JPH02113352A - File recovery system for multiple system - Google Patents

Abstract

PURPOSE:To easily recovery a file of the multiple system by eliminating hardware common use by the providing the file with a copy attribute and a time stamp. CONSTITUTION:The file 3 is provided with the attribute flag 4 consisting of the copy attribute and time stamp and the flag is transferred and stored on a disk device (external storage device) 1 of a CPU whose copy attribute is on. If trouble occurs to one CPU, the copy attribute is referred to and the system is recovered with the file having the latest time stamp obtained from the CPU having the same file. Consequently, the multiple system is structured without having any common-use file and the recovery is easily made.

Description

[Detailed Description of the Invention] [Summary] The present invention provides a file recovery method in a multiplexed system that solves the problem of complexity due to the necessity of improving the reliability of the common section when sharing files by providing a common section. In order to solve problems such as readiness, a file has a copy attribute that indicates whether copying is necessary and a timestamp that indicates the time it was updated, and when a failure occurs, a timestamp is obtained from the CPU that has the same file by referring to the copy attribute. By restoring using the latest file, it is possible to eliminate the need for hardware sharing and restore files in a multiplexed system.

[Industrial application field]

The present invention relates to a file recovery method in a multiplex system.

[Prior Art and Problems to be Solved by the Invention] Conventionally, a duplex or multiplex system has been adopted because the stoppage of a total of X machine systems would lead to serious losses. If one common file is accessed by a plurality of CPUs in order to construct this multiplexed system, there is a problem in that it becomes essential to improve the reliability of the common file. Furthermore, when individual files are used as shared files, there is a problem in that it is difficult to efficiently restore and reinstall the files into the system.

The present invention provides files with copy attributes and timestamps, and when a failure occurs, refers to the copy attributes and timestamps and recovers with the latest file, thereby eliminating the need for hardware sharing and recovering files in a multiplexed system. The aim is to make it easy to do.

[Means to solve problems]

Means for solving the problem will be explained with reference to FIG.

In FIG. 1, a disk device 1 stores a file 3 and an attribute flag 4. As shown in FIG.

The attribute flag 4 is made up of a copy attribute indicating whether copying of the file 3 is necessary and a timestamp indicating the time when the file was updated, and is held in the file 3.

[Effect]

As shown in FIG. 1, the present invention provides a file 3 with an attribute flag 4 consisting of a copy attribute and a timestamp, and transfers this copy attribute to a disk device (external storage device) l of a CPU for which the copy attribute is on and saves it. However, if a failure occurs in one of the CPUs, the copy attribute is referenced to restore the file using the latest time stamp obtained from the CPU that has the same file.

Therefore, it is possible to construct a multiplex system without having common files, and it is also possible to perform recovery easily.

〔Example〕

Next, the configuration and operation of one embodiment of the present invention will be explained in detail using FIGS. 1 and 2.

In FIG. 1(a), the disk device 1 has CPU#1
, CPU2#, CPU#3, etc., and stores files (A, B, C...N)3. File 3 has attribute Flag (attribute flag) 4.

This attribute flag 4 is for each CPU#1, CPU#2, CP
Copy attributes associated with U#3... and Tim, which is the time when file 3 was created (or latest updated)
It consists of e-stamps (time stamps). This attribute Flag4 is provided only for the file 3 that needs to be shared in order to facilitate multiplexing.

In FIG. 1 (b), CPU#i CPU#2, CP
U#3... is a central processing unit that performs processing.

LANf, II'4n adapter 2 is LAN7! l< Due to functions such as simultaneous broadcasting and group broadcasting, file 3 and attribute flag 4 are sent to the CPU with attribute flag 4 on.
It is used to transfer etc.

Next, the operation of the configuration of this embodiment will be explained with reference to FIG. 1(b).

In Figure 1 (b), ■ indicates C when creating File.
Determine the opy attribute. This means that when creating file 3, CPU #1 on the execution system determines which of the copy attributes in attribute flag 4 in Figure 1 (a) should be turned on to transfer the file. ing.

(2) goes into operation, and it becomes necessary to change Fi Is.

■ is the File (data and attributes Fl) of the own system (execution system)
ag's Time stamp). As a result, the file 3 and the attribute FIag4 are saved in the active disk device l.

■'' further transfers file 3 and attribute FIag4 to the standby system with the copy attribute of attribute F1ag4 set to ON.

(2) The standby system that received the transfer of file 3 and attribute Flag 4 saves it in its own disk device.

Through the above processing, when file 3 is created or updated during normal operation, it is saved in the active disk device 1, transferred to the standby CPU with the copy attribute on, and saved in the disk device 1. Ru.

Next, a recovery method when an execution system failure occurs will be explained using FIG.

In FIG. 2, ■ indicates that a failure has occurred in the active system CPU #1, and the standby system CPU #2 has been switched to the active system.

In ■, CPU #2, which has become the new execution system, uses file 3.
Refers to the C0PY attribute of attribute FIag4 and searches for a CPU that has the same file that is turned on.

■ is the file attribute Fl from the CPU system found in ■.
Collect ag4 Time stamps.

■ collects Time stamps of the same file in ■, and if the Time stamp of the new executable file is the newest, the self-system data (file 3) is collected.
(Restore file 3 of own system to other applicable CPU)
(to make it the latest file 3).

(2) If the other system has the latest data (file 3), it is obtained via the LAN, and based on this data, it is transferred to the CPU of the own system and the corresponding other system to create the latest file 3. At this time, if there are multiple CPU systems that have the latest data (file 3), the data is obtained from the CPU with the lowest priority to try to average the load as much as possible.

Through the above processing, even if a failure occurs in the active CPU, the standby CPU retrieves the attribute flags (copy attributes and timestamps) and restores the latest file.

Note that if a failure occurs in the standby system, the standby system is ignored.

〔Effect of the invention〕

As explained above, according to the present invention, the file 3 is given the attribute flag (copy attribute, time stamp) 4 and saved in the disk device 1 of the CPU with the copy attribute on,
If a failure occurs, we use a configuration that refers to these attribute flags 4 and restores to the latest file 3.
There is no need to have a common file, a multiplexed system can be easily constructed, and the latest file (data) can be easily restored in the event of a failure.

[Brief explanation of drawings]

FIG. 1 is a block diagram of one embodiment of the present invention, and FIG. 2 shows a recovery method in the event of an execution system failure according to the present invention. In the figure, l is a disk device, 3 is a file, and 4 is an attribute Fl.
ag (copy attribute + timestamp). Patent applicant: BFN Co., Ltd.

Claims (1)

[Claims] In a file recovery method in a multiplex system, a plurality of CPUs connected via a communication path are each provided with an external storage device for storing files, and these plural CPUs are
The file has a copy attribute indicating whether the file needs to be copied and a timestamp indicating the updated time in association with the PU, and the file with the copy attribute and timestamp is sent to a CPU with the copy attribute on. If a file is transferred and saved to an external storage device, and one of the CPUs fails, the file will be recovered using the latest timestamp obtained from the CPU that has the same file by referring to the copy attribute. A file recovery method for a multiplexed system characterized by the following configuration.