JP5516202B2 - Data monitoring program, data monitoring apparatus and data monitoring system - Google Patents

Description

  The present invention relates to a technique for monitoring data consistency between databases.

  Some computer systems maintain consistency between databases that are independently updated. For this reason, a technique is used for maintaining data consistency by a reflection process in which an update to data in one database is reflected in data in another database.

  As another technique for maintaining consistency, a monitoring device that monitors jobs executed on a computer monitors the actual operating status of jobs at regular intervals, and the operating status recognized by the monitoring status and the actual operating status. When there is a mismatch between the two, a technique for matching the two is used.

Japanese Patent Laid-Open No. 10-240600 JP-A-5-12205

  By the way, in the process of reflecting the update between databases, a failure may occur due to various factors, and the process may fail. For this reason, it is conceivable to monitor the consistency of data in each database to be reflected.

  However, the update reflection process between databases is not always performed immediately after the data is updated in one database. For this reason, there may be a time difference from when data is updated in one database until the update is reflected in data in another database. In this case, if the monitoring process immediately detects an error when the data inconsistency is detected, the data is inconsistent but the reflection process is not yet completed, i.e. There is a possibility that the error state is detected as an original error.

  In view of the above-described problems, the disclosed technique aims to suppress useless error detection by monitoring data consistency by distinguishing pseudo errors from original errors.

  In the disclosed technique, a plurality of databases, storage means for storing the relationship of data items of data stored in each of the plurality of databases, and a computer capable of accessing the first table include storage means. If the monitoring data corresponding to the data items related to each of the plurality of databases is extracted at each predetermined monitoring interval and the extracted monitoring data of each database does not match, the first table reads the previous data When the process of reading the monitoring data is executed and the previous monitoring data cannot be read from the first table, or when the previous monitoring data read and the extracted monitoring data do not match , Storing the extracted monitoring data in the first table, and reading the previous monitoring data and the extracted monitoring data; When you match, and outputs an error.

  According to the disclosed technology, since the monitoring is performed by distinguishing the pseudo error from the original error, it is avoided that the pseudo error is processed as the original error although it is in the pseudo error state. Therefore, useless error detection in the monitoring process is suppressed.

1 is an overall configuration diagram of an embodiment of a data monitoring system. It is explanatory drawing of one Example of the common item between DB. It is explanatory drawing of one Example of a reflection interval table. It is explanatory drawing of one Example of the monitoring item table. It is explanatory drawing of one Example of the monitoring status table. It is a flowchart of one Example of the process in a monitoring server. FIG. 6 is an explanatory diagram of a specific example of one example of a monitoring state, where (A) is a normal monitoring result, (B) is a normal monitoring result after a monitoring data inconsistency occurs, C) shows a case where the monitoring result is an error.

  FIG. 1 is an overall configuration diagram of an embodiment of a data monitoring system 10. The data monitoring system 10 includes an individual system 12 having a monitoring server 11 and a server including a database to be monitored (hereinafter referred to as DB). The data monitoring system 10 includes a component selection system 12A, a circuit mounting design system 12B, a test analysis system 12C, and a purchasing system 12D, which are systems used in each process such as product design, as a breakdown of the individual system 12. The monitoring server 11, the component selection system 12A, the circuit mounting design system 12B, the test analysis system 12C, and the purchase system 12D are connected to each other via a network. The network is, for example, a local area network (LAN) or a wide area network (WAN).

The monitoring server 11 is a computer including at least a CPU (Central Processing Unit) and a storage device such as a memory or a storage.
Each of the component selection system 12A, the circuit mounting design system 12B, the test analysis system 12C, and the purchase system 12D is a computer system that has at least one server and is operated independently. The server of the component selection system 12A includes a common component DB 120A. Similarly, the server of the circuit packaging design system 12B includes a CAD DB 120B. The server of the test analysis system 12C includes an analysis DB 120C. Further, the server of the purchase system 12D includes a purchase DB 120D. These DBs 120 </ b> A to 120 </ b> D hold data about articles that are objects such as product design for each article. Then, the data in the DBs 120A to 120D are updated independently in the individual systems 12A to 12D.

  Here, the DBs 120A to 120D include data indicating contents common to at least two DBs. Such data is data of common items that are specific items indicating common contents among the items included in the DBs 120A to 120D. The individual systems 12A to 12D cooperate with each other and associate common item data between the DBs. Further, the individual systems 12A to 12D perform a reflection process for reflecting the data of the common item updated in the DB of the own system in the data of the common item related to the same article in the DB of the other system at every predetermined reflection interval. . The reflection interval is an interval set in advance by a system administrator or the like according to the combination of the DB that is the reflection source and the DB that is the reflection destination in the reflection process.

  FIG. 2 shows a specific example of common items in the DBs 120A to 120D. For example, the common item “standardization flag” is included in the common component DB 120A, the CAD DB 120B, and the analysis DB 120C. Specifically, the item “handling classification” in the common parts DB 120A, CAD DB 120B, and analysis DB 120C is an item corresponding to the “standardization flag”, and indicates the content of the common item “standardization flag”. Here, for example, when the “handling classification” data in the common parts DB 120A is updated, the update is performed by the reflection processing performed at predetermined reflection intervals, and the “handling classification” data in the CAD DB 120B and the analysis DB 120C is updated. It is reflected in.

Hereinafter, the monitoring server 11 will be further described.
As shown in FIG. 1, the monitoring server 11 includes a monitoring interval determination unit 11A and a monitoring unit 11B. Furthermore, the monitoring server 11 includes a reflection interval table 11C, a monitoring item table 11D, and a monitoring state table 11E. In the monitoring server 11, a data monitoring program is executed by the CPU, and by cooperating with hardware devices such as a storage device, an input / output device, and a port for realizing communication, the monitoring interval determination unit 11A and the monitoring unit 11B Function is realized. The data monitoring program can be recorded on a computer-readable recording medium such as a magnetic tape, a magnetic disk, a CD-ROM, or a DVD-ROM. A system administrator or the like can execute the data monitoring program by installing the data monitoring program recorded on the recording medium in the monitoring server 11. The reflection interval table 11C, the monitoring item table 11D, and the monitoring state table 11E are stored in a storage device provided in the monitoring server 11.

  The monitoring interval determination unit 11A determines the monitoring interval for the DBs 120A to 120D based on the reflection interval at which the reflection process is performed between the DBs 120A to 120D. Note that the monitoring interval determination unit 11A is an example of a monitoring interval determination function and a monitoring interval determination unit.

  The monitoring unit 11B monitors the data consistency of the common items for the same article data stored in the DBs 120A to 120D for each monitoring interval determined by the monitoring interval determination unit 11A. The monitoring unit 11B is an example of a reference function, a determination function, a comparison function, an error processing function, and a storage function, or a reference unit, a determination unit, a comparison unit, an error processing unit, and a storage unit.

  The reflection interval table 11C is a table in which the reflection interval at which the reflection processing is performed between the DBs 120A to 120D is stored in accordance with the combination of the reflection source DB and the reflection destination DB. As shown in FIG. 3, the reflection interval table 11C includes a pattern ID that is an identifier of a reflection interval pattern that is uniquely determined according to the DB that is the reflection source in the reflection process, and reflection intervals between the DBs 120A to 120D. In the reflection interval column corresponding to the reflection source DB, information indicating the reflection source is stored, while in the reflection interval column corresponding to the reflection destination DB, from the reflection source DB. The time indicating the reflection interval of the reflection process is stored. Further, when a certain DB is a reflection source, information indicating that there is no reflection target is stored in the column of the reflection interval corresponding to the DB having no common item to be reflected. The data in the reflection interval table 11C is set in advance by a system administrator or the like. The reflection interval table 11C is an example of the second table.

  The monitoring item table 11D is a table in which common item names of the databases, that is, item names to be monitored are stored. As shown in FIG. 4, the monitoring item table 11D includes a common item name, a DB name in which the common item is stored, and a corresponding item name of the common item in each DB. The monitoring item table 11D indicates which DB contains the common item and plays a role of associating the common item name with the actual corresponding item name in each of the DBs 120A to 120D. The data of the monitoring item table 11D is set in advance by a system administrator or the like.

  The monitoring state table 11E is a table for temporarily storing the data of the DBs 120A to 120D for each article and for each common item. As shown in FIG. 5, the monitoring state table 11E includes an article number that is an identifier for uniquely identifying an article, a common item name, and data of the common item in the DBs 120A to 120D. Note that information indicating that there is no common item is stored in the data column corresponding to the DB not including the common item. The monitoring state table 11E is used to hold data of each DB at the previous monitoring time. The monitoring status table 11E is an example of the first table.

Hereinafter, the process in the monitoring server 11 is demonstrated using the flowchart shown in FIG.
The monitoring interval determination unit 11A reads reflection intervals between the DBs in all reflection interval patterns from the reflection interval table 11C. The monitoring interval determination unit 11A determines the longest reflection interval as the monitoring interval from the read reflection intervals (S1). In the case of the data example of the reflection interval table 11C in FIG. 3, the longest reflection interval is 30 minutes among the reflection intervals in which the common component DB 120A is the reflection source and the purchase DB 120D is the reflection destination. Therefore, in the present embodiment, the monitoring interval determination unit 11A sets 30 minutes as the monitoring interval.

The monitoring unit 11B selects one of the target articles. (S3).
Next, the monitoring unit 11B selects one target common item (S4).
The monitoring unit 11B reads the common item name and the DB including the common item among the DBs 120A to 120D (hereinafter referred to as the target DB in the description of this process) from the monitoring item table 11D. Furthermore, the monitoring unit 11B reads the corresponding item name corresponding to the common item among the items of each target DB from the monitoring item table 11D. Then, the monitoring unit 11B accesses each target DB based on the read common item name and corresponding item name, and refers to the common item data of each target DB as monitoring data (S5).

  Here, the monitoring unit 11B determines whether or not all the monitored monitoring data matches between the target DBs (S6). Then, when the monitoring data match all the target DBs as a result of the determination in S6 (Y), the monitoring unit 11B does not perform the processes in S7 to S10, and performs the process in S11. Migrate to

  On the other hand, if the monitoring data differs between the target DBs as a result of the determination in S5, the monitoring unit 11B (N) indicates that the monitoring data of the target DB at the previous monitoring for the common item currently being processed is the monitoring state table 11E. It is determined whether or not it is stored in (S7). If the previous monitoring data is not stored as a result of the determination in S7 (N), the monitoring unit 11B proceeds to the process in S9. If the previous monitoring data is not stored in the monitoring state table 11E as described above, the monitoring data are all consistent between the target DBs at the time of the previous monitoring, and the monitoring result is normal. On the other hand, if the previous monitoring data is stored as a result of the determination in S7 (Y), the monitoring unit 11B reads the previous monitoring data. The monitoring unit 11B determines whether or not the previous monitoring data read out matches the monitoring data referred to in S5 in all target DBs (S8).

  If the result of determination in S8 is that the previous monitoring data differs from the monitoring data referenced in S4 in any of the target DBs (N), the monitoring unit 11B proceeds to the processing in S9. That is, the monitoring unit 11B stores the monitoring data referred to in S5 as a warning state in the monitoring state table 11E in association with the common item name (S9). At this time, if the previous monitoring data is already stored in the monitoring state table 11E, the monitoring unit 11B deletes the previous monitoring data and stores the monitoring data referred to in S5 instead.

  On the other hand, when the result of determination in S8 is that the previous monitoring data matches the monitoring data referenced in S5 in all the target DBs (Y), the monitoring unit 11B has a failure in the reflection process between the target DBs. As a result, the monitoring result is regarded as an error (S10).

  Here, the monitoring unit 11B determines whether all common items have been processed (S11). If the result of determination in S10 is that there is a common item that has not yet been processed (N), the monitoring unit 11B returns to S4 and proceeds to processing for the next common item. On the other hand, when the process is performed for all common items (Y), the monitoring unit 11B further determines whether the process is performed for the data of all articles (S12). If the result of determination in S12 is that processing has not been performed on all the article data (N), the monitoring unit 11B returns to S3 and proceeds to processing of the next article data. On the other hand, when all the articles are processed (Y), the monitoring unit 11B returns to S2 and starts the monitoring process again.

  Next, the monitoring process according to the content of common item data in the common parts DB 120A, CAD DB 120B, analysis DB 120C, and purchase DB 120D will be described with a specific example.

  FIG. 7A is an example of “component shape” data of article number 1 stored in each target DB including the common item “component shape” at the time of a certain monitoring process. The common item “component shape” is included in the common component DB 120A and the CAD DB 120B as shown in the data example of the monitoring item table 11D in FIG. The “component shape type” in the common component DB 120A and the “component shape type name” in the CAD DB 120B correspond to the common item “component shape”. For this reason, the data of these items in each target DB are set as monitoring data.

  Here, as shown in FIG. 7A, the data in the common parts DB 120A and the CAD DB 120B are both “not set” and match. That is, in this case, the monitoring data matches in all target DBs. Therefore, the monitoring unit 11B processes the monitoring result as normal.

  On the other hand, FIG. 7B is an example of the “regulation flag” data of the article number 2 stored in each target DB including the common item “regulation flag” during a certain monitoring process. The common item “restriction flag” is included in the common component DB 120A, the CAD DB 120B, and the analysis DB 120C, as indicated by the data in the monitoring item table 11D of FIG. The “RoHS product number determination flag” in the common parts DB 120A, CAD DB 120B, and analysis DB 120C corresponds to the common item “restriction flag”. For this reason, the data of these items in each target DB are set as monitoring data.

  Here, in the first monitoring, the data in the common parts DB 120A is “conforming”, while the data in the CAD DB 120B and the analysis DB 120C are “in process”, and the data does not match between the DBs. That is, in this case, the monitoring data is different between the target DBs. For this reason, the monitoring unit 11B stores the monitoring data in the monitoring state table 11E as a warning state.

  Furthermore, in the second monitoring, all of the data of the common parts DB 120A, the CAD DB 120B, and the analysis DB 120C are “conforming” and are the same. That is, in this case, the monitoring data matches in all target DBs. For this reason, the monitoring unit 11B processes the monitoring result as normal in the second monitoring. At this time, the monitoring unit 11 deletes the monitoring data stored in the monitoring state table 11E in the first monitoring.

  FIG. 7C is an example of “purchase price” data of the article number 3 stored in each target DB including the common item “purchase price” during a certain monitoring process. The common item “purchase price” is included in the common parts DB 120A, the CAD DB 120B, and the purchase DB 120D, as indicated by the data in the monitoring item table 11D of FIG. The “part unit price (latest update month)” in the common parts DB 120A and the “part unit price” in the CAD DB 120B and the purchase DB 120D correspond to the common item “purchase price”. For this reason, the data of these items in each DB are set as monitoring data.

  Here, in the first monitoring, the data of the common parts DB 120A and the CAD DB 120B is “not set”, while the data of the purchase DB 120D is “800 yen”, and the value of the monitoring data differs between the target DBs. ing. However, in this monitoring, the warning state is the first time, and the previous monitoring data is not stored in the monitoring state table 11E. For this reason, the monitoring unit 11B stores the monitoring data in the monitoring state table 11E without making the monitoring result an error.

  Next, in the second monitoring, the data in the CAD DB 120B is “not set”, while the data in the common parts DB 120A and the purchase DB 120D is “800 yen”, and the values of the monitoring data are different between the target DBs. ing. However, in comparison with the previous monitoring data stored in the monitoring status table 11E at the previous monitoring, the data in the common parts DB 120A does not match between “not set” and “800 yen”. That is, in this case, the monitoring data does not match between the target DBs, but the previous monitoring data and the current monitoring data are changed. Therefore, the monitoring unit 11B does not make the monitoring result an error, and stores the current monitoring data in the monitoring state table 11E instead of the previous monitoring data.

  Further, in the third monitoring, the data in the CAD DB 120B is “not set”, but the data in the common parts DB 120A and the purchase DB 120D is “800 yen”, and the value of the monitoring data differs between the target DBs. ing. This monitoring data also matches the data of all target DBs in comparison with the previous monitoring data stored in the monitoring status table 11E at the previous monitoring. For this reason, the monitoring unit 11B sets the monitoring result as an error. At this time, the monitoring unit 11 deletes the monitoring data stored in the monitoring state table 11E during the previous monitoring.

  According to the monitoring server 11, the monitoring unit 11 </ b> B performs the following process when monitoring the consistency of data of common items between DBs. That is, the monitoring unit 11B, when there is inconsistency in the monitoring data, when the previous monitoring data is not stored in the monitoring state table 11E, that is, when there is no inconsistency during the previous monitoring, Do not immediately set the monitoring result as an error. In addition, the monitoring unit 11B performs monitoring if there is a change in the monitoring data of any DB from the previous monitoring time even if the monitoring data is inconsistent continuously with the previous monitoring time. The result is not an error. In this way, when the monitoring data has changed from the previous monitoring time, it is in the middle of the reflection process, and the monitoring data is inconsistent but the reflection process is simply incomplete That is, there is a high possibility of a pseudo error state that is not an error. On the other hand, if the monitoring data is inconsistent and the monitoring data has not changed in all DBs in comparison with the previous monitoring, the monitoring unit 11B has failed in the reflection process. It is determined that there is an error, and the monitoring result is regarded as an error. In this way, the monitoring unit 11B can process the state of a pseudo error that is not an original error while distinguishing it from the original error. Therefore, the monitoring unit 11B avoids detecting the status of a pseudo error occurring during that time as an original error even if there is a time difference until an update to a DB is reflected in another DB. Can do.

  Further, the monitoring interval determination unit 11A determines a time longer than the longest interval among the data reflection intervals between the DBs as the monitoring interval. For this reason, when data is updated in a certain DB, even if the reflection process is not completed at the next monitoring time, the data update is reflected in all the DBs to which the data update should be reflected at the latest at the latest monitoring time. On the other hand, a reflection process is performed. Since the monitoring unit 11B performs the monitoring process with a sufficient monitoring interval in this way, it is possible to avoid the monitoring result as an error in a pseudo error state with a higher probability.

  In the case of a system environment in which computational resources and network resources are limited, it is not desirable to perform reflection processing frequently because it affects the original business processing. Further, depending on the contents of data to be reflected, there is a case where there is no problem in business even if the reflection process is not necessarily performed immediately. In such a case, since the reflection interval is generally set to be long, the time from the data update to the reflection processing becomes long, and the state of the pseudo error is particularly frequent. However, since the monitoring mechanism of the monitoring server 11 performs monitoring at a monitoring interval longer than the reflection interval, as described above, even if the pseudo error state frequently occurs as described above, the monitoring result is regarded as an error. It can be avoided.

  Here, the monitoring interval determination unit 11A can determine the monitoring interval based on not only the reflection interval but also other factors. For example, the monitoring interval determination unit 11A may determine a time longer than the reflection time as a monitoring interval by referring to a table storing the reflection time required to reflect data between DBs. In this way, for example, if a system administrator or the like sets a sufficient reflection time in the table assuming a batch update of a large amount of data, even if the reflection processing itself takes a long time. Thus, it is possible to avoid the detection of a pseudo error during the reflection process as an original error.

  Furthermore, the monitoring unit 11B may perform monitoring processing at a monitoring interval that is set in advance in the monitoring server 11 by a system administrator or the like. In this case, by setting a sufficient monitoring interval by the system administrator or the like, it is possible to obtain the same effect as when the monitoring interval determining unit 11A determines the monitoring interval as described above. In this case, the monitoring interval determining unit 11A does not need to determine the monitoring interval.

  The monitoring mechanism of the monitoring server 11 is also applied to monitoring data consistency between DBs in a load balancing system or the like in which a plurality of servers have DBs having the same structure and perform synchronization processing with each other. be able to. In such a case, the entire data record is set as monitoring data, not only the data of the common item that is a specific item among the items of each DB.

  Further, in the data monitoring system 10, the monitoring mechanism of the monitoring server 11 monitors the consistency between DBs updated in the servers of different individual systems 12. However, the monitoring mechanism can be applied to any DB as long as it is a plurality of DBs subject to data reflection processing, regardless of the system configuration. For example, the monitoring mechanism can also be applied to monitoring consistency between DBs provided in different servers in the same system. In addition, since the reflection interval is generally short between DBs included in the same server, a pseudo error state is unlikely to occur. However, the monitoring mechanism can also be applied to monitoring consistency between DBs.

The monitoring mechanism is not limited to a system used in each process such as product design as described above, but can be applied to a DB of a system used for any business.
Further, the monitoring mechanism of the monitoring server 11 is not a separate device independent of the system including the monitoring target DB, but may be implemented on the server side included in the system including the monitoring target DB, for example.

  Further, the monitoring unit 11B can notify the system administrator of the error as a process when the monitoring result is an error. The monitoring unit 11B can use any method as an error notification method. For example, the monitoring unit 11B can send an error content to the system administrator by e-mail. The monitoring unit 11B can also display error contents for a display device such as a display connected to the monitoring server 11 or print the error contents on a paper medium or the like. Furthermore, the monitoring unit 11B may accumulate the content of the error or the content of the warning state in the storage device as an error log. By accumulating the error log in this way, the system administrator can use the error log to determine the occurrence tendency of errors, such as the occurrence of errors and warning conditions frequently occurring in the data reflection process between DBs. It becomes possible to analyze.

DESCRIPTION OF SYMBOLS 10 Data monitoring system 11 Monitoring server 11A Monitoring interval determination part 11B Monitoring part 11C Reflection interval table 11D Monitoring item table 11E Monitoring state table 12 Individual system 12A Component selection system 120A Common component DB
12B Circuit packaging design system 120B CAD DB
12C test analysis system 120C analysis DB
12D Purchasing System 120D Purchasing DB

Claims (6)

A plurality of databases, storage means for storing a relationship between data items stored in each of the plurality of databases, and a computer capable of accessing the first table;
An extraction function for extracting monitoring data corresponding to data items related to each of the plurality of databases at a predetermined monitoring interval with reference to the storage means;
When the monitoring data of each database extracted by the extraction function does not match, a read function for executing a process of reading the previous monitoring data from the first table;
When the previous monitoring data could not be read from the first table by the reading function, or when the previous monitoring data read and the monitoring data extracted by the extraction function do not match, A control function for storing the extracted monitoring data in the first table and outputting an error when the previous monitoring data read out and the extracted monitoring data match;
A data monitoring program characterized by realizing the above. The plurality of databases perform a process of reflecting data of other databases at a reflection time interval defined in each database for each data item stored in the storage unit in each database. ,
The data monitoring program according to claim 1, wherein the predetermined monitoring interval is an interval equal to or greater than a reflection interval defined in each of the databases.   Read the reflection intervals in all combinations from the second table in which the reflection intervals corresponding to the combination of the database that is the reflection source and the database that is the reflection destination are stored in the reflection process, and among the read reflection intervals 3. The data monitoring program according to claim 2, further causing the computer to realize a monitoring interval determining function for determining an interval equal to or longer than the longest reflection interval as a monitoring interval.   The data monitoring program according to any one of claims 1 to 3, wherein the related data items are common items indicating common contents among items of the plurality of databases. A plurality of databases, storage means for storing the relationship of data items of data stored in each of the plurality of databases, and a data monitoring device capable of accessing the first table,
Extracting means for extracting monitoring data corresponding to data items related to each of the plurality of databases at a predetermined monitoring interval with reference to the storage means;
When the monitoring data from each of the databases extracted by the extraction unit does not match, a reading unit that executes a process of reading the previous monitoring data from the first table;
When the previous monitoring data could not be read from the first table by the reading unit, or when the previous monitoring data read out and the monitoring data extracted by the extracting unit do not match, Control means for storing the extracted monitoring data in the first table and outputting an error when the previous monitoring data read out and the extracted monitoring data match;
A data monitoring device comprising: In each of the data items stored in the storage means, a plurality of databases that perform processing of reflecting data of other databases at a reflection time interval defined by each,
Storage means for storing the relationship of data items of data stored in each of the plurality of databases;
Extracting means for extracting monitoring data corresponding to data items related to each of the plurality of databases at a predetermined monitoring interval with reference to the storage means;
If the monitoring data of the databases extracted by the extraction means do not match, a reading means for executing a process of reading the previous monitoring data from the first table;
When the previous monitoring data could not be read from the first table by the reading unit, or when the previous monitoring data read out and the monitoring data extracted by the extracting unit do not match, Control means for storing the extracted monitoring data in the first table and outputting an error when the previous monitoring data read out and the extracted monitoring data match;
A data monitoring system comprising:

Images