JP3309198B2 - Database multiplexing method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to a database multiplexing method using a main database and a duplicate database.

[0002]

2. Description of the Related Art Conventionally, this type of database multiplexing method includes a main database and a duplicate database, and accumulates data in the main database. The data stored in the database is regularly copied to a duplicate database. In addition, a history of operations performed when data is stored in the main database is recorded in an operation log. The replica database (backup data of the main database at a certain point in time) and the current operation log of the main database can be used even if the device that holds the main database fails.
It is backed up by hardware so that it can always be read. When a failure occurs, the backup data of the main database at a certain point in time is read from the duplicate database, the operation after that point is read from the operation log, and the operation is re-executed to completely restore the same thing as the main database. This operation is made possible by dual hardware and a large-capacity secondary recording device.

[0003]

However, such a conventional database multiplexing method requires double hardware and a large-capacity secondary recording device, resulting in high cost. That is, in a business such as a financial business, when a failure occurs, it is necessary to completely restore the same database as the main database, and even if the cost becomes high, it is unavoidable. However, in a system such as a building management system whose main purpose is monitoring, it is more important that data can be continuously monitored and maintenance is easier than data loss. In such a system, when the above-described database multiplexing method is applied, the main database is completely restored even though some data may be lost, resulting in excessive quality and cost. Up.

The present invention has been made to solve such a problem, and an object of the present invention is to provide a system for monitoring a building management system or the like, which has dual hardware and a large capacity. It is an object of the present invention to provide a database multiplexing method which does not require a secondary recording device and can reduce costs.

[0005]

In order to achieve the above object, according to the present invention , data is stored in a main database while the data is stored in the main database between a previous copy cycle and a current copy cycle. In a database multiplexing method in which accumulated data is periodically copied to a duplicate database, if a failure occurs in the device that holds the main database, the duplicate database is used as a new main database, while a failure has occurred since the previous copy. Until it happens
Data accumulated in the old primary database during
Old main data from which mismatched data is removed
The database is reused as a duplicate database .

[0006]

[0007]

According to the present invention, data newly stored in the main database is copied to the duplicate database at regular intervals. When a failure occurs in the device holding the main database, the duplicate database is used as a new main database, and the system is continuously operated. In this case, the data stored in the old main database (main database held by the failed device) between the previous copy and the occurrence of the failure is not included in the new main database, but the copy cycle , The amount of data loss can be reduced. In this case, the data accumulated in the old main database from the previous copy to the occurrence of the failure is regarded as inconsistent data, and the old main database from which the inconsistent data has been removed is reused as the duplicate database.

[0009]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, the present invention will be described in detail based on embodiments. FIG. 2 is a system configuration diagram of a building management system to which the present invention is applied. In the figure,
1, 2 and 3 are main control units (MCU: hereinafter referred to as monitoring devices), 4 is an area controller master (ACM), and 5-1 and 5-2 are area controllers (A
IC), 6-1 to 6-6 are remote stations (I-
DGP), the monitoring devices 1, 2 and 3 provide an operator interface, the area controller master 4 provides a gateway function, and the area controller 5-
1, 5-2 perform control in the area, and the remote stations 6-1 to 6-6 perform control on a device basis.

In this building management system, the monitoring apparatuses 1, 2, and 3 have a monitoring application (hereinafter, referred to as a monitoring AP) for supporting a monitoring operation. FIG. 1 shows a system configuration of the monitoring apparatuses 1, 2, and 3. Monitoring device 1,
2, 3 are monitoring APs 1-1, 2-1 and 3-1; switching processing units 1-2; 2-2; 3-2;
-3, 2-3, 3-3 and database trigger 1-4,
2-4, 3-4 and change logs 1-5, 2-5, 3-5
And system revision tables 1-6, 2-6, 3-6, revision tables 1-7, 2-7, 3-7,
It has module status monitoring units 1-8, 2-8, and 3-8. In this embodiment, the database 1-3 is currently used as the main database and the database 2-
It is assumed that 3, 3-3 is used as a duplicate database. Although not shown in the figure, the monitoring devices 1 and
2 and 3 have CPUs.

The monitoring APs 1-1, 2-1 and 3-1 support a monitoring operation performed by the respective monitoring devices 1, 2 and 3.
The switching processing unit 1-2, 2-2, 3-2 is the monitoring AP-1.
-Relay the update / selection request from -1, 2-1 and 3-1 to the main database 1-3. The main database 1-3 stores data for monitoring operations. Duplicate database 2-
3, 3-3 stores the duplicate data of the main database 1-3 at a certain point in time. Database trigger 1-4, 2-4
In step 3-4, an updated location for the database 1-3, 2-3, 3-3 is written. Change log 1-5, 2-5, 3-5
Holds the history of the changed part of the database 1-3, 2-3, 3-3. System revision table 1-6, 2
-6 and 3-6 hold the number of times the main database is switched as the system revision. Revision table 1-7,
2-7, 3-7 are databases 1-3, 2-3, 3-3
Stores revision indicating version of. The module status monitoring units 1-8, 2-8, 3-8 are monitoring devices 1, 2,
Monitor if 3 is working properly.

Next, the characteristic operation (backup logic) of the building management system will be described with reference to the functions of each unit in the monitoring devices 1, 2, and 3.

[Accumulation of Update Contents from Monitor AP] Monitor A
The update content from P • 1-1 is sent to the main database 1-3 by the switching processing unit 1-2. Monitoring AP 2-1
Are updated by the switching processing unit 2-2 to the main database 1-3. The contents of updates from the monitoring AP 3-1 are sent to the main database
Sent to 3. Thereby, in the main database 1-3,
The data from the monitoring APs 1-1, 2-1 and 3-1 are updated and accumulated.

When data is updated and stored in the main database 1-3, the database trigger 1-4 is updated every time the data is updated.
Is started, and the update location is described in the change log 1-5 using the table name and the ROWID (row). In this case, if the revision (the current revision of the main database 1-3) in the entry of the revision table 1-7 specified by the system revision in the system revision table 1-6 is N, the change part is N + 1. It is described in the entry of the specified change log 1-5.

[0015] For example, the changed part is described as "200 in the table".
3, if the system revision in the system revision table 1-6 is "2" as shown in FIG. 3, the system revision "2" designates the entry "2" of the revision table 1-7. The revision “200” in the entry “2” is read as the current revision N of the main database 1-3, and N +
The entry of the change log 1-5 designated by 1, ie, 2
The changed part in the 01th entry is “20 in the table”.
0 line ". When the changed part is described in the change log 1-5, the revision in the entry “2” of the revision table 1-7 is incremented to “20”.
1 ".

The entries of the change log 1-5 are used cyclically, and when the current revision of the main database 1-3 indicates the maximum entry of the change log 1-5, the change log 1-5 is used. The changed part is described in the first entry.

[Copy Process to Replica Database] FIG. 4
2 illustrates a state of the monitoring device 1 and the monitoring device 2 before the copy processing is performed. In this example, the main database 1-
3 and a copy database 2-3 have a table as a data update storage location. Further, the current revision of the main database 1-3 in the monitoring device 1 is “201”, and the current revision of the duplicate database 2-3 in the monitoring device 2 is “199”.
And In addition, it is assumed that the main database 1-3 and the duplicate database 2-3 match by the previous copy process except for the 200th line of the table and the 1500th line of the table.

When a certain period of time has elapsed since the previous copy processing, the CPU of the monitoring apparatus 2 starts a copy processing program, and the copy processing program is updated and stored in the main database 1-3 from the previous copy cycle to the current copy cycle. Copy the data to the duplicate database 2-3. That is, the monitoring device 2
CPU compares the current revision of the main database 1-3 with the current revision of the duplicate database 2-3, and the current revision of the main database 1-3 is ahead of the current revision of the duplicate database 2-3. Is copied to the replication database 2-3 as one transaction.

In FIG. 4, the current revision of the main database 1-3 is "201" and the duplicate database 2-3 is
Is "199", and the current revision of the main database 1-3 is "200", "201".
Of the change log 1-5, the contents of the “1500th line of the table” indicated by the entry “200” and the entry “20 of the change log 1-5”
The contents of the “line 200 of the table” indicated by “1” are copied to the “line 1500 of the table”
And "line 200 of the table". Reading from the main database 1-3 at the time of copying is performed by a method (database function) in which the table contents are retained at the contents at the time of starting the transaction of the copy.

The contents of the current revision of the duplicate database 2-3 and the contents of the change log 2-5 are updated by updating the contents of the duplicate database 2-3 by the copy process, and the database trigger 2-4 is activated. The update is performed using the same logic as in the case of “accumulation of update content from monitoring AP”. As a result, the revision in the entry “2” of the revision table 2-7 becomes “201”, and the change log 2
-5 entry "200" contains "150
The “line 0” is described in the entry “201” of the change log 2-5.

Although the copy process of the monitoring device 2 to the copy database 2-3 has been described here, the copy process is similarly performed to the copy database 3-3 of the monitoring device 3. In this case, the CPU of the monitoring device 3 performs a copy process to the replication database 3-3, and the monitoring devices 2, 3
By causing the CPU to perform the copy process, the load on the CPU in the monitoring device 1 having the main database 1-3 can be reduced.

[Processing when a failure occurs in the monitoring device holding the main database] When a failure occurs in the monitoring device 1 holding the main database 1-3, the module status monitoring units 1-8 of the monitoring devices 1, 2, and 3 , 2-8, and 3-8 detect an abnormality, and the system with the most advanced revision among the duplicate databases 2-3 and 3-3 is set as a new main database, and the system is upgraded.

For example, if the revision of the copy database 2-3 is the most advanced, the copy database 2-3
Is the new primary database. As a result,
The new main database 2-3 stores the monitoring APs 1-1 and 1-2.
The data from -1 and 3-1 are updated and accumulated, and the system is operated continuously.

Here, when a failure occurs in the monitoring device 1,
The current revision of the main database 1-3 is "20
3 "and the current revision of the duplicate database 2-3 is" 201 ", the update data corresponding to the revisions" 202 "and" 203 "of the main database 1-3 are included in the duplicate database 2-3. Not. That is, when the duplicate database 2-3 is the new main database 2-3, the update data corresponding to the revisions "202" and "203" of the old main database 1-3 is lost.

However, in this building management system, it is more important that data can be continuously monitored and that maintenance is easy than data loss. Therefore, the revision “20” of the old main database 1-3 is
Even if the update data corresponding to “2” and “203” is lost,
Not much problem. For example, historical data (daily / monthly reports, meter readings, etc.) is guaranteed by the ACM by vertical dispersion, so that a loss due to this mechanism does not result in missing data. In addition, a normal operator operation (change of schedule or change of set value of temperature) is not canceled due to loss in this mechanism because the controller has master data. The amount of data loss can be reduced by shortening the copy cycle.
Can be reduced.

In the system revision up process, the monitoring devices other than the monitoring device 1 having the old main database 1-3 in which the failure has occurred, that is, the monitoring devices 2 and 3,
The system revision in the system revision table 2-6, 3-6 is incremented, and the current revision of the duplicate database 2-3, 3-3 at that time is updated to the revision table 2- specified by the incremented new system revision. Copy to the entry in 7, 3-7.

This system revision up processing will be described by taking the monitoring device 2 as an example. As shown in FIG. 5, the system revision in the system revision table 2-6 is incremented to "3", and this incremented value is set. Entry "3" of revision table 2-7 specified by new system revision "3"
The current revision “201” of the duplicate database 2-3 at that time is copied.

The revision table 2-7 (3-
The entry of 7) is used cyclically, and when the current system revision indicates the maximum entry of the revision table 2-7 (3-7), the revision table 2 is used.
-7 (3-7) copies the current revision of the duplicate database 2-3 (3-3) to the first entry.

[Process for Reusing Main Database]
When a failure occurs in the monitoring device 1 having the main database 1-3, the monitoring device 1 is automatically restarted and the monitoring function is restored as described above. At this time, for example, the replication database 2
-3 is the new main database, and the system has already started operation. At this time, a new main database 2-
In No. 3, update data after the occurrence of the failure is added to data up to the copy processing immediately before the occurrence of the failure in the old main database 1-3. That is, the updated data after the occurrence of the failure is added to the new main database 2-3 without including the data updated and accumulated in the old main database 1-3 from the previous copy to the occurrence of the failure. .

Here, in order to reuse the old main database 1-3 as a duplicate database, the old main database 1-3 is copied from the previous copy until the failure occurs.
After the data accumulated and updated in the new main database 2-3 is removed as unmatched data (hereinafter, this processing is referred to as data matching processing), the updated data after the occurrence of the failure that has been updated and accumulated in the new main database 2-3 is replaced with the old main database 1-3 Need to be copied to

[Data Matching Process] The data matching process will be specifically described with reference to FIGS. FIG. 6 shows a state of the monitoring device 1 and the monitoring device 2 immediately before the occurrence of the failure. The 200th line of the table of the main database 1-3 and the 1500th line of the table have not yet been copied to the duplicate database 2-3.

FIG. 7 shows a state where a failure has occurred in the monitoring apparatus 1. The database 2-3 of the monitoring device 2 becomes a new main database, and the system revision of the system revision table 2-6 is set to "3" by the system version-up processing, and the database 2-3 is entered in the entry "3" of the revision table 2-7. Has been copied.

FIG. 8 shows a state in which the monitoring apparatus 1 which has been down due to a fault has returned. In this state, the new main database 2-3 is different from the old main database 1-3 in the 200th line of the table and the 1500th line of the table. The data matching process recovers this state.

In this data matching process, the system revision “2” in the system revision 1-6 of the monitoring device 1 is used.
, The revision “199” in the entry “2” of the revision table 2-7 in the monitoring device 2 designated by the system revision “2” is read, and the revision “1” is added to the revision “199”. 20
From “0” to the current revision “201” of the old main database 1-3, “the 1500th row of the table” and the “200th row of the table” described in the change log 1-5 are changed to the new main database 2 -3 overwrites the old main database 1-3.

After performing the data matching process in this way, the monitoring device 1 performs a system revision up process, and reuses the old main database 1-3 as a duplicate database. That is, in the monitoring device 1, the system revision in the system revision table 1-6 is incremented, and the revision "1" at the time of the previous copy is set.
99 "into the entry" 3 "in the revision table 1-7 designated by the new system revision" 3 ".

In this embodiment, a revision and a system revision are used cyclically. For this reason, it is not possible to judge that the revision or the system revision is “advanced” just because the numerical value is large. Therefore, in this embodiment, it is assumed that the revision and the system revision are not separated by more than half a lap, and if the revision is more than half a lap, it is determined that the revision is late.

[0037]

As is apparent from the above description, according to the present invention, if data newly stored in the main database is copied to the duplicate database at regular intervals and a failure occurs in the device holding the main database, The replicated database will be used as the new primary database. In this case, the data stored in the old primary database between the previous copy and the failure will not be included in the new primary database, but will be copied. By shortening the cycle, it is possible to reduce the amount of data loss , and failures have occurred since the last copy
Data stored in the old primary database until
Is the mismatched data, and the old
The primary database is reused as a duplicate database
In a system such as a building management system whose main purpose is monitoring, redundant hardware and a large-capacity secondary recording device are not required, and costs can be reduced.

[Brief description of the drawings]

FIG. 1 is a diagram showing a system configuration of each monitoring device in the building management system shown in FIG.

FIG. 2 is a system configuration diagram of a building management system to which the present invention is applied.

FIG. 3 is a diagram for explaining accumulation of update contents from a monitoring AP;

FIG. 4 is a diagram illustrating a copy process to a duplicate database;

FIG. 5 is a diagram for explaining a system revision up process;

FIG. 6 is a diagram for explaining data matching processing (immediately before a failure occurs).

FIG. 7 is a diagram for explaining data matching processing (when a failure occurs).

FIG. 8 is a diagram for explaining data matching processing (at the time of failure recovery).

[Explanation of symbols]

1, 2, 3 ... monitoring device, 1-1, 2-1, 3-1 ... monitoring AP, 1-2, 2-2, 3-2 ... switching processing unit, 1-
3,2-3,3-3 ... database, 1-4,2-4
3-4: Database trigger, 1-5, 2-5, 3-5
... Change log, 1-6, 2-6, 3-6 ... System revision table, 1-7, 2-7, 3-7 ... Revision table, 1-8, 2-8, 3-8 ... Module status monitoring Department.

────────────────────────────────────────────────── ─── Continuation of the front page (56) References JP-A-3-250257 (JP, A) JP-A-5-307576 (JP, A) JP-A-7-262068 (JP, A) JP-A-5-262068 73393 (JP, A) (58) Field surveyed (Int. Cl. ⁷ , DB name) G06F 12/00 531 G06F 12/16 310

Claims (1)

(57) [Claims] 1. A database multiplexing method for storing data in a main database and periodically copying data stored in the main database between a previous copy cycle and a current copy cycle to a duplicate database. In the case where a failure occurs in an apparatus holding the main database, the duplicate database is used as a new main database while the failure occurs after the previous copy.
Data stored in the old main database until
Data before and after removing this mismatched data.
Reuse old master database as duplicate database
Database multiplexing method characterized by characterized by the the to.