JP2003263353A - Fault occurrence cause analytical device, fault occurrence cause analytical method, and fault occurrence cause analytical program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a fault occurrence cause analytical device, an analytical method and an analytical program capable of analyzing efficiently the fault occurrence cause. <P>SOLUTION: A processing device 11 has an access log recording part 11a for recording the content of an SQL sentence transmitted to a database management system program in an access log file 14 together with transmission source client information at the start time and the normal finish time of processing by the transmitted SQL sentence, an access log retrieval processing part 11b for retrieving and acquiring a log record having a strong possibility of a fault occurrence cause from the access log file 14, and a fault log record recording part 11c for recording the retrieved log record in a fault log record file 15. <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus for analyzing a cause of a failure, which identifies a process related to the failure by analyzing an access log from a client to the database when a failure occurs in the database. , Analysis method and analysis program.

[0002]

2. Description of the Related Art Conventionally, when investigating the cause of a failure occurring in a database, a standard log file of a database management system and a log file of an operating system (hereinafter referred to as OS) are first investigated. By analyzing the survey results obtained here, I was able to understand the nature of the failure occurring in the database. Further, in the conventional technique for analyzing the cause of the failure, when the failure occurs due to the SQL statement of the function processed in the database management system, the SQL statement of the function being processed is generated. Some have the function to output to the log. Conventionally, the standard log file of the database management system and O
It was possible to analyze the cause of failure by using the S log file.

[0003]

However, when the database management system is in a state in which it cannot be recorded in the log due to a hang-up of the database management system, etc. It was very difficult to identify the process that became the above from the log file.

Therefore, an object of the present invention is to perform processing when a failure occurs in the database management system, especially when the information related to the failure cannot be output to the log file. An object of the present invention is to provide a failure occurrence cause analysis apparatus, an analysis method, and an analysis program capable of identifying a SQL statement having a certain function and efficiently analyzing the cause of the failure occurrence based on the result.

[0005]

In order to achieve the above-mentioned object, a failure occurrence cause analysis apparatus of the present invention includes a processing device, a program storage section for storing a database management system program, and management by the database management system program. A database, an access log file for which log management is performed by the processing device, and a failure log record file for storing a log record that is likely to be the cause of failure, and the processing device
An access log recording unit for recording the content of the SQL statement transmitted to the database management system program in the access log file together with the sender client information at the start and normal end of the processing by the transmitted SQL statement; The access log file includes an access log search processing unit that searches and acquires a log record having a high possibility of causing a failure, and a failure log record recording unit that records the searched log record in the failure log record file. ing.

In addition, the access log search processing unit in the failure cause analysis apparatus of the present invention searches for a record in which "start" is recorded in the "state" item of the access log from the access log file, and then, “Start time”, “Client process ID” and “S” in the record
It is possible that the record that has the same content as that recorded in the "QL sentence" item and that is not recorded as "End" in the "Status" item is searched for in the access log file and that the record is the cause of failure. Is obtained as a high log record.

The method of analyzing the cause of failure according to the present invention is
From the access log file, search for the record in which "Start" is recorded in the "Status" item of the access log and retrieve the corresponding record, and the "start time" of the record obtained in the first process, "Client process I
Search the records in the access log file for the records that have the same contents as those recorded in the "D" and "SQL sentence" items and do not record "End" in the "Status" item, and find the cause of the failure. And a third step of storing the log record acquired in the second step in the failure log record file.

Also, the failure cause analysis program of the present invention records the contents of the SQL statement in the access log file together with the sender client information at the start and normal end of the processing by the sent SQL statement. A recording function, an access log search processing function that searches and acquires a log record that has a high possibility of causing a failure from the access log file, and a failure log record recording function that records the searched log record in the failure log record file. Is realized by a computer.

[0009]

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, an apparatus for analyzing a cause of failure in one embodiment of the present invention will be described with reference to FIG. FIG. 1 is a block diagram showing the configuration of a failure occurrence cause analysis device as a server according to an embodiment of the present invention.

<Configuration of Failure Cause Analysis Device> In FIG. 1, a server computer (failure cause analysis device) 1
0 and the client computers 20 and 21 are connected via a network 30 such as the Internet.
The server computer 10 includes a processing device 11 that performs various processing operations, a program storage unit 12 that stores a database management system program, a database 13 that is managed by the database management system program, and log management by the processing device 11. The access log file 14 for performing the above is provided, and the failure log record file 15 for storing the log record which is highly likely to be the cause of the failure as described later. In this embodiment, the two client computers 20, 2
Although one is connected, this number is not limited to two.

In the processing device 11, the contents of the SQL statement transmitted from the client computers 20 and 21 to the database management system program are transmitted to the processing source 11 at the start and the normal end of the processing by the transmitted SQL statement. Access log recording unit 11a for recording in the access log file 14 together with the transmission source client information such as the client process ID, and the access log file 1
4, an access log search processing unit 11b that searches for and acquires a log record that has a high possibility of causing a failure, and a failure log record recording unit 11c that records the searched log record in the failure log record file 15. Has been.

<Access Log Recording Format> The recording format of the access log recorded in the access log file 14 in the server computer 10 of the present embodiment having the above configuration will be described with reference to FIG. FIG. 2 is a diagram showing an example of a recording format of an access log stored in the access log file 14.

The recording format of the access log shown in FIG. 2 has items of "state", "start time", "end time", "client process ID", and "SQL sentence". Each item will be described below. The "state" item is an item indicating whether the corresponding log record is before or after the execution of the SQL statement function. This item includes "start" and "after" if the function is before execution. If so, “end” is recorded. In addition, the time when the processing of the corresponding SQL statement is started (the SQL statement is received) is recorded in the “start time” item, and the time when the processing of the SQL statement is normally ended is recorded in the “end time” item. It Further, the “client process ID” item records the process ID given to the client process that executes the process of the corresponding SQL statement, and the “SQL statement” item records the content of the SQL statement.

<Processing Operation in Normal Time> In the client-server system in which the failure cause analysis device having the above-described configuration is used as the server computer 10, the processing operation by the failure cause analysis device 10 in normal time is as follows. Has been done.

When the SQL statement is transmitted from the client computers 20 and 21 to the server computer 10 in the normal state, the client process ID of the transmission source at the start and the normal end of the processing by the transmitted SQL statement.
The access log recording unit 11a in the processing device 11 records the contents of the SQL sentence together with the transmission source client information such as “4.

<Processing Operation by the Present Device When a Failure Occurs> Next, the processing operation by the failure occurrence cause analysis apparatus 10 of the present embodiment when a failure occurs in the database management system will be described with reference to FIG. . Figure 3
3 is a flowchart showing a procedure for analyzing an access log having the configuration shown in FIG.

First, the access log search processing unit 11b in the processing device searches the access log file 14 for a record in which "start" is recorded in the "state" item of the access log (step 101). As a result of this search, if the applicable record does not exist in the access log file, this processing operation ends, and if the applicable record exists, the applicable record is acquired from the access log file 14 (step 102). , 103). next,
“Start time” of the record acquired in step 101,
A record in which the same contents as those recorded in the “client process ID” and “SQL sentence” items are recorded and “end” is recorded in the “state” item is set to the step 10
Similar to 1, the access log search processor 11b searches the access log file 14 (step 104).

As a result of the search, if the corresponding record exists, it is judged in the processing device 11 that the process related to the acquired record has been completed normally and is not the cause of the failure, and the step is restarted. The processing operations from 101 are sequentially performed. If there is no corresponding record, there is no “end” log record corresponding to the “start” log record recorded in the “state” item. It is judged to have abnormally ended. That is, since the process related to the acquired record is determined to be the process related to the fault, in other words, the process that is likely to be the cause of the fault, the fault log record recording unit 11c outputs the log record to the access log. The error log record file 15 is stored separately from the file 14 (steps 105 and 106).

After the processing operation up to step 106 is completed as described above, the processing operation from step 101 is executed again, and the processing is repeated for every log record in the access log file 14 one record at a time. . In this way, only the log record of the process that has abnormally ended during the process can be extracted and specified from the access log file 14, and the cause of the failure can be easily analyzed based on the extracted log record. Becomes
Specifically, for example, when the failure occurrence cause analysis apparatus 10 of the present invention performs an analysis process on an access log as shown in FIG. 2, it is the second from the top in FIG. Only the log record of is stored in the failure log record file 15. As a result, in the present embodiment, the second log record is the log record of the abnormally terminated process, and the process related to this log record can be specified as the cause of the failure. Further, even when a plurality of log records are stored in the failure log record file 15, the log records related to the processing that causes the failure occurrence are narrowed down, so that the failure cause analysis work can be performed efficiently.

<Other Embodiments> In the above embodiment,
Identify the SQL statement of the function that was being processed when the failure occurred,
Failure cause analysis device 10 capable of analyzing the cause of failure
The failure cause analysis procedure by the apparatus 10 has been described, but the means for facilitating the failure cause analysis operation is not limited to the failure cause analysis apparatus 10 described above, but the processing operation of the present invention. And a processing device 11 in the server computer 10
Alternatively, the processing operation may be executed based on this program.

Specifically, the client computer 2
An access log recording function that records the contents of the SQL statement in the access log file 14 together with the sender client information such as the client process ID of the sender at the start and normal end of the processing by the SQL statement sent from 0, 21. And an access log search processing function of searching and acquiring a log record having a high possibility of a failure occurrence from the access log file 14, and a failure log record recording function of recording the searched log record in the failure log record file 15. By storing a program that causes the processing apparatus to realize the processing apparatus in the program storage unit 12 and configuring the server computer 10 so that the processing apparatus 11 performs the processing operation of the present invention based on the program, as in the above-described embodiment. A log record of the processing that terminated abnormally during processing from the access log file 14. Can be identified removed Only in, it becomes possible to easily analyze the cause of the failure based on the retrieved log record.

[0022]

As described above, according to the present invention,
Even if the database management system falls into a hang-up state, it is possible to identify the SQL statement of the log record that was being processed when the fault occurred, and to efficiently analyze the cause of the fault. Becomes

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a failure occurrence cause analysis device as a server according to an embodiment of the present invention.

FIG. 2 is a diagram showing an example of a recording format of an access log stored in an access log storage unit according to the present invention.

FIG. 3 is a flowchart for explaining a procedure for analyzing an access log according to the present invention.

[Explanation of symbols]

10 ... Server computer (failure cause analysis device),
11 ... Processor, 11a ... Access log recording unit, 11b
Access log search processing unit, 11c ... Fault log record recording unit, 12 ... Program storage unit, 13 ... Database, 14 ... Access log file, 15 ... Fault log record file, 20, 21 ... Client computer, 30 ... Network.

Claims (4)

[Claims] 1. A processing device, a program storage section for storing a database management system program, a database managed by the database management system program, an access log file for log management by the processing device, and a cause of failure occurrence. And a failure log record file for storing a log record that is likely to be, wherein the processing device uses the contents of the SQL statement transmitted to the database management system program according to the transmitted SQL statement. An access log recording unit that records in the access log file together with sender information at the start and normal end of the process, and an access log search that retrieves and acquires a log record that is likely to cause a failure from the access log file Processing part and retrieved log record The failure cause analysis apparatus characterized by comprising: a failure log records recording unit for recording the fault log record file, a. 2. The access log search processing unit searches the access log file for a record in which "start" is recorded in the "status" item of the access log, and then, the "start time" and "client" of this record are searched. Process I
Search the records in the access log file for the records that have the same contents as those recorded in the "D" and "SQL sentence" items and do not record "End" in the "Status" item, and find the cause of the failure. The failure cause analysis apparatus according to claim 1, wherein the failure occurrence cause analysis apparatus acquires the log record as a log record having a high possibility. 3. A first step of searching for a record in which “start” is recorded in the “status” item of the access log from the access log file and acquiring the corresponding record, and a record of the record acquired in the first step. Search the access log file for a record in which the same contents as those recorded in the "start time", "client process ID", and "SQL sentence" items are recorded and "end" is not recorded in the "status" item. Then, it has a second step of acquiring the relevant record as a log record that is highly likely to be the cause of the failure and a third step of storing the log record acquired in the second step in the failure log record file. Characteristic failure cause analysis method. 4. An access log recording function for recording the contents of the SQL statement in an access log file together with the information of the sender at the start and normal end of processing by the sent SQL statement, and a failure occurs from the access log file. The feature is that the computer implements an access log search processing function that searches and acquires log records that are likely to be the cause and a failure log record recording function that records the searched log records in a failure log record file. Failure cause analysis program.