variety of businesses today, and is indispensable for mission-critical and daily tasks.

[0003] Such a down (failure) of the server poses a serious problem in business execution. To deal with such a server failure, a specialized system administrator or a network system engineer is required. However, the system administrator or the like is usually not near the server, and it is generally the user of the client that finds a defect.

Techniques for automatically managing such a computer network system such as a client / server system are disclosed in, for example, JP-A-10-247169, JP-A-11-4222 and JP-A-11-222.
-31114.

[0005] Japanese Patent Application Laid-Open No. Hei 10-247169 discloses that management information of a managed system such as a server, a terminal (client), and a router is collected in a short time by the management system, and the load concentrated on the management system is reduced as much as possible. For this purpose, a DM agent that collects management information in a management system according to a predetermined management scenario is generated and moved to each managed system, and managed by the DM agent in an execution environment of each managed system according to the management scenario. It describes techniques for collecting information and reporting it to a management system.

[0006] Japanese Patent Application Laid-Open No. 11-4222 discloses that
In the management of a network using SNMP (Simple Network Management Protocol), there is described a technology for shortening the time from occurrence of a state change of a management agent to detection of the change without overloading a management manager or a network. I have.

Japanese Patent Application Laid-Open No. 11-31114 discloses that a management server connected to a network is accessed from a remote place, start / stop control of this management server is performed, information to be managed is collected, and a Web browser is provided. Describes a technique for centrally and efficiently managing a network by displaying the information.

However, according to these conventional techniques, it is not possible to quickly detect an abnormality of an application for each server (server application) in a server that is operating normally at first glance. That is, there are various unexpected patterns in the server down phenomenon, and it is difficult to cope with the simple work.

For example, in a network configuration in which a groupware database server and a groupware client are connected to an intranet, the database server is a server that is always used in daily work.
There is no problem if this server operates for 24 hours, but sometimes it goes down, and when it goes down, there are various problems described below.

(1) There is a phenomenon that a server failure of groupware cannot be detected quickly. (2) When a part of the OS (Operating System) of the server is down, an abnormality of the server application cannot be noticed. (3) Even if the network function in the server is normal, the server application function may be down,
In that case, server down cannot be detected from the network management information.

(4) In the case where the network function is operating and there is no function for detecting that the server application is operating normally, the network administrator cannot detect the malfunction immediately. (5) When some functions of the OS of the server, for example, console (keyboard, mouse and display device) functions are down, the failure cannot be detected unless the server is used. (6) It cannot be detected that the OS of the server and the server application are functionally synchronized (operation check of necessary functions).

(7) If a part of the server function is down, the server cannot be detected from the client unless the function is used (failure cannot be found). (8) A frequently occurring phenomenon is that the agent (a program that automatically starts a server application without an operator) cannot start the server application properly when started. In addition, agents
It has a function to shut down and start up the server application, and controls the startup of the server application.

FIG. 9 is a flowchart showing an example of conventional server management control. When the main switch is turned on, P
A self-diagnosis test by OST (power on test) and a network function test by SNMP etc. are sequentially performed. If the respective test results are normal (steps 901 to 904), the server application is started. Load the program (step 9
05), the server system is started (step 90)
6). If a fatal error is detected in the POST and the network function test (steps 907 and 908),
The contents of each error are displayed (steps 909, 9
10) Stop the startup.

[0014] However, only the result of the test in the POST and the network function that is normal indicates that a failure that can be recognized for the first time at the time of use, such as an abnormality of a server application, is detected. Even if it looks like it may be in a state where it cannot actually operate.

[0015]

The problem to be solved is that, in the prior art, a server machine (server machine) or a network failure is managed by POST, network monitoring protocol (SNMP) or the like. Therefore, it is possible to detect an abnormality, but it is not possible to detect an abnormality such as whether or not all tasks necessary for starting a server application are operating.

An object of the present invention is to solve these problems of the prior art, and to provide a total that the server machine itself is operating, the network function is operating, and the server application is functioning. To monitor and detect and manage failures as a whole server. In particular, only server applications or OS
It is an object of the present invention to provide a server management method capable of promptly detecting a partial abnormal failure of the above, and automatically performing a response process according to the content of the failure.

[0017]

In order to achieve the above object, the server management method of the present invention requires that the server machine itself operates and that the network function operates in order for the server function to operate correctly. Focusing on the fact that it is necessary that the server application is functioning, a "status file" that can manage the entire server system is provided. The “status file” includes the self-diagnosis result of the server machine itself, the monitoring result of the network function by SNMP, the operation state detection result of the server application, and the operation state detection result of the OS module used by the server application. Is recorded and all results are "OK".
If the status is (normal), the system is started up as if it is operating normally, but if one or more of them is different from the normal status ("NG"), it is determined as "error". As a result, it is possible to promptly detect an abnormality of the application in the server that is operating normally at first glance.

In the "status file",
Record all functions that must be started for each server machine or network etc. to operate properly, for example, record the function name as `` normal value data '' with character code for each server machine or network, When the server is actually started, the function names of the sequentially started functions are recorded as “current data” in the “status file”, and all the function names previously recorded as “normal value data” are referred to as “current data”. ”, It is determined that the server machine itself and the network have started up normally,
If even one is missing, it is determined as an error.

If it is determined as "error", it is confirmed which check result is abnormal. For example, when only a certain server application is abnormal, the server application is shut down and restarted. I do. This makes it possible to automatically resolve startup errors that often occur when the server application is started, and it is possible to easily find out only server application errors. Can be resolved quickly.

When it is determined (determined) as “error (failure, abnormality)”, the contents of the “status file” are stored in a logging database file such as a server application. It is embedded in a logging file where event processing is recorded, so that the client user can determine the status when the database access function is operating. As a result, the server administrator or the client user can access the logging file of the server application to be checked from the client machine without using the console of the server machine itself, that is, without going to the installation location of the server machine. You can check the situation.

If the determined “error” is a failure other than the network function such as a failure of the server application or the OS module, the server administrator is notified of the server application by using an e-mail function independent of the server application. Notify an error. As a result, it is possible to detect an event that has not been detected unless a server abnormality is checked beforehand by the arrival of an electronic mail.

If the determined “error” is a failure of the OS module used in the server application,
Server machine console (keyboard / mouse / C
In this case, the on / off control (restart) of the main switch (power switch) of the server machine is performed. As a result, it is possible to automatically perform the operation in which the server administrator conventionally goes to the installation location of the server machine, determines that the system is locked, and controls the power of the server machine (off / on).

[0023]

Embodiments of the present invention will be described below in detail with reference to the drawings. FIG. 1 is a flowchart illustrating an example of a processing operation according to the server management method of the present invention. FIG. 2 is a block diagram illustrating a configuration example of a server machine that performs a processing operation according to the server management method of the present invention. , FIG.
FIG. 4 is a block diagram showing a configuration example of a client-server system performing the processing operation in FIG. 3, and FIG. 4 is an explanatory diagram showing a configuration example of a status file created by the server machine in FIG.

In FIG. 3, 1 is a server machine provided with a groupware server function, 2 to 4 are client machines (hereinafter, referred to as “client” in the figure), and 5 is L
The client is an AN, and the clients 2 to 4 access the server machine 1 via the LAN 5 and use a server application (groupware).

The server machine 1 is normally operated for 24 hours, and the clients 2 to 4 are provided with the PO only when the user is in business.
The WER (power supply) switch is turned on to operate. In such a system, “hardware (server machine body)”, “OS (operating system)”, “network function”, “server application”, “server attached function (backup device / power failure prevention device / remote device) Access modem / others) ”etc. are all in operation, and the correct operation state is achieved.

As for the “hardware (server machine body)”, an abnormality in the hardware is detected by a power-on test (POST) performed at the time of power-on. In addition, about "network function",
Conventionally, a network monitoring protocol (SNMP: Sim
ple Network Management protocol) to detect errors.

However, there is no reliable "OS error detection" or "server application error detection". Conventionally, a server administrator checks hardware, hard disk, task status, memory status and the like as necessary. Operation confirmation.

In this example, a program for creating a status file (status file) 18 having the configuration shown in FIG. 4 for managing the operating states of the server machine 1 and a program for monitoring the contents of the status file are provided in the server machine 1. Is provided, and the current data and the normal data are compared to detect which portion has an abnormality.

As shown in FIG. 2, the server machine 1
CRT (Cathode Ray Tube) and LCD (Liquid Crystal)
(Display), an input device 1b including a keyboard and a mouse, an external storage device 1c including a hard disk drive, and a CPU (Central Processi).
ng Unit) 10, a main memory 11, and the like, and an information processing apparatus 1 that performs computer processing by a storage program method
d, an optical disk 1e on which a processing program and data according to the present invention are recorded, a driving device 1f for reading the optical disk 1e, an uninterruptible power supply for supplying power to a battery cell when a commercial power supply such as a power failure is stopped (in the figure) And hereinafter referred to as “UPS: Uninterruptible Power Supply”).

By installing a processing program or data recorded on the optical disk 1e and reading it into the main memory 11 and executing it by the CPU 10, the functions according to the present invention are implemented as the server management processing section 12 in the information processing apparatus 1d. You.

The server management processing unit 12 includes a POST processing unit 13, a network test processing unit 14, an application test processing unit 15, an OS test processing unit 16, a result recording processing unit 17, a status file 18, a verification processing unit 19, The operation is described below.

The server machine 1 includes a server management processing unit 1
The POST processing unit 13 in 2 reads the current data relating to the “hardware state” from the POST self-diagnosis function result information. OS test processing unit 1
6, the current data on the "OS state" is read from the application state information of the task manager, the current data on the "network state" is read from the SNMP information by the network test processing unit 14, and the " The current data on the "application state" is read from the logging data.

Each of these reading results is recorded in the status file 18 by the result recording processing unit 17. The status file 18 can be generally viewed as a text file format. Then, the result recording processing unit 17 compares each read result with the data read at normal time to determine (judge) each current operating state, and records the self-diagnosis result as “OK / NG”. I do.

That is, in the status file 18, as shown in FIG. 4, all functions that must be started up for the server machine and the network to operate properly, for example, the function names are represented by character codes ( aaa, bbb, ccc, ..., AAA, BB
B,...), Is recorded as “normal value data” for each server machine or network, and the function names of the sequentially activated functions at the time of actual startup of the server are referred to as “current data” in the status file 18. If all function names previously recorded as "normal value data" are recorded in "current data",
It is determined that the server machine body (hardware state) and the network state have started up normally (“OK”), and if any of them is missing, it is determined that an error has occurred (“NG”).

Thereafter, the verification processing unit 19 determines the operation state based on the self-diagnosis result (“OK / NG”) in the status file 18 by the result recording processing unit 17. That is, if the contents of each item are all “OK (same as normal operation)”, it is determined that the system is operating normally, but one or more of them are “NG” and different from the normal state. In this case, it is determined that the error is “ERROR”.

Here, if it is determined that the error is “ERROR”, the server machine 1
The error processing unit 20 checks which item in the status file is abnormal, and performs error processing corresponding to each item. For example, when only the server application is abnormal, the server application is shut down and restarted.

By doing so, it is possible to detect an abnormality in the server application before the clients 2 to 4 actually use it, and to automatically eliminate a startup abnormality that often occurs when the server application is started. it can.

The error processing unit 20 stores the status file 18 in a logging database file of a server application (a logging in which the event processing of the application since the application was started is recorded) as shown in the following sample. File).

<Sample logging file>"The server has been closed. Release 4.5.5 is running.-Event Interceptor (Version 4.6)-Event Interceptor started-Database Replicator started-Mail Router started for domain A・ Router: Internet SMTP host nts_s_pro07 in domain ・ Router: Shutdown is in progress ・ Router: Beginning mailbox file compaction ・ Router: Completed mailbox file compaction ・ Start of index update processing ・ Stats agent started ・ Agent manager started ・ AMgr: Starting execution number 2 ・ AMgr: Starting execution number 1 ・ Query / Set Hand1er started ・ Query / Set Handler (Version 4.6) ・ Reporter started ・ Reporter: Sending statistics to 'statrep.nsf'eve
ry720 minutes.Running ana1ysis dai1y.Event Dispat
cher started.-Starting the database server "

Thus, when the database access function is operating, the server application of the server application can be viewed by viewing the logging database from the client without using the console (keyboard, mouse, display device, etc.) of the server machine 1. Abnormality can be detected. That is, the status can be determined from each user of the clients 2 to 4. Also, the server administrator can check the situation without going to the installation location of the server machine 1.

The server machine 1 has an e-mail function independent of the server application. The error processing unit 20 refers to the status file 18 and determines that the network function is normal. Using the e-mail function, the server application notifies the pre-registered e-mail address of the server administrator of the abnormality of the server application. Thus, the server administrator can detect a server abnormality even when not at the installation location of the server machine 1.

The error processing unit 20 creates an agent to be activated when an abnormality is determined with reference to the status file 18, and collects data and a logging file necessary for analyzing a failure of the server application by this agent. If the abnormality is a server application failure or OS failure,
Using the function of the UPS 1g, the power supply control (power off / on, etc.) of the information processing apparatus 1d (server machine 1) is performed.

Thus, it is possible to automatically cope with a failure of the console function of the server machine 1. Conventionally, a server administrator goes to the installation location of the server machine 1, determines that a system lock failure has occurred, and turns off the power supply of the server machine 1. The work of controlling (off / on) is unnecessary. In particular, in order to shorten such server downtime, the server administrator needs to check before working hours, which has been a large workload, but the workload can be reduced.

Also, by performing this operation when the server application is started up, if the start-up is usually at about 5:00 am, the server system is brought into the operating state at 8:00 am or the like without any problem. be able to.

Hereinafter, the server management control operation in the server machine 1 will be described with reference to FIG. When the main switch is turned on (step 101), first,
Perform POST (power-on test) (step 10
2) and the result (“OK / NG”) (step 10)
3), and record it in the hardware status column of the status file 18 (steps 104 and 105).

Next, the network function is confirmed by SNMP (step 106), and as a result (“OK / N
G ”) (step 107), the status file 18
Is recorded in the network status column in (Step 10)
8, 109).

Further, the server application is activated, its activation state is confirmed and the activation module is checked (step 110), and as a result ("OK / NG")
Is recorded in the server application status column in the status file 18 (steps 112 and 113).

Further, the operating state of the OS module used in the server application is confirmed (step 11).
4) and the result (“OK / NG”) (step 11
5), and record it in the OS status column of the status file 18 (steps 116 and 117).

Then, the confirmation result of each status column recorded in the status file 18 is verified (step 118). If all the states are recorded as "OK", the server machine is started without any abnormality (step 119), and if any one of "NG" is recorded, "Error notification output" etc. Or the error processing corresponding to the state where the "NG" is recorded is determined by determining which state is "NG" as shown in FIGS. 5 to 8 (step 121). ).

As described above, the POST, the network function test, the server application activation state test, the OS
If at least one of the startup tests has an abnormality, processing such as error display is performed, so that at first glance an abnormality of the server application that appears to be operating normally can be promptly detected.

FIG. 5 is a flowchart showing a first error processing operation example in the server management method of the present invention.
This example shows a detailed example of the processing in step 121 in FIG. 1, and the status file 1 shown in FIG.
In step 8, if "NG" is recorded only in the server application state (steps 501 to 50)
4) Shut down and restart the server application (step 505).

If the error is other than that of the server application, an error message is displayed and startup is stopped (step 506). By doing so, it is possible to automatically cope with the "startup trouble" that often occurs when the server application is started, and it is possible to quickly resolve the abnormality of the server application that is first identified at the time of use / access. .

FIG. 6 is a flowchart showing a second example of the error processing operation in the server management method of the present invention.
This example shows a detailed example of the processing in step 121 in FIG. 1, and the status file 1 shown in FIG.
If "NG" is recorded in any of the states in step 8 (steps 114 to 119), it is determined which is "NG" and in what state (step 601). It is set in the logging data of the application (step 602).

By doing so, the server administrator and the client user can view the logging data from the client without using the console of the server machine, that is, without going to the installation location of the server machine. In this way, a server error can be detected.

FIG. 7 is a flowchart showing a third example of the error processing operation in the server management method of the present invention.
This example shows a detailed example of the processing in step 121 in FIG. 1, and the status file 1 shown in FIG.
In step 8, if "NG" is recorded only in the server application state (steps 501 to 50)
4) In other words, if the network function is normal, the information of the status file 18 is converted to text and the system administrator is notified by e-mail operating independently of the server application (step 701). The application is shut down and restarted (step 505).

The status file 18 shown in FIG.
If "NG" is recorded as the error of the OS (steps 501 to 503), that is, if the network function is normal, the status file 18
Of the information as text, and notifies the system administrator by e-mail that operates independently of the server application (step 702).
If it is a T error or an error in the network function (steps 501 and 502), an error is displayed (step 50).
6).

By doing so, the server administrator can detect the phenomenon that could not be detected unless the server abnormality was checked before by the arrival of the e-mail, and quickly detect the occurrence of a failure. it can.

FIG. 8 is a flowchart showing a fourth error processing operation example in the server management method of the present invention.
This example shows a detailed example of the processing in step 121 in FIG. 1, and the status file 1 shown in FIG.
In step 8, if "NG" is recorded only in the server application state (steps 501 to 50)
4) Shut down and restart the server application (step 505), and if "NG" is recorded as an OS error (steps 501 to 5).
03), the main power switch of the server machine is turned off and on using the function of the UPS 1g in FIG. 2 to restart (step 801).

If "NG" is recorded in the status file 18 shown in FIG. 4 if the error is a POST error or a network function error (step 501,
502), an error is displayed (step 506).

By doing so, conventionally, the server administrator goes to the installation location of the server machine, judges that the system is locked, turns off the power of the server machine, and restarts it. Can be done That is, when the OS causes an error, the keyboard and mouse of the server machine usually do not operate, and the operator cannot perform any operation, so that the power supply is turned “off / on”. The power off / on control function of the server machine is used to automatically determine the power off / on of the server machine based on the error content.

As described above with reference to FIGS. 1 to 8, in the server management method of the present embodiment, it is determined whether not only the server machine itself and the network function but also the server application and the OS module are functioning normally. Record
A status file 18 that can be managed is provided. The self-diagnosis result of the machine body, the monitoring result of the network function by SNMP, the result of detecting the operation state of the server application, and the OS module used by the server application are recorded in the status file 18. If all the operation state detection results are “OK (normal)”, the system is started up normally, but one or more of them are different from the normal state (“NG”). )), It is determined as “error”. As a result, it is possible to quickly detect the abnormality of the server application that is operating normally at first glance.

When it is determined as “error”, it is determined which check result is abnormal. For example, when only the server application is abnormal,
Shut down and automatically restart the server application. As a result, it is possible to automatically solve the startup abnormality that often occurs when the server application is started.

When it is determined that an error has occurred, the contents of the “status file” are embedded in the logging database file of the server application.
When the database access function is operating, enable the client user to determine the status. This allows the server administrator or client user to
Even without going to the installation location of the server machine and using the console, the situation can be confirmed by accessing the logging file of the server application to be confirmed from the client machine.

If the determined “error” is a failure other than the network function such as a failure of the server application or the OS module, the server administrator is notified of the server application by using an e-mail function independent of the server application. Notify an error. As a result, it is possible to detect an event that has not been detected unless a server error has been seen in the past by the arrival of an e-mail.

If the determined “error” is a failure of the OS module used in the server application,
The main switch (power switch) of the server machine is turned on / off (restarted) using the function of the UPS.
As a result, it has become possible for the server administrator to automatically go to the installation location of the server machine, determine that the system is locked, and control the power of the server machine (off / on).

The present invention is not limited to the examples described with reference to FIGS. 1 to 8 and can be variously modified without departing from the gist thereof. For example, in this example,
Although specific processes in the error process are individually performed, they may be performed in combination with each other. That is, after the error processing (logging data setting) in step 602 in FIG. 6 and the error processing (main switch off / on) in step 801 in FIG.
The procedure (mail notification) in steps 701 and 702 in the above may be performed.

In this embodiment, the groupware server has been described as an example. However, the present invention can be applied to server management of a database server, a file server, and a mail server. In this example, a LAN is used for the network, but a WAN may be used.

In this example, the status file 18 shown in FIG. 4 is recorded as a file in a text format (character code), but any other format that can be read from a wide range of applications may be used. Instead of a character code, a bit flag, a byte flag, an image flag, or the like may be used.

In this example, the status file 18
Is embedded in the logging file of the running server application, but any logging file that can be easily accessed may be used. Further, in this example, a UPS is provided, and when the operating system is abnormal, the power of the server (the information processing apparatus 1d) is turned off and on using the function of the UPS.
Rebooting (restarting) using the function of the information processing apparatus 1d itself may be performed.

[0070]

According to the present invention, a failure detection of a server machine body by POST and a network monitoring protocol are provided.
(SNMP) along with network function failure detection,
Abnormality detection, such as whether all tasks necessary to start the server application are running, is also performed at system startup, and that the server machine itself is running, the network function is running, and By monitoring the functioning of the server application as a whole and detecting and managing failures in the entire system, it is possible to quickly detect and deal with abnormalities in seemingly normal server applications. ,
For example, it is possible to support efficient operation of a server system that integrally manages a network and a database, such as groupware.

In particular, the identification information of all functions necessary for normal operation of each of the hardware, the network, the application, and the operating system, which are the targets of the abnormality detection, is stored in a text format in advance. Time data, and the identification information of each function activated when the power switch is turned on is recorded as “current data”, and the “current data” is compared with the “normal data”. Otherwise, an abnormality is detected, so that the abnormality can be determined without being affected by the order of the function items.

Further, a response process corresponding to the detected fault content is automatically performed. For example, if the server application is abnormal, the server application is automatically restarted, so that a startup abnormality that often occurs when the server application is started can be automatically resolved.
It is possible to improve the reliability of the server system.

Further, by embedding the contents of the fault in the logging database file of the server application, the situation can be confirmed by accessing the logging file from the client machine, even if the server machine is not used. This eliminates the need for the user to go to the installation location of the server machine, and can reduce the work load.

If the failure is a failure other than the network function, the failure is notified by using the e-mail function, so that the failure can be easily detected by the server administrator and the response to the failure can be expedited. It is possible to

If the failure is a failure of the OS module used in the server application, the main switch (power switch) of the server machine is turned off and on, the server machine is restarted, and the system lock is automatically set. , It is possible to reduce the workload of the server administrator.

[Brief description of the drawings]

FIG. 1 is a flowchart illustrating an example of a processing operation according to a server management method of the present invention.

FIG. 2 is a block diagram illustrating a configuration example of a server machine that performs a processing operation according to the server management method of the present invention.

FIG. 3 is a block diagram illustrating a configuration example of a client-server system that performs the processing operation in FIG. 1;

FIG. 4 is an explanatory diagram showing a configuration example of a status file created by a server machine in FIG. 3;

FIG. 5 is a flowchart showing a first error processing operation example in the server management method of the present invention.

FIG. 6 is a flowchart showing a second error processing operation example in the server management method of the present invention.

FIG. 7 is a flowchart showing a third error processing operation example in the server management method of the present invention.

FIG. 8 is a flowchart showing a fourth error processing operation example in the server management method of the present invention.

FIG. 9 is a flowchart showing a conventional server management control example.

[Explanation of symbols]

1: server machine, 1a: display device, 1b: input device, 1c: external storage device, 1d: information processing device, 1e:
Optical disk, 1f: drive device, 1g: UPS (uninterruptible power supply), 2 to 4: client, 5: LAN, 10:
CPU, 11: main memory, 12: server management processing unit, 1
3: POST processing unit, 14: network test processing unit, 15: application test processing unit, 16: OS test processing unit, 17: result recording processing unit, 18: status file, 19: verification processing unit, 20: error processing unit .

Claims (6)

[Claims] 1. A server management method for detecting an abnormality of a server constituting a client-server system, wherein a hardware abnormality is detected by a self-diagnosis unit and a network function abnormality is detected by a network management unit when a power switch is turned on. Performing an error detection of application startup and an error detection of the operating system, recording a result of each detection, and determining an error if any one of the detection results is abnormal. Server management method. 2. The server management method according to claim 1, wherein all functions required for normal operation of each of the hardware and the network and the application and the operating system are associated with each other. The identification information is registered in a text format in advance, the identification information of each of the functions activated upon turning on the power switch is recorded, and the recorded identification information is compared with the previously registered identification information. A server management method characterized by detecting an error if they do not match. 3. The server management method according to claim 1, wherein when an error is determined, which of the detection results is abnormal is confirmed based on the record, and only the application is determined. A server management method characterized by shutting down and restarting the application if abnormal. 4. In the server management method according to any one of claims 1 to 3, when an error is determined,
A server management method, wherein each of the recorded detection results is written to a logging file. 5. In the server management method according to any one of claims 1 to 4, when an error is determined,
A server that checks which detection result is abnormal based on the record and, if the network function is normal, sends abnormality occurrence information to a preset mail address using an electronic mail. Management method. 6. In the server management method according to any one of claims 1 to 5, when an error is determined,
A server management method comprising: confirming which detection result is abnormal based on a record; and if the operating system is abnormal, performing a power switch off / on control and restarting the server.