JPH0589060A - Data base input output processing system - Google Patents

Abstract

PURPOSE:To solve the complexity of a file control at the time of moving from individual operation to duplex operation to occur by providing a data base at respective computers at a loose coupling duplex computer system. CONSTITUTION:At a data base processor 3, a buffer memory 32 is provided. At the time of writing, the coincidence of data outputted from computers 1 and 2 is confirmed by using the buffer memory 32, and after confirmation, written into a data base file 31. At the time of reading, to the computer requesting earlier, the reading data from the data base file 31 are transferred, stored in the buffer memory 32, and next, to the computer requesting later, the content of the buffer memory 32 is transferred.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a database input / output processing system, and more particularly to a database input / output processing system for a loosely coupled redundant computer system in which processing results are output only from a main computer for the same input.

[0002]

2. Description of the Related Art Conventionally, a database input / output processing system of this type is provided with a database for each of a main computer and a slave computer, and data input / output processing is carried out for each system, so that the master system and the slave system are processed. The processing is performed so that the contents of the database held in 1) are the same.

[0003]

The above-mentioned conventional database processing method has a database in each of the main system and the sub system, and therefore, the main system manages when operating from the independent operation to the duplex operation. It is necessary to copy the contents of the database to the database managed by the subordinate system. Therefore, in a system having a large-capacity database, the copying time of the database becomes enormous and it takes too much time to shift to the redundant operation. Therefore, it is difficult to adopt the redundant computer system in the system having the large-capacity database. There is. In addition, since both systems have databases, there is a problem that the management of the file device becomes complicated.

[0004]

According to the database input / output processing method of the present invention, a buffer is used in a database input / output processing method of a loosely coupled redundant computer system in which a processing result is output only from a main computer for the same input. Provide a database processing device that has a memory and communication means with each of the redundant computers to perform input / output processing of the database file, and when outputting from the main and sub computers to the database file, data from both systems Output to the data file after confirming the agreement of the two, and when reading the database file requested by the main computer and the subordinate computer, the request was made according to the earlier read request of both systems. Transfer to the computer of the system and temporarily store it in the buffer memory, and then calculate the system that received the request later. The constructed by transferring the contents stored temporarily in the buffer memory.

[0005]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram of an embodiment of the present invention. In the embodiment shown in FIG. 1, computers 1 and 2 of the same type are connected by inter-computer communication devices 12 and 21 to intermittently collate data to form a loosely coupled redundant computer system. Further, the computers 1 and 2 are connected to the database processing device 3 via the data transfer devices 13 and 23 capable of high-speed transfer, respectively, and the database processing device 3
Has a buffer memory 32 having a capacity capable of accumulating transfer data at least once, and performs input / output processing of the database file 31.

FIG. 2 is a sequence diagram at the time of writing from the computer of FIG. 1 to the database file and at the time of reading from the database file to the computer. The operation of FIG. 1 will be described below with reference to FIG. Now calculators 1 and 2
It is assumed that one is in the main system and the other is in the slave system, which is operating in duplicate.

First, at the time of writing, since the main system and the slave system perform the same processing by loose coupling, the same write request is generated from both systems at substantially the same time. Upon receiving the transfer of the main system write data requested earlier, the data is written in the buffer memory 32. Then, in response to the transfer of the slave write data, the buffer memory 3
The contents of the buffer memory 32 are written to the database file 31 if there is no mismatch.

Next, at the time of reading, the same read request is generated almost at the same time as at the time of writing, so that the database processing device 3 receives, for example, the slave read request requested earlier, The data is read from the database file 31, the read data is transferred to the slave system that received the request, and this data is stored in the buffer memory 32. Then, since the read request should be the same for the main system that has received the request later, the data accumulated in the buffer memory 31 is read and transferred.

At the time of writing as described above, the database processing device 3 receives the write data of the previously requested system and writes it in the buffer memory 32 first. Alternatively, the contents of the buffer memory 32 may be written to the database file 31 after the data is accumulated in the buffer memory 32 while confirming the coincidence of the transferred data, and the accumulation is completed without error. ..

[0011]

As described above, the present invention provides a loosely-coupled duplex system having a large-capacity database by providing a database processor capable of performing input / output processing between the loosely-coupled redundant computer system and a centralized database. This has the effect of facilitating the structure of the computer system.

[Brief description of drawings]

FIG. 1 is a block diagram of an embodiment of the present invention.

FIG. 2 is a sequence diagram of the operation of FIG.

[Explanation of symbols]

 1, computer 3 database processing device 12, 21 inter-computer communication device 13, 23 data transfer device 31 database file 32 buffer memory

Claims (1)

[Claims] 1. In a database input / output processing method of a loosely coupled redundant computer system in which processing results are output only from a main computer for the same input, a buffer memory and communication means for communicating with each of the redundant computers are provided. A database processor is installed to perform input / output processing of the database file, and when outputting to the database file from the master and slave computers, it is checked whether the data from both systems match before outputting to the data file. When reading the database file requested by the master and slave computers, the read is performed according to the earlier read request of both systems, and the data is transferred to the computer of the system that received the request and is temporarily stored in the buffer memory. Then, the contents temporarily stored in this buffer memory can be transferred to the computer of the system that received the request later. A database input / output processing method characterized by: