JPH08110840A - Restoration system for magnetic disk device of double redundancy constitution - Google Patents

Abstract

PURPOSE: To make the necessary time for copy in a needed disk device shorter than before by referring to a file update management table when a magnetic disk device of double redundancy constitution is put back to double redundancy operation from single independent operation. CONSTITUTION: At the time of restoration from the single independent operation state of only a system 10A to the double operation state wherein a system 10B serves as a stand-by system, only updated files are copied from a magnetic disk device 3A to a magnetic disk device 3B while referring to the file update management table 4 as a list of the names of files updated during the single independent operation on the side of the system 10A, thereby synchronizing the pieces of information on the magnetic disk devices 3A and 3B with each other.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a recovery system for a magnetic disk unit having a redundant redundant structure, and particularly in an information communication system requiring high reliability in which no service interruption is allowed.
The present invention relates to a recovery system of a redundant magnetic disk device having a redundant redundant structure, which is used when a failure occurs in one system of the redundant system and the system is restored.

[0002]

2. Description of the Related Art As shown in FIG. 2, a conventional dual redundant configuration magnetic disk device is provided with a main processor 1A.
And 1B, main memories 2A and 2B, and magnetic disk devices 3A and 3B as auxiliary storage devices, and systems 11A and 11B are connected by a synchronous bus 5. In the dual redundant configuration system, when the main processor 1A of the system 11A is operating, the main processor 1B of the system 11B is in the standby state. In that case, the magnetic disk device 3B in the system 11B maintains exactly the same contents as the magnetic disk device 3A in the system 11A by redundant writing from the main processor 1A of the system 11A, and always stores the latest information. . When the main processor 1B of the system 11B is operating, the main processor 1A of the system 11A is in a standby state, and the magnetic disk device 3A in the system 11A uses the redundant writing from the main processor 1B to perform magnetic recording in the system 11B. The same contents as the disk device 3B are maintained, and the latest information is always stored.

For example, when the system 11A is stopped during operation due to a failure or maintenance work, the system 11A and the system 11B are disconnected from each other. Meanwhile, the system 11B continues the service operation by the independent operation. Under such a single-sided system operation, the information of the magnetic disk device 3A in the system 11A and the system 1
The contents differ from the information of the magnetic disk device 3B in 1B. The redundant copy process for eliminating such a difference in information content starts from the moment when the shift from the single operation to the double operation becomes possible. In this redundant copy process, volume copy is executed from the magnetic disk device 3B on the system 11B side to the magnetic disk device 3A of the system device 11A. After the redundant copy process is completed,
The magnetic disk devices 3A and 3B of the systems 11A and 11B have the same information.

[0004]

In the above-mentioned conventional redundant copy processing, the magnetic disk devices 3A and 3B are used.
Since volume copying is performed between files, there is no file update during single (single) operation, and copy processing is performed even for files that do not need to be copied. There is a problem that processing time is wasted for copying.

[0005]

According to the recovery system of the present invention, each main processor of two redundant systems configures writing of the contents of the magnetic disk device of its own system and the same contents are applied to the magnetic disk devices of other systems. In the recovery method of a dual redundant configuration magnetic disk device having a writing process, the file name of the file in the magnetic disk of the own system updated during one-side operation of only the own system due to suspension of operation of the other system is updated. The file information corresponding to the file name is transferred from the own system to the other system by storing it in the management table and referring to the file name stored in the file update management table when the operation of the other system is restored. And copying to the magnetic disk of another system.

[0006]

In the conventional recovery method, the volume copy is a standard specification, and particularly in the case of a large capacity disk, it takes a lot of time to copy. In the present invention, when the operation is switched from the single operation to the double operation, the redundant copy for synchronization is performed only for the files that have been updated during the single operation, so the operation load required for the files is the minimum necessary. In a short time.

[0007]

The present invention will be described in detail with reference to the drawings.

FIG. 1 is a block diagram showing an embodiment of the present invention. In this embodiment, the file update management table 4 is provided in the magnetic disk device 3A. Although the file update management table 4 does not exist on the magnetic disk device 4A during normal duplex operation, it is created on the magnetic disk device on the operating side when the single-duplex independent operation state is caused by a failure or maintenance. (The magnetic disk device 3A in this embodiment). Then, for the file whose file content is updated during the single operation alone, the name of the file is written in the file update management table 4. further,
Needed while referring to the file update management table 4 for the magnetic disk device (3B) on the recovery side from the magnetic disk device (3A) on the operating side at the moment when it becomes possible to change from the single operation to the dual operation Execute redundant copies for as much as you want.

FIG. 3 shows that, in the present embodiment, from the moment when the duplex operation state is changed to the single operation state, the second operation is performed again.
It is a sequence chart which illustrates the program operation until the moment it restores to a heavy duty operation system. At the moment when the hard failure occurs on one side during the duplex operation, the system where the hard failure occurs is automatically disconnected and the single (single) operation is started. The parent process of the program learns from the system call (process number 1) that the opposite system has failed, and notifies the child process of that (process number 2).
The child process notified by the parent process creates the file update management table 4 (process number 3). And 1
When the file with the file name "a" is updated during the duplex operation, the parent process sends the name of the updated file "a" to the child process (process number 4). The child process writes the received file name "a" in the file update management table 4 (process number 5).
Similarly, when the file "b" is updated, the file name "b" is transmitted to the child process (process number 6), and the received child process adds it to the file update management table 4 (process number 7). After that, at the moment when the hardware failure is repaired and the duplex operation becomes possible, a system call informs the parent process of the recoverability (process number 8). In response to this, the parent process sends a message to the child process that the opposite system is recoverable (process number 9).
To do. In response to this message, the child process executes file copy from the operating magnetic disk device (3A) to the recovering magnetic disk device (3B) in order to synchronize the file contents of the dual magnetic disk device ( Processing number 1
0), and after completion of copying, returns a response to the parent process (process number 11). Since the files copied at this time are only the file names in the file update management table 4, the time required for the synchronous copy for restoration is the minimum necessary, and the load on the main processor is not necessary. The increase can be suppressed.

[0010]

According to the present invention as described above,
The name of each file updated during duplex operation is stored in the file update management table, and at the moment when the system is restored to the duplex operation ready state, synchronous copy is performed with reference to this file update management table. By doing
The time required for copying in the transition state to the duplication can be suppressed to be shorter than before.

[Brief description of drawings]

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

FIG. 2 is a block diagram illustrating a conventional method.

FIG. 3 is a sequence chart illustrating the operation of redundant copy processing according to the embodiment of this invention.

[Explanation of symbols]

 1A, 1B Main Processor 2A, 2B Main Memory 3A, 3B Magnetic Disk Unit 4 File Update Management Table 5 Synchronous Bus 10A, 10B, 11A, 11B System

Claims (2)

[Claims] 1. A dual redundant magnetic structure having a process in which each main processor of two redundant systems writes the contents of a magnetic disk device of its own system and also writes the same contents to a magnetic disk device of another system. In the recovery method of the disk device, the file name of the file in the magnetic disk of the own system that was updated during one-side operation of only the own system due to the suspension of operation of the other system is stored in the file update management table, and By referring to the file name accumulated in the file update management table when the operation is restored, the file information corresponding to the file name is transferred from the own system to the other system and copied into the magnetic disk of the other system. A method for recovering a redundant magnetic disk device having a dual redundant structure including the steps. 2. The recovery method of a dual redundant magnetic disk device according to claim 1, wherein the file update table is set in the magnetic disk of its own system.