JP2000020248A - Disk fault recovering method, disk driver, and recording medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a disk fault recovering method, a disk driver and a disk which are applicable to even an inexpensive stand-alone type computer device by dividing the disk into areas and automatically duplexing data. SOLUTION: The disk driver 2 sends a command to a disk device and receives a stylus from the disk device to control access to the disk 1. The disk 1 is divided (partitioned) into a system block where an OS and application programs are stored, a data block 12 where data are stored, and a backup block 13 where the same data as the data written to the data block 12 are stored and the disk driver 2 actualizes mirroring for even a disk device which does not have disk array structure by writing the data to the data block 12 and backup block 13 to make a fault recovery of data.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to a disk failure recovery method, a disk driver (disk control program) for realizing the disk failure recovery method, and a disk (recording medium).

[0002]

2. Description of the Related Art Conventionally, as a speed improvement idea, there has been a disk array system in which a plurality of disks are arranged in an array to shorten a seek time and a rotation waiting time to increase the speed. In configuring a plurality of disks in an array, the number of components is inevitably increased, and as a result, there has been a problem that reliability is reduced. Including redundant information in the array configuration to solve such a problem, RAID (Redundant Redundancy) is a technology that has both high-speed access to disk devices and failure prevention measures.
ant Arrays of Independent (Inexpensive) Disks). (1) RAID1 RAID is a technology that divides data and stores it on a plurality of disks to realize high-speed access and high reliability. RAID is classified into several levels depending on the difference between the divided storage and the method of troubleshooting. Dr. Patterson and others at the University of California, Berkeley (Patterson)
son, DA, Gibson, G, Katz, RH, "A Case for Redundant Arr
ays of Inexpensive Disks (RAID), "Report No.UCB / CSD
87/391, Computer Science Division, University of Cal
ifornia Berkeley, Dec 1987) defines five types of RAID. Although RAID0 is not included therein, RAID0 is considered as a type of RAID because of its usefulness. Therefore, RAID is currently classified into six levels of RAID0 to RAID5 as shown below. The efficiency of disk capacity and the efficiency of data access differ depending on the level. RAID0 striping RAID1 mirroring (duplexing) RAID2 striping (byte) and ECC by Hamming code RAID3 striping (byte) and parity drive fixed RAID4 striping (block) and parity drive fixed RAID5 striping (block) and parity drive distribution For example, in RAID1, disk Performs duplexing (mirroring). That is, as shown in FIGS. 9 (a) and 9 (b), two sets of data having exactly the same contents (two or more sets may be used).
To the discs 94-1 and 94-2. This allows
Even if a certain disk fails, the processing can be executed using the normal disk. It should be noted that RAID1 includes software RAID for performing duplication by processing by software such as a disk driver 92 of the host computer 91 as shown in FIG. 9A, and a disk device as shown in FIG. 9B. There is a hardware RAID that performs duplication using hardware such as the controller 93.

In both cases, writing is performed simultaneously on both disks 44-1 and 44-2 as shown in FIG.
Reading is performed on one disk 4 as shown in FIG.
Perform from 4-1. When one disk 44-2 fails as shown in FIG. 10C, the operation is performed only with the other disk 44-1. That is, in the software RAID of FIG. 9A, the disk driver 42 writes the same data to the disks 44-1 and 44-2 at the same time to perform the duplication. Also, the hardware RAID shown in FIG.
Then, the controller 43, which has received a request or command from the host computer 41, writes the same data to the disks 44-1 and 44-2 and simultaneously duplicates the data.

[Single Disk System] In a single disk system, as shown in FIG. 11, data is written / read using a single disk, and the reliability of data written on the disk is taken into consideration separately. Not. In the single disk system, the device driver 92 'of the host computer 91 issues a command (write / read command) to the disk device,
Data transfer is performed to write / read data to / from the disk 94. (2) Others As another disk array failure recovery method,
Japanese Patent Application Publication No. 230903 discloses a technique for recovering a disk array from a failure by securing an empty area as a failure recovery area when a failure occurs. The technology described in the above publication is a technology for improving failure recovery in RAIDs 3 to 5 (a method using striping and a parity drive). When a failure occurs in a disk array to which any of RAIDs 3, 4, and 5 is applied, a normal drive is used. Is secured as a failure recovery area, the data of the failed drive is distributed and stored in the recovery area, and the recovered data 2
The next parity is generated to prevent data loss when a new drive fails.

[0005]

SUMMARY OF THE INVENTION In recent years, Windows
With the adoption of an operating system (OS) such as Windows 95 or Windows NT (both are registered trademarks of Microsoft Corporation in the United States), the amount of access to the hard disk (HDD) and the amount of data increase, and the frequency of hard disk failures increases. The damage such as data loss due to is increasing. Disk failures include a disk crash due to a failure of a mechanical mechanism such as a write / read head, a sector error when the disk is scratched, a data error due to erroneous writing due to noise, and the like.

[0006] RAID 1 can protect important data in the event of a hardware failure of a disk.
Since the same data is written to two (or 2n) disks, the cost is higher than in a single system, and the usable disk capacity is substantially reduced to half or less. However, in the above-described software RAID (FIG. 9A), it is necessary to access the disk twice at the time of data writing, so that the writing time is more than twice as long as the normal access, and the hardware RAID (FIG. 9B) )) Has a problem that the cost of the controller is further required.

[0007] Further, in the single disk system, since there is no redundancy in the disk, there is a disadvantage that data cannot be restored when an error occurs in the disk.

In the disk array failure recovery method described in JP-A-6-230903, a temporary vacant area is secured as a failure recovery area when a failure occurs, and the penalty at the time of failure occurrence is large. There is a problem.

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems and disadvantages, and divides and stores data in each disk of a disk array in order to realize high-speed access and high reliability in the disk array. Unlike a RAID, a disk device that does not have a disk array configuration is provided with a disk failure recovery method, a disk driver, and a disk that can recover data with a high probability when a disk failure occurs. It is an object of the present invention to provide a disk failure recovery method, a disk driver, and a disk that can be applied to a low-cost stand-alone computer device by automatically duplicating data.

[0010]

In order to solve the above problems, a disk failure recovery method according to a first aspect of the present invention divides a disk area into three blocks including two blocks having the same capacity.
The same data is written to the two blocks having the same capacity, and when a disk error occurs in one of the blocks having the same capacity, access is made to a block in which no error has occurred.

A disk driver according to a second aspect of the present invention includes:
A disk driver comprising a program for controlling access to a disk device which operates internally or externally connected to a computer device, wherein three blocks including two blocks of the same capacity when a data write request to a disk is requested. After writing of data to one block of the same capacity of the divided disk is completed, a write command for writing the same data to another block of the same capacity is further issued to the disk device, and the data is written to the disk device. When there is a data read request to the disk, data is read exclusively from a predetermined block of blocks of the same capacity, and when an error occurs in a predetermined block, data is read from another block of the same capacity. Reading is performed.

According to a third aspect of the invention, in the disk driver of the second aspect, when writing data to two blocks having the same capacity, the sector for which access has been completed is held, and the current writing is performed by the sector. It is characterized in that writing control is performed from a block including the block.

According to a fourth aspect of the present invention, in the disk driver according to the second or third aspect, the other block having the same capacity is used as a backup block, and a request by an application is prohibited in the backup block. And

A recording medium according to a fifth aspect of the present invention is a recording medium in which data is written / read by a disk device which operates by being connected to a computer device internally or externally. It is divided into three blocks including two blocks, and one of the two blocks having the same capacity is set as a data block in which data reading has priority over the other block, and the other of the two blocks having the same capacity is used as the data block. A backup block to be written is provided, and blocks of the same capacity among the three blocks are divided into system blocks for reading / writing program or interpreter data.

[0015]

DESCRIPTION OF THE PREFERRED EMBODIMENTS <Disk Partition> FIG.
FIG. 1 is a diagram showing an embodiment of a disk area configuration to which a disk failure recovery method of the present invention can be applied. In the present invention, a disk is configured by dividing into three blocks (partitions).

In FIG. 1, a disk 1 stores a program such as a control program, an application program and a utility program, an interpreter code, a system block 11 for storing intermediate values and the like when the program is executed, and a data file. It has a data block 12 to be stored and a backup block 13 for storing the same data as the data written in the data block 12. In this embodiment, a hard disk (HDD) is used as the disk 1, but the disk 1 is not limited to this.

These blocks 11 to 13 are shipped by dividing the physical continuous area (usually in units of sectors) of the disk by the manufacturer in advance or allocated when formatting the disk (the data block 12 and the backup block 13 are the same). Size is assigned). Further, a log area for writing the error occurrence location of the disk 1 and its type (whether a write error or a read error, etc.) may be secured on the disk 1 (in the present embodiment, the log area is set to the computer apparatus main body). (This is secured in a nonvolatile memory such as a flash memory provided in the (host).)

In this embodiment, as described above, a case where the computer device has one disk device 100 built-in or externally attached is taken as an example, and one disk is composed of a system block, a data block, and a backup block. Although divided into blocks, if two or more disk devices can be built-in or externally attached to a computer device, the area of each disk (recording medium) driven by two or more disk devices is contiguous. It may be divided into three blocks considering the target area. Also, the number n of disk devices built in or external to the computer device
Is set to n = a + 2m, a units can be regarded as system blocks, m units as data blocks, and other m units as backup blocks (a, m ≧ 1).

<Overview of Disk Backup System>
2A and 2B are explanatory diagrams of a disk failure recovery method according to the present invention. FIG. 2A shows the state of access to a disk during normal operation of an application program, and FIG. 2B shows the state when a failure occurs in a data block. It is a figure showing an operation state.
In the following description, an example is described in which one disk device is built in or externally attached to a computer device (hereinafter, referred to as a host). However, as described above, a plurality of disk devices are built in or externally attached to a host computer. It may be. That is, in a disk array, data is divided and stored discretely on each disk (parallel storage), but in the present invention, data is written and read in a continuous block (serial storage).

In FIG. 2, the disk driver 2 gives a command to the disk device 100,
And a disk control program group for controlling access to the disk 1 while exchanging status with the OS, and includes a disk backup program group having the following functions based on the present invention for disk backup. . Note that the disk driver may be configured with hardware (disk controller) including a disk backup module). Further, the disk driver 2 may be located on the host side or on the disk device side.

In FIG. 2, access to the disk 1 is performed by writing to both the data block 12 and the backup block 13 during writing, and reading is performed from the data block 12 during normal operation. When an error (sector error or the like) occurs in one block, reading / writing is performed in the other block.

[Overview of Function of Disk Driver] 1. Blocks with high access frequency, ie, system blocks 11 are not mirrored (duplicated data), and data (application programs) other than data written to system blocks 12 Mirroring of data written / read to / from a data file created or updated by 2. A backup block 13 of the same size (capacity) as the data block 12 is secured on the disk 1 for mirroring the data. 3. Depending on the user (that is, user application program or user operation), access to the backup block 13 cannot be read or written. 4. Mirroring is performed in data units, not file units. 5. An access request for reading data from the OS,
That is, when there is a data read access request, the highest priority is given to reading data from the data block 12. 6. When there is an access request for data writing from the OS, that is, a data write request, access is made from a block close to the previously accessed sector in order to increase the speed. For example, the backup block 13
In the current access, the data is written to a position in the backup block 13 close to the sector (the next sector), and then the same data is written to the data block 12. 7. Data block 12 or backup block 1
In the case where an error occurs in any one of the disk drives 2 and 3, the disk drive 2 does not return an error occurrence status (a flag indicating an error occurrence state) to the OS as in the case of the single system (FIG. 11). First, the data to be written is written in the block where no error has occurred, and in the case of reading, the data block 12 is used.
Data to be read from the backup block 1
3 and then read the error notification (for example,
Error notification to the user by warning lamp display, warning message display, warning sound output, warning message output, etc.)
Perform 8. Data block 12 or backup block 1
In the case where an error occurs in any one of the blocks No. 3, as described above, in the case of writing, the data to be written is written to an error-free block.
Thereafter, transfer of the write data is performed only to the block where no error has occurred. In the case of reading, data block 1
2 is read from the backup block 13, and thereafter, read access is performed only to the backup block 13. 9. The disk error log at the time of error occurrence (disk error information indicating the date and time when the error occurred, the disk in which the error occurred, the sector in which the error occurred, the type of error, etc.) is stored in the non-volatile memory of the host. Displayed when an error has occurred on the disc. 10. When an error occurs in the system block 11, an error occurrence status (a flag indicating an error occurrence state) is returned to the OS. Subsequent processing is based on an instruction from the OS. 11. If an error occurs in both the data block 12 and the backup block 13, an error status (a flag indicating an error occurrence state) is returned to the OS. Subsequent processing is based on an instruction from the OS. In this case, the disk driver 2 sets the current block as an access block so as not to prohibit access to the current block (that is, to enable access to the current block according to an instruction from the OS). Returns the error status.

<Operation of Disk Driver> [During Normal Operation] FIG. 3 shows normal operation of the disk device 100 (operation when a disk error has not occurred).
FIG. 7 is an explanatory diagram of a basic operation of the disk driver 2 when the operation is performed.
Read / write access to 1 or data block 1
FIG. 2B is an explanatory diagram of the basic operation of the disk driver 2 when there is a data read request for reading data from the data block 2, and FIG. 2B shows data access of the disk driver 2 when there is a data write request to the data block 12. FIG. 4 is an explanatory diagram of a basic operation at the time. FIG. 4 is a flowchart showing the basic operation of the disk driver 2 when accessing data. Hereinafter, a basic operation at the time of data access by the application program of the disk driver 2 will be described with reference to FIGS. 3 (a) and 3 (b), the symbols S0 to S18 of the arrows indicate the steps in the flowchart of FIG.

3 and 4, when the computer device receives a disk access request from the application program 3, the OS 4 issues a disk access request to the disk driver 2 (S0).

When the disk driver 2 receives the disk access request from the OS 4, it checks the request (S1). If the request is a read request, the disk driver 2 reads the read command and the read start sector of the data block 12, the number of transfer sectors, and the like. When a disk access request from the OS 4 is a write request, a write command and a write command including a write start sector and the number of transfer sectors of the data block 12 are issued to the disk device 100.
(S2).

The disk driver 2 is a disk device 100
After issuing the command to the OS 4, an interrupt wait process is performed, and the interrupt wait status (status flag, etc.) is passed to the OS 4 (OS
When 4 receives the interrupt wait status, the OS 4 notifies the application program 3 that the disk device 100 is waiting for an interrupt. In this case, if the processing of the application program 3 does not involve data transfer to (or from) the disk 1, the processing can be executed. Otherwise, the processing is performed until a transfer start notification is received from the OS 4. Wait for execution). When the disk device 100 becomes transfer interruptible (accessible), it returns an interrupt status to the disk driver 2 (or the disk driver 2 searches for a status byte provided in the disk device 100 at a predetermined timing, and It is determined whether or not the interrupt equivalent bit has been turned on) (S3).

The disk driver 2 executes the above-mentioned step S0
Examines the disk access request received from the OS 4 (S4), and returns an interrupt status to the OS 4 if the disk access request is a read request (S4).
5) issuing a data transfer permission to the disk device 100,
The data (read data) is transferred from the data block 12 of the disk 1 to the read buffer (memory) on the application side (S6), and after the final interrupt processing is completed (= transfer completed: the number of transfer sectors is determined by counting). (S7), a request end process is performed (S7).
8) A termination notification is issued to the OS 4 (S9).

If the disk access request received from the OS 4 in step S0 is a write request, the disk driver 2 returns an interrupt status to the OS 4 (S10), sends a data transfer notification to the disk device 100, (Write data) from the read buffer (memory) on the side to the data block 12 of the disk 1 (S1).
1) When the writing to the data block 12 is completed (S1)
2) The disk device 1 without performing the final interrupt processing
00, a write command to the backup block 13 (including a write command, the number of write start sectors of the backup block 13, the number of transfer sectors, etc.) is issued (S1).
3) When the disk device 100 returns an interrupt status indicating that data can be written (transferred), the status is set to O.
The data is transferred to S4 (S14), a data transfer notification is issued to the disk device 100, and the same data (write data) as written in the data block 12 from the read buffer (memory) on the application side is stored in the backup block 1
3 (S15).

When the final processing of the data transfer to the disk device 100 (= writing to the backup block 13) is completed (S16), the disk driver 2 performs a request end processing (S17) and passes an end status to the OS 4 (S18). ).

By the above operation, the disk device 100 reads data from the data block 12 of the disk 1 during normal operation, and writes the same data to the data block 12 and the backup block 13 for writing. That is, mirroring can be performed by a computer device having a minimum disk configuration in which one disk device is built in or externally mounted.

Although the basic operation at the time of data access by the application program has been described in the above description of FIG. 4, the operation at the time of program access by the OS, that is, the access of the system block 11, is described in the flowchart. S0, steps S12 and S16 are omitted, and step S
The blocks accessed in steps 1 to S11 and steps S16 to S18 may be replaced by the system block 13 instead of the data block 12, the application program 3 may be the OS 4, and the data may be the program.

[Operation when Error Occurs] FIG. 5 is an explanatory diagram of the basic operation of the disk driver 2 when an error occurs in the disk device 100. FIG. FIG. 7B is an explanatory diagram of a basic operation of the disk driver 2 in the case where the error occurs, and FIG. 7B is an explanatory diagram of a basic operation of the disk driver 2 when an error occurs in the data block 12 or the backup block 13. (: System block error) In FIG. 5 or FIG. 6, the disk driver 2 checks the status sent from the disk device 100 (T1). If an error of the system block 11 occurs, the disk driver 2 passes the error status to the OS4. Error recovery processing is entrusted to the OS 4 (T2).

(Errors in Data Block and Backup Block) If an error occurs in the data block 12 or the backup block 13 in step T1, the data block 12 and the backup block 13
It is determined whether an error has occurred in both the data block 12 and the backup block 13 (if an error has occurred in the data block 12 and the backup block 13 at the same time, The transition to T4 occurs when an error has occurred and another error has occurred in the other one), and the transition to T6 occurs when one of the data block 12 and the backup block 13 has an error. (T3).

(: Both the data block and the backup block have errors) In step T3, the data block 1
When an error occurs in both the backup block 2 and the backup block 13, the error status is passed to the OS 4, and the error processing is entrusted to the OS 4 or the application program 3. In this case, the disk driver 2 sets the current block as an access block so as not to prohibit access to the current block (that is, to allow access to the current block according to an instruction from the OS). An error occurrence status is returned to the OS (T4),
Notify the user of the error. In the embodiment, since the error notification is performed by the error notification program 5, the disk driver 2 transits to the error notification program in this case (T5).

(Either Data Block or Backup Block Error) In step T3, if an error occurs in either the data block 12 or the backup block 13 and the error occurs during read access, the data is deleted. Read from backup block (data transfer). That is, when an error occurs in the data block 12, the current transfer sector number is added to the read start address (sector) of the backup block 13 and given to the OS 4 as a read restart address, and data is read from the backup block 13. From (T6), the disk driver 2 transitions to the error notification program and notifies the user of a read error (T7). In step T3, when an error occurs during the write access, when the error occurs in the data block 12, the backup block 1 writes the same data to the backup block.
The write start address (sector) of No. 3 is given to the OS 4 and a write command is given to the disk device 100 to start rewriting data in the backup block 13 (T8), and transit to T11 (T9). If an error occurs in the middle of the write access and the error occurs in the backup block 13, the write interruption status is passed to the OS 4 to terminate the write access (T10). The disk driver 2 transitions to the error notification program and causes the user to perform a write error notification (T11).

[Operation for Improving Disk Access Efficiency] In the disk failure recovery method of the present invention, since the disk driver 2 performs mirroring for writing the same data to the data block 12 and the backup block 13, the writing to one block is completed. When the access to the other block is started, a seek time is required when the write head is moved each time as in RAID1, and when a single disk is used, a single disk system (FIG. 1) is used.
In the present invention, it takes twice as long as 1). However, according to the present invention, when the computer device is started or an application is executed, the OS or the program is read from the system block 11, the set value is written, and the data from the data block 12 are written. At the time of reading, the same access as in the case of the single system is performed as will be described later, so that a decrease in time efficiency due to mirroring is minimized. Also,
As for the disk capacity, the system block 11 is not duplicated to increase the disk capacity ratio between the data block and the backup block.

(: Data Read Operation) FIGS. 7A and 7B are explanatory diagrams of the data read operation control of the disk device 100 by the disk driver 2. FIG.
FIG. 4 is an explanatory diagram of a read operation for the system block 11 when the disk drive 10 is operating normally.
FIG. 8 is an explanatory diagram of a read operation for the data block 12 when 0 is operating normally, and FIG.
FIG. 9 is an explanatory diagram of a read operation (when an error occurs) for the backup block 13 when an error occurs in FIG.

As shown in FIGS. 7A and 7B, the read operation when the disk device 100 is operating normally is as follows.
Only the operation of transferring the data of the block requested by the OS is performed, and the access speed in this case is similar to that of the single disk system. FIG. 7 (c)
When an error occurs in the data block 12 as shown in (1), the transfer from the data block 12 is not performed.
Since the data is transferred from the backup block 13, the access speed in this case is the same as that of the single disk system.

(: Data Write Operation) FIGS. 8A and 8B are diagrams for explaining the write operation control of the disk device 100 by the disk driver 2. FIG. 8A shows the case where the disk device 100 is operating normally and the data block 12 shown in FIG. FIG. 8B is an explanatory diagram of an access block in the next write when data is written to the backup block 13. FIG. 11B shows an access in the next write when data is written to the backup block 13 when the disk device 100 is operating normally. It is explanatory drawing of a block.

As shown in FIGS. 8 (a) and 8 (b), the write operation when the disk device 100 is operating normally is performed in a single system in order to shorten the seek time (movement time of the head). Similarly, by accessing the first sector to be written in this access from a block close to the previously accessed sector, the data write time is shortened as a result.

More specifically, when data writing is started from the data block 12 in FIG. 8A, the same data is written from the backup block 13 after writing to the data block 12 is completed to perform data mirroring. Then, the writing is completed at a certain sector (p) of the backup block 13, and the position of the writing end sector (p) is held in the memory. Here, if writing is started from the data block 12 in the next access, a seek time is required. Therefore, in order to reduce the seek time to almost zero, the write end sector (p) of the backup block 13 which is the end position of the previous access. (P +
1) Writing is started from a block including the sector or the end sector (p) (in this case, the backup block 13), and when writing to the backup block 13 is completed, writing of the same data to the data block 12 is started and the data block is started. The writing is completed in a certain sector (q) of twelve. The next access operation is performed in the data block 12.
Is the position closest to the write end sector (q) of (q +
1) When writing is started from a block including the sector or the end sector (in this case, the data block 12) and writing to the data block 12 is completed, writing of the same data to the backup block 13 is started. By repeating such an access operation, the seek time is reduced.

That is, assuming that the time for moving the header from the backup block 13 to the data block 12 is t, in the method of starting writing from the data block 12 for each access, n-times access is performed (n-1).
It takes a seek time from the backup block 13 to the data block. On the other hand, in the method of starting the current access from the sector near the previous access end position as in the write operation of the present invention, the header is moved from the sector near the previous access end position to the position where the current data is written. When the time is t ′ and the access is performed n times, (n−
1) Although a seek time of t 'is required, since t'<< t, (n-1) t '<< (n-1) t, and an access start seek time can be greatly reduced.

Further, by providing a log area for storing error information at the time of occurrence of an error in the nonvolatile memory, it is possible to notify the user of the occurrence of the error without accessing the disk at the next power-on.

[0044]

As described above, according to the present invention,
When protecting data from hard disk failures,
Disk error is not intended to avoid or prevent hard disk failure, but is to mirror data for the purpose of accepting and recovering from hard disk failure as a trivial event. Data can be recovered with a high probability in the event of an occurrence. In particular, the present invention can be applied to a low-cost stand-alone type computer device, for example, a computer device which is configured to be operable with one disk device connected or built therein.

In a disk array, data is divided and stored discretely in each disk. However, in the present invention, data is written and read in continuous blocks, so that the configuration of a disk driver is simpler than in the case of RAID. Can be Further, a plurality of disks may be built in or external to the host computer.

Since the current write operation is performed while holding the previous access position, high-speed mirroring can be realized.

[Brief description of the drawings]

FIG. 1 is a diagram showing an embodiment of a disk area configuration to which a disk failure recovery method of the present invention can be applied.

FIG. 2 is an explanatory diagram of a disk failure recovery method according to the present invention.

FIG. 3 is an explanatory diagram of a basic operation of the disk drive when the disk device is performing a normal operation.

FIG. 4 is a flowchart showing a basic operation of the disk driver when accessing data.

FIG. 5 is an explanatory diagram showing a basic operation of a disk driver when an error occurs in a disk device.

FIG. 6 is a flowchart showing a basic operation of the disk driver when accessing data.

FIG. 7 is an explanatory diagram of data read operation control of a disk device by a disk driver.

FIG. 8 is an explanatory diagram of write operation control of a disk device by a disk driver.

FIG. 9 is an explanatory diagram of a dual disk (mirroring) system using RAID1.

FIG. 10 is an explanatory diagram of a data recovery operation at the time of a disk failure by RAID1.

FIG. 11 is an explanatory diagram of a single disk system.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Disk (recording medium) 2 Disk driver 3 Application program (application) 11 System block 12 Data block 13 Backup block 100 Disk device

Claims (5)

[Claims] 1. A disk area is divided into three blocks including two blocks of the same capacity, and the same data is written in the two blocks of the same capacity. When a disk error occurs in one of the blocks of the same capacity, no error is detected. An error recovery method for a disk, comprising accessing an occurred block. 2. A disk driver comprising a program for controlling access to a disk device that operates internally or externally connected to a computer device, wherein when a data write request to a disk is requested, the disk driver has a capacity of two. After writing data to one block of the same capacity of the disk divided into three blocks including blocks, a write command for writing the same data to another block of the same capacity is further issued to the disk device. When a request to read data from the disk is made, data is exclusively read from a predetermined block of the blocks having the same capacity, and an error occurs in the predetermined block. Reading data from another block of the same capacity. I disk driver. 3. When data is written to the two blocks having the same capacity, a sector which has been accessed last time is held, and writing control is performed from a block including the sector in writing this time. The disk driver according to claim 2. 4. The disk driver according to claim 2, wherein another block of the same capacity is used as a backup block, and a request by an application is prohibited in the backup block. 5. A disk drive which operates by being connected to a computer device internally or externally to write / read data.
A recording medium from which reading is performed, wherein the area of the disc is divided into three blocks including two blocks of the same capacity, and one of the two blocks of the same capacity is given priority over data reading over the other block. Data blocks, the other of the two blocks having the same capacity being a backup block for writing the same data as the data block, and the non-capacity block of the three blocks being a system block for reading / writing program or interpreter data. A recording medium characterized by being divided.