JPH02118737A - System for restoring and processing fault of duplexing computer system - Google Patents

Abstract

PURPOSE:To easily and accurately restore a fault by automatically copying and transmitting the content of a disk file from an electronic computer in the middle of operating to an electronic computer which newly comes to be waiting and equalizing the contents of the disk files in both electronic computers at the time of restoring the fault. CONSTITUTION:At the time of restoring the fault, the electronic computer 10 in the middle of operation copies all stored data from the disk file 11 by a copy means 14, gives them to an equivalence request means 15 and stores them in a copy data storage means 16. The equivalence request means 15 gives given data to an equivalence performance means 13 and gives them to an inter- computer communication means 28 in the electronic computer 20 which newly comes to be waiting by the restoration of the fault through an inter-computer communication means 18. In the waiting electronic computer 20, given data are stored in the disk file 21. When the fault occurs in one of the duplexed electronic computers and is restored after detachment, the disk files of both electronic computers can equally and automatically be restored while data are collected on a real time basis.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention is applicable to a redundant electronic computer, and particularly in a system that connects a cash register inline and collects data in real time. When a computer fails and is disconnected from the system and then restored, the system collects data and
This invention relates to a failure recovery processing method for a redundant computer system that makes the disk files of both calculators the same.

〔overview〕

The present invention provides a fault recovery processing method for a redundant computer system in which computers equipped with disk files are duplicated. Then, by copying and sending the contents of the disk file to make the contents of the disk files of both computers the same, disaster recovery can be easily and accurately performed.

[Conventional technology]

Conventionally, in order to make the disk files of duplicated computers identical, it was necessary to copy the disk files using batch processing.

[Problem that the invention seeks to solve]

In the conventional redundant computer system mentioned above, in order to make each disk file the same, it is necessary to copy the disk files by batch processing, so after a failure occurs in the redundant computer system, it is necessary to restore the redundant computer system again. had the disadvantage that it was necessary to temporarily stop collecting real-time data from cash registers or wait until nighttime when the cash registers were finished, making it impossible to easily and accurately perform disaster recovery work.

An object of the present invention is to provide a fault recovery method for a redundant computer system that can automatically perform fault recovery work simply and accurately by eliminating the above-mentioned drawbacks.

[Means for solving problems]

The present invention provides a failure recovery processing method for a redundant computer system in which computers are duplicated, each of which includes a data storage means for storing data and a data storage control means for controlling the storage of data in the data storage means. The electronic computer includes a copying means for copying the data stored in the data storage means, a copy data storage means for storing the copied data, and a copying means for copying the data from the operating computer to the standby computer. an equivalence requesting means for relaying the data copied by the equivalence requesting means; an equivalence execution means for receiving the data relayed by the equivalence requesting means and passing it to the data storage means; and an equivalence execution means for communicating between the two computers and transmitting and receiving the data. The present invention is characterized by comprising an intercomputer communication means for mutually monitoring faults.

[Effect]

At the time of recovery from a failure, an operating computer uses its copying means to copy all data stored in its data storage means, such as a disk file, and passes it to its equivalence requesting means, as well as storing it in its copy data storage means. The equivalence requesting means passes the passed data to its equivalence execution means. The equivalent execution means passes the passed data via its intercomputer communication means to the intercomputer communication means of the computer that has newly become on standby due to failure recovery. The electronic computer on standby stores the transferred data in its disk file via its equivalent execution means and data storage control means.

Therefore, if one of the duplicated computers fails and is disconnected and then restored, it is possible to automatically restore the same disk files on both computers while collecting data in real time. Become.

Note that copying of disk files during disaster recovery is performed in physical units, and if writing occurs to the disk file during copying, if data is in an area that has already been copied, it will be copied to the disk file of the standby computer. If the data is in an uncopied area, it will be written only to the active disk file.

〔Example〕

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing an embodiment of the present invention, in which human data from a cash register is stored in real time.

This embodiment includes disk files 11 and 21 as data storage means for storing data, respectively, and disk control means 12 and 22 as data storage control means for controlling the storage of data in these disk files 11 and 21, respectively. In a failure recovery processing method for a redundant computer system in which computers (1) 10 and (2) 20 are duplexed, computers (1) 10 and (2) 20 are stored in disk files 11 and 21. Disk file copying means 14 and 24 as copying means for copying data, copy tables 16 and 26 as copy data storage means for storing the copied data, and computer (1) 10 in operating state. or (2) equivalence requesting means 15 and 25 for relaying the data copied by the disk file copying means 14 or 24 from the electronic computer (2) 20 or (1) 10, respectively, from the computer 20 in the standby state, and the equivalence requesting Equivalent execution means 13 and 23 which receive the data relayed by means 15 and 25 respectively and pass them to disk files 11 and 21 communicate with the two electronic calculations 1 (1) 10 and (2) 20 to process the data. It is provided with calculation inter-communication means 18 and 28 for performing delivery and mutual fault monitoring, respectively.

In FIG. 1, 17 and 27 are inline communication means, 19 and 29 are communication line switching means, and 30
is a communication line switching device, and 40 is a cash register.

The feature of the present invention is that, in FIG. 1, equivalent execution means 13 and 23, disk file copy means 14 and 24
, equivalence requesting means 15 and 25, copy tables 16 and 26, and calculation communication means 18 and 28 are provided.

Next, the operation of this embodiment will be explained. First, we will provide an overview of the operation of each part.

Disk control means 12 and 22 read and write data to and from disk files 11 and 21, respectively. The inline communication means 17 and 18 are
Communication is performed between the electronic computers (1) 10 and (2) 20 and the cash register 40, respectively. The communication means 18 and 28 between the calculation units perform communication between the electronic calculation II (1) 10 and (2) 20, and perform data exchange and mutual down monitoring.

The communication line switching means 19 and 29 each issue a command to the communication line switching device 30 to switch the communication line.

The equivalence requesting means 15 and 25 are the disk files 11 of the duplicated electronic computers (1) 10 and (2) 20.
In order to equalize the contents of disk files 11 or 21, the commands and data of the disk files 11 or 21 are transferred from the operating computer (1) 10 or (2) 20 to the standby computer (2) 20 or (1) 10. Relay. The equivalent execution means 13 and 23 receive the command and data and execute the disk control means 12 and 2, respectively.
Pass it to 2. Disk file copy means 14 and 24
When computer (1) LO or (2) 20, respectively, is restored after a failure, the computer (1) 10 or (2) 20 that is in operation status is restored to disk file 1.
1 or 21 is read out, and the data is sent to the computer (1) 10 or (2) 20 which has newly entered the save state. Normally, data is read from or written to the disk files 11 and 21 in units of logically defined records, but the disk file copy means 14 and 24 read from or write data to the disk files 11 and 21, respectively. The time is determined in physical units (sector units).

The copy tables 16 and 26 are used to copy the disk file 1 by the disk file copy means 14 and 24, respectively.
When copying 1 and 21, record how far the copy was made. Copy tables 16 and 26
There is an area that corresponds one-to-one to the physical read/write unit (sector unit) of the disk, and a write flag is set each time copying is executed.

Next, the overall operation of this embodiment will be explained separately: (1) when the system is normal, (2) when the system is abnormal, and (2) when the system is restored.

■ When the system is normal Figure 2 is an explanatory diagram showing the flow of information when the system is normal. Information generated by the cash register 40 is sent to the operating electronic computer (1) 10 via the communication line switching device 30 by inline communication. Electronic computer (in IHO, from inline communication means 17 to disk control means 12
and is written to the disk file 11. Also,
Inline communication means 17 is used to write the same information to disk file 21 of standby computer (2) 20.
The information is sent to the equivalence requesting means 15, and the computer (2) 20 on standby is sent to the computer (2) 20 by the computing communication means 18 and 28.
sent to. In the electronic computer (2) 20 on standby, the data is transferred from the equivalent execution means 23 to the disk control means 22 and written to the disk file 21.

(2) At the time of system abnormality FIG. 3 is an explanatory diagram showing the flow of information when the computer (1) 10 in operation has a failure and is switched to the computer (2) 20 on standby. Information generated at the cash register 40 is sent to the switched electronic computer (2) 20 via the communication line switching device 30 by inline communication. In the electronic computer (2) 20, the data is sent to the disk control means 22 via the inline communication means 27 and written to the disk file 21, as in normal operation.

■ At the time of system recovery FIG. 4 is an explanatory diagram showing the flow of information when the failed computer (1) 10 becomes normal and is performing recovery processing as a system. To restore disk file 11,
The disk file copying means 24 is the disk file 2
1 is read out, and the data is written to the disk file 11 by the disk control means 12 through the equivalence requesting means 25, the intercomputer communication means 28, the intercomputer communication means 1B, and the equivalence execution means 13. At this time, the copy status of the disk file 21 is recorded in the copy table 26.

When data to be written to the disk file 21 is generated by the cash register 40 during disk file copying, the data is sent to the computer (2) 20 via the communication line switching device 30 by inline communication. In the electronic computer (2) 20, the data is written to the disk file 21 via the inline communication means 27 and the disk control means 22, as in normal operation. In order to write the same information to the disk file 11 of the recovered computer (1) 10, it is sent from the inline communication means 27 to the equivalence request means 25. The equivalence requesting means 25 refers to the copy table 26 and does not send the information to the intercomputer communication means 28 if the write data is to an area that has not been copied yet. If the data is in an area that has already been copied, the inter-computer communication means 28
and 18, the data is sent to the restored electronic computer (1) 10, and written to the disk file 11.

FIGS. 5(a), 11b) and (C) are explanatory diagrams showing the method of registering write data in this case, and FIG.
) indicates a normal disk file copy, and (c) in the same figure
indicates the case where data is written to an uncopied area,
(C) of the same figure shows a case where write data is generated in an already copied area.

In other words, as mentioned above, when disk file write data occurs during copying during disaster recovery, if the data is in an area that has not been copied, it is written only to the disk file of the computer that is running, and even if it is data in an area that has already been copied. For example, it also writes to the disk file of a computer that is on standby.

〔Effect of the invention〕

As explained above, in a duplex computer system, when one computer is disconnected due to a failure and then recovered, the disk files of both child computers are automatically created to be the same while collecting data in real time. It has a restorative effect. Therefore, according to the present invention, not only can failure recovery be easily performed, but also data can be guaranteed, which is highly effective.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of the present invention. FIG. 2 is an explanatory diagram showing the flow of information during normal operation. FIG. 3 is an explanatory diagram showing the flow of information at the time of the abnormality. FIG. 4 is an explanatory diagram showing the flow of information at the time of restoration. FIGS. 5(a), 5(a) and 5(c) are explanatory diagrams showing a registration method when data written to a disk file occurs during disk file copying. FIG. 10...Electronic computer (1), 11.21...Disk file page 12.22...Disk control means, 13
．． 23...Equivalence execution method L 14.24...Disk file copying means, 15.25...Equivalence requesting means,
16.26... Copy table, 17.27... Inline communication means, 18.28... Inter-computer communication means, 19.29... Communication line switching means, 20...
Electronic computer (2), 30... communication line switching device,
40... Cash register machine.

Claims (1)

[Scope of Claims] 1. Failure recovery processing for a redundant computer system in which computers are duplicated, each of which includes a data storage means for storing data and a data storage control means for controlling the storage of data in the data storage means. In the method, the electronic computer includes a copying means for copying data stored in the data storage means, a copy data storage means for storing the copied data, and a copying means for copying the data stored in the data storage means, and a copying means for copying the data stored in the data storage means, and a copying means for copying the data stored in the data storage means, and a copying means for copying the data stored in the data storage means, and a copying means for copying the data stored in the data storage means. and an equivalence requesting means for relaying the data copied by the copying means;
Equivalence execution means that receives the data relayed by the equivalence requesting means and passes it to the data storage means, and intercomputer communication means that communicates between two computers to exchange the data and monitor mutual failures. A failure recovery processing method for a redundant computer system characterized by the following.