WO2024098696A1 - Data recovery method, apparatus and device, and readable storage medium - Google Patents

Abstract

The present application discloses a data recovery method, apparatus and device, and a readable storage medium. The method comprises: backing up data of a master node in a secondary node of a multi-control node so as to use data in the secondary node as mirror image data; monitoring whether the number of faulty disks of a RAID array in the master node exceeds a fault tolerance of the RAID array; if the number of faulty disks of the RAID array exceeds the fault tolerance of the RAID array, sending information of lost data in the faulty disks to the secondary node, and the secondary node acquiring, from the mirror image data, data corresponding to the information and sending the data; receiving the data corresponding to the information, and writing the data corresponding to the information into corresponding partitions of hot spares corresponding to the faulty disks; and replacing the corresponding faulty disk with the hot spares so as to reconstruct the RAID array with normal disks. Mirror image data is added to a secondary node, and the mirror image data in the secondary node is used for data recovery after the number of faulty discs exceeds a fault tolerance, thereby improving the data reliability of a storage system.

Description

Data recovery method, device, equipment and readable storage medium

CROSS-REFERENCE TO RELATED APPLICATIONS

This application claims priority to a Chinese patent application filed with the China Patent Office on November 11, 2022, with application number 202211409858.4 and application name “A data recovery method, device, equipment and readable storage medium”, all contents of which are incorporated by reference in this application.

Technical Field

The present application relates to the technical field of data storage, and more specifically, to a data recovery method, device, equipment and readable storage medium.

Background technique

Society has entered the era of big data, where massive amounts of data need to be stored securely and reliably. In many industries, not only is the amount of data that needs to be stored growing exponentially, but the reliability of stored data has reached an extreme. Losing a small amount of data can often lead to fatal business disasters, such as in the banking and military fields. Therefore, storage systems have been looking for breakthroughs in increasing data reliability and improving I/O (Input/Output) performance.

In terms of increasing data reliability, the industry currently uses RAID (Redundant Arrays of Independent Disks) technology to improve data reliability, and uses redundant disks in the RAID array to recover data from failed disks. In terms of improving I/O performance, the industry has used multi-control nodes to form a cluster. Specifically, in order to ensure the high availability of the system, at least two nodes will be used to form an IOGROUP (input and output group), at least two nodes are connected to one port of a dual-port hard disk respectively, at least two nodes in the IOGROUP are each other's peer nodes, and one or more IOGROUPs form a cluster, and the nodes in the cluster can communicate with each other. Among them, the master node is responsible for processing the I/O requests of the host, and the auxiliary node is responsible for the background tasks of the storage system (such as RAID array initialization, inspection and reconstruction tasks, etc.), so as to provide storage I/O performance. However, when the master node uses RAID technology for I/O data storage, if the number of failed disks in the storage system exceeds the maximum number of failed disks that the RAID array can recover, the failed disk data cannot be recovered through the internal mechanism of the RAID array, resulting in low data reliability of the storage system.

Summary of the invention

In view of this, the purpose of the present application is to provide a data recovery method, device, equipment and readable storage medium for recovering failed disk data to improve data reliability.

In order to achieve the above objectives, this application provides the following technical solutions:

A data recovery method, comprising:

Backing up data in the main control node in the auxiliary node among the multiple control nodes, so that the data in the auxiliary node is used as mirror data of the main control node;

Monitor whether the number of failed disks in the RAID array in the master control node exceeds the fault tolerance of the RAID array;

If the number of failed disks in the RAID array exceeds the fault tolerance of the RAID array, the information about the lost data in the failed disks is sent to the auxiliary node, which obtains the data corresponding to the information from the mirror data and sends it;

Receive data corresponding to the information, and write the data corresponding to the information into a corresponding partition of a hot spare disk corresponding to the failed disk;

Use the hot spare disk to replace the corresponding failed disk to re-form the RAID array with the normal disk.

In some embodiments of the present application, backing up data in a master control node in a secondary node among multiple control nodes includes:

Back up the logical volume in the master node in the auxiliary node to obtain the corresponding mirror logical volume; the logical volume and the mirror logical volume contain multiple blocks, and there is a mapping relationship between the blocks and the blocks in the disk. The mapping relationship is stored in both the master node and the auxiliary node;

Send information about lost data in the failed disk to the secondary node, including:

According to the blocks and mapping relationship contained in the failed disk, data loss metadata corresponding to the logical volume is formed, and the data loss metadata is sent to the auxiliary node; the data loss metadata includes information on whether the blocks contained in the logical volume are lost;

The auxiliary node obtains the data corresponding to the information from the mirror data and sends it, including:

The auxiliary node obtains the block number of the lost data according to the mapping relationship and the data loss metadata, and sends the data corresponding to the block number in the mirrored logical volume to the master node.

In some embodiments of the present application, data loss metadata corresponding to a logical volume is formed according to the blocks and mapping relationships contained in the failed disk, including:

According to the blocks and mapping relationship contained in the failed disk, data loss metadata corresponding to the logical volume is formed with bitmap as the data organization method; wherein each bit in the bitmap indicates whether the corresponding block contained in the logical volume is lost.

In some embodiments of the present application, before using the hot spare disk to replace the corresponding failed disk, the method further includes:

Determine whether each stripe in the RAID array needs to recalculate the check blocks;

If there is a stripe for which the check blocks need to be recalculated, the check blocks are calculated according to the blocks in the stripe, and the calculated check blocks are written to the corresponding partition of the hot spare disk corresponding to the failed disk.

In some embodiments of the present application, receiving data corresponding to information, and writing the data corresponding to the information into a corresponding partition of a hot spare disk corresponding to a failed disk, includes:

The data corresponding to the information is received block by block, and the data corresponding to the information is written block by block into a corresponding partition of the hot spare disk corresponding to the failed disk.

In some embodiments of the present application, if the number of failed disks in the RAID array exceeds the fault tolerance of the RAID array, the method further includes:

Redirect the host's I/O requests to the secondary node;

After the hot spare disk is used to replace the corresponding failed disk to re-form the RAID array with the normal disk, the following steps are also included:

Redirect the host's I/O requests to the master node.

In some embodiments of the present application, when redirecting the I/O request of the host to the auxiliary node, it also includes:

The data newly sent by the host is stored in the master control node, so that the data newly sent by the host stored in the master control node is used as the mirror data of the auxiliary node.

In some embodiments of the present application, if the number of failed disks in the RAID array does not exceed the fault tolerance of the RAID array, the method further includes:

The blocks of the normal disks contained in each stripe in the RAID array are used for calculation to obtain the blocks of the hot spare disk corresponding to the failed disk;

After all data lost on the failed disk is recovered, the hot spare disk is used to replace the failed disk to re-form the system with the normal disk. RAID array.

In some embodiments of the present application, monitoring whether the number of failed disks in the RAID array in the master control node exceeds the fault tolerance of the RAID array includes:

Regularly monitor whether the number of failed disks in the RAID array in the master control node exceeds the fault tolerance of the RAID array.

In some embodiments of the present application, it also includes:

If the number of failed disks in the RAID array does not exceed the fault tolerance of the RAID array, the process returns to the step of periodically monitoring whether the number of failed disks in the RAID array in the master control node exceeds the fault tolerance of the RAID array.

In some embodiments of the present application, if the number of auxiliary nodes is greater than 1, information about lost data in the failed disk is sent to the auxiliary node, including:

The information about lost data in the failed disk is sent to an auxiliary node selected from a plurality of auxiliary nodes according to a preset selection strategy.

In some embodiments of the present application, after the hot spare disk is used to replace the corresponding failed disk to re-form a RAID array with the normal disk, the following is further included:

Regularly clean up the mirror data in the secondary node.

A data recovery device, comprising:

A backup module, used to back up the data in the main control node in the auxiliary node among the multiple control nodes, so as to use the data in the auxiliary node as the mirror data of the main control node;

A monitoring module is used to monitor whether the number of failed disks in the RAID array in the master control node exceeds the fault tolerance of the RAID array;

A sending module is used to send information about lost data in the failed disks to the auxiliary node if the number of failed disks in the RAID array exceeds the fault tolerance of the RAID array, and the auxiliary node obtains data corresponding to the information from the mirror data and sends it;

A writing module, used for receiving data corresponding to the information, and writing the data corresponding to the information into a corresponding partition of the hot spare disk corresponding to the failed disk;

The first replacement module is used to replace the corresponding failed disk with the hot spare disk to re-form a RAID array with the normal disk.

A data recovery device, comprising:

Memory for storing computer programs;

A processor is used to implement the steps of any of the above data recovery methods when executing a computer program.

A non-volatile readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of any of the above data recovery methods are implemented.

The present application provides a data recovery method, device, equipment and readable storage medium, wherein the method comprises: backing up data in a main control node in a secondary node in a multi-control node, so as to use the data in the secondary node as mirror data of the main control node; monitoring whether the number of failed disks in a RAID array in the main control node exceeds the fault tolerance of the RAID array; if the number of failed disks in the RAID array exceeds the fault tolerance of the RAID array, sending information about lost data in the failed disk to the secondary node, and the secondary node obtaining data corresponding to the information from the mirror data and sending the information; receiving data corresponding to the information, and writing the data corresponding to the information into a corresponding partition of a hot spare disk corresponding to the failed disk; replacing the corresponding failed disk with the hot spare disk, so as to re-form a RAID array with normal disks.

The above technical solution disclosed in the present application backs up the data in the main control node in the auxiliary node in the multi-control node, so that the data backed up in the auxiliary node is used as the mirror data of the main control node. When the number of failed disks in the RAID array in the main control node exceeds the fault tolerance of the RAID array, the information of the lost data in the failed disk is sent to the auxiliary node, and the auxiliary node obtains the data corresponding to the information of the lost data from the mirror data and sends the data out. After the data is recovered, the data is written to the corresponding partition of the hot spare disk corresponding to the failed disk, and then the hot spare disk is used to replace the failed disk and re-compose the RAID array with the normal disk. It can be seen that the application can improve the data reliability of the storage system by adding mirror data in the auxiliary node and using the mirror data in the auxiliary node for data recovery after the number of failed disks exceeds the fault tolerance.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are merely embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

FIG1 is a flow chart of a data recovery method provided by an embodiment of the present application;

FIG2 is a flow chart of another data recovery method provided in an embodiment of the present application;

FIG3 is a schematic diagram of the structure of a master control node provided in an embodiment of the present application;

FIG4 is a schematic diagram of a reconstruction process when the number of failed disks exceeds the fault tolerance provided by an embodiment of the present application;

FIG5 is another schematic diagram of the structure of a master control node provided in an embodiment of the present application;

FIG6 is a schematic diagram of an auxiliary node provided in an embodiment of the present application;

7 is a schematic diagram of reconstructing and restoring data of a failed disk when the number of failed disks does not exceed the fault tolerance provided by an embodiment of the present application;

FIG8 is a schematic diagram of the structure of a data recovery device provided in an embodiment of the present application;

FIG. 9 is a schematic diagram of the structure of a data recovery device provided in an embodiment of the present application.

Detailed ways

The core of the present application is to provide a data recovery method, device, equipment and readable storage medium for recovering failed disk data to improve data reliability.

The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

Referring to FIG. 1 and FIG. 2 , FIG. 1 shows a flow chart of a data recovery method provided in an embodiment of the present application, and FIG. 2 shows a flow chart of another data recovery method provided in an embodiment of the present application. A data recovery method provided in an embodiment of the present application may include:

S11: backing up the data in the main control node in the auxiliary node among the multiple control nodes, so as to use the data in the auxiliary node as the mirror data of the main control node.

It should be noted that the execution subject of the present application can be a master control node among multiple control nodes (including a master control node and auxiliary nodes), or it can be a storage system. The present application is explained using the execution subject as a multiple control node as an example.

First, the master control node can process the I/O request sent by the host and store the data corresponding to the I/O request (specifically, I/O data). At the same time, the data in the master control node can be backed up in the auxiliary node in the multi-control node, that is, the data stored in the master control node corresponding to the I/O request sent by the host can be backed up in the auxiliary node. The data backed up in the auxiliary node can be specifically stored in the memory of the auxiliary node to facilitate rapid data acquisition and transmission.

By backing up the data in the master node in the auxiliary node, the data backed up in the auxiliary node can be used as the mirror data of the master node, so that the radial data in the auxiliary node can be used to reconstruct and restore the data in the failed disk in the master node, thereby effectively improving the data reliability of the storage system.

S12: Monitor whether the number of failed disks in the RAID array in the master control node exceeds the fault tolerance of the RAID array; if the number of failed disks in the RAID array exceeds the fault tolerance of the RAID array, execute step S13; if the number of failed disks in the RAID array does not exceed the fault tolerance of the RAID array, return to step S12.

While processing the I/O request sent by the host, or after processing the I/O request sent by the host, the master control node can monitor the number of failed disks in the RAID array in the master control node, and determine whether the number of failed disks in the RAID array exceeds the fault tolerance of the RAID array (that is, the maximum number of failed disks that the RAID array can recover).

Among them, the RAID array level has its own internal data redundancy. For example, the RAID5 array provides a P check block data redundancy. The RAID5 array uses the single redundancy of the P check block (the P check block in the stripe is obtained by the XOR operation of the data block, and the data block is the valid data sent by the host) to restore the data of a faulty disk, that is, the fault tolerance of the RAID5 array is 1, that is, when the number of faulty disks is 1, the internal data redundancy can be used to restore the data of the faulty disk; the RAID6 array provides P check blocks and Q check blocks (the Q check blocks and P check blocks in the RAID6 array are used together to restore the two faulty disks in the RAID6 array) data dual redundancy. The RAID6 array uses the dual redundancy of P check and Q check to restore the data of two faulty disks, that is, the fault tolerance of the RAID6 array is 2, that is, when the number of faulty disks does not exceed 2, the internal data redundancy can be used to restore the data of the faulty disk. It should be noted that the RAID array mentioned in this application can be a RAID5 array or a RAID6 array, of course, it can also be other RAID arrays, and this application does not limit this.

If the number of failed disks in the RAID array exceeds the fault tolerance of the RAID array, the data redundancy mechanism inside the RAID array cannot be used to recover the data of the failed disks, because the number of failed disks that fail at the same time exceeds its fault tolerance. In this case, the master control node needs to use the mirror data of the auxiliary node to reconstruct the lost data in the failed disk. If the number of failed disks in the RAID array does not exceed the fault tolerance of the RAID array, it can return to step S12, that is, continue to monitor whether the number of failed disks in the RAID array in the master control node exceeds the fault tolerance of the RAID array.

S13: Send the information of the lost data in the failed disk to the auxiliary node, and the auxiliary node obtains the data corresponding to the information from the mirror data and sends it.

In step S12, if the number of failed disks in the RAID array exceeds the fault tolerance of the RAID array, the external redundancy provided by the auxiliary node can be used to recover the data of the failed disk. Specifically, the master node can obtain information about the lost data in the failed disk, so as to retrieve the lost data of the master node from the auxiliary node based on the information about the lost data. For example, the information about the lost data can be which strips of data in the failed disk are lost, etc., wherein a strip is a partition of the physical storage medium on the disk, and is used for the granularity of data reconstruction of the RAID array.

After obtaining the information about the lost data in the failed disk, the master node can send the information about the lost data in the failed disk to the auxiliary node. After receiving the information about the lost data, the auxiliary node can obtain the data corresponding to the information about the lost data from the mirror data stored in itself, and send the obtained information about the lost data to the master node, so that the master node can reconstruct and recover the data based on the data.

In addition, the auxiliary node only retrieves the data corresponding to the information of the lost data and transmits the data to the master node for data reconstruction, which can minimize the amount of data that needs to be retrieved and transmitted, and also minimize the time window for data reconstruction.

S14: receiving data corresponding to the information, and writing the data corresponding to the information into a corresponding partition of the hot spare disk corresponding to the failed disk.

Based on step S13, the master node may receive data corresponding to the information about lost data sent by the auxiliary node. And the data corresponding to the information of the lost data can be written into the corresponding partition of the hot spare disk corresponding to the failed disk. Among them, the hot spare disk is also called spare, a spare storage drive.

S15: Use the hot spare disk to replace the corresponding failed disk to re-form a RAID array with normal disks.

After all lost data in the failed disk is recovered and written to the corresponding partition of the corresponding hot spare disk, the hot spare disk can be used to replace the corresponding failed disk to re-form the RAID array with the normal disk, thereby completing the recovery of the failed disk data.

For example, as shown in Figures 3 and 4, Figure 3 shows a schematic diagram of the structure of the master node provided by the embodiment of the present application, and Figure 4 shows a schematic diagram of the reconstruction process when the number of failed disks exceeds the fault tolerance provided by the embodiment of the present application. In Figures 3 and 4, the master node has a RAID5 array composed of five hard disks, and Figure 4 shows the data reconstruction process when two disks of the master node of Figure 3 fail at the same time. In Figure 3, disk 1 and disk 2 fail at the same time (exceeding the fault tolerance of RAID5), so the data of strips 1-2, 5, 9, 13-14 and 17-18 are lost, and the internal mechanism of the RAID5 array is exceeded for data recovery. Therefore, the above scheme in the present application is used for data recovery, and a hot spare disk 1 corresponding to disk 1 and a hot spare disk 2 corresponding to disk 2 are obtained, and the corresponding disk 1 is replaced by the hot spare disk 1, and the corresponding disk 2 is replaced by the hot spare disk 2, so as to re-form the RAID array with the normal disks (disk 3, disk 4, disk 5).

Through the above process, it can be known that, in response to data loss events with a high probability in the storage system, the present application adds mirror data in the auxiliary node in the cluster under the cluster multi-controller storage system. When the number of failed disks exceeds the maximum number of failed disks that can be recovered by the RAID array, the mirror data of the auxiliary node is used to reconstruct and restore the data in the failed disk. Specifically, the main control node uses the data of the normal disk and the data transmitted from the auxiliary node (the data originally needs to be written to the failed disk, but the disk fails and the writing is not successful, so the mirror data of the auxiliary node corresponding to the data is transmitted to the main control node) for offline reconstruction, so as to effectively improve the data reliability of the storage system.

The above technical solution disclosed in the present application backs up the data in the main control node in the auxiliary node in the multi-control node, so that the backed-up data in the auxiliary node is used as the mirror data of the main control node. When the number of failed disks in the RAID array in the main control node exceeds the fault tolerance of the RAID array, the information of the lost data in the failed disk is sent to the auxiliary node, and the auxiliary node obtains the data corresponding to the information of the lost data from the mirror data and sends the data out. After receiving the data sent by the auxiliary node, the data is written to the corresponding partition of the hot spare disk corresponding to the failed disk, and then the hot spare disk is used to replace the failed disk and re-compose the RAID array with the normal disk. It can be seen from this that the present application can improve the data reliability of the storage system by adding mirror data in the auxiliary node and using the mirror data in the auxiliary node for data recovery after the number of failed disks exceeds the fault tolerance.

In some embodiments of the present application, backing up data in a master control node in a secondary node among multiple control nodes may include:

Back up the logical volume in the master node in the auxiliary node to obtain the corresponding mirror logical volume; the logical volume and the mirror logical volume contain multiple blocks, and there is a mapping relationship between the blocks and the blocks in the disk. The mapping relationship is stored in both the master node and the auxiliary node;

Send information about lost data in the failed disk to the secondary node, which may include:

According to the blocks and mapping relationship contained in the failed disk, data loss metadata corresponding to the logical volume is formed, and the data loss metadata is sent to the auxiliary node; the data loss metadata includes information on whether the blocks contained in the logical volume are lost;

The auxiliary node obtains the data corresponding to the information from the mirror data and sends it, which may include:

The auxiliary node obtains the block number of the lost data according to the mapping relationship and the data loss metadata, and sends the data corresponding to the block number in the mirrored logical volume to the master node.

In this application, considering that the host performs I/O access through logical volumes rather than blocks, the master control node When storing the data corresponding to the I/O request, it will be stored in the logical volume first. Therefore, when backing up the data in the master node in the auxiliary node, the logical volume in the master node can be backed up in the auxiliary node to obtain the corresponding mirror logical volume. Among them, the logical volume in the master node and the mirror logical volume both contain multiple blocks (block is the granularity of the host I/O access data, which is a logical unit. There is a certain mapping relationship between block and strip. Through this mapping relationship, it can be located to the specific block in the physical disk. Multiple blocks constitute a volume). There is a mapping relationship between the block and the block in the disk.

The mapping relationship is stored in both the master node and the auxiliary node, that is, the mapping relationship between blocks and sub-blocks is maintained in both the master node and the auxiliary node. It should be noted that blocks and sub-blocks can be one sub-block corresponding to multiple blocks, so that a sub-block will be equally divided into several blocks, or one sub-block corresponding to one block.

For details, please refer to Figures 5 and 6. Figure 5 shows another structural schematic diagram of the master control node provided in the embodiment of the present application, and Figure 6 shows a schematic diagram of the auxiliary node provided in the embodiment of the present application. Figure 5 shows a block diagram of the master control node configured with a RAID5 array. The mirror volume A in the auxiliary node in Figure 6 corresponds to the volume A in the master control node in Figure 5, and the mirror volume B in the auxiliary node in Figure 6 corresponds to the volume B in the master control node in Figure 5. The master control node processes the I/O request of the host as the main server, and the data of the I/O request is distributed in a RAID5 array composed of five hard disks. In Figure 5, the master control node has a RAID5 array composed of five hard disks. Strip is the granularity size used for data reconstruction of the RAID array, and block is the granularity size used for host I/O access to data. The block is also the granularity size used for data mirroring between the master control node and the auxiliary node. From another perspective, a strip is a physical data unit of a disk, and a block is a logical data unit. A volume consists of multiple blocks. Both the master node and the slave node maintain a mapping relationship between blocks and strips, where the mapping relationship can be one strip corresponding to multiple blocks, so that a strip is equally divided into several blocks, or one strip corresponding to one block. For easier understanding, in Figures 5 and 6, one strip is designed to correspond to one block, such as block 0-99 corresponds to strip 1A in Figure 5. The left side of Figure 5 provides an example of a mapping relationship organized by volume. The data in volume A is distributed on five disks, as shown in the rectangular box containing the letter "A", and the data in volume B is also distributed on five disks, as shown in the rectangular box containing the letter "B". Specifically, the data in volume A is distributed in strips 1-2, 4-5, 9, 11-12, 16, and 19-20. The data in volume B is distributed in strips 3, 6-8, 10, 13-15, and 17-18.

On the basis of the above, when the master control node sends the information about lost data in the failed disk to the auxiliary node, it can first form the data loss metadata corresponding to the logical volume in the master control node according to the blocks contained in the failed disk and the mapping relationship between the stored blocks and the blocks. Among them, the data loss metadata is also metadata (a data structure for managing stripes, which can be a bitmap or a hash table, etc.), but the data loss metadata not only manages stripes but also identifies the data unit of lost data in the failed disk (the data unit is a block, that is, the data loss metadata is for the blocks in the logical volume). Specifically, the data loss metadata contains information on whether the blocks contained in the logical volume in the master control node are lost (the information on whether all blocks contained in the logical volume are lost is included in the data loss metadata). It should be noted that the data loss metadata involved in this application is a global variable used to implement a specified logical function, which can be implemented in C language or C++.

Then, the master node can send the formed data loss metadata to the auxiliary node. Accordingly, after receiving the data loss metadata, the auxiliary node can scan the data loss metadata, and obtain the block number of the lost data according to the mapping relationship between the stored blocks and the blocks and the received data loss metadata. For example, in FIG6 , the block numbers of the lost data finally obtained by the auxiliary node are 1-2, 5, 9, 13-14 and 17-18. Then, the auxiliary node can send the data corresponding to the block numbers in the mirrored logical volume to the master node. Taking FIG6 as an example, the data corresponding to the block numbers 1-2, 5, 9, 13-14 and 17-18 are sent to the master node.

Through the above, not only the mirror logical volume is formed in the auxiliary node, but also the mutual The data mirroring is block-sized, which is better for the host and the user. What is transmitted between the master node and the auxiliary node is the data loss metadata, and the auxiliary node only retrieves the lost data blocks (specifically the data corresponding to the blocks in the failed disk) based on the data loss metadata and transmits it to the master node for data reconstruction. This minimizes the amount of data that needs to be retrieved and transmitted, and also minimizes the time window for data reconstruction.

In some embodiments of the present application, data loss metadata corresponding to a logical volume is formed according to the blocks and mapping relationships contained in the failed disk, which may include:

According to the blocks and mapping relationship contained in the failed disk, data loss metadata corresponding to the logical volume is formed with bitmap as the data organization method; wherein each bit in the bitmap indicates whether the corresponding block contained in the logical volume is lost.

In the present application, data loss metadata corresponding to the logical volume and organized in a bitmap format can be formed based on the blocks and mapping relationships contained in the failed disk, and each bit in the bitmap indicates whether the corresponding block contained in the logical volume is lost. For example, when the bit is designed to be 0, it indicates that the corresponding block is a lost block, and when the bit is 1, it indicates that the corresponding block is a non-lost block (of course, other designs can also be adopted as needed).

Specifically, taking FIG. 6 as an example, both disk 1 and disk 2 fail, so strips 1-2, 5, 9, 13-14, and 17-18 are blocks where data is lost, and the bits of the blocks corresponding to each other are set to 0. According to the one-to-one relationship between strips and blocks, the data loss metadata of volume A in the master node and the data loss metadata in volume B are represented as follows: the first row of bitmap (representing volume A): {0(block 1), 0(block 2), 1(block 4), 0(block 5), 0(block 9), 1(block 11), 1(block 12), 1(block 16), 1(block 19), 1(block 20)}, the second row of bitmap (representing volume B): {1(block 3), 1(block 6), 1(block 7), 1(block 8), 1(block 10), 0(block 13), 0(block 14), 1(block 15), 0(block 17), 0(block 18)}. It should be pointed out that the bitmap metadata organization method designed by the present application is a two-dimensional bitmap metadata organization method, in which each row represents the bit position of each block in a logical volume, and the bit positions of the blocks in different logical volumes are on each column. There are multiple rows for multiple logical volumes, thus forming a two-dimensional bitmap metadata organization method, that is, the data loss metadata is a two-dimensional bitmap metadata organization method. In summary, for the logical volume of the master control node shown in FIG5 , the final data loss metadata (that is, bitmap metadata) can be obtained. The first line is: 0 0 1 0 0 1 1 1 1 1, and the second line is: 1 1 1 1 1 1 0 0 1 0 0. Of course, the bitmap metadata organization method in the data loss metadata can also be adjusted according to actual needs, and the present application does not limit this.

The data loss metadata using bitmap as the data organization method is not only simple and clear, but also convenient for the auxiliary node to quickly determine the block number of the lost data based on it.

Of course, the data loss metadata can also be other data organization methods such as hash tables, as long as it can indicate whether the blocks contained in the logical volume are lost and enable the auxiliary node to determine the block number of the lost data based on this information. In addition, it should be noted that the two-dimensional bitmap metadata organization method designed in this application can be applied not only in this business scenario, but also in other business scenarios.

In some embodiments of the present application, before using the hot spare disk to replace the corresponding failed disk, the following steps may also be included:

Determine whether each stripe in the RAID array needs to recalculate the check blocks;

If there is a stripe for which the check blocks need to be recalculated, the check blocks are calculated according to the blocks in the stripe, and the calculated check blocks are written to the corresponding partition of the hot spare disk corresponding to the failed disk.

In the present application, before using the hot spare disk to replace the corresponding failed disk, the master control node can also determine whether each stripe in the RAID array needs to recalculate the check blocks, wherein the master control node can make judgments on each stripe in turn. It should be noted that a stripe is a collection of position-related blocks on different disks of the array, and is a unit for organizing blocks on different disks. For details, see FIG. 5 , in which a dotted rectangular box 101 represents a stripe. Taking the stripe as a unit, using the blocks in the stripe The blocks are XORed to reconstruct the data and calculate the P check block. Therefore, in the RAID array, the redundancy of the RAID array is maintained with stripes as units. As shown in FIG5 , stripe 101 is composed of strip1A, strip2A, strip3B, strip4A and P check block (i.e., Parity1). Strip1A-4A can be the data blocks sent by the host, and the P check block is the redundant block obtained by the XOR operation of strip1A-4A of stripe 101, which stores redundant data.

When the master control node determines whether the stripe in the RAID array needs to recalculate the check blocks, it can specifically determine whether the check blocks in the stripe are located on the faulty disk, or in other words, whether the check blocks are missing in the stripe. If the check blocks in the stripe are located on the faulty disk, or in other words, the check blocks are missing in the stripe, it is determined that the check blocks in the stripe need to be recalculated; if the check blocks in the stripe are not located on the faulty disk, or in other words, the check blocks are not missing in the stripe, it is determined that the check blocks in the stripe do not need to be recalculated.

If it is determined that the stripe needs to recalculate the parity block, the parity block is calculated based on each block in the stripe (the blocks mentioned here specifically refer to the blocks in the normal disks in the stripe and the blocks in the hot spare disk corresponding to the faulty disk in the stripe), specifically, the data blocks in the stripe are used to perform an XOR operation to calculate the parity block. For example, the stripe 102 in Figures 4 and 6 can be based on strip6A, strip7B, strip8B and the restored strip5A (that is, strip5A in the hot spare disk 2 corresponding to the faulty disk 2) to perform an XOR operation to obtain P parity block 2 (Parity2). After the parity block of the stripe is calculated, the calculated parity block can be written into the corresponding partition of the hot spare disk corresponding to the faulty disk. Corresponding to Figure 6, the obtained Parity2 is written into the corresponding partition of the hot spare disk 1. Through the above process, the corresponding stripe can be completely recovered (that is, not only the lost block data is recovered, but also the lost parity block is recovered), so as to improve the reliability of data recovery and the reliability of RAID array reorganization.

It should be noted that FIG. 2 shows a method for performing check block judgment and writing data to the corresponding partition of the hot spare disk, that is, after the master control node writes the block data received from the auxiliary node to the corresponding partition of the hot spare disk, it judges whether the corresponding stripe in the RAID array (specifically, the stripe corresponding to the block data writing) needs to recalculate the check block. If recalculation is required, the check block is calculated, and the recalculated check block is written to the corresponding partition of the hot spare disk, so that the corresponding stripe is completely restored; if it is determined that the check block does not need to be recalculated, or the recalculated check block is written After the verification block is written to the corresponding partition of the hot spare disk, it is determined whether all the data of the failed disk has been restored and reconstructed. If all the data of the failed disk has been restored and reconstructed, all the lost data is restored and written to the corresponding partition of the hot spare disk, and the failed disk is replaced with the hot spare disk, and the RAID array is reconstructed with other non-faulty disks. If all the data of the failed disk has not been restored and reconstructed, it is moved to the next stripe to continue writing the received block data to the corresponding partition of the hot spare disk to perform the block reconstruction in the next stripe, and this cycle is repeated until all the lost data is restored. Of course, it is also possible to write the data corresponding to the information of all the lost data to the corresponding partition of the hot spare disk corresponding to the failed disk, and then determine whether the verification block needs to be recalculated stripe by stripe.

In some embodiments of the present application, receiving data corresponding to information and writing the data corresponding to the information into a corresponding partition of a hot spare disk corresponding to the failed disk may include:

The data corresponding to the information is received block by block, and the data corresponding to the information is written block by block into a corresponding partition of the hot spare disk corresponding to the failed disk.

In the present application, when the master control node receives the data corresponding to the information sent by the auxiliary node, it can receive the data corresponding to the information block by block, that is, receive the data corresponding to each block in the faulty disk in a serial manner, so as to achieve orderly data reception. In addition, when writing the data corresponding to the information into the corresponding partition of the hot spare disk corresponding to the faulty disk, it can also write the data corresponding to the information block by block into the corresponding partition of the hot spare disk corresponding to the faulty disk, that is, write the data corresponding to each block in the faulty disk into the corresponding partition of the hot spare disk corresponding to the faulty disk in a serial manner, so as to achieve orderly data writing and recover.

It should be noted that the master control node can receive and write data corresponding to the blocks at the same time, that is, while receiving data corresponding to a block, the previously received data corresponding to the block can be written into the corresponding partition of the corresponding hot spare disk. Of course, the master control node can also receive data corresponding to a block, and after writing the data into the corresponding partition of the corresponding hot spare disk, receive data corresponding to another block. This application does not limit this.

In some embodiments of the present application, if the number of failed disks in the RAID array exceeds the fault tolerance of the RAID array, the following may also be included:

Redirect the host's I/O requests to the secondary node;

After the hot spare disk is used to replace the corresponding failed disk to re-form the RAID array with the normal disk, the following steps may also be performed:

Redirect the host's I/O requests to the master node.

In the present application, if the number of failed disks of the RAID array exceeds the fault tolerance of the RAID array, the I/O request of the host can be redirected to the auxiliary node, that is, the auxiliary node receives and processes the I/O request of the host, and the auxiliary node stores the data corresponding to the I/O request of the host to ensure that the I/O request of the host can be processed normally. Moreover, redirecting the I/O request of the host to the master control node can make the auxiliary node effectively play a role as the data mirror of the master control node. In addition, while redirecting the I/O request of the host to the master control node, the master control node can go offline (that is, no longer receive and process the I/O request of the host), in which case the master control node can reconstruct and recover the data offline.

In addition, after replacing the corresponding failed disk with the hot spare disk to re-form the RAID array with the normal disk, the master control node can redirect the host's I/O requests to the master control node, that is, the master control node continues to receive, process and execute I/O requests from the host, so that the storage system returns to normal.

In some embodiments of the present application, when redirecting the I/O request of the host to the auxiliary node, the following steps may also be included:

The data newly sent by the host is stored in the master control node, so that the data newly sent by the host stored in the master control node is used as the mirror data of the auxiliary node.

In the present application, when redirecting the I/O request of the host to the auxiliary node, the data newly sent by the host can also be stored in the main control node (that is, the data newly sent by the host corresponding to the I/O request is stored not only in the auxiliary node, but also in the main control node), so that the data newly sent by the host stored in the main control node can be used as the mirror data of the auxiliary node, so that if the auxiliary node writes data to the disk and the number of failed disks exceeds the fault tolerance, the mirror data in the main control node can be used to recover the data of the failed disk in the auxiliary node, that is, to implement a process similar to the failed disk data recovery of the main control node, so as to improve the reliability of data storage in the auxiliary node.

In some embodiments of the present application, if the number of failed disks in the RAID array does not exceed the fault tolerance of the RAID array, the method may further include:

The blocks of the normal disks contained in each stripe in the RAID array are used for calculation to obtain the blocks of the hot spare disk corresponding to the failed disk;

After all the data lost on the failed disk is recovered, the hot spare disk is used to replace the failed disk to re-form the RAID array with normal disks.

In the present application, in step S12, if the number of failed disks in the RAID array in the master control node is monitored to be less than the fault tolerance of the RAID array, the data redundancy mechanism within the RAID array can be used to recover data. Specifically, the blocks of the normal disks contained in each stripe in the RAID array are used to perform XOR calculations to obtain the blocks of the hot spare disk corresponding to the failed disk. Then, after all the data lost on the failed disk is recovered, that is, the number of failed disks in all stripes in the RAID array is restored. After the data is recovered, the hot spare disk corresponding to the failed disk is used to replace the failed disk to re-form the RAID array with the normal disk. In this way, the data of the failed disk can be recovered when the number of failed disks does not exceed the fault tolerance, thereby improving the data reliability of the storage system.

Specifically, please refer to Figure 7, which shows a schematic diagram of reconstructing and recovering the data of a failed disk when the number of failed disks does not exceed the fault tolerance provided by an embodiment of the present application. In Figure 7, the master control node has a RAID5 array consisting of five disks. If only one disk of the master control node fails, data recovery can be performed through the internal mechanism of the RAID5 array. Among them, Figure 7 describes that disk 1 in the master control node fails, and the spare hot spare disk is used to reconstruct the lost data of disk 1. For example, strip1 in the hot spare disk is obtained by XOR operation of strip2A in disk 2, strip3B in disk 3, strip4A in disk 4, and P check block (parity1) in disk 5. Similarly, the data of disk 1 in other stripes can be recovered. When the hot spare disk recovers all the lost data of disk 1, the hot spare disk will replace disk 1 and form a new RAID5 array with disks 2-5 to process host I/O requests.

In some embodiments of the present application, monitoring whether the number of failed disks in the RAID array in the master control node exceeds the fault tolerance of the RAID array may include:

Regularly monitor whether the number of failed disks in the RAID array in the master control node exceeds the fault tolerance of the RAID array.

In the present application, the master control node can monitor whether the number of faulty disks in the RAID array in the master control node exceeds the fault tolerance of the RAID array. Specifically, a timer can be used to monitor regularly, so as to facilitate timely discovery of the relationship between the number of faulty disks and the fault tolerance, and then facilitate timely adoption of different data reconstruction and recovery methods for data recovery, so as to improve the data reliability of the storage system. Among them, the time interval of the timing can be set according to actual experience, and this application does not limit this.

Of course, the master control node can also monitor in real time whether the number of failed disks in the RAID array in the master control node exceeds the fault tolerance of the RAID array.

In some embodiments of the present application, the following may also be included:

If the number of failed disks in the RAID array does not exceed the fault tolerance of the RAID array, the process returns to the step of periodically monitoring whether the number of failed disks in the RAID array in the master control node exceeds the fault tolerance of the RAID array.

In the present application, when the master control node is performing timed monitoring, if the number of failed disks in the monitored RAID array does not exceed the fault tolerance of the RAID array, the master control node can return to the step of performing timed monitoring to determine whether the number of failed disks in the RAID array in the master control node exceeds the fault tolerance of the RAID array, that is, performing timed cyclic monitoring, so as to timely discover the relationship between the number of failed disks and the fault tolerance, and then facilitate timely adoption of different data reconstruction and recovery methods for data recovery, so as to improve the data reliability of the storage system.

It should be noted that while returning to execute the step of periodically monitoring whether the number of failed disks in the RAID array in the master control node exceeds the fault tolerance of the RAID array, the master control node can also use the data redundancy mechanism inside the RAID array to recover data to ensure the data reliability of the storage system.

In some embodiments of the present application, if the number of auxiliary nodes is greater than 1, sending information about lost data in the failed disk to the auxiliary node may include:

The information about lost data in the failed disk is sent to an auxiliary node selected from a plurality of auxiliary nodes according to a preset selection strategy.

In the present application, if the number of auxiliary nodes is greater than 1, the master node may select an auxiliary node from the plurality of auxiliary nodes according to a preset selection strategy (for example, a selection strategy with the smallest workload or a selection strategy with the best working performance (lowest failure rate, etc.), etc.), and then the master node may send the information about the lost data to the selected auxiliary node. Click to use the mirror data in the auxiliary node to reconstruct and restore the data in the failed disk.

Through the above method, one of the multiple auxiliary nodes can be involved in data recovery, so that stable and orderly data recovery can be performed, thereby improving data reliability.

If the number of auxiliary nodes is 1, the master control node directly sends the information of lost data in the failed disk to the auxiliary node, so that the auxiliary node participates in the recovery of the failed disk data.

In some embodiments of the present application, after the hot spare disk is used to replace the corresponding failed disk to re-form a RAID array with the normal disk, the following method may also be included:

Regularly clean up the mirror data in the secondary node.

In the present application, after the hot spare disk is used to replace the corresponding failed disk to re-form the RAID array with the normal disk, the mirror data in the auxiliary node can also be cleaned up regularly (specifically, the mirror data in the auxiliary node can be cleaned up by overwriting) to reduce memory usage and enable the auxiliary node to use more data sent by the host as mirror data to participate in the recovery of the failed disk data.

The embodiment of the present application further provides a data recovery device. Referring to FIG. 8 , a schematic diagram of the structure of a data recovery device provided by the embodiment of the present application is shown, which may include:

A backup module 81, used to back up the data in the main control node in the auxiliary node among the multiple control nodes, so as to use the data in the auxiliary node as the mirror data of the main control node;

A monitoring module 82, used to monitor whether the number of failed disks in the RAID array in the master control node exceeds the fault tolerance of the RAID array;

A sending module 83 is used to send information about lost data in the failed disk to the auxiliary node if the number of failed disks in the RAID array exceeds the fault tolerance of the RAID array, and the auxiliary node obtains data corresponding to the information from the mirror data and sends it;

A writing module 84, used for receiving data corresponding to the information, and writing the data corresponding to the information into a corresponding partition of the hot spare disk corresponding to the failed disk;

The first replacement module 85 is used to replace the corresponding failed disk with a hot spare disk to re-form a RAID array with normal disks.

In some embodiments of the present application, the backup module 81 may include:

The backup unit is used to back up the logical volume in the master node in the auxiliary node to obtain the corresponding mirror logical volume; the logical volume and the mirror logical volume contain multiple blocks, and there is a mapping relationship between the blocks and the blocks in the disk, and the mapping relationship is stored in the master node and the auxiliary node;

The sending module 83 may include:

A forming unit, used to form data loss metadata corresponding to the logical volume according to the blocks and mapping relationship contained in the failed disk, and send the data loss metadata to the auxiliary node; the data loss metadata includes information on whether the blocks contained in the logical volume are lost;

The auxiliary node is specifically used to obtain the block number of the lost data according to the mapping relationship and the data loss metadata, and send the data corresponding to the block number in the mirrored logical volume to the main control node.

In some embodiments of the present application, the forming unit may include:

A subunit is formed, which is used to form data loss metadata corresponding to the logical volume and organized in a bitmap according to the blocks and mapping relationships contained in the failed disk; wherein each bit in the bitmap indicates whether the corresponding block contained in the logical volume is lost.

In some embodiments of the present application, the following may also be included:

The judgment module is used to judge whether each stripe in the RAID array is The check blocks need to be recalculated;

The first calculation module is used to calculate the check block according to each block in the stripe if there is a stripe for which the check block needs to be recalculated, and write the calculated check block into the corresponding partition of the hot spare disk corresponding to the failed disk.

In some embodiments of the present application, the writing module 84 may include:

The writing unit is used to receive data corresponding to the information block by block, and write the data corresponding to the information block by block into the corresponding partition of the hot spare disk corresponding to the failed disk.

In some embodiments of the present application, the following may also be included:

A first redirection module, configured to redirect the host's I/O request to the auxiliary node if the number of failed disks in the RAID array exceeds the fault tolerance of the RAID array;

The second redirection module is used to redirect the I/O request of the host to the master control node after the hot spare disk replaces the corresponding failed disk to re-compose the RAID array with the normal disk.

In some embodiments of the present application, the following may also be included:

The storage module is used to store the data newly sent by the host in the main control node when redirecting the I/O request of the host to the auxiliary node, so as to use the data newly sent by the host stored in the main control node as the mirror data of the auxiliary node.

In some embodiments of the present application, the following may also be included:

A second calculation module is used to calculate the blocks of the hot spare disk corresponding to the failed disk by using the blocks of the normal disk contained in each stripe in the RAID array if the number of failed disks in the RAID array does not exceed the fault tolerance of the RAID array;

The second replacement module is used to replace the failed disk with a hot spare disk after all data lost on the failed disk is recovered, so as to re-form a RAID array with normal disks.

In some embodiments of the present application, the monitoring module 82 may include:

The timing monitoring unit is used to periodically monitor whether the number of failed disks in the RAID array in the master control node exceeds the fault tolerance of the RAID array.

In some embodiments of the present application, the monitoring module 82 may also include:

The return execution unit is used to return to the step of periodically monitoring whether the number of failed disks in the RAID array in the master control node exceeds the fault tolerance of the RAID array if the number of failed disks in the RAID array does not exceed the fault tolerance of the RAID array.

In some embodiments of the present application, if the number of auxiliary nodes is greater than 1, the sending module 83 may include:

The sending unit is used to send the information of lost data in the failed disk to an auxiliary node selected from multiple auxiliary nodes according to a preset selection strategy.

In some embodiments of the present application, the following may also be included:

The scheduled cleaning module is used to regularly clean up the mirror data in the auxiliary node after the hot spare disk replaces the corresponding failed disk to re-compose the RAID array with the normal disk.

The embodiment of the present application further provides a data recovery device. Referring to FIG. 9 , a schematic diagram of the structure of a data recovery device provided by the embodiment of the present application is shown, which may include:

A memory 91, used for storing computer programs;

The processor 92, when used to execute the computer program stored in the memory 91, can implement the following steps:

Back up the data in the master control node in the auxiliary node in the multi-control node so that the data in the auxiliary node can be used as the mirror data of the master control node; monitor whether the number of failed disks in the RAID array in the master control node exceeds the fault tolerance of the RAID array; if the number of failed disks in the RAID array exceeds the fault tolerance of the RAID array, send the information of the lost data in the failed disk to the auxiliary node; The auxiliary node obtains the data corresponding to the information from the mirror data and sends it; receives the data corresponding to the information, and writes the data corresponding to the information into the corresponding partition of the hot spare disk corresponding to the failed disk; uses the hot spare disk to replace the corresponding failed disk to re-form the RAID array with the normal disk.

The embodiment of the present application further provides a non-volatile readable storage medium, in which a computer program is stored. When the computer program is executed by a processor, the following steps can be implemented:

Back up the data in the main control node in the auxiliary node in the multi-control node to use the data in the auxiliary node as the mirror data of the main control node; monitor whether the number of failed disks in the RAID array in the main control node exceeds the fault tolerance of the RAID array; if the number of failed disks in the RAID array exceeds the fault tolerance of the RAID array, send the information of the lost data in the failed disk to the auxiliary node, and the auxiliary node obtains the data corresponding to the information from the mirror data and sends it; receive the data corresponding to the information, and write the data corresponding to the information into the corresponding partition of the hot spare disk corresponding to the failed disk; use the hot spare disk to replace the corresponding failed disk to re-form the RAID array with the normal disk.

The non-volatile readable storage medium may include: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and other media that can store program codes.

The description of the relevant parts of a data recovery device, equipment and readable storage medium provided in the present application can refer to the detailed description of the corresponding parts of a data recovery method provided in an embodiment of the present application, and will not be repeated here.

It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that there is any such actual relationship or order between these entities or operations. Moreover, the term "include", "comprise" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method, article or equipment that includes a series of elements are inherent to the elements. In the absence of more restrictions, the elements limited by the sentence "comprise one..." do not exclude the presence of other identical elements in the process, method, article or equipment that includes the elements. In addition, the above-mentioned technical solution provided in the embodiment of the present application is consistent with the corresponding technical solution in the prior art in principle, and the part is not described in detail, so as not to repeat too much.

The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims (20)

A data recovery method, comprising: Backing up data in the main control node in an auxiliary node among the multiple control nodes, so as to use the data in the auxiliary node as mirror data of the main control node; Monitoring whether the number of failed disks in the RAID array in the master control node exceeds the fault tolerance of the RAID array; If the number of failed disks in the RAID array exceeds the fault tolerance of the RAID array, information about lost data in the failed disks is sent to the auxiliary node, and the auxiliary node obtains data corresponding to the information from the mirror data and sends the data; receiving data corresponding to the information, and writing the data corresponding to the information into a corresponding partition of a hot spare disk corresponding to the failed disk; The hot spare disk is used to replace the corresponding failed disk to re-form a RAID array with normal disks. The data recovery method according to claim 1, characterized in that backing up the data in the main control node in the auxiliary node among the multiple control nodes comprises: The logical volume in the master node is backed up in the auxiliary node to obtain a corresponding mirrored logical volume; the logical volume and the mirrored logical volume contain a plurality of blocks, and there is a mapping relationship between the blocks and the blocks in the disk, and the mapping relationship is stored in both the master node and the auxiliary node; Sending information about lost data in the failed disk to the auxiliary node includes: According to the blocks contained in the failed disk and the mapping relationship, data loss metadata corresponding to the logical volume is formed, and the data loss metadata is sent to the auxiliary node; the data loss metadata includes information on whether the blocks contained in the logical volume are lost; The auxiliary node obtains data corresponding to the information from the mirror data and sends the data, including: The auxiliary node obtains the block number of the lost data according to the mapping relationship and the data loss metadata, and sends the data corresponding to the block number in the mirrored logical volume to the main control node. The data recovery method according to claim 2 is characterized in that, according to the blocks contained in the failed disk and the mapping relationship, forming the data loss metadata corresponding to the logical volume comprises: According to the blocks contained in the failed disk and the mapping relationship, the data loss metadata corresponding to the logical volume and organized in a bitmap is formed; wherein each bit in the bitmap indicates whether the corresponding block contained in the logical volume is lost. The data recovery method according to claim 1, characterized in that before using the hot spare disk to replace the corresponding failed disk, it also includes: Determine whether each stripe in the RAID array needs to recalculate the check blocks; If there is a stripe for which the check blocks need to be recalculated, the check blocks are calculated according to the blocks in the stripe, and the calculated check blocks are written into the corresponding partition of the hot spare disk corresponding to the failed disk. The data recovery method according to claim 4, characterized in that determining whether each stripe in the RAID array needs to recalculate the check block comprises: Determine whether the check block in the stripe is located on the failed disk; or, Determine whether the parity block is missing in the stripe. The data recovery method according to claim 5, further comprising: If the parity block in the stripe is located on the failed disk, or the parity block is missing in the stripe, it is determined that the stripe needs Recalculate the checksum block; If the parity block in the stripe is not located on the failed disk, or the parity block is not missing in the stripe, it is determined that the stripe does not need to recalculate the parity block. The data recovery method according to claim 4, characterized in that calculating the check block according to each block in the stripe comprises: The blocks in the stripe are used to perform an XOR operation to calculate the check block. The data recovery method according to claim 1, characterized in that receiving the data corresponding to the information and writing the data corresponding to the information into a corresponding partition of the hot spare disk corresponding to the failed disk comprises: The data corresponding to the information is received block by block, and the data corresponding to the information is written block by block into a corresponding partition of the hot spare disk corresponding to the failed disk. The data recovery method according to claim 1, characterized in that if the number of failed disks in the RAID array exceeds the fault tolerance of the RAID array, further comprising: Redirecting the I/O request of the host to the auxiliary node; After the hot spare disk is used to replace the corresponding failed disk to re-form a RAID array with a normal disk, the method further includes: Redirecting the I/O request of the host to the master control node. The data recovery method according to claim 9, characterized in that when redirecting the I/O request of the host to the master control node, it also includes: The I/O request of the host is no longer received and processed. The data recovery method according to claim 9, characterized in that when redirecting the I/O request of the host to the auxiliary node, it also includes: The data newly sent by the host is stored in the master control node, so that the data newly sent by the host stored in the master control node is used as the mirror data of the auxiliary node. The data recovery method according to claim 1, characterized in that if the number of failed disks in the RAID array does not exceed the fault tolerance of the RAID array, it further comprises: Using the blocks of the normal disks contained in each stripe in the RAID array to perform calculations, obtain the blocks of the hot spare disk corresponding to the failed disk; After all the lost data of the failed disk is recovered, the hot spare disk is used to replace the failed disk to re-form a RAID array with normal disks. The data recovery method according to claim 1, characterized in that monitoring whether the number of failed disks in the RAID array in the master control node exceeds the fault tolerance of the RAID array comprises: Regularly monitor whether the number of failed disks in the RAID array in the master control node exceeds the fault tolerance of the RAID array. The data recovery method according to claim 1, characterized in that monitoring whether the number of failed disks in the RAID array in the master control node exceeds the fault tolerance of the RAID array comprises: Real-time monitoring is performed to determine whether the number of failed disks in the RAID array in the master control node exceeds the fault tolerance of the RAID array. The data recovery method according to claim 13 or 14, characterized in that it also includes: If the number of failed disks in the RAID array does not exceed the fault tolerance of the RAID array, the process of periodically monitoring whether the number of failed disks in the RAID array in the master control node exceeds the fault tolerance of the RAID array is returned. step. The data recovery method according to claim 1, characterized in that if the number of the auxiliary nodes is greater than 1, sending the information of lost data in the failed disk to the auxiliary node comprises: The information about lost data in the failed disk is sent to an auxiliary node selected from the plurality of auxiliary nodes according to a preset selection strategy. The data recovery method according to claim 1, characterized in that after using the hot spare disk to replace the corresponding failed disk to re-form a RAID array with a normal disk, it also includes: The mirror data in the auxiliary node is cleaned up regularly. A data recovery device, comprising: A backup module, used to back up data in the main control node in the auxiliary node among the multiple control nodes, so as to use the data in the auxiliary node as the mirror data of the main control node; A monitoring module, used to monitor whether the number of failed disks in the RAID array in the master control node exceeds the fault tolerance of the RAID array; A sending module, configured to send information about lost data in the failed disks to the auxiliary node if the number of failed disks in the RAID array exceeds the fault tolerance of the RAID array, and the auxiliary node obtains data corresponding to the information from the mirror data and sends the data; A writing module, used for receiving data corresponding to the information, and writing the data corresponding to the information into a corresponding partition of the hot spare disk corresponding to the failed disk; The first replacement module is used to use the hot spare disk to replace the corresponding failed disk, so as to re-form a RAID array with normal disks. A data recovery device, comprising: Memory for storing computer programs; A processor, configured to implement the steps of the data recovery method according to any one of claims 1 to 17 when executing the computer program. A non-volatile readable storage medium, characterized in that a computer program is stored in the non-volatile readable storage medium, and when the computer program is executed by a processor, the steps of the data recovery method according to any one of claims 1 to 17 are implemented.