CN104407815A - Method and device for bad sector processing in RAID5 disk array - Google Patents

Abstract

The invention provides a method and device for processing bad sectors in a RAID5 disk array, belonging to the technical field of data fault-tolerant storage. The RAID5 disk array of the invention adds a reserved disk, and the array bitmap indicates the bad sectors of the disk. When reading and writing encounter a bad sector, transfer the bad sector data to the same position on the reserved disk, and set the bitmap; when the bitmap scan encounters a bad sector, read and write operations read from the same position on the reserved disk . This improves the storage performance and reliability of the RAID5 array.

Description

Method and device for bad sector processing in RAID5 disk array

technical field

The invention relates to a data fault-tolerant storage technology, in particular to a method and a device for repairing bad sector lost disks and fast reconstruction in a RAID5 disk array. the

Background technique

RAID technology is to use relatively small-capacity disks to form a large-capacity disk array according to certain rules to provide external storage resources. RAID (Redundant Array of Independent Disks) is called an independent redundant disk array.

The main purposes of using RAID are:

(1) Redundant storage of data on the disk to achieve fault tolerance;

(2) When there is an I/O request, the disk can be accessed concurrently to improve the throughput.

RAID5 supports redundancy, and stores verification information on all disks in the array. Its verification and technology uses XOR verification in disk stripes, which is called "using block cross access and uniform distribution of verification information" mode .

A stripe is the division of the storage space of a RAID member disk into smaller spaces of equal size.

XOR operation is characterized by A⊕B = C ==> B⊕C = A, A⊕C = B;

Therefore, if A and B are data blocks, and C is a check block, when the disk where B is located fails, B can be calculated through the XOR check of A and C. This is the most basic principle of RAID5 fault tolerance, which can tolerate a disk failure.

RAID5 read and write operations take stripes as the basic unit, the width of the stripe is 4KB, and the length of the stripe is the number of member disks. For the write operation of the disk data in the stripe, the parity disk data must be updated to ensure the consistency of the stripe data.

At present, when a RAID5 read and write operation encounters a bad sector, the disk with the bad sector will be removed from the RAID5 disk array, and the RAID5 will lose its fault tolerance when it is downgraded. The data on the deleted disk is obtained through the XOR check of other data disks, and written into the hot spare disk. After the reconstruction is completed, the hot spare disk replaces the deleted disk, and the fault tolerance mechanism of the entire RAID5 array is restored. However, when a RAID5 array is fully reconfigured, it is equivalent to performing a synchronous check on all the data in the entire array space. The whole work is time-consuming and labor-intensive, which not only affects the read and write performance of the array, , When another disk has a bad sector, the disk with the bad sector will be removed from the RAID5 array, resulting in the loss of two disks in RAID5, the entire array will fail, and all data will be lost.

The full-disk reconstruction operation of the hot spare disk in RAID5 not only affects the read and write performance of the array, but also affects the reliability of the array. the

Contents of the invention

The implementation of the present invention proposes a method and device for processing bad sectors in a RAID5 disk array, which can effectively reduce the performance impact caused by the reconstruction of the entire disk of the RAID5 hot spare disk, and quickly restore the data on the bad sectors. The embodiment of the invention also provides a corresponding device.

The first aspect of the present invention provides a kind of method of bad sector processing in RAID5 disk array, comprising:

1. The RAID5 disk array supports (X+1) hard disks, where X is the number of data disks, and 1 is the disk reserved for bad sector data transfer.

2. The RAID5 bitmap mark records which area in the disk is a bad sector.

3. Before performing read and write operations on RAID5, first check the bitmap indication of this area. If there is no indication on the bitmap, enter normal read and write operations.

4. When the read operation encounters a bad sector of the disk, mark this area as a bad sector with a bitmap, and verify the data on the bad sector through the stripe XOR operation of RAID5, and write this data into the reserved the same position on the disk.

5. When the write operation encounters a bad sector of the disk, mark this area as a bad sector with a bitmap, and write the data to be written into the same position on the reserved disk.

6. When the bitmap area is checked, read and write from the same position on the reserved disk.

A second aspect of the present invention provides a device for bad sector processing in a RAID5 disk array, including:

1), RAID5 disk array is composed of data disk and reserved disk;

2) The bitmap scanning unit is responsible for scanning the bitmap indication before reading and writing operations, and setting the bitmap indication when encountering a bad sector;

3) The data migration unit, when reading and writing encounters a bad sector, migrates the verified data or the data to be written to the same position on the reserved disk. the

Description of drawings

Figure 1 is a RAID5 structure diagram.

FIG. 2 is a schematic diagram of a RAID5 bad sector processing method.

Detailed ways

The content of the present invention is described in more detail below:

The embodiment of the present invention provides a method for processing bad sectors in a RAID5 disk array. When the RAID5 array encounters a bad sector, the hard disk with the bad sector will not be removed, and the hot spare disk will not be reconfigured, and the bad sector will be removed. The data is transferred to the same position on the reserved disk, and the second bad sector operation is shielded, which reduces the performance loss caused by the reconstruction of the RAID5 hot spare disk and increases the reliability of the array.

Its operation steps are:

When the RAID5 disk array is created, a reserved disk for bad sector transfer is added. The bitmap area is used to mark the bad sectors of the RAID5 disk array. When reading and writing to RAID5, the bitmap scanning unit first scans Bitmap area, check the mark:

1. If the read-write area is not marked, read and write directly;

When a read operation encounters a bad sector, the bitmap scanning unit sets the bitmap corresponding to this area, and obtains the data on the bad sector through stripe verification, and writes the data to the same position on the reserved disk superior. When a write operation encounters a sector, the bitmap scanning unit sets the bitmap corresponding to this area, and writes the data to be written to the same position on the reserved disk;

2. If the read-write area is marked, read and write data from the same position on the reserved disk.

Claims (7)

1. a method for bad sector process in RAID5 disk array, is characterized in that, comprising:

The reserved dish of data batchmove is fanned containing 1 piece for bad in RAID5 disk array;

In RAID5 bitmap label record disk, which block region is bad sector;

Before read-write operation is carried out to RAID5, first check that the bitmap in this region indicates, if bitmap is without sign, just enter normal read-write operation;

When read operation runs into a disk bad sector, marking this region with bitmap is bad sector, by the band XOR of RAID5, verifies out the data in bad sector, by the equivalent locations of reserved for the write of these data dish;

When write operation runs into a disk bad sector, marking this region with bitmap is bad sector, the equivalent locations of the reserved dish of the data that will write write;

When inspection bitmap region has sign, just read from reserved dish co-located and carry out read-write operation.

2. method according to claim 1, is characterized in that,

RAID5 disk array containing X+1 block hard disk, X is data disks quantity, and 1 fans data batchmove for bad

Reserved dish.

3. according to the requirement of right 1, it is characterized in that, also comprise:

RAID5 bitmap is present in the end of each member dish, coils corresponding bad sector for indicating this member.

4. method according to claim 1, is characterized in that,

First sweep bitmap region before RAID5 read-write operation, the bitmap reading and writing region corresponding is 1, shows that this region is bad sector, if bitmap is 0, shows that this region is normal sector.

5. method according to claim 1, is characterized in that,

The treatment scheme of bad sector is met in read-write: running into bad sector when reading RAID5 operation, indicating the bitmap of corresponding region, and the data in exclusive or check bad sector, and by this data batchmove to reserved disk peer area; When write operation runs into a disk bad sector, marking this region with bitmap is bad sector, the peer area of the reserved dish of the data that will write write.

6. method according to claim 1, is characterized in that,

Bitmap scans the treatment scheme of bad sector: when inspection bitmap region has sign, just reads from reserved dish co-located and carries out read-write operation.

7. a device for bad sector process in RAID5 disk array, is characterized in that, comprising:

(1), RAID5 disk array, be made up of data disks and reserved dish;

(2), bitmap scanning element, be responsible for before read-write operation, sweep bitmap indicate, in time running into bad sector, set bitmap indicate;

(3), data migration unit, when read and write run into bad sector time, by verification to data or the Data Migration that will write coil equivalent locations to reserved.