JPH0281224A - Backup device for magnetic disk device - Google Patents

Abstract

PURPOSE:To restore data of a failure magnetic disk device in a short time by writing the same data in a magnetic tape device as soon as data is written in the magnetic disk device. CONSTITUTION:As soon as write to a magnetic disk device 3 is executed, the magnetic disk device 3 records its recording position and data in a magnetic tape device 4. Accordingly, when the magnetic disk device 3 causes a failure, which data on a magnetic tape corresponds to data on the magnetic disk device 3 is known easily, therefore, data retrieving speed on the magnetic tape is improved. In such a way, when an arbitrary disk device 3 causes a failure, and it is tried to restore the data in the failure disk device 3 to other disk device, a data transfer to the magnetic disk device from the magnetic tape is executed at a high speed and easily.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to an information processing system including a magnetic disk device and its backup device, and in particular includes a plurality of magnetic disk devices, which reduces reliability. The present invention relates to a backup device for a magnetic disk device suitable for an information processing device.

[Conventional technology]

Conventionally, in information processing apparatuses having one or more magnetic disk drives, measures for improving reliability and early recovery from failure of magnetic disk drives have been proposed as seen in Japanese Patent Application Laid-Open No. 59-180765. How to write the same data on two disks, Nikkei Electronics no. 437 P48-49 (1987, 12, 28)
As seen in the example above, in case any magnetic disk device fails, one spare disk device is provided for every number of disk devices used, and if any magnetic disk device fails, the spare disk device is installed. A method was used to immediately switch to a disk device.

[Problem that the invention sought to solve]

Among the conventional techniques mentioned above, when using a method of writing exactly the same data to two magnetic disk drives, the advantage is that the recovery time is virtually zero, and that the magnetic disk drives writing the same data fail at the same time. The reliability of the magnetic disk device is extremely high, since the magnetic disk device will not be in an unusable state unless it is used. However, this method requires twice as many magnetic disk drives, which leads to an increase in cost, and also significantly increases the volume of the entire magnetic disk drive. Furthermore, since data is attempted to be written to two magnetic disk devices at once, the average access time becomes longer than when writing is performed only to - magnetic disk devices.

In addition, the method of having one extra spare disk device in case a disk device fails causes a small increase in cost and device volume, but it is possible to back up the magnetic disk device and transfer data to the disk device. Since data is not written at the same time as writing, it is impossible to completely restore data in the event of a disk failure, resulting in poor reliability.There is no special consideration given to the high speed of data restoration.

An object of the present invention is to provide a highly reliable magnetic disk device.
The objective is to realize a low-cost, small-scale backup device that can restore data at high speed.

[Means to solve the problem]

The above purpose is to simultaneously write to the magnetic disk device.
Means for recording the written data along with the magnetic disk device and position where the data was recorded on magnetic tape, and a spare disk device that can be immediately switched to the failed disk if one of the magnetic disk devices fails. This is accomplished by means for notifying a control device that controls the magnetic disk device and magnetic tape device which magnetic disk device has failed when any magnetic disk device fails.

[Effect]

At the same time as writing to a magnetic disk device, the magnetic disk device, its recording position, and data are recorded on a magnetic tape.If the magnetic disk device fails, which data on the magnetic tape will be lost to the magnetic disk device. Since it is easy to know what corresponds to the data, the data retrieval speed on the magnetic tape is improved. As a result, when any disk device fails and an attempt is made to restore data on the failed disk device to another disk device, data transfer from the magnetic tape to the magnetic disk device becomes fast and easy.

Data restoration on the failed disk device can be further speeded up by having a spare disk device. If any magnetic disk device fails, a spare disk device can immediately replace the failed disk device.

Furthermore, by notifying the control device that controls the magnetic disk device and magnetic tape of the failed disk device, the control device starts moving the data that was in the failed magnetic disk device from the magnetic tape to the spare disk device, and Restoration of the above data is done quickly.

〔Example〕

An embodiment of the present invention will be described below with reference to FIGS. 1 to 6.

FIG. 2 shows an embodiment of a system in which the present invention is used.This embodiment includes a host computer 1.file control table 11!2. Magnetic disk device fi3. Magnetic disk device 0 consisting of backup magnetic tape devices 4 and 5
What corresponds to the present invention is a file control system! ! 2. This is a magnetic tape device 4.

Magnetic disk installation from host computer 1! ! I3 no 3
When there is a request to write data to one of the magnetic disk devices 32, for example, the magnetic disk device 32, the file control device 2 transfers the data to the magnetic disk device 3 and records the same data on the magnetic tape 4 at the same time. The contents of the magnetic disk drive [3 are recorded on the magnetic tape 5 at regular intervals. When the contents of 3 are copied, the contents of magnetic tape 4 are all cleared at the same time.

The processing procedure when the magnetic disk device 33 fails is described in the third section.
If the magnetic disk device 33 shown in the figure fails, the file control device i! 2 indicates that the magnetic disk device 33 is in a read/write disabled state, or the magnetic disk device 3 is in a read/write disabled state.
Detected by hardware error signal etc. from 3 (300
), the file control device 2 detecting a failure in the magnetic disk device 33 writes the data that was in the magnetic disk device 33 in the magnetic tape device 5 to the spare magnetic disk device 30 (301). After the data copy from the magnetic tape device 5 is completed, the file control device 2 writes the data recorded in the magnetic disk device 33 among the information recorded in the magnetic tape device 4 to the spare magnetic disk device 30 (302 ) - All the data that was in the magnetic disk device 33 on the magnetic tape device 4 is transferred to the spare magnetic disk device 30.
When the transfer is completed, the magnetic disk device 33 is completely restored. File control device i! 2 changes the address of the spare magnetic disk device 30 so that it can be treated as a new magnetic disk device 33 (303), and the data in the magnetic disk device 33 in the magnetic tape devices 4 and 5 remains unchanged on each magnetic tape. left on top.

The failed magnetic disk device 33 is the file control device i.
! 2, the disk is disconnected from the bus and itt source line (304), making it possible to replace it with a new disk. When the replacement is detected by a hardware reset of the magnetic disk device (305), the magnetic disk control device 2 connects the replaced new magnetic disk device to the path, initializes the new disk device, and restarts the new disk device. Allow it to be used as a spare magnetic disk device (
306), and when the processing of the two magnetic disk devices is completed, the failure processing ends (307).

If the magnetic disk device fails due to the above processing,
A failed magnetic disk device is quickly and completely restored.

FIG. 1 shows an embodiment of a file control device 2 and a magnetic tape device 4, which are the main parts of the present invention.

The file control device 2 has an input/output port 21 . Input/output port 2 with file devices 3 and 5
2. An instruction comparator 23 that selects only write instructions from among the instructions sent from the host computer l to the magnetic disk device. The magnetic tape control table M24. only sends write commands to the magnetic tape device 4 during normal operation, and sends data from the magnetic tape device 4 to the spare magnetic disk device 30 of the magnetic disk device 3 during failure processing. and a magnetic disk failure detector 25.

An example of actual operation is shown in FIG. 4. FIG. 4a shows the flow of instructions and data during normal operation.

File control table that received commands through input/output port 21! ! 2 sends the command to the magnetic disk device 3 and simultaneously sends the same command to the command comparator 23 (■). Upon receiving the command, the command comparator 23 sends the command to the magnetic tape controller 24 only if it is a write command. The magnetic tape controller 24 records information on the magnetic tape device in the form shown in FIG. 5 (■). At this time, the magnetic tape device has
The information sent is the address of the magnetic disk device (magnetic disk device number + cylinder, head, sector number)
Unlike magnetic disk devices, there is no mechanical waiting time until information is written, so if the data recording speed of a magnetic tape device is equal to or higher than that of a magnetic disk device, it can be used as a tape device. 4 does not effectively slow down the data writing process from the host computer 1.

FIG. 4 shows an example of operation during failure handling. The failure detector 25 detects that an arbitrary magnetic disk device 3x has failed (■), and transmits the information that the failure has been detected to the command comparator 23.
Then, set the information that the failure occurred and the number 3x of the failed magnetic disk device in the magnetic tape control table v124 (
■). When this information is set, the command comparator 23 allows the host computer 1 to access the magnetic disk device 3.

and prohibits write command transfer to the magnetic tape device 4. At this time, the magnetic disk device 3 is connected to the host computer 1.
continues to issue a signal that it is in use or that it is in the process of recovering from a failure.

By setting the information that the magnetic disk unit W13x has failed, the magnetic tape control device 24 first instructs the magnetic tape device 5 to transfer the data of the failed magnetic disk device 3x to the spare magnetic disk device 30 (■). The device 5 issues a data write command to the spare magnetic disk device 1230, and sends out the data. Malfunctioning magnetic disk device! 3x, the magnetic tape device 5 sends a copy end signal to the input/output port 22.
When this signal is sent to the magnetic tape control device 24 through (■), the magnetic tape control device 24 transfers the data recorded in the failed magnetic disk device 3n from the magnetic tape device 4 to the failed magnetic disk device. Sends a command to read along with the command and address information (■), and receives and receives commands and data from the magnetic tape device 4 (■).
■), sending to the spare magnetic disk device 130 (■),
When all the data has been sent, the magnetic tape control device sends a transfer end signal to the failure detection device (■).

The failure detection device 25 that received this is the input/output boat 22.
Through magnetic disk device! 30, sends a device number entry signal and a disk address 3x (■), and thereafter treats the spare disk device 30 as 3x.Finally, the failure detector 25
Then, the command comparator 23 and magnetic tape control device 24 are unset.@l) The failure recovery process of - series ends.

In the embodiment mentioned above, the magnetic tape devices 1t4 and 5 are written separately, but they can be combined.If the contents of the magnetic tape 4 are added to 3 after the magnetic tape 5, the magnetic tape device i ! There is no need to switch from 5 to 4.

According to the embodiments described above, when a failure occurs in a magnetic disk drive, data on the failed magnetic disk drive can be automatically restored without placing any burden on the host computer.

In the embodiment described above, access from the host computer 1 is prohibited as soon as the failure recovery process starts, but this may be allowed by the instruction comparator 23 having a mode that allows interrupts from the host computer 1. Realized. This interrupt occurs when the host computer 1 accesses the magnetic disk device 3. When this interrupt occurs, the instruction comparator 23 commands the magnetic tape control device 24 to stop failure recovery processing, and resumes normal operation. return. When the access to the magnetic disk drive [3 is completed, the failure recovery process is continued again. However, access to the failed magnetic disk device and writing to the magnetic tape device 4 are prohibited.

As a method of permitting the host computer 1 to access the magnetic disk device 3 during the failure recovery process, the following embodiments may also be cited.

FIG. 6 shows an example thereof.

Host computer 1. Magnetic disk device! It is assumed that each of the 3.degree. magnetic tape devices 6 (the combination of the magnetic tape devices 4 and 5 of the previous embodiment) has a built-in interface having a bus arbitration function. When the magnetic disk device 13x fails, the information is notified to the host computer 1 (■), and the host computer 1 that receives the notification,
A malfunctioning magnetic disk unit [3
A command is given to transfer the data of x to the spare magnetic disk device 30. After receiving the command, the magnetic tape device 6 performs an operation to acquire the bus 7, and after acquiring the bus, transfers data to the magnetic disk device 30 (■), and divides the data transfer into appropriate lengths. This is done so that the magnetic tape device 6 does not continue to monopolize the bus 7. When a request from the host computer 1 occurs during data transfer, this data transfer is temporarily stopped after a certain length, and bus arbitration is started. By winning the bus arbitration, the host computer 1 can access all magnetic disk devices except for the failed one.

When the data transfer from the magnetic tape device 6 to the magnetic disk device 3o is completed, the magnetic tape device @6 notifies the host computer 1 of the completion of data transfer (■). Change the device number to the number of the failed magnetic disk device, and
Failure recovery processing ends. This embodiment is 1 e.g. SC8
I (Small Computer System)
The application can be considered in a system equipped with a m-interface.

〔Effect of the invention〕

According to the present invention, if one of the magnetic disk drives that is composed of a plurality of magnetic disk drives fails, the data that was on the failed magnetic disk drive can be transferred to other magnetic disk drives in a short time. It is possible to completely restore it. The hardware that must be added to provide this function is no more than an improvement to the control device of a conventional backup magnetic tape device, and there is no need for a significant increase in cost or volume.

[Brief explanation of the drawing]

Fig. 1 is a schematic diagram of a file control device that is the main part of the present invention, Fig. 2 is a schematic diagram showing an embodiment of a system using the present invention, and Fig. 3 is a schematic diagram of a file control device that is the main part of the present invention. FIG. 4(a) is a flowchart showing a failure handling procedure of the present invention, and FIG.
b) is a schematic diagram showing an example of an operation during failure handling, and FIG. 5 is a schematic diagram showing one implementation of a recording method on a magnetic tape. FIG. 6 is a schematic diagram showing an embodiment in which a path usage right acquisition capability is added to a magnetic disk device and a magnetic tape device. DESCRIPTION OF SYMBOLS 1...Host computer, 2...File control device, 23...Instruction comparator, 24...Magnetic tape control device, 25...Failure detector, 3...Magnetic disk device. 30... Spare magnetic disk device, 4,5.6... Magnetic tape device. muscle country country

Claims (1)

[Claims] 1. A host computer, and a magnetic disk device that is controlled by the host computer and performs data loading and unloading;
A backup device for a magnetic disk device, characterized in that, in an information processing device comprising a magnetic tape device for backing up a magnetic disk device, data is written to the magnetic disk device and at the same time the same data is written to the magnetic tape device. 2. In the information processing device, if an irrecoverable failure occurs in any magnetic disk device, it is detected and the data corresponding to the data on the failed magnetic disk device is transferred from the magnetic tape device to the spare magnetic tape device. 2. A backup device for a magnetic disk device according to claim 1, wherein the magnetic disk device is restored by writing data to the magnetic disk device.