JP2000039970A - System for controlling double failure prevention of disk array system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To normally interpolate data by correcting the data of a defective disk by using the data of another disk and the first parity when any failure is generated in one part of disks, and correcting an error from a part other than the error when any error is generated in the data and a second parity. SOLUTION: At the time of initializing an array disk(A/D) 2, an array controller(A/C) 1 prepares and initializes vertical parities in a block in addition to parities among each disk 20-23. The vertical parities are preserved in the A/C1 and each disk 20-23. When a write command is issued from a CPU 3 to the A/C1, the A/C1 reads data before update from an A/D2 at first. Then, the parities among each disk 20-23 are calculated, and the vertical parities are prepared from a difference between the data read from each disk at first and data to be written, and the data and the parities are written in the A/D2. Thus, the double failure of the array system can be reduced to the minimum.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to a double failure prevention control system for a disk array system.

[0002]

2. Description of the Related Art Conventionally, a disk array system stores valuable data such as a database and is used in a situation where data loss due to a disk failure is not allowed.

A related prior art is disclosed in
"Error Correction Device for Data Storage Device" described in JP-A-9-110098, "Failure Recovery Method and Disk Array Device for Disk Array Device" described in JP-A-6-230903, JP-A-6-289998 "A disk array device" described in JP-A-7-28608, "Improvement of a highly secure disk array" described in JP-A-7-28608, and "Error correction device for storage device" described in JP-A-9-167120. .

[0004]

When the disk array system is also operated in a normal state, the system is safe against a disk failure and a media error. However, if one disk fails, the reverse occurs. Can be a very fault prone system.

SUMMARY OF THE INVENTION An object of the present invention is to provide a system in which data can be normally interpolated even when one disk fails and data is reconstructed on a substitute disk even if a media error occurs on the remaining disks. .

[0006]

According to the double failure prevention control method of a disk array system according to the present invention, a first parity generated based on data written to another disk is written to some disks, An array disk in which a second parity generated based on data in the same disk is written to the same disk; first parity generating means for generating the first parity; A second parity generating means for generating the data, and when a failure occurs in some of the array disks,
A first correcting unit for correcting data of a failed disk by using data of another disk or the first parity, and an error occurring in data of the disk having the array disk or the second parity. And a second correcting means for correcting an error from a portion other than the error sometimes.

The double failure prevention control method for a disk array system according to the present invention is the above-mentioned double failure prevention control method for a disk array system, wherein the array disk is configured by correcting corrected data to first and second parity. Is provided with a spare disk for writing.

Further, in the double failure prevention control method for a disk array system according to the present invention, in the double failure prevention control method for a disk array system described above, the disk in which the first parity is written has the first parity. It is characterized in that it differs for each included codeword.

Further, a double failure prevention control method for a disk array system according to the present invention is the above-mentioned double failure prevention control method for a disk array system, further comprising storage means for storing the second parity. The parity generation means is configured to determine the difference between the current data and the data newly written to the array disk and the second data in the storage means.
The second parity in the storage means is updated on the basis of the parity.

[0010]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The double failure prevention control method for a disk array system according to the present invention minimizes a double failure by adding redundant data to the disk array system.

In FIG. 1, when the array controller 1 initializes the array disk 2, in addition to the parity between the disks 20 to 23, the array controller 1 defines the blocks in the disks 20 to 23 as shown in FIG. Creates and initializes the vertical parity of This vertical parity is stored in the array controller 1
And stored in each of the disks 20 to 23. When a write command is issued from the CPU 3 to the array controller 1, the array controller 1 first reads data before being updated from the array disk 2. Next, each disk 20
Simultaneously with the calculation of the parities between ２３ and ２３, a vertical parity is first created from the difference between the data read from each disk and the data to be written, and then the data and the parity are written to the array disk 2. This makes it possible to use the vertical parity to recover data on the disk when a media error occurs in another disk used for recovery during recovery work in the event of an array failure due to a disk failure. Therefore, double failure of the array system can be minimized.

Referring to FIG. 1, an array controller 1
Are connected to a CPU 3 that issues write / read instructions and an array disk 2 that is a group of several disks that hold data.

Next, a specific configuration of the array controller 1 and a method of using each disk will be described.

FIG. 2 is a block diagram showing a configuration example of the array controller 1 according to the embodiment of the present invention. FIG.
In FIG. 1, the array control circuit 10 is a central circuit and controls each device and circuit. The array control circuit 10 includes a disk controller 11, a vertical parity holding memory 1,
2. The nonvolatile memory 13 is connected. The vertical parity holding memory 12 and the nonvolatile memory 13 are used for the purpose of reducing the amount of access to the array disk 2.

FIG. 3 shows a configuration example of the array disk 2.
Taking the data written at the head of the array disk as an example, data A11, data B11, and data C11 are one set of data, and the parity for the set is parity D11. The following data is similarly configured by changing the parity disk in order. The above is an example of the data configuration of a disk array currently generally used. In the present embodiment, a block (eg, block A1) is defined only in each disk, and a parity (example) for the block is defined. : The vertical parity A1) is also stored in the disk.

Next, the operation of the system shown in FIGS. 1 to 3 will be described with reference to the drawings.

First, when using the disk array system of this embodiment, it is necessary to initialize the disks. At this time, the parity between the disks 20 to 23 is generated, and the blocks in each of the disks 20 to 23 are generated. Definition and generation of vertical parity in the block.

In operating the disk array system,
When the power is turned on, first, the array controller 1 reads all the vertical parities of the disks 20 to 23 from the vertical parity holding memory 12.

When a data write command is issued from the CPU 3 to the array controller 1, first, the array control circuit 10 reads data from the area to be subjected to the data write from the array disk 2. The vertical parity of the block corresponding to the area to be subjected to the data write is also read from the vertical utility holding memory 12 at the same time.
Next, a new vertical parity is calculated from the difference between the data before the update and the data to be written and the current vertical parity.
The parity between the disks is also created from the data in the area as before, and then the data and the parity between the disks are written into the array disk 2 and the vertical parity is written into the vertical parity holding memory 12, respectively. When the data in the disk array is updated after the power is turned on, a data update bit is set in the nonvolatile memory 13. This is used to indicate that data and vertical parity do not match when a power failure or the like occurs during operation of the disk array. When a read command is issued from the CPU 3, the parity and the vertical parity are not referred to or changed, but are only read from the array disk 2.

To terminate the operation, the vertical parity is written back to the disk, the data update bit in the nonvolatile memory 13 is lowered, and the power is turned off. When the data update bit of the non-volatile memory is set when the system is started next time, the data of the disk is read before the operation, the vertical parity is generated again, and the operation is performed thereafter. If the data update bit of the nonvolatile memory is not set, the vertical parity read when the system is started can be used as it is.

When one of the disks 20 to 23 of the array disk 2 fails during operation, the data of the failed disk 2n is reconstructed from the remaining normal disks to the replacement disk 24 (not shown). When a media error occurs in the remaining disks, the data is restored using the data and the vertical parity of the blocks in the same disk to prevent a double failure.

[0022]

Conventionally, if one of the disks constituting the disk array system becomes unusable due to a failure or the like, the possibility of occurrence of a failure is much higher than that of a general disk configuration. If any disk fails, the data on the failed disk can be reproduced using the parity between the remaining disks, but if a media error occurs on the remaining disks, the data cannot be reproduced.
In a system in which a replacement disk is prepared, data for the replacement disk must be reconstructed from all the data on the remaining disks, and even if one media error occurs in the remaining disks, the data to be written to the replacement disk is lost. Reproduction becomes impossible. Since this media error occurs at a certain probability due to the physical characteristics of the disk, there is no way to avoid it.

According to the present invention, by adding vertical parity, even if one media error occurs in a block defined in the same disk during data reconstruction, data can be interpolated from vertical parity. The potential for severe injury can be greatly reduced.

Therefore, according to the present invention, there is an effect that the double failure can be minimized in the disk array system.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a double failure prevention control method for a disk array system according to an embodiment of the present invention.

FIG. 2 is a block diagram showing a configuration of an array controller shown in FIG.

FIG. 3 is a diagram showing a recording format of an array disk 2 shown in FIG.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Array controller 2 Array disk 3 CPU 10 Array control circuit 11 Disk controller 12 Vertical parity holding memory 13 Non-volatile memory 20-23 disks

Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat II (reference) G11B 20/18 570 G11B 20/18 570Z 572 572F

Claims (4)

[Claims] 1. A first parity generated based on data written to another disk is written to some disks, and a second parity generated based on data in the same disk is written to some disks.
An array disk in which parity of the array disk is written to the same disk; a first parity generation unit for generating the first parity; a second parity generation unit for generating the second parity; A first correction unit for correcting data of the failed disk by using data of another disk or the first parity when a failure occurs in some of the disks; A second correcting means for correcting an error from a portion other than the error when an error occurs in the data or the second parity. A double failure prevention control method for a disk array system, comprising: 2. The double failure prevention control method for a disk array system according to claim 1, wherein said array disk includes a spare disk for writing corrected data to first and second parities. Characteristic Double failure prevention control method for disk array system. 3. The double failure prevention control method for a disk array system according to claim 1, wherein a disk to which the first parity is written is different for each code word including the first parity. Double failure prevention control method for disk array systems. 4. The double failure prevention control system for a disk array system according to claim 1, further comprising a storage unit for storing the second parity, wherein the second parity generation unit is configured to store the second parity. Updating a second parity in the storage means based on a difference between current data and data newly written to the array disk and the second parity in the storage means. Double failure prevention control method for the system.