JPS6376151A - Magnetic disk device - Google Patents

Abstract

PURPOSE:To automatically back up data and to eliminate the need for the time of a backing-up by exciting simultaneously a pair of magnetic heads and writing simultaneously the same data into at least two areas. CONSTITUTION:When input data 1 are stored into a writing data memory 2, the data stored into the writing data memory 2 are converted to a serial NRZ signal by a parallel/serial converter 3. An error detecting code generating circuit 4 generates an error detecting code 22 with data 21 and further, adds and outputs an error detecting code 21 to the end of the data 21. A modulator 5 executes the MFM modulation to the NRZ signal to add the error detecting code 22 to the data 21. The MFM modulated signal is simultaneously inputted to a first writing amplifier 6 and a second head amplifier 2, the first writing amplifier 6 excites a first magnetic head 8, and simultaneously, a second writing amplifier 7 excites a second magnetic head 9. The data are written onto a first data surface 16 by the first magnetic head 8, and simultaneously, the data are written onto a second data surface 17 by the second magnetic head 9.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an external storage device for a computer, and more particularly to a data buffer lag device for a single external storage device.

[Conventional technology]

A conventional device uses two magnetic disk devices as an online disk device and an offline disk device, as described in the prior art section of Japanese Patent Application Laid-Open No. 59-225870, and after writing data to the online disk device,
By writing the same data to an offline disk device, data was backed up and reliability was increased.

In contrast, in the invention described in Japanese Patent Application Laid-open No. 59-223870, an external storage device is provided with a plurality of storage units different from magnetic disks, such as bubble memory and core memory,
Reliability and transfer time have been improved by simultaneously writing data to and reading data from the plurality of storage units.

[Problem that the invention attempts to solve]

Among the above conventional techniques, in the method using two magnetic disk drives, the same data must be transferred twice in order to write data to the two magnetic disk drives, and the time required for data transmission is This is twice as much as when using a magnetic disk drive.

However, bubble memory and core memory were more expensive than magnetic disk devices, and had a smaller storage capacity per device volume.

Furthermore, in the above two conventional examples, data is read from the bank amplifier side only when an error occurs in the data on the main side, so if a data error occurs only on the backup side, the There wasn't. Furthermore, no consideration was given to recovery of data in which an error has occurred.

The object of the present invention is to provide a single magnetic disk device.

To provide a magnetic disk device capable of automatically performing backup and eliminating backup time.

A further object of the present invention is to automatically perform backup without special backup time, automatically check for errors in backup data, and recover erroneous data using a single magnetic disk device. The purpose of the present invention is to provide a magnetic disk device with a high level of performance.

[Means for solving problems]

The above object is to divide a recording area of a magnetic disk that rotates on the same rotating shaft into a plurality of areas, and to divide at least two areas out of the plurality of divided areas.

a set of a plurality of magnetic heads for reading and writing the same data at the same time, means for simultaneously making the set of magnetic heads four times old; and means for simultaneously exciting the set of magnetic heads with the same data; , 111 of the above set of throats at once
This is achieved by providing means for detecting errors in the read data or means for detecting errors in the read data, and means for selecting the read data in which the error detecting means has not detected an error.

[Operation J: When writing data to a magnetic disk, a set of magnetic heads is simultaneously excited to simultaneously write the same data in at least two areas. With the above method, data backup is automatically performed, and the backup amplifier time is no longer a concern.

For the purpose of the following explanation, it is assumed that the same data is written in the recording areas of A and B among the plurality of recording areas.

When reading data from a magnetic disk, the magnetic disk device usually selects and outputs the data written in the recording area of the first side. At the same time, the presence or absence of errors in the output data is checked. If an error is detected in the data read from Party A's recording area, the device automatically switches to the data that extends from Party B's recording area and outputs the data. At this time, the magnetic head that reads the recording area B moves at the same time as the magnetic head that reads the recording area A.

Since it is located in the recording area of B, there is no need to take the time to move the magnetic head of B again.

In addition, since the magnetic disk in which Party A's recording area exists and the magnetic disk in which Party B's recording area exists are attached to the same rotating shaft and rotate, are at the same position and do not change.

Therefore, Party A's recording area and Party B's recording area can always be read and written at the same time.In addition, it is possible to check whether there are errors in the data read from Party A's recording area and data read from Party B's recording area. By testing at the same time.

Errors in data on the blow-up side can be detected. If there is no error in the data read out from the recording area of the first side, the data read from the recording area of the first side is selected and output.

If an error is detected in the data read from Party A's recording area, the device automatically switches to the data read from Party B's recording area and outputs the data.

Furthermore, the data read from the recording area of B is stored in the semiconductor memory. After data reading is completed, the data stored in the semiconductor memory is used to rewrite the recording area of A. Conversely, if an error is detected in the data read from the recording area of Party B, the data is read from the recording area of Party A, temporarily stored in the semiconductor memory, and rewritten to the recording area of Party B. By rewriting the above.

Recovery of erroneous data can be performed.

〔Example〕

The following five examples of the first embodiment of the present invention are shown in Fig. 1, Fig. 2, Fig. 3,
This will be explained with reference to FIGS. 4, 5, and 6.

In FIG. 1, the data writing circuit of this embodiment includes a sheet data converter 2, a parallel-to-serial converter 5, an erroneous detection code generation circuit 4, a modulator 5, a first write amplifier 6, a second write amplifier 7, and a first write amplifier 7. It is composed of a magnetic head 8 and a second magnetic head 9. The first magnetic head 8 and the second magnetic head 9 write data on an a-air disk 10. The magnetic disk 10 is attached to a rotating shaft 12 and rotated by a spindle motor 11.

In FIG. 2, in this embodiment, the first head 8 and the second head 9 are on the same surface of the magnetic disk and are attached to a head arm 13. The head arm 13 is rotated by a stepping motor 14.

The first magnetic head 8 and the second magnetic head 9 are connected to the magnetic disk 1.
Move in the radial direction of 0.

In FIG. 6, there are 2N tracks on one side of the magnetic disk 100, and the tracks are numbered starting from 0 from the outermost track to the inner track. In this embodiment, the 2N tracks are divided into two, and the recording area of the N trunks from the 0 track 194 on the outer circumferential side to the N-1 track 19h is defined as the first data surface 16. 1) The recording area of N tracks up to rack 19 ct is used as the second data surface 17, and the recording area is divided into two. The first data surface 16 is read and written by the first magnetic head 8.
The two magnetic heads 8 and 9 are arranged so that the second data surface 17 is read and written by the second magnetic head 9. For example, the first
When the magnetic head 8 is located on the recording cell 184 of the M sector 20a of the 0 track 19a, the second magnetic head 9
Track 19cI7) Recording cell 18A of M sector 204
located above. As shown in FIG. 4, the data recorded in the recording cell 18 includes an error detection code 22 at the end of the data 21 to detect errors in the data when reading the data.
is added.

In FIG. 5, the read circuit of this embodiment includes a first magnetic head 8, a second magnetic head 9, a head switch 25, a read amplifier 24, a demodulator 25, and so on. It consists of an error detector 26°, a serial parallel converter 27, and a read data memory 2 days. The head switch 26 is internally connected as shown in FIG. 5 except when reading data with the first magnetic head 8, but when reading data with the second magnetic head 9, the internal connections are as shown in FIG. The connection shown changes.

In this embodiment, the operation when inputting data will be explained. First, the stepping motor 14 shown in FIG. 2 rotates the head arm 13 to move the first magnetic head 8 and the second magnetic head 9 to the data writing position. Here, it is assumed that the first magnetic head 8 is on the O track 194 in FIG. 3, and the second magnetic head 9 is on the N track 19G.

Next, in FIG. 1, when the input data 1 is transferred, the input data 1 is temporarily stored in the data memory 2. When input data 1 is stored in write data memory 2, data writing is started. The data stored in the write data memory 2 is converted into a serial NRZ signal by a parallel-to-serial converter 3. The false detection code generation circuit 4 generates the false detection code 22 using the data 21 as shown in FIG. 4, and further adds the false detection code 21 to the end of the data 21 and outputs the resultant code. The modulator 5 performs MFM modulation on the NRZ signal obtained by adding the false detection code 22 to the data 21. The MFM modulated signal is simultaneously input to the first write angle 6 and the second head angle, and the first write angle 6 excites the first magnetic head 8, and at the same time, the second write angle excites the second magnetic head 9. . The first magnetic head 8 writes data on the first data surface 16, and the second magnetic head 9 writes data on the second data surface 17, thereby writing the same data in two locations at the same time. For example, M sector 2 of 0 track 19a
The data written in the recording cell 186 of 0a and the data written in the recording cell 187 of the M sector 204 of the N track 190 are the same. That is, the data on the first data surface 16 is backed up to the second data surface 17 at the same time as the data is written to the first data surface 16.

Conversely, in this embodiment, the operation when reading data will be explained. As in the case of writing data above, first, the stepping motor 14 shown in FIG. 2 is rotated, and the first
The magnetic head 8 and the second magnetic head 9 are moved to the data reading position. At this time, the first magnetic head 8 and the first magnetic head 9 move while being fixed to the head arm 15, so that, for example, the first magnetic head 8 is connected to the recording cell 184.
When located above, the second magnetic head 9 is always located above the recording cell 18h. Therefore, the first magnetic head 8 and the second magnetic head 9'li can always read the same data previously written.

In this embodiment, as shown in FIG. 5, normally the head switch 26 amplifies the signal from the first magnetic head 8 by the lead ag 24. The output of the readag 24 is converted back from the MFM modulated signal by a demodulator 25 to an NRZ signal to which an erroneous detection code 22 has been added. The error detector 26 reads data 2 read by the error production code 22.
Check whether there are any errors in 1. At the same time, the data 21 of the NRZ signal outputted from the demodulator 25 except for the error detection code 22 is converted back from serial to parallel by the serial/parallel converter 27 and stored in the read data memory 28 . When the first magnetic head 8 finishes reading the data from the first data surface and the error detector 26 detects no error in the data, the data in the read data memory 28 is output as output data 29. However, the false detector 26
If an error is detected in the data, the connection within the head switch 26 is changed as shown in FIG. 6, and the second magnetic head 9 reads out the booter from the g2 data surface 17.

The signal from the second magnetic head 9 is sent to the read data memory 172 via the same route as the readout by the first magnetic head described above.
8 is stored as read data, and the output data 29
is output as

According to this embodiment, since the moving distance of the magnetic head is the same as that of the conventional method, the time required for moving the magnetic head can be halved.

Next, a second embodiment of the present invention will be explained with reference to FIGS. 7 and 8. FIG. 7 is a configuration diagram of mechanical parts of this embodiment, and FIG. 8 is a diagram showing the arrangement of tracks on a magnetic disk. In FIG. 8, in this embodiment, a first magnetic head 8 and other three magnetic heads 9a, 8k
, 8c, and the four magnetic heads a, aa, sb.
, the head arm 15a that supports the ac, and the magnetic disk 1
04. io! and an enclosure 154 that houses the magnetic head 9 and the other five magnetic heads 9α and 9b, which have the same configuration as above.

9G and the above 411! magnetic heads 9, 9a, 9
A head arm 15b supporting the magnetic disks 10G and 9c, and a magnetic disk 10G.

10ct, a stepping motor 14, and a spindle motor 11 having a rotating shaft 12 projecting upward and downward. And 2
The two enclosures 154, 154 are preferably sealed together. Magnetic disk IQa, 1oa.

1(lc, jQtt) are mounted on the same rotating shaft 12 and rotated by one spindle motor 11. The head arms 1sa and 154 are rotated by one stepping motor 14. 15
A magnetic head S, aa attached to the tip of a.

8k, 8(=) and the magnetic heads 9, 9a, l, 90 attached to the tip of the head arm 15 move simultaneously. For example, when the first head 8 moves to the outermost track 194 of the first data surface 16, Recording cell 18a of M sector 20↓
When in the upper position, the second head 9 is positioned above the second data surface 17.
Recording cell 1 of M sector 206 of outermost track i9c of
Located on aj. Further, the first magnetic head 8 moves from the outermost track 19a to the innermost track 1 of the first data surface 16.
9b, the second magnetic head 9 simultaneously moves to the second magnetic head 9b.
The data surface 17 is moved from the outermost track 19G to the innermost track 19ct. Further, the A-th data surface A6 is in the enclosure 154, the second data surface 17 is in the enclosure 15h, and when pledging data,
Magnetic disks 101 and 10b in enclosure 154
According to this embodiment, the number of enclosures storing the magnetic disks is 21t! Since the same data is written to the magnetic disks in each enclosure at the same time, dust may get into one of the enclosures, or a crash of the magnetic head may cause dust to damage the data on the internal magnetic disks. Even if there is a problem with reading data from the magnetic disk in the other enclosure, the data on the magnetic disk in the other enclosure will not be affected, so it is possible to prevent data from being lost due to contamination with dust or data corruption.

Next, regarding the fifth embodiment of the present invention, FIGS.
This will be explained using figures. FIG. 9 is a configuration diagram of the mechanical parts of this embodiment, and FIG. 10 is a diagram showing the arrangement of tracks on a magnetic disk. Note that the electric circuit is the same as that shown in FIGS. 1 and 5 in the first embodiment. In Figure ag9, the first magnetic head 8 and the other six magnetic heads 86. t3j,
80, helad arm 164, and magnetic disk 1oa,
The enclosure 15α that houses the ICI&, the second magnetic head 9, and the other 51i! magnetic heads 94, 9h, 9
G, head arm 15, and magnetic disk IOC, 1
o* and the enclosure 15b which houses the same are preferably sealed from each other. And magnetic disk 1OA,
10! , 1oc, and 10j are attached to the same rotating shaft 12 and rotated by one spindle motor 11. The stepping motor 144 rotates the head arm 154, and the stepping motor 1jl rotates the head arm 13k.

In this embodiment, the stepping motor 14 is used to move the first magnetic head 8 and the second magnetic head 9 in opposite directions.
6 and 11. For example, when the first magnetic head 8 moves in the movement direction 304, the second magnetic head 9 moves in the movement direction 30G. Conversely, when the first magnetic head 8 moves in the moving direction sob, the second magnetic head 9 moves in the moving direction 50d.

The first data surface 16 and the second data surface 1 which write the same data at the same time are in different enclosures, and the first
As shown in Figure Q, when data is written in the recording cell 18eL of the M sector 204 of the outermost track 191& on the first data surface 16, on the second data surface 17.

The same data is simultaneously written into the recording cell 6 of the M sector 202 of the innermost track 19i. Conversely, the innermost trunk 19b of the first data surface 16 and the second data surface 17
The same data is simultaneously written to the outermost track 190 of .

According to this embodiment, when one magnetic head is located on the jiiJfgJ1 of the magnetic disk, the other magnetic head is located on the outer circumferential side of the magnetic disk and reads the same data. It is possible to compensate for the deterioration in lead margin caused by

Next, as a further embodiment of the present invention, the data succession circuit is divided into two systems, and two magnetic heads are used to generate data simultaneously.
! - An embodiment that enables readout will be described using FIG. 11. In this figure, elements having the same reference numerals as in FIG. 1 are indicated by the same numbers as in FIG.

In the figure, 110 is a first lead amplifier, 111 is a second lead amplifier, 112 is a first demodulator, 116 is a second demodulator, 1
14 is a first false detector, and 115 is a second false detector.

116 is a first serial/parallel converter, 117 is a second serial/parallel converter, 118 is a wJ1 read data memory, 119 is a second read data memory, 120 is a read data selector, 122 is a comparator, and 126 is an internal controller 23. be. Note that in an actual application device, there are eight magnetic heads, for example, as shown in Figure i7, but for the sake of explanation,
Only two magnetic heads are shown. The following description will be made using the pair of magnetic heads shown in FIG. 7 to explain the field recording and reading of signals. Also, lead apricot 110,
111 karariido data memo I7118, 119-,
[The circuit in [] uses two systems of the same circuit '16.

In this embodiment, the readable recording surface of the first magnetic head 8 is the first data surface (upper surface) of the magnetic disk 1Oa in FIG. 7, and the recording surface on which the second magnetic head 9 reads and writes is the magnetic disk 10j. The operation when writing data on the second data surface (lower surface) will be explained. First, the stepping motor 14 simultaneously rotates the first head arm 15a and the second head arm 13h, and simultaneously moves the first head 8 and the second head 9 to the data writing position.

Next, in FIG. 11, when the input data 1 is transferred, the input data 1 is temporarily stored in the write data memory 2. When input data 1 is input to write data memory 2, internal controller 126 sends a command to parallel-serial converter 6 to start writing data. The write data stored in the write data memory 2 is converted into a serial NRZ signal by a parallel-to-serial converter 5. The erroneous detection code generation circuit 4 generates an erroneous detection code and adds it to the end of the write data 21 as shown in FIG. The modulator 5 performs MFM modulation on the combined data of the write data 21 and the erroneous detection code 22. The MFM modulated data is simultaneously input to the first write amplifier 6 and the second write amplifier, and the first write amplifier 6 excites the first magnetic head 8, and at the same time, the second write amplifier 7 excites the second magnetic head 9. do. Data is written to the M1 data surface by the first magnetic head 8,
By writing data to the second data surface by the second magnetic head 9, the same data is written into two locations at the same time. That is, the data on the first data surface is backed up on the second data surface at the same time as the data is written on the first data surface.

When all the data in the write data memory 2 has been erased, the internal controller 123 starts verification. Verification is to extract the data written just before,
This is an operation of comparing the data with the data in the write data memory 2 to confirm that the data has been written without error. In verifying, first the data written immediately before is read.

The data committed to the first data surface is transferred to the first magnetic head 8.
The signal is read out by the first read amplifier 110 and amplified by the first read amplifier 110. The signal output from the first read amplifier 110 is converted back from the MFM modulated signal by the first demodulator 112 to an NRZ signal to which an erroneous detection code has been added. First demodulator 112
The data returned to the NRZ signal by the first error detector 11
4, the error detection code 22 is used to check whether there is an error, and at the same time, the first serial/parallel converter 116 returns the part excluding the error detection code 22 from serial to parallel.
The data is stored in the read data memory 118. The data on the second data surface is.

At the same time as reading the first data surface, the data is transferred from the second magnetic head 9 to the second read angle 111 and the second demodulator 1 through the same route.
13. Second false detector 115, second serial-parallel converter 117
The read data is then stored in the second read data memory 119.

When data reading is completed, the internal controller 12
3, the comparator 122 compares the data in the first read data 118 and the data in the second read data 119 with the data in the write data memory 2. The data in the first read data memory 118 is stored in the write data memory 2.
The second read data memory 118
When the data in the write data memory 2 matches the data in the write data memory 2, the first false detector 114 and the second false detector 115
- The internal controller 2126 determines that the writing was successful only when neither detects an error.
Finish blue-filling and verification.

However, if the data in the first read data memory 118 does not match the data in the write data memory 2, or if an error is detected by the first error detector 114, the same Data is written again at the position and verification is performed on the first data surface.

On the other hand, if the data in the second read data memory 119 does not match the data in the write data memory 2,
If an error is detected by the error detector 115, data is written again at the same position on the second data surface where the data was previously written, and the second data surface is verified.

If the data in the first read data memory 118 and the second
If both data in the read data memory 119 are abnormal, rewriting and verifying are performed on both the first data surface and the second data surface at the same time. Furthermore, if the data does not become normal even after rewriting and verifying the data several times, it is determined that there is an abnormality in the recording location, and the internal controller 123 changes the recording location and writes and verifies the data. Record the recording location where the abnormality occurred and prevent the abnormal recording location from being used from now on. When reading data, first, the first magnetic head 8 and the second magnetic head 9 are activated by the stepping motor 14 shown in the tIXZ diagram. Move to the location where the data to be read is written. Here, the position of the data to be read on the first data surface is temporarily assumed to be the midpoint, and the position of the data on the second data surface is temporarily assumed to be the point O. The data at the middle point is extracted by the first magnetic head 8, and the data at the point A is read out by the second magnetic head 9. In the electric circuit shown in FIG. 11, data at the middle point is stored in the first read data memory 118 and data at the second point is stored in the second read data memory 119 in the same way as when verifying the electric circuit described above. Since the first magnetic head 8 and the second magnetic head 9 move simultaneously and write the same data, the data in the first read data memory 118 and the data in the second read data memory 119 are error-free. are the same as far as possible. After reading data from the first and second data planes to the first and second read data memories 118 and 119, if no error is detected by the first error detector 114, the j7g1 read data memory The data in 118 is output as output data 1214 through the read data selector 120. If an error is detected by the first error detector 114, the internal controller 125 selects the data in the second read data memory 119 using the read data selector 120 and outputs it as output data 121h. Further, the data in the second read data memory 119 is transferred to the write data memory 2, and rewritten and verified at the midpoint on the first data surface. On the other hand, if the second error detector 115 detects an error, the internal controller 123 transfers the data in the first read data memory 118 to the write data memory 2.
Transfer it to.

Rewrite and verify at point O on the second data surface. In rewriting and verifying due to a data error during reading, as in the rewriting and verifying during writing, if an error is detected in the data even after repeating numerical rewriting and 7 eyes, Change the recording position, rewrite and verify, record the recording location where the error occurred, and prevent the abnormal recording location from being used from now on. According to this embodiment, data is read from two data surfaces at the same time. , is stored in the memory, so after an error occurs in one read data, there is no need for time to read the other data anew.

In addition, the internal controller of the magnetic disk device performs verification after writing data, rewrites when an error occurs during verification, and manages abnormalities on the data surface.
The software load on the host 7 system is reduced. Furthermore, because the enclosure that stores the magnetic disk is divided into two, dust may get into one of the enclosures, or a crash of the magnetic head may cause dust, which interferes with reading data from the internal magnetic disk. Even if this occurs, since the same data is recorded on the magnetic disk in the other enclosure, it is possible to prevent data from being lost due to contamination with dust or Heradfluff islands.

〔Effect of the invention〕

According to the present invention, since the same data can be written to multiple locations at the same time, there is no need for special processing for data backup, and no extra data transfer time is required for backup. Also, when reading data, multiple magnetic heads move simultaneously and can reach multiple locations where the same data has been written at the same time, so it takes time to move the magnetic heads again to read the backup data. Not necessary.

Furthermore, when reading data, data on the backup side is also read out at the same time and errors can be detected, so the reliability of the backup data is improved without requiring extra time and operations to read the backup data.

Further, when an error is detected in the read data, the other error-free data is automatically read out and rewritten, so there is no need to transfer the data again from the outside.

As described above, according to the present invention, the reliability of a magnetic disk device can be improved without requiring time and operations for processing backup data.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a data writing circuit according to the first embodiment of the present invention, FIG. 2 is a mechanical configuration diagram used in the embodiment of FIG. 1, and FIG. 3 is a recording area in the first embodiment. FIG. 4 is a diagram showing the positional relationship between data and error detection codes in the first embodiment. FIG. 5 is a block diagram of the data reading circuit in the first embodiment. The figure is the first
FIG. 3 is a diagram showing an example of internal connections of a head switching device in the embodiment. FIG. 7 is a mechanical configuration diagram of the second embodiment, FIG. 8 is a tosic layout diagram of the second embodiment, FIG. 9 is a mechanical configuration diagram of the third embodiment, and FIG. 10 is a mechanical configuration diagram of the sixth embodiment. FIG. 11 is a block diagram of the data writing and continuous output circuit of the fourth embodiment. 1...Input data, 2...Write data memory. 6... Parallel-serial converter, 4... Erroneous detection code generation circuit, 5... Modulator, 6...
...First ride door/gun, 7.Second write amplifier, 8.First magnetic head, 9.Second magnetic head, 10.Magnetic disk, 11.Spindle motor% 12.・
... Rotating shaft, 13... Head arm, 14... Stepping motor. 15...Enclosure τ% 16...1st y
'-It plane. 17... Second data surface, 18... Recording cell, 19... Track, 20... Sector, 21... Data, 22... Erroneous detection code. 26...Head switcher, 24...Lead amplifier, 2
5... Demodulator, 26... False detector, 2
7...Series-parallel converter, 2nd...Read data memory, 29...Output data, 30-#*b-J tso, -off 4.4.19 Ken No. 1 Figure Atsushi 2 Figure 3 Figure Ivy Cow Figure 5 What Figure 2 Output day! Atsushi 7 Figure 8 Figure 9 Figure 7 Theta

Claims (1)

[Scope of Claims] 1. In a magnetic disk device, means that can simultaneously excite a plurality of magnetic heads with the same write data, and means that can simultaneously move the plurality of magnetic heads that can be simultaneously excited by the excitation means; One or more magnetic disks that are attached to the same rotating shaft and rotate are provided, and the plurality of magnetic heads that can be simultaneously excited are arranged in the recording area of the magnetic disk, and the recording area of the magnetic disk is A magnetic disk characterized in that it is divided into a plurality of areas in which each of a plurality of simultaneously excitationable magnetic heads writes the same data at the same time, and one or more of the plurality of divided recording areas is used for backup. Device. 2. means for detecting an error in one or more pieces of data out of a set of data read simultaneously by the plurality of simultaneously excitationable magnetic heads; 2. A magnetic disk drive according to claim 1, further comprising means for selecting one or more pieces of data from a set of data. 3. A memory means is provided for storing one or more pieces of data out of a set of data read simultaneously by the plurality of simultaneously excitationable magnetic heads, and when an error is detected in the data by the error detection means, 3. The magnetic disk device according to claim 2, wherein error-free data is rewritten from the memory means to the magnetic disk. 4. A plurality of enclosures for storing the magnetic disks and the magnetic heads are provided, and the same data is simultaneously recorded on the magnetic disks in different enclosures. The magnetic disk device according to item 3. 5. When data is written from the outer circumferential side of the magnetic disk to the inner circumferential side of the magnetic disk in Party A's recording area among the plurality of recording areas in which the same data is simultaneously written, the data is written in the recording area of Party B on the magnetic disk. A magnetic disk device according to claim 1, 2, or 3, characterized in that writing is performed from the inner circumferential side to the outer circumferential side.