WO2023102784A1

WO2023102784A1 - Data access method and apparatus, disk controller, disk and data storage system

Info

Publication number: WO2023102784A1
Application number: PCT/CN2021/136563
Authority: WO
Inventors: 黄飞; 郭林; 徐浩浩; 王浩博; 朱文翰; 方照勋
Original assignee: 华为技术有限公司
Priority date: 2021-12-08
Filing date: 2021-12-08
Publication date: 2023-06-15
Also published as: CN116583827A

Abstract

Disclosed in the present application are a data access method and apparatus, a disk controller, a disk and a data storage system. The data access method comprises: a disk controller acquiring first data to be written to a target disk and sending the first data to the target disk; when it has been determined that the writing of the first data needs to be delayed, the disk controller sending the target disk a delayed writing instruction for the first data, so as to prevent the target disk from frequently executing data writing operations to reduce the resource consumption caused by said operations; or the disk controller determining a target storage address and a target disk for prefetched data, and sending the target disk a data prefetching instruction generated on the basis of the target storage address, so that the target disk stores in a cache unit the data stored in a non-volatile memory, enabling the data to subsequently be directly read from the cache unit without needing to be read from the non-volatile memory, and thereby greatly reducing the reading latency, enhancing the data reading efficiency and improving the performance of the data storage system.

Description

Data access method, device, disk controller, disk and data storage system

technical field

The present application relates to the technical field of computer storage, in particular to a data access method, device, disk controller, disk and data storage system.

Background technique

A magnetic disk is a memory used to store computer programs or data, and it has the characteristic of keeping data intact after power failure. In a data storage system, a computer or an application server can save processed data to a disk through a write disk command, or read required data from a disk through a read disk command.

However, if the disk receives the data that needs to be written by the computer or application server, it executes the data write operation, which will lead to frequent data write operations, increasing system power consumption and resource consumption; Performing a data read operation to read the required data from the disk will increase the read delay and reduce the efficiency of reading data. The above data reading and writing process will affect the performance of the data storage system.

Contents of the invention

The embodiments of the present application provide a data access method, device, disk controller, disk and data storage system, so as to improve the performance of the data storage system.

In the first aspect, the embodiment of the present application provides a data access method, the method includes: the disk controller in the data storage system obtains the first data to be written into the target disk, and sends the first data to the target disk; When a piece of data satisfies the following setting conditions, the disk controller sends a delayed write instruction for the first data to the target disk; wherein the setting conditions include: writing to the data to be written in the disk array after the first data Including the second data; the disk array includes a plurality of disks controlled by the disk controller; the second data is the data to be written in the target stripe; the target stripe includes multiple stripe units located on different disks, and the multiple stripes The unit includes a target stripe unit used to save the first data in the target disk.

The disk controller sends a delayed write instruction for the first data to the target disk, so that the target disk delays writing the first data into the non-volatile memory, avoids the disk from frequently performing data write operations, and saves resources brought about by writing data Consumption, improve the performance of the data storage system.

Compared with the target disk judging whether to delay writing data to the non-volatile memory, the disk controller can not only know the data written to the target disk, but also know the data written to other disks in the disk array. Therefore, the disk controller can more accurately determine whether it is necessary to delay writing data to the non-volatile memory of the target disk in conjunction with the situation of writing data to multiple disks.

Exemplarily, after acquiring the first data, the disk controller determines a target disk, a target stripe unit, and a target stripe for storing the first data. Wherein, the target stripe unit is a stripe unit in the target disk, the target stripe is a stripe including the target stripe unit, and the target stripe includes not only the target stripe unit, but also the stripe unit located in different disks of the disk array. multiple strip units. The disk controller can also predict the disks and stripes corresponding to the data to be written into the disk array after the first data.

For example, if the first data is the data to be written to the first stripe unit of the target disk, the disk controller predicts that after the first data, the data to be written to the disk, including the second stripe The second data of the unit, where the second stripe unit and the first stripe unit belong to the same target stripe, and the second stripe unit and the first stripe unit belong to different disks, the disk controller can predict that the When inputting the second data, it is also necessary to read back the first data, and update the parity data of the target stripe based on the first data and the second data. Therefore, the disk controller determines that the first data is to be read back later. return the data, so as to determine that the first data satisfies the set condition, the disk controller sends a delayed write instruction of the first data to the target disk, so that the target disk saves the first data in the cache unit, and the delayed write non-volatile In storage, reduce the number of write data operations performed by the target disk. After the disk controller obtains the second data, it can obtain the first data from the target disk, determine the updated parity data based on the first data and the second data, and then send a data write instruction to the target disk again, and the data write The instruction instructs the target disk to write the first data into the target storage address in the non-volatile memory.

For another example, if the first data is the parity data to be written to the third stripe unit of the target disk, the disk controller predicts that after the first data, the data to be written to the disk also includes The second data of the second stripe unit in the target stripe needs to update the parity data when writing the second data, so the disk controller can determine that the first data is the data to be updated that needs to be updated, and then send The target disk sends a delayed write indication, so that the target disk delays writing the first data. After acquiring the second data, the disk controller may determine updated check data based on the second data, which is hereinafter referred to as updated data. The disk controller sends a data write instruction for updating data to the target disk, wherein the data write instruction for updating data is used to instruct the target disk to write the update data into the non-volatile memory of the target disk. At this time, the target disk can Directly write the update data without writing the first data into the non-volatile memory, which can reduce the repeated execution of data write operations for the same stripe unit in the non-volatile memory, reduce the number of erases and writes that consume the disk, and extend the length of the disk. service life.

In a possible implementation, the disk controller can be connected between the host and the disk, the host sends the data to be written to the disk to the disk controller through the write disk command, and the disk controller sends the data write command to the disk . In the embodiment of the present application, the first data may be the data carried in the disk write command received by the disk controller, and the disk controller may judge the data writing mode according to the logical addresses in the multiple write disk commands received continuously Whether it is sequential writing, if the data writing mode is sequential writing, and there are other free stripe units in the target stripe besides the target stripe unit, the disk controller can speculate that the data to be written includes the The second data written into the target stripe can further determine that the first data satisfies the set condition.

Since the disk controller is connected between the host and the disk, multiple disks can be managed. Compared with the disk determining whether the first data is data that can be written delayed, the disk controller can write data to multiple disks in combination. In a case, it is more accurately determined whether the first data is data that can be written in a delayed manner, thereby reducing misoperation of the disk.

In a possible implementation manner, after the disk controller obtains the first data to be written to the target disk, it sends a data write command containing the first data to the target disk, and the disk controller may carry the delayed write instruction in the In the data write command, send to the target disk. Carrying the delayed write indication in the data write instruction can reduce signaling overhead. In another possible implementation manner, the disk controller may also carry the delay write instruction in an instruction other than the data write instruction of the first data, and send it to the target disk. For example, the disk controller may send a separate operation instruction including a delayed write instruction to the target disk, so as to increase the degree of freedom in sending the delayed write instruction. Regardless of whether the delayed write instruction is sent to the target disk through a data write command or through a separate operation command, the delayed write instruction may include a delayed write identifier and a target storage address of the first data, wherein the first The target storage address of a piece of data may be the address of a target stripe unit used to store the first data, so that the target disk can accurately determine which address the data can delay writing.

If the disk controller sends the delayed write instruction to the target disk in an instruction other than the data write instruction of the first data, the instruction containing the delayed write instruction can be sent to the target disk in the data write instruction of the first data. After sending the command, the time interval between the sending time and the sending time of the data write command of the first data is less than the set time threshold, so as to ensure that before the target disk writes the first data into the non-volatile memory, the A delayed write indication for first data is received.

In the second aspect, the embodiment of the present application provides a data access method, the method includes: the disk controller in the data storage system determines the target storage address of the prefetched data and the target disk, and sends the data generated based on the target storage address to the target disk. The prefetch data instruction, so that the target disk pre-reads the data stored in the target storage address of the non-volatile memory into the cache unit based on the prefetch data instruction, and the subsequent target disk receives the first data for the data When reading instructions, the data can be read directly from the cache unit without having to read the data in the non-volatile memory, which can greatly reduce the read delay, improve the efficiency of reading data, and improve data storage. system performance.

In a possible implementation, the disk controller can determine whether the data read mode is sequential read according to the logical addresses in multiple disk read instructions received continuously, and if the data read mode is sequential read, then according to the current The logical address in the received disk read command determines the target storage address and target disk of the prefetched data. In another possible implementation manner, during the process of repairing the disk to be repaired, the disk controller may determine the target storage address and the target disk of the prefetched data according to the target stripe to which the stripe unit to be repaired belongs. Wherein, the stripe unit to be repaired is the stripe unit in the disk to be repaired, the target stripe includes a plurality of stripe units located on different disks, and the target storage address can be the stripe unit in the target stripe except the disk to be repaired The address of any stripe unit other than .

Since the disk controller is connected between the host and the disk, it can manage multiple disks. Compared with the data stored in which address is prefetched by the disk, the disk controller can combine the management of multiple disks, which is more accurate. It can accurately determine the data stored in which address needs to be prefetched, thereby reducing the misoperation of the disk.

In a possible implementation, after the disk controller determines the target storage address of the prefetched data and the target disk, it may carry the prefetch data indication in the second data read command sent to the target disk, and send it to the target disk ; The second data read command is used to instruct the target disk to read data from the specified storage address. Carrying the prefetch data instruction in the data write instruction or the data read instruction can reduce signaling overhead. In another possible implementation manner, the disk controller may send a separate operation instruction including the prefetch data indication to the target disk, so as to increase the degree of freedom in sending the prefetch data indication. Wherein, the individual operation instruction may be any access instruction sent by the disk controller to the target disk.

In the third aspect, the embodiment of the present application also provides a data access method, which is applied to a disk, and the method includes:

receiving the first data sent by the disk controller, and receiving a delayed write indication of the first data sent by the disk controller; the delayed write indication is used to instruct the disk to temporarily store the first data in a cache unit of the disk;

Temporarily store the first data in the cache unit of the disk.

In a possible implementation manner, the delayed writing indication includes a delayed writing identifier and a target storage address of the first data.

In a possible implementation manner, the disk may receive a data write instruction sent by the disk controller, and obtain a delayed write instruction carried in the data write instruction; the data write instruction includes the above-mentioned first data. In another possible implementation manner, the disk may receive an individual operation command sent by the disk controller including a delayed write instruction.

In a possible implementation, if the first data is verification data, after the first data is temporarily stored in the cache unit of the disk, the disk receives a data write instruction for updating data sent by the disk controller; wherein, the update The data is the updated verification data, and the data write instruction of the updated data instructs the disk to write the updated verification data into the target storage address, and the target storage address is the same as the target storage address of the first data; according to the target storage address, the The update data is written into the non-volatile memory of the disk, and the first data is deleted in the cache unit.

In the fourth aspect, the embodiment of the present application also provides a data access method, which is applied to a disk, and the method includes:

receiving the prefetch data instruction sent by the disk controller; the prefetch data instruction is generated by the disk controller based on the target storage address of the prefetch data;

Reading the prefetch data stored in the target storage address carried by the prefetch data indication into the cache unit of the disk;

receiving a first data read instruction for prefetched data sent by the disk controller;

Get the prefetched data from the cache unit, and send the prefetched data to the disk controller.

In a possible implementation manner, the disk may receive the second data read instruction sent by the disk controller, and acquire the prefetch data instruction carried in the second data read instruction; the second data read instruction is used to instruct the disk to read from Specify the storage address to read data. In another possible implementation manner, the disk may receive an individual operation command sent by the disk controller including the instruction to prefetch data.

In the fifth aspect, the embodiment of the present application also provides a data access device, which is applied to a disk controller, and the device includes:

a data acquisition unit, configured to acquire the first data to be written into the target disk;

A sending unit, configured to send the first data to the target disk when it is determined that the first data satisfies the set condition, and send a delayed write indication of the first data to the target disk; the delayed write indication is used to indicate the target The disk temporarily stores the first data in the cache unit of the target disk.

In the sixth aspect, the embodiment of the present application also provides a data access device, which is applied to a disk controller, and the device includes:

A prediction unit for determining a target storage address and a target disk for prefetching data; the target disk includes a cache unit;

The instruction sending unit is used to send to the target disk a prefetch data instruction generated based on the target storage address; the prefetch data instruction is used to instruct the target disk to read the prefetch data from the target storage address into the cache unit; and, to the target disk Send a data read command for prefetched data;

The data receiving unit is configured to receive the prefetch data obtained and sent by the target disk from the cache unit.

In the seventh aspect, the embodiment of the present application also provides a data access device, which is applied to a magnetic disk, and the device includes:

The receiving unit is configured to receive the first data sent by the disk controller, and receive the delayed writing instruction of the first data sent by the disk controller; the delayed writing instruction is used to instruct the disk to temporarily store the first data in the cache unit of the disk ;

The execution unit is configured to temporarily store the first data in the cache unit of the disk.

In the eighth aspect, the embodiment of the present application also provides a data access device, which is applied to a magnetic disk, and the device includes:

The instruction receiving unit is used to receive the prefetch data instruction sent by the disk controller; the prefetch data instruction carries the target storage address of the prefetch data;

an instruction execution unit, configured to read the prefetched data stored in the storage space corresponding to the target storage address into the cache unit of the disk;

The data transmission unit is configured to receive a first data read instruction for the prefetched data sent by the disk controller, acquire the prefetched data from the cache unit, and send the prefetched data to the disk controller.

In the ninth aspect, the embodiment of the present application also provides a disk controller, including a memory and a processor, the memory stores a computer program that can run on the processor, when the computer program is executed by the processor When executed, the processor is made to implement any one of the methods described in the above-mentioned first aspect, or any one of the methods described in the above-mentioned second aspect.

In the tenth aspect, the embodiment of the present application also provides a magnetic disk, including a memory and a processor, the memory stores a computer program that can run on the processor, when the computer program is executed by the processor , so that the processor implements any one of the methods described in the above third aspect, or any one of the methods described in the above fourth aspect.

In the eleventh aspect, the embodiment of the present application further provides a data storage system, including the disk controller provided in the ninth aspect and the disk provided in the tenth aspect.

In a twelfth aspect, an embodiment of the present application provides a computer-readable storage medium, where computer-executable instructions are stored in the computer-readable storage medium, and the computer-executable instructions are used to make a computer perform any of the above-mentioned first aspects. A method, or any method provided in the second aspect above.

In a thirteenth aspect, an embodiment of the present application provides a computer-readable storage medium, where computer-executable instructions are stored in the computer-readable storage medium, and the computer-executable instructions are used to make a computer perform any of the above-mentioned third aspects. A method, or any method provided in the fourth aspect above.

In a fourteenth aspect, the embodiment of the present application provides a computer program product, which includes computer-executable instructions, and the computer-executable instructions are used to make a computer execute any method provided in the above-mentioned first aspect, or the method provided in the above-mentioned second aspect. any of the methods.

In the fifteenth aspect, the embodiment of the present application provides a computer program product, which includes computer-executable instructions, and the computer-executable instructions are used to make a computer execute any method provided in the third aspect above, or the method provided in the fourth aspect above. any of the methods.

For the technical effects that can be achieved by any one of the third aspect to the fifteenth aspect above, reference can be made to the description of the beneficial effects in the first aspect or the second aspect above, and will not be repeated here.

Description of drawings

FIG. 1 is a schematic structural diagram of a data storage system provided by an embodiment of the present application;

FIG. 2 is a schematic structural diagram of a disk array provided by an embodiment of the present application;

FIG. 3 is a schematic structural diagram of another data storage system provided by an embodiment of the present application;

FIG. 4 is an interaction diagram between a disk controller and a target disk in the process of writing data provided by an embodiment of the present application;

FIG. 5 is an interaction diagram between a disk controller and a target disk in another data writing process provided by an embodiment of the present application;

FIG. 6 is an interaction diagram between a disk controller and a target disk in a process of prefetching data provided by an embodiment of the present application;

FIG. 7 is an interaction diagram between a disk controller and a target disk in another process of prefetching data provided by an embodiment of the present application;

FIG. 8 is a flow chart of a data access method performed by a disk controller provided in an embodiment of the present application;

FIG. 9 is a flow chart of a data access method performed by a disk provided in an embodiment of the present application;

FIG. 10 is a flow chart of another data access method performed by a disk controller provided in an embodiment of the present application;

FIG. 11 is a flow chart of another disk-executed data access method provided in the embodiment of the present application;

FIG. 12 is a structural block diagram of a data access device provided by an embodiment of the present application;

FIG. 13 is a structural block diagram of another data access device provided by the embodiment of the present application;

FIG. 14 is a structural block diagram of another data access device provided by the embodiment of the present application;

FIG. 15 is a structural block diagram of another data access device provided by the embodiment of the present application;

FIG. 16 is a schematic structural diagram of a disk controller provided by an embodiment of the present application;

FIG. 17 is a schematic structural diagram of a magnetic disk provided by an embodiment of the present application.

Detailed ways

In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the embodiments of the present application will be described in detail below in conjunction with the accompanying drawings. The terms used in the embodiments of the present application are only used to explain specific embodiments of the present application, and are not intended to limit the present application. Apparently, the described embodiments are only some of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the scope of protection of this application.

Before introducing the specific solutions provided by the embodiments of this application, some terms in this application will be explained to facilitate understanding by those skilled in the art, and the terms in this application will not be limited.

(1) Redundant array of independent disks (RAID) system: a disk array composed of multiple disks (also known as physical disks), also known as a disk group, multiple disks as a whole Provides storage functions and is the basis of virtual disks. A virtual disk can be understood as a continuous data storage unit divided by a disk group, which is equivalent to one or more independent disks. A RAID system uses a RAID controller to manage multiple disks, and one RAID controller can manage multiple disks.

The RAID controller obtains redundancy protection by writing the same data to multiple disks at the same time or writing the verification results corresponding to the written data to the disks, thereby improving the reliability of stored data and providing users with higher storage performance and more High data reliability.

"Multiple" in the embodiment of the present application refers to two or more, in view of this, "multiple" can also be understood as "at least two" in the embodiment of the present application. "At least one" can be understood as one or more, such as one, two or more. For example, including at least one means including one, two or more, and does not limit which ones are included, for example, including at least one of A, B and C, then what is included can be A, B, C, A and B, A and C, B and C, or A and B and C. "And/or" describes the association relationship of associated objects, indicating that there may be three types of relationships, for example, A and/or B may indicate: A exists alone, A and B exist simultaneously, and B exists independently. In addition, the character "/", unless otherwise specified, generally indicates that the associated objects before and after are in an "or" relationship.

Unless otherwise specified, ordinal numerals such as "first" and "second" mentioned in the embodiments of the present application are used to distinguish multiple objects, and are not used to limit the sequence, timing, priority or importance of multiple objects.

In order to improve the performance of a data storage system, an embodiment of the present application provides a data access method. The data access method can be applied to the data storage system shown in FIG. 1 or FIG. 3 .

FIG. 1 is a schematic structural diagram of a data storage system. Due to the ever-increasing business volume that needs to be processed, the data that needs to be run on a single server is also increasing. When a single disk is limited by capacity and security and cannot support system services on the server, multiple disks need to be combined and used as a virtual disk to meet the actual needs of the server.

The data storage system shown in FIG. 1 may include a server 100, a disk array 300 composed of multiple disks, and a disk controller 200 for managing the disks. Wherein, the disk controller 200 may be a RAID controller connected between the server 100 and multiple disks. The server 100 may be referred to as a host (host), and the disk may be a solid state disk (solid state disk, SSD) or a mechanical hard disk, such as a serially attached small computer system interface (serial attached small computer system interface, SAS) disk or Serial advanced technology attachment (SATA) disk. A RAID controller and a plurality of disks form a RAID system. The RAID controller receives a disk access command sent by a host, and obtains a logical address in the disk access command. The logical address may include a logical unit number (LUN). The RAID controller converts the logical address in the disk access instruction into the physical address of the disk, and accesses the target disk corresponding to the physical address, and writes data to or reads data from the target disk according to the disk access instruction. In the application scenario shown in Figure 1, the host is a server, and in other application scenarios, the host may also be a computer or other electronic equipment.

Exemplarily, as shown in FIG. 2 , it is assumed that the disk array 300 includes three disks, namely disk 301 , disk 302 and disk 303 , and the three disks are managed by the disk controller 200 . For example, the disk controller may be a RAID controller, and the RAID controller and three disks form a RAID system, and the RAID controller may manage the three disks according to RAID5 level rules.

In a RAID system, the RAID controller obtains redundancy protection by writing the data and the check results calculated on the data to multiple disks. for different RAID levels. Different RAID levels correspond to different data redundancy strategies; for example, a RAID level 5 RAID system can support one disk failure, and a RAID 6 level RAID system can support two disk failures at the same time. When a disk fails, as long as the number of failed disks does not exceed the tolerance range of the RAID system, the RAID system can restore the data in the failed disk according to the data and parity information in the remaining disks, thus ensuring data consistency. Currently, seven standard RAID levels are RAID0 level, RAID1 level, RAID2 level, RAID3 level, RAID4 level, RAID5 level, and RAID6 level. Standard RAID levels can be combined to form new extended levels, such as RAID50 level and RAID60.

As shown in FIG. 2, the storage space in the disk 301, the disk 302 and the disk 303 can be divided into one or more virtual disks, wherein the logical address corresponding to a virtual disk can be LUN x. A virtual disk may include multiple stripes. The virtual disk shown in FIG. 2 includes five stripes, namely stripe 1 , stripe 2 , stripe 3 , stripe 4 and stripe 5 . Each strip includes a plurality of strip units. The multiple stripe units that make up a stripe are located on different disks. Taking strip 1 as an example, strip 1 includes a strip unit 11 , a strip unit 12 and a strip unit 13 . The stripe unit 11 is located on the disk 301 , the stripe unit 12 is located on the disk 302 , and the stripe unit 13 is located on the disk 303 . Taking the stripe unit 11 as an example, the stripe unit 11 located on the disk 301 means that one or more storage units in the disk 301 form the stripe unit 11 . In addition to the stripe unit 11, the disk 301 further includes a stripe unit 21, a stripe unit 31, a stripe unit 41, and the like. Different stripe units in disk 301 are composed of different storage units. The storage unit refers to the smallest unit in the disk that can perform read and write operations. Different types of disks have different storage units. For example, the smallest storage unit of an SSD is a page. The smallest storage unit of a mechanical hard disk is a sector.

The size of the stripe unit, which may also be referred to as the stripe length (strip length), can be set according to requirements. For example, each stripe unit shown in FIG. 2 can store 128KB of data. Multiple stripe units forming the same stripe can be divided into stripe units for storing data and stripe units for storing check values. For example, the stripe units used to store data in stripe 1 can be stripe unit 11 and stripe unit 12, and the stripe unit used to store check values can be stripe unit 13; The stripe unit for storing data can be stripe unit 21 and stripe unit 23, the stripe unit for storing check value can be stripe unit 22; the stripe unit for storing data in stripe 3 can be The stripe unit 32 and the stripe unit 33, the stripe unit used to store the check value may be the stripe unit 31; and so on. Taking stripe 1 as an example, data LBA0-127 can be stored in stripe unit 11, a total of 128KB of data, data LBA128-255 can be stored in stripe unit 12, and data LBA0-127 and LBA128 can be stored in stripe unit 13. ~255 The check result P obtained by XOR calculation.

The RAID controller and a plurality of disks can communicate with each other through an internal connection path to transmit control and/or data signals, wherein the internal connection path includes a bus, for example, a peripheral component interconnect express (PCIe) or other A bus used to connect different devices. The RAID controller stores stripe status information, and the stripe status information is used to record the status of each stripe unit in the RAID system. The state of each stripe unit can include an idle state and a valid state. A stripe unit in a valid state can also be called a valid stripe unit. A valid stripe unit refers to a stripe unit that stores data; a stripe unit in an idle state is also It may be referred to as an idle stripe unit, and an idle stripe unit refers to a stripe unit that currently does not store data. For example, for the stripe unit 11, when no data is stored in the stripe unit 11, the state of the stripe unit 11 is idle, and after the RAID controller sends a data write command for the stripe unit 11 to the disk 301, the stripe The state of the belt unit 11 is changed from the idle state to the active state.

The data storage system shown in FIG. 3 may include a computer 400 and a disk array 300 composed of multiple disks. Wherein, the computer 400 can be used as a disk controller for managing multiple disks. Wherein, the computer 400 may also be replaced by a server or other electronic devices. The structure of the disk array 300 may also be shown in FIG. 2 . The process of managing multiple disks by the computer 400 is the same as the process of managing multiple disks by the RAID controller shown in FIG. 1 , and will not be repeated here.

In the embodiment of the present application, the performance of the data storage system is improved through the cooperation between the disk controller and the disk. For ease of understanding, the following takes the RAID controller shown in FIG. 1 as an example of the disk controller to describe the specific implementation manners of the data access method provided by the embodiment of the present application in the process of writing data and prefetching data.

Fig. 4 shows an interaction diagram between the disk controller and the disk during the process of writing data. As shown in Figure 4, in the process of writing data, a data access method provided by an embodiment may include the following steps:

S401. The disk controller acquires first data, and determines a target disk and a data type corresponding to the first data.

The first data can be understood as data to be written to the target disk. The data type of the first data may include delayed write type and non-delayed write type, wherein, delayed write type data may include but not limited to data that needs to be used later, or data that needs to be updated immediately; non-delayed write The type of data can also be referred to as data that does not need to be delayed for writing, for example, data that does not need to be used temporarily.

The RAID controller can determine the target disk and the target stripe to which the target stripe unit for saving the first data belongs, if the data to be written is to be written to one or more disks controlled by the RAID controller after the first data , including the second data to be written into the target stripe, it is determined that the data type of the first data is the delayed write type.

Exemplarily, in some embodiments, the RAID controller receives a disk write command from the host, and determines the first data according to data carried in the disk write command. The RAID controller can determine the target disk for saving the first data according to the logical address in the disk write command. In one embodiment, the first data may be the data carried in the above-mentioned write-to-disk instruction. At this time, the stripe unit corresponding to the logical address in the write-to-disk instruction is the target stripe unit, and the disk to which the target stripe unit belongs is target disk. In another embodiment, the first data may be a verification result determined according to the data carried in the write to disk command. At this time, the target stripe to which the stripe unit corresponding to the logical address in the write to disk command belongs may be determined. And determine the target stripe unit used to save the verification result in the target stripe, and the disk to which the target stripe unit belongs is the target disk, which will be described in detail below.

The RAID controller can also determine whether the current data writing mode is sequential write according to the logical addresses in the multiple write disk commands received continuously, wherein the sequential write refers to the logical addresses in the multiple write disk commands received continuously It is continuous, that is to say, multiple disk write instructions indicate to sequentially write data to each stripe unit in a virtual disk. For example, in an application scenario where a user finishes editing a document and saves the document, the host sends multiple disk write commands to the RAID controller continuously, and the logical addresses in the multiple disk write commands are continuous. It is considered that the current data writing mode is sequential writing, that is, the document data can be sequentially written into each stripe unit indicated by the logical address in each write-to-disk command. If the logical addresses in multiple disk write instructions are discontinuous, it can be determined that the current data writing mode is not sequential writing; for example, in the application scenario where the user resets or modifies the parameters of multiple different functions, The parameters of each function may be stored in different virtual disks or in non-adjacent stripe units in the same virtual disk. When saving the parameters of each function, the host sends multiple disk write commands to the RAID controller, and multiple The logical addresses in the write-to-disk command are discontinuous. At this time, it can be considered that the current data writing mode is not sequential writing.

It has been explained above that the RAID controller can determine the target stripe unit and the target stripe for storing the first data. In an optional embodiment, if the RAID controller determines that the current data writing mode is sequential writing , the RAID controller can determine the state of each stripe unit in the target stripe used to save the first data according to the saved stripe status information, if the target stripe includes an idle stripe other than the target stripe unit unit, it may be determined that the first data is data of a delayed write type, and if there is no other free stripe unit in the target stripe except the target stripe unit, it may be considered that the first data is data of a non-delayed write type. For example, assuming that the current data writing mode is sequential writing, the data carried in the disk writing command received by the RAID controller from the host includes data LBA0-127. At this time, the data to be written may include data LBA0-127 and the verification result P ₀ corresponding to the data LBA0-127.

When writing data LBA0-127, the data LBA0-127 is the first data, and the RAID controller can determine the storage location corresponding to the data LBA0-127 as the stripe unit 11 according to the logical address in the write disk instruction. Band unit 11 is a target stripe unit, stripe 1 to which stripe unit 11 belongs is a target stripe, and disk 301 to which stripe unit 11 belongs is a target disk. According to the saved stripe status information, the RAID controller can determine that the stripe unit 12 in the target stripe is currently in an idle state. Since it is currently written sequentially, the RAID controller can determine that data will be written in the stripe unit 12 in the future, That is, the data to be written includes the second data to be written into the target stripe, and it can be determined that when writing data into the stripe unit 12, the data LBA0-127 and the data to be written into the stripe unit 12 need to be used The verification result P is re-determined. Therefore, it can be determined that the data LBA0-127 are data to be used later, that is, the RAID controller determines that the first data is data of the delayed write type.

When writing the verification result P ₀ , the verification result P ₀ is the first data, and the RAID controller can determine the storage location corresponding to the data LBA0-127 carried in the write disk command according to the logical address in the write disk command as the stripe Strip unit 11, the stripe to which stripe unit 11 belongs is stripe 1, and the verification result P ₀ and the data LBA0-127 need to be stored in the same stripe, so stripe 1 is the target stripe, and the RAID controller can determine, In the stripe 1, the stripe unit 13 is used to store the verification result, therefore, the stripe unit 13 is the target stripe unit, and the disk 303 to which the stripe unit 13 belongs is the target disk. According to the saved stripe status information, the RAID controller can determine that the stripe unit 12 in the target stripe is currently in an idle state. When writing data to the stripe unit 12 later, it needs to use the data LBA0-127 and the data to be written. The data in the stripe unit 12 re-determines the verification result P, and then it can be determined that the verification result P ₀ is data that needs to be updated immediately, that is, the RAID controller determines that the first data is delayed write data.

In another optional embodiment, assuming that the current data writing mode is sequential writing, the RAID controller has sent a data writing instruction to write data LBA0-127 to the stripe unit 11 to the disk 301, and saves In the stripe status information of the stripe unit 11, the state of the stripe unit 11 is set to the valid state, and the RAID controller has sent the verification result P ₀ to the disk 303 to write the data into the stripe unit 13. In the band status information, the status of the stripe unit 13 is set to be valid. When the data carried by the RAID controller in the disk write command received from the host is data LBA128-255, the RAID controller can determine the storage location corresponding to the data LBA128-255 as the stripe unit 12 according to the logical address carried in the write disk command. , the verification result P can be calculated according to the data LBA128-255 and the data LBA0-127 in the stripe unit 11. At this time, the data to be written may include the data LBA128-255 and the verification result P.

When writing data LBA128-255, the data LBA128-255 is the first data, the storage location corresponding to LBA128-255 is the stripe unit 12 as the target stripe unit, and the stripe 1 to which the stripe unit 12 belongs is the target stripe, The disk 302 to which the stripe unit 12 belongs is the target disk. According to the saved stripe state information, the RAID controller can determine that all the stripe units except stripe unit 12 in the target stripe are in valid state. At this time, the RAID controller can determine that the data LBA128-255 are temporarily unnecessary The used non-delayed write type data, that is, the RAID controller may consider the first data as non-delayed write type data.

When writing the verification result P, the verification result P is the first data, and the RAID controller can determine that the storage location corresponding to the verification result P is the stripe unit 13 in the stripe 1, and the stripe 1 is the target stripe, The stripe unit 13 is the target stripe unit, and the disk 303 to which the stripe unit 13 belongs is the target disk. The RAID controller can determine that data is stored in both the stripe unit 11 and the stripe unit 12 in the target stripe, and all the other stripe units in the target stripe except the stripe unit 13 are valid. At this time, the RAID The controller may determine that the verification result P is non-delayed write data, that is, the RAID controller may consider the first data as non-delayed write data.

When performing the business of repairing the disk 302, if the RAID controller determines that each stripe unit in the next stripe of the target stripe corresponding to the first data is in an idle state according to the saved stripe state information, it can be determined that There is no need to continue writing data to the disk 302 for the time being, and it can be considered that the current first data does not need to be delayed in writing, and is non-delayed writing data.

S402. If it is determined that the first data is data of a delayed write type, the disk controller generates a data write instruction including a delayed write instruction.

S403, the disk controller sends a data write instruction to the target disk.

In addition to the delayed write instruction, the data write instruction also includes the above-mentioned first data and a storage address of the first data. The storage address of the first data may be an address of a stripe unit for storing the first data. In other embodiments, the data carried in the data write instruction may not be the above-mentioned first data, but other data to be written and storage addresses of other data to be written.

Exemplarily, in the data structure of the data write instruction, data bits at certain fixed positions may be preset as operation identification bits, and the operation identification bits may occupy one data bit or multiple data bits. The following is an example of an operation flag occupying 2 data bits. If the operation flag is "10", it indicates that the flag is a delayed write flag, and the data write command carries a delayed write instruction; if the operation The flag bit is "01", indicating that the flag is an immediate write flag, and the data write instruction carries an immediate write instruction, which may also be called a non-delayed write instruction.

In one embodiment, in the data structure of the data write instruction, the preset N data bits may carry a target storage address corresponding to a delayed write indication or an immediate write indication, and the target storage address may be is the address of the target stripe unit for saving the first data. For example, assuming that the target disk is disk 301, in the data write command sent by the RAID controller to disk 301, the set N data bits carry the address information of the stripe unit 11, and the operation in the data write command If the flag is “10”, the data writing instruction indicates that the disk 301 can delay writing the first data corresponding to the stripe unit 11 .

In another embodiment, in the data structure of the data write instruction, the preset M data bits may carry the disk identifier and the target entry of the corresponding target disk indicating the delayed write indication or the immediate write indication. with the address of the unit. For example, in the data write instruction sent by the RAID controller to the disk 301, the set M data bits carry the disk identification of the disk 301 and the address of the stripe unit 11, and the operation identification bit in the data write instruction If it is “10”, the data writing instruction indicates that the disk 301 can delay writing the first data corresponding to the stripe unit 11 .

In some other embodiments, if it is determined that the first data is non-delayed write data, the RAID controller may generate a data write instruction including an immediate write instruction, and send the data write instruction to the target disk. Wherein, the immediate write instruction may also be referred to as a non-delayed write instruction. For example, assuming that the target disk is disk 301, in the data write command sent by the RAID controller to disk 301, the set M data bits carry the disk identification of disk 301 and the address of stripe unit 11, and the data write If the operation identification bit in the write command is “01”, the data write command indicates that the disk 301 can write the first data corresponding to the stripe unit 11 immediately. Alternatively, if it is determined that the first data is non-delayed write data, the RAID controller may directly send a data write instruction to the target disk, and the data write instruction may not carry a delayed write instruction or an immediate write instruction.

In other embodiments, if the RAID controller cannot determine whether the first data is delayed write type data or non-delayed write type data, the data write command sent by the RAID controller to the target disk may not carry delay Write instructions or write instructions now.

In some other embodiments, the disk controller may also include the data delay writing instruction in the data read instruction sent to the target disk, and send it to the target disk. In addition to the delayed write instruction, the data read instruction may also include the storage address of the data to be read.

S404, the target disk executes the delayed writing instruction carried in the data writing instruction.

The disk includes non-volatile memory, and each stripe unit is located in the non-volatile memory. In order to reduce the time delay for the RAID controller to access the disk, a cache unit is also provided in the disk, and the time for accessing the cache unit is much shorter than the time for accessing the non-volatile memory. The cache unit may be used to temporarily save data that needs to be written into the non-volatile memory, that is, first data. Exemplarily, the cache unit may use a cache memory (cache), and the non-volatile memory may use a flash memory (flash).

Taking the target disk as the disk 301 as an example, in one embodiment, the disk 301 receives the data write instruction sent by the RAID controller, and reads the N data bits set in the data write instruction to the stripe unit 11, and read that the operation identification bit in the data write command is "10", then the disk 301 can execute the delay write instruction carried in the data write command, and delay writing to the corresponding stripe unit 11 a data. Exemplarily, the disk 301 may temporarily store the first data corresponding to the stripe unit 11, that is, the data LBA0-127, in the cache unit, wait for other data to be written to be received, and then perform a data write operation. The data are written into the non-volatile memory together, and multiple data are written to the non-volatile memory each time a write operation is performed, thereby avoiding frequent write operations to the non-volatile memory. If within the set duration, no other data to be written is received, the write data operation is performed, and the first data is written into the stripe unit 11 in the non-volatile memory; or, the cache unit of the magnetic disk 301 is full, When part of the data needs to be cleared, the disk 301 executes a data write operation, and writes the first data into the stripe unit 11 in the nonvolatile memory. In some optional embodiments, the disk executes the delayed writing instruction, and after temporarily storing the first data in the cache unit, it may also wait for the first data to be refreshed, and then perform a write data operation, and write the refreshed data to into non-volatile memory.

In another embodiment, the disk 301 receives the data write instruction sent by the RAID controller, and reads the disk identifier of the disk 301 and the stripe unit 11 from the M data bits set in the data write instruction. Address information, and read the operation flag in the data write command is "10", then the disk 301 can execute the delay write instruction carried in the data write command, and delay write to the first corresponding stripe unit 11 data.

In some other embodiments, if the disk receives a data write instruction sent by the RAID controller, and the data write instruction carries a delayed write instruction, the disk can determine whether to execute the delayed write instruction according to its own state; for example, if the disk If the data storage ratio of the cache unit reaches the set maximum ratio threshold, the disk may not execute the delayed write instruction, but execute the write data operation, and write the first data together with other data stored in the cache unit into the non-volatile in memory. The data storage ratio refers to the ratio between the amount of data stored in the cache unit and the data capacity of the cache unit. If the disk receives the data write instruction sent by the RAID controller, and the data write instruction carries an immediate write instruction, the disk can determine whether to execute the immediate write instruction according to its own state; for example, if the data storage ratio of the cache unit of the disk is If the set minimum ratio threshold is not reached, the disk may not execute the immediate write instruction, and temporarily save the first data in the cache unit, and may execute writing data after the data storage ratio of the cache unit reaches the set minimum ratio threshold The operation is to write the data saved in the cache unit into the non-volatile memory together.

In some embodiments, if the first data is verification data, after sending the delayed writing instruction of the first data to the target disk, if the disk controller obtains the second data, then determine the updated verification data based on the second data. verification data, and send a data write instruction for the updated verification data to the target disk; the data write instruction is used to instruct the target disk to write the updated verification data into the non-volatile memory of the target disk. After the target disk temporarily saves the first data in the cache unit, it waits for a set time. If within the set time, the target disk receives a data write instruction of the updated verification data sent by the disk controller; the data write instruction Instruct the disk to write the updated verification data into the target storage address, the target storage address is the same as the target storage address of the first data; according to the target storage address, write the updated verification data into the non-volatile memory of the disk , and delete the first data in the cache unit. If the target disk does not receive the data write instruction of the updated verification data within the set time, the target disk will write the first data into the non-volatile memory.

FIG. 5 shows another interaction diagram between the disk controller and the disk during the process of writing data. As shown in Figure 5, in the process of writing data, the data access method provided by another embodiment may include the following steps:

S501. The disk controller determines a target disk and a data type corresponding to the first data.

S502. If it is determined that the first data is data of a delayed write type, the disk controller generates a delayed operation instruction.

For example, assuming that the first data is the above-mentioned check result P ₀ , the check result P ₀ is determined according to the data LBA0-127 carried in the write disk command received by the RAID controller from the host. Referring to the description in step S401 above, the target disk corresponding to the verification result P ₀ is disk 303, the target stripe unit is the stripe unit 13 of stripe 1, and the RAID controller can determine that the verification result P ₀ is a delayed write For input type data, the RAID controller generates a delayed operation instruction, where the delayed operation instruction refers to an individual operation instruction including a delayed write instruction.

S503, the disk controller sends a delayed operation instruction to the target disk.

In an embodiment, the disk controller may send the delay operation instruction to the target disk after sending the data write instruction of the first data to the target disk.

The delay operation instruction only includes a delay write instruction, and does not include the data to be written and the storage address of the data to be written. Exemplarily, in the delayed operation instruction, an operation identification bit and an address identification bit are included. For example, in the delayed operation instruction, the operation identification bit may be "10", indicating that the operation instruction carries a delayed write instruction. If the operation identification bit in the operation instruction is "01", it can be considered that the operation instruction is an immediate operation instruction.

In one embodiment, in the delayed operation instruction, N data bits can be preset as address identification bits, and the information in the address identification bits is used to indicate the corresponding target storage address of the delayed write instruction, that is, for saving Address information of the target stripe unit of the first data. For example, in the delay operation instruction sent by the RAID controller to the disk 303, what N data bits carry is the address information of the stripe unit 13, then the delay operation instruction indicates that the disk 303 can delay writing to the corresponding first stripe unit 13. One data, that is, the verification result P ₀ mentioned above.

In another embodiment, in the delayed operation instruction, M data bits can be preset as address identification bits, and the information in the address identification bits is used to indicate that the delayed writing indicates the disk identification and target of the corresponding target disk. Address information of the stripe unit. For example, in the delayed operation instruction sent by the RAID controller to the disk 303, what M data bits carry is the disk identification of the disk 303 and the address information of the stripe unit 13, then the delayed operation instruction indicates that the disk 303 can delay writing the stripe. Band unit 13 corresponds to the first data.

If the disk controller sends the delayed operation instruction to the target disk, the delayed operation instruction may be sent after sending the data write instruction of the first data to the target disk. The time interval between the sending time of the delayed operation instruction and the sending time of the data write instruction of the first data is less than the set time threshold, so as to ensure that the first data can be received before the target disk writes the first data into the non-volatile memory. A delayed operation instruction for the first data.

S504, the target disk executes the received delayed operation instruction.

Both the delayed operation instruction and the immediate operation instruction may be referred to as independent operation instructions.

Taking the target disk as the disk 303 as an example, in one embodiment, the disk 303 receives an individual operation instruction sent by the RAID controller, and reads that the operation identification bit in the individual operation instruction is "10", indicating that the individual operation The instruction is a delay operation instruction, and the disk 303 reads the address information of the stripe unit 13 from the N data bits set in the separate operation instruction, then the disk 303 can execute the delay operation instruction to delay writing to the stripe unit 13 corresponds to the first data.

In another embodiment, the disk 303 receives the individual operation instruction sent by the RAID controller, and the operation flag read in the individual operation instruction is "10", indicating that the individual operation instruction is a delayed operation instruction, and, The magnetic disk 303 reads the disk identification of the magnetic disk 303 and the address information of the stripe unit 13 from the M data bits set in the separate operation instruction, then the magnetic disk 303 can execute the delayed operation instruction, and the delayed write to the stripe unit 13 corresponds to the first data of .

In some other embodiments, the target disk receives the delayed operation instruction sent by the RAID controller, and may also determine whether to execute the delayed operation instruction according to its own state.

After executing the received delayed operation instruction, the target disk may choose to return the instruction execution result to the RAID controller, or may not return the instruction execution result to the RAID controller.

S505. The disk controller updates the first data, and determines a data type of the updated first data.

Optionally, when the RAID controller receives the disk write instruction from the host again, the data carried in the disk write instruction is data LBA128-255. Referring to the description in step S401 above, the RAID controller can determine the data LBA128-255 corresponding to The target stripe is also stripe 1, which is the same as the target stripe corresponding to the first data, that is, the verification result P ₀ . At this time, the RAID controller can update the verification result P ₀ according to the data LBA128-255.

In some embodiments, the update process may include: the RAID controller may acquire data LBA0-127 from the disk 301 . It has been explained in the above-mentioned embodiments that the disk 301 executes the delay write instruction issued by the RAID controller, temporarily stores the data LBA0-127 in the cache unit, and receives the read instruction of the RAID controller for the data LBA0-127, the disk 301 The 301 can directly obtain the data LBA0~127 from the cache unit, and transfer the data LBA0~127 to the RAID controller; without reading data from the non-volatile memory, it can greatly save the time for reading data and reduce the RAID level. Latency for the controller to access the disk. After the RAID controller acquires the data LBA0-127, the verification result P calculated according to the data LBA0-127 and the data LBA128-255 is used as the updated first data.

In some other embodiments, the update process may include: the RAID controller may acquire the verification result P ₀ from the disk 303 . It can be seen from the above that the disk 303 executes the delayed write instruction issued by the RAID controller, temporarily stores the verification result P ₀ in the cache unit, and receives the read instruction of the RAID controller for the verification result P ₀ , the disk 303 The verification result P ₀ can be obtained directly from the cache unit, and the verification result P ₀ can be transmitted to the RAID controller; without reading data from the non-volatile memory, it can greatly save the time for reading data and reduce the Latency for the RAID controller to access disks. After the RAID controller obtains the verification result P ₀ , it determines the verification result P according to the verification result P ₀ and data LBA128-255; for example, the RAID controller can reversely calculate the data LBA0-127 according to the _verification result P The verification result P obtained by calculating the data LBA0-127 and the data LBA128-255 uses the verification result P as the updated first data.

After the updated first data is determined, the RAID controller can determine that the verification result P is data of the non-delayed write type, since the data corresponding to each stripe unit of the target stripe, that is, stripe 1, has been determined.

S506, the disk controller sends a data write instruction including the updated first data to the target disk.

For example, for the verification result P, the corresponding target disk is the disk 303 . The RAID controller sends a data write instruction to the disk 303, the data write instruction includes the verification result P, and the address information of the target stripe unit used to save the verification result P; wherein, the target stripe unit is a stripe unit 13.

S507. If it is determined that the updated first data is non-delayed write data, the disk controller generates an immediate operation instruction.

S508, the disk controller sends an immediate operation instruction to the target disk.

Since the verification result P is non-delayed write data, the RAID controller sends an immediate operation command to the disk 303 . Exemplarily, the immediate operation instruction includes an operation identification bit and an address identification bit. The operation identification bit may be "01", indicating that the operation instruction carries an immediate write instruction.

In one embodiment, in the immediate operation instruction, N data bits may be preset as address identification bits, and the information in the address identification bits is used to indicate that the address information indicating the corresponding target stripe unit is immediately written or The address information of the target storage address. For example, in the immediate operation instruction sent by the RAID controller to the disk 303, what N data bits carry is the address information of the stripe unit 13, and then the immediate operation instruction indicates that the disk 303 can immediately write into the corresponding first stripe unit 13. One data, that is, updated first data: verification result P.

In another embodiment, in the immediate operation instruction, M data bits can be preset as address identification bits, and the information in the address identification bits is used to indicate that the disk identification and target of the corresponding target disk are written immediately. The address information of the stripe unit, or the disk identification of the target disk and the address information of the target storage address. For example, in the immediate operation instruction sent by the RAID controller to the disk 303, what M data bits carry is the disk identification of the disk 303 and the address information of the stripe unit 13, and then the immediate operation instruction indicates that the disk 303 can write the stripe immediately. The data to be written corresponding to the tape unit 13, that is, the verification result P.

In some other embodiments, after the disk controller determines that the updated first data is non-delayed write data, it may not send an immediate operation instruction to the target disk, so as to save signaling overhead.

S509, the target disk executes the received immediate operation instruction.

Taking the target disk as the disk 303 as an example, in one embodiment, the disk 303 receives an individual operation instruction sent by the RAID controller, and reads that the operation identification bit in the individual operation instruction is "01", indicating that the individual operation The instruction is an immediate operation instruction, and the disk 303 reads the address information of the stripe unit 13 from the N data bits set in the separate operation instruction, then the disk 303 can execute the immediate operation instruction, and the stripe unit 13 corresponds to The data to be written, that is, the verification result P is written into the stripe unit 13 of the non-volatile memory.

In another embodiment, the disk 303 receives the individual operation instruction sent by the RAID controller, and the operation identification bit read in the individual operation instruction is "01", indicating that the individual operation instruction is an immediate operation instruction, and, The magnetic disk 303 reads the disk identification of the magnetic disk 303 and the address information of the stripe unit 13 from the M data bits set in the separate operation instruction, and then the magnetic disk 303 can execute the immediate operation instruction, and the corresponding standby of the stripe unit 13 Writing data, that is, the verification result P is written into the stripe unit 13 of the non-volatile memory.

In some other embodiments, the disk receives the immediate operation instruction sent by the RAID controller, and may also determine whether to execute the immediate operation instruction according to its own state.

S510, the target disk sends an instruction execution result to the disk controller.

After the RAID controller sends an immediate operation command to the target disk, it waits for the target disk to feed back an execution result of the command. The target disk determines that the received immediate operation command is an individual operation command, and after executing the immediate operation command sent by the RAID controller, sends the command execution result to the RAID controller, and the command execution result can include information that the command has been executed or the command has been successfully executed . After the RAID controller receives the command execution result fed back by the target disk, it can end the data writing process.

In some other embodiments, the target disk may not send the command execution result to the disk controller.

The above description of the data access method shown in Figure 5 focuses on the differences from the data access method shown in Figure 4, and the same place as the data access method shown in Figure 4 can be referred to in Figure 4 Execution of the data access method shown, no more details.

In the embodiment of the present application, the disk controller that manages the disk determines the data type of the data to be written. If the disk controller determines that the data to be written is data that needs to be used later or data that is updated immediately, it can send a delay to the disk. The write instruction indicates that the disk can delay writing data into the non-volatile memory, thereby avoiding frequent disk write operations, helping to reduce the occupation of memory bandwidth resources, and can also reduce power consumption. If the data to be written is data that will be updated immediately, by delaying the writing of data, the operation of repeatedly writing data to the same storage address can be reduced, and the service life of the disk can be improved. Compared with determining whether to delay data writing to the non-volatile memory by the disk, since the disk controller has a better perception capability for upper-layer services, it may be more accurate to perform this step by the disk controller.

For example, assuming that the disk receives 64KB of data to be written and executes a disk write operation once, if the host needs to write 1TB of data into the three disks of the RAID system in a sequential manner, using the method provided in the embodiment of this application, at most It can reduce the repeated writing of 0.5TB data. It can be seen that through the method provided by the embodiment of the present application, the data writing operation of the disk can be greatly reduced, the power consumption can be reduced while the performance is improved, and the service life of the disk can be improved.

Fig. 6 shows an interaction diagram between the disk controller and the disk during the process of prefetching data. As shown in FIG. 6, in the process of prefetching data, a data access method provided by an embodiment may include the following steps:

S601. The disk controller determines a target storage address and a target disk of prefetched data.

In some embodiments, the RAID controller can determine whether the current data reading mode is a sequential read according to the logical addresses in the multiple read disk instructions received continuously, wherein the sequential read refers to multiple read disks received continuously The logical addresses in the instructions are continuous, that is to say, multiple disk read instructions indicate to read data from each stripe unit in a virtual disk sequentially. For example, when a user opens a file, the host sends multiple read disk commands to the RAID controller continuously, and the logical addresses in the multiple read disk commands are continuous. At this time, the current data read mode can be considered as sequential Read means that the data of the document can be sequentially read from each strip unit indicated by the logical address in each disk read command, and sent to the host for display. If the logical addresses in multiple read disk commands are discontinuous, it can be determined that the current data read mode is not sequential read; for example, when the user operates multiple different functions at the same time, the host needs to read the parameters of each function, The parameters of each function may be stored in different virtual disks, so it is necessary to read data from different virtual disks or non-adjacent stripe units in the same virtual disk. At this time, the host sends multiple read disks to the RAID controller command, and the logical addresses in multiple disk read commands are discontinuous, at this time, it can be considered that the current data reading mode is not sequential reading.

In an optional embodiment, if the RAID controller determines that the current data reading mode is sequential reading, then according to the logical address in the currently received disk read command, determine the target storage corresponding to the subsequent data that needs to be read. The address and the target disk are the target storage address and the target disk of the prefetched data, and the target storage address can also be understood as the address of the target stripe unit. In another optional embodiment, if the RAID controller determines that the current data reading mode is sequential reading, then determine the disk corresponding to the currently read data according to the logical address in the currently received disk read instruction; The disk corresponding to the currently read data is used as the target disk, and the target storage address of the data that still needs to be read in the future is determined for the target disk, that is, the target storage address of the prefetched data. For example, assume that the currently read data is data LBA0-127, and the data LBA0-127 are stored in the stripe unit 11 of the magnetic disk 301. The disk 301 is the target disk. Since the current read is sequential, the RAID controller can determine that the data to be read later is the data stored in the stripe unit 21, and can determine that the stripe unit 21 is the target stripe unit.

In some other embodiments, the RAID controller starts the background service, and determines the target storage address and the target disk of the data to be prefetched according to the currently executed service. The RAID controller can start the service of repairing the disk according to the operation of repairing the disk by the user, and can also automatically start the service of repairing the disk when it detects that a certain disk fails. Exemplarily, it is assumed that the RAID controller detects that the disk 302 fails, and starts the business of repairing the disk 302, that is, according to the data stored in each stripe unit of the disk 301 and the disk 303, determine each stripe unit of the disk 302 correct data stored in the disk 302, and write the correct data into each stripe unit of the disk 302. For example, for stripe 1, the RAID controller can determine the stripe unit 12 of the disk 302 according to the data LBA0-127 stored in the stripe unit 11 of the disk 301 and the verification result P stored in the stripe unit 13 of the disk 303. correct data stored in . At this time, the disk to be repaired is disk 302, the stripe unit to be repaired is stripe unit 12, the target stripe to which stripe unit 12 belongs is stripe 1, and stripe 1 includes stripe unit 11 and stripe unit 13, The disk 301 may be used as the target disk, the address of the stripe unit 11 may be used as the target storage address, and the disk 303 may be used as the target disk, and the address of the stripe unit 13 may be used as the target storage address. The process of executing the data prefetch instruction when the disk 303 is used as the target disk is the same as the process of executing the data prefetch instruction when the disk 301 is used as the target disk.

S602. The disk controller generates a data access instruction including a prefetch data instruction according to the address information of the target storage address.

S603, the disk controller sends a data access instruction to the target disk.

The data access command may be a data read command or a data write command.

In some embodiments, the RAID controller receives a disk read command from the host, and sends a data access command to the target disk, or, the RAID controller can actively send a data read command to the target disk according to a currently executed service. The data read instruction may include address information of the currently read data, and may also include a prefetch data instruction. Exemplarily, in the data structure of the data read instruction, data bits at certain fixed positions may be preset as operation identification bits, and the operation identification bits may occupy one data bit or multiple data bits. In the following, the operation identification bit occupies 2 data bits as an example for illustration. If the operation identification bit is "00", it indicates that the data read instruction carries a prefetch data instruction.

In an embodiment, in the data structure of the data read instruction, pre-set N data bits may also carry address information indicating a corresponding target storage address of the prefetched data. For example, assuming that the target disk is disk 301, in the data read instruction sent by the RAID controller to disk 301, the set N data bits carry the address of the stripe unit 11, and the operation identifier in the data read instruction If the bit is “00”, the data read instruction instructs the disk 301 to prefetch the data stored in the stripe unit 11 .

In another embodiment, in the data structure of the data read instruction, pre-set M data bits may carry address information indicating the disk ID of the corresponding target disk and the target storage address of the prefetched data. For example, in the data read instruction sent by the RAID controller to the disk 301, the set M data bits carry the disk identification of the disk 301 and the address of the stripe unit 11, and the operation identification bit in the data read instruction is “00”, the data read instruction instructs the disk 301 to prefetch the data stored in the stripe unit 11 .

In some other embodiments, the RAID controller receives a disk write command from the host and sends the data write command to the target disk, or the RAID controller can actively send the data write command to the target disk according to the currently executed service. The data writing instruction includes the data to be written and the address information of the stripe unit to be written for saving the data to be written. The data write instruction may also include a prefetch data instruction. Exemplarily, in the data structure of the data write instruction, data bits at certain fixed positions may be preset as operation identification bits, and the operation identification bits may occupy one data bit or multiple data bits. For example, the operation identification bit occupies 2 data bits, and if the operation identification bit is "00", it indicates that the data write instruction carries a prefetch data instruction.

Same as the data read instruction, in an embodiment, in the data structure of the data write instruction, pre-set N data bits may also carry address information indicating the corresponding target storage address of the prefetched data. In another embodiment, in the data structure of the data write instruction, pre-set M data bits may carry address information indicating the disk ID of the corresponding target disk and the target storage address of the prefetched data.

S604, the target disk executes the prefetch data instruction carried in the data access instruction.

Taking the target disk as the disk 301 as an example, in one embodiment, the disk 301 receives the data access instruction sent by the RAID controller, and reads the N data bits set in the data access instruction to the stripe unit 11 address information, and read that the operation identification bit in the data access instruction is "00", then the disk 301 can determine that the data access instruction carries a prefetch data indication, indicating that the data stored in the stripe unit 11 Data is prefetched.

In another embodiment, the disk 301 receives the data access instruction sent by the RAID controller, and reads the disk identifier of the disk 301 and the stripe unit 11 from the M data bits set in the data access instruction. address information, and read the operation flag in the data access instruction as "00", then the disk 301 can determine that the data access instruction carries a prefetch data instruction, indicating that the data stored in the stripe unit 11 is to be prefetched. Prefetching.

When executing the prefetch data instruction, the disk 301 can first judge whether the data stored in the stripe unit 11 is stored in the cache unit, and if it has been stored in the cache unit, the operation can be ended; if not stored in the cache unit, The data stored in the stripe unit 11 is then read from the non-volatile memory and stored in the cache unit. When the subsequent disk controller reads the data, the disk 301 can directly obtain the data from the cache unit without reading the data from the non-volatile memory, so the read delay of the disk can be reduced and the data can be quickly returned to the disk controller. For example, assuming that the disk reads data from the non-volatile memory and returns it to the disk controller, it takes 60us each time; the disk reads data from the cache unit and returns it to the disk controller, and it takes 10us each time. Then every time data is read, the read delay of 50us can be reduced.

FIG. 7 shows another interaction diagram between the disk controller and the disk in the process of prefetching data. As shown in FIG. 7, in the process of prefetching data, another embodiment provides a data access method that may include the following steps:

S701. The disk controller determines a target storage address and a target disk of data to be prefetched.

S702. The disk controller generates a prefetch data instruction according to the address information of the target storage address.

S703, the disk controller sends a data prefetch instruction to the target disk.

Only the prefetch data indication is included in the prefetch data instruction. Exemplarily, in the data prefetching instruction, an operation identification bit and an address identification bit are included. For example, in the prefetch data instruction, the operation identification bit may be "00", indicating that the operation instruction carries a prefetch data instruction.

In one embodiment, in the prefetch data instruction, N data bits can be preset as address identification bits, and the information in the address identification bits is used to indicate that the prefetch data indicates the address information of the corresponding target stripe unit . The prefetch data instruction is one of the independent operation instructions. For example, in the individual operation instruction sent by the RAID controller to the disk 301, the N data bits carry the address information of the stripe unit 11, and the operation identification bit in the individual operation instruction is "00", then the individual operation The command is a data prefetch command, which is used to instruct the disk 301 to prefetch the data stored in the stripe unit 11 .

In another embodiment, in the prefetch data instruction, M data bits can be preset as address identification bits, and the information in the address identification bits is used to indicate that the prefetch data indicates the disk identification and the corresponding target disk. Address information of the target stripe unit. For example, in the individual operation instruction sent by the RAID controller to the disk 301, the M data bits carry the disk identification of the disk 301 and the address information of the stripe unit 11, and the operation identification bit in the individual operation instruction is "00 ”, then the single operation command is a data prefetch command, which is used to instruct the disk 301 to prefetch the data stored in the stripe unit 11 .

S704, the target disk executes the received prefetch data instruction.

Taking the target disk as the disk 301 as an example, in one embodiment, the disk 301 receives an individual operation instruction sent by the RAID controller, and reads that the operation identification bit in the individual operation instruction is "00", indicating that the individual operation The instruction is a prefetch data instruction, and the magnetic disk 301 reads the address information of the stripe unit 11 from the N data bits set in the separate operation instruction, then the magnetic disk 301 can execute the prefetch data instruction, and the stripe unit The data saved in 11 is prefetched.

In another embodiment, the disk 301 receives an individual operation instruction sent by the RAID controller, and the operation identification bit read in the individual operation instruction is "00", indicating that the individual operation instruction is a data prefetch instruction, and , the magnetic disk 301 reads the disk identification of the magnetic disk 301 and the address information of the stripe unit 11 from the M data bits set in the separate operation instruction, and then the magnetic disk 301 can execute the prefetch data instruction, and the data in the stripe unit 11 Saved data is prefetched.

S705, the target disk sends an instruction execution result to the disk controller.

After the RAID controller sends the data prefetch command to the target disk, it waits for the target disk to feed back an execution result of the command. The target disk determines that the received prefetch data command is an independent operation command. After executing the prefetch data command sent by the RAID controller, it sends the command execution result to the RAID controller. The command execution result can include that the command has been executed or the command has been successfully executed. Information. After the RAID controller receives the instruction execution result fed back by the target disk, it can end the process of prefetching data.

The above description of the data access method shown in Figure 7 focuses on the differences from the data access method shown in Figure 6, and the same place as the data access method shown in Figure 6 can be referred to in Figure 6 Execution of the data access method shown, no more details.

In the embodiment of the present application, the disk controller that manages the disk determines the data that needs to be prefetched, sends a prefetch data indication to the disk, and instructs the disk to read the data that needs to be prefetched into the cache unit, so as to reduce the number of read data. The read latency of the disk can be shortened to improve system performance. Compared with the data that needs to be prefetched determined by the disk, because the disk controller has a better perception of the upper-layer business, the data that needs to be prefetched by the disk controller can be determined more accurately and reduce the misjudgment caused by prefetch. Power consumption generated by additional data movement, thus reducing system power consumption.

For example, suppose it takes 60us for the disk to read 64KB of data from the non-volatile memory. If the 1TB capacity RAID system needs to be expanded from 6 disks (stripe unit size is 64KB) to 7 disks, each 1280KB For example, if the read operation is reduced by 60us, the time to complete 1TB disk expansion can be saved by about 50s. It can be seen that the method provided by the embodiment of the present application can effectively save time in the disk expansion business scenario.

The data access method provided in the embodiment of the present application can not only be applied to RAID controllers, but also can be used by other nodes that can identify data that needs to be used later, including but not limited to hosts, smart network cards, computers, etc.

In the embodiment of the present application, during the process of writing data, the method performed by the disk controller can refer to the method shown in Figure 8, including the following steps:

S801. Acquire first data to be written into a target disk.

S802. When it is determined that the first data satisfies the set condition, send the first data and a delayed writing indication of the first data to the target disk.

Wherein, the delayed writing indication is used to instruct the target disk to temporarily store the first data in the cache unit of the target disk. The first data meets the following setting conditions: the first data is the data to be written to the first stripe unit of the target disk, and the data to be written after the first data includes the second stripe unit to be written to the second stripe unit. Data; the second stripe unit and the first stripe unit belong to the same target stripe, and the second stripe unit and the first stripe unit belong to different disks.

In a case where it is determined that the first data satisfies the above-mentioned setting condition, the disk controller sends an indication of delayed writing of the first data to the target disk.

In an optional embodiment, the setting condition includes that the first data is data to be read back; whether the first data satisfies the setting condition is determined by the following method: if the first data is the data to be written into the target disk The data of the first stripe unit, and it is predicted that the data to be written after the first data includes the second data to be written into the second stripe unit, then determine that the first data is the data to be read back; the second The strip unit and the first strip unit belong to the same target strip, and the second strip unit and the first strip unit belong to different disks; or, the setting conditions include that the first data is data to be updated; the first Whether the data meets the set condition: if the first data is the verification data of the third stripe unit to be written to the target disk, and it is predicted that the data to be written after the first data includes the second stripe to be written If the second data of the unit is the second data, it is determined that the first data is the data to be updated; the second stripe unit and the third stripe unit belong to the same target stripe, and the second stripe unit and the third stripe unit belong to different disks.

In an optional embodiment, the first data may be the data carried in the received write disk instruction. The disk controller can determine whether the data write mode is sequential write according to the logical addresses in multiple write disk commands received continuously. If the data write mode is sequential write, and the target stripe includes If there is no free stripe unit outside, it is determined that the data to be written includes the second data to be written into the target stripe.

In some embodiments, after the disk controller obtains the first data to be written to the target disk, it sends a data write command containing the first data to the target disk, and the disk controller may carry the delayed write instruction in the data write command. command, sent to the target disk. In some other embodiments, the disk controller may send a first operation instruction including a delayed write instruction to the target disk, and the first operation instruction may be any instruction other than the data write instruction of the first data. The delayed writing indication includes a delayed writing identifier and a target storage address of the first data. If the disk controller sends the delayed write instruction to the target disk in an instruction other than the data write instruction of the first data, the instruction containing the delayed write instruction can be sent to the target disk in the data write instruction of the first data. After sending the command, the time interval between the sending time and the sending time of the data write command of the first data is less than the set time threshold, so as to ensure that before the target disk writes the first data into the non-volatile memory, the A delayed write indication for first data is received.

In some embodiments, if the first data is verification data, after sending the delayed writing instruction of the first data to the target disk, if the disk controller obtains the second data, then determine the updated verification data based on the second data. The test data, that is, the update data, sends a data write instruction for the update data to the target disk; the data write instruction for update data is used to instruct the target disk to write the update data into the non-volatile memory of the target disk.

In the embodiment of the present application, during the process of writing data, the method performed by the disk can refer to the method shown in Figure 9, including the following steps:

S901. Receive first data sent by a disk controller, and receive a delayed writing indication of the first data sent by the disk controller.

Wherein, the first data satisfies the following setting conditions: the first data is the data to be written to the first stripe unit of the target disk, and the data to be written after the first data includes the data to be written to the second stripe unit. The second data: the second stripe unit and the first stripe unit belong to the same target stripe, and the second stripe unit and the first stripe unit belong to different disks.

In a possible embodiment, the disk may receive a data write instruction of the first data sent by the disk controller, and obtain the first data and the delayed write indication of the first data from the data write instruction; the data write The command contains the data to be written to the disk. In another possible embodiment, after receiving the data write command of the first data, the disk may receive the data write command of other data sent by the disk controller again, and obtain the first data write command carried in the data write command. A data delay write instruction; the data write instruction includes data to be written to the disk. In another possible embodiment, a separate operation instruction including a delayed write instruction sent by the disk controller is received.

S902. Temporarily store the first data in the cache unit of the disk.

The delayed writing indication includes a delayed writing identifier and a target storage address of the first data. The magnetic disk executes the delayed writing instruction, temporarily stores the first data in the cache unit of the magnetic disk according to the target storage address of the first data, and delays writing into the non-volatile memory of the magnetic disk.

In some embodiments, after the first data is temporarily stored in the cache unit of the disk, if the disk receives a data write instruction for updating data sent by the disk controller; the updating data is updated check data; the data writing of updating data The input command instructs the disk to write the update data into the target storage address, which is the same as the target storage address of the verification data; according to the target storage address, the update data is written into the non-volatile memory of the disk, and stored in the cache unit Delete the verification data.

In the embodiment of the present application, in the process of prefetching data, the method performed by the disk controller can refer to the method shown in Figure 10, including the following steps:

S1001. Determine a target storage address and a target disk of prefetched data.

In some embodiments, if the disk controller determines that the data read mode is sequential reading according to the logical addresses in multiple disk read instructions received continuously, then it can be determined according to the logical addresses in the currently received disk read instructions The target storage address and target disk of the prefetched data.

In some other embodiments, during the process of repairing the disk to be repaired, the target storage address and the target disk of the prefetched data may be determined according to the target stripe to which the stripe unit to be repaired belongs, wherein the stripe to be repaired The unit is a stripe unit in the disk to be repaired, and the target stripe includes multiple stripe units located on different disks.

S1002. Send the prefetch data instruction generated based on the target storage address to the target disk.

Wherein, the target disk includes a cache unit, and the prefetch data indication is used to instruct the target disk to read the prefetch data from the target storage address into the cache unit.

In some embodiments, the disk controller may carry the prefetch data indication in the data read instruction sent to the target disk, and send it to the target disk; the data read instruction includes the storage address of the data to be read. In some other embodiments, the disk controller may send a separate operation command including an instruction to prefetch data to the target disk.

S1003. Send a first data read instruction for the prefetched data to the target disk.

S1004. Receive the prefetched data acquired and sent by the target disk from the cache unit.

In the embodiment of this application, during the process of prefetching data, the method performed by the disk can refer to the method shown in Figure 11, including the following steps:

S1101. Receive a prefetch data instruction sent by a disk controller.

Wherein, the prefetch data indication carries a target storage address of the prefetch data.

In some embodiments, the disk may receive the second data read instruction sent by the disk controller, and acquire the prefetch data instruction carried in the data read instruction; the second data read instruction includes the storage address of the data to be read. In some other embodiments, the disk may receive a second operation instruction including a prefetch data instruction sent by the disk controller, and the second operation instruction is an instruction other than the second data read instruction.

S1102. Read the prefetched data stored in the storage space corresponding to the target storage address into the cache unit of the disk.

S1103. Receive a first data read instruction for the prefetched data sent by the disk controller.

S1104. Acquire the prefetched data from the cache unit, and send the prefetched data to the disk controller.

Based on the same inventive concept as the above-mentioned method embodiment, the embodiment of the present application also provides a data access device, which can be applied to a disk controller, and the disk controller can be the above-mentioned RAID controller, host, equipment such as computers or servers. The data access device can be used to implement the functions of the data access method embodiments executed by the above-mentioned disk controller, and thus can realize the beneficial effects of the above-mentioned data access method embodiments.

In one embodiment, as shown in FIG. 12 , the data access device 1200 provided in the embodiment of the present application may include a data obtaining unit 1201 and a sending unit 1202 . The data access device is used to implement the functions in the method embodiment shown in FIG. 8 above. When the data access device is used to implement the functions of the method embodiment shown in FIG. 8: the data acquiring unit 1201 may be used to perform S801, and the sending unit 1202 may be used to perform S802. For example: the data obtaining unit 1201 is used to obtain the first data to be written into the target disk; the sending unit 1202 is used to send the first data to the target disk when it is determined that the first data meets the set conditions, and send the first data to the target disk. The disk sends a delayed write indication of the first data, wherein the setting conditions include: the first data is the data to be written to the first stripe unit of the target disk, and the data to be written after the first data includes the data to be written Enter the second data of the second stripe unit; the second stripe unit and the first stripe unit belong to the same target stripe, and the second stripe unit and the first stripe unit belong to different disks.

In a possible implementation manner, the first data is the data carried in the received write disk instruction; the data acquisition unit 1201 can be specifically configured to: determine the Whether the data writing mode is sequential writing, if the data writing mode is sequential writing, and the target stripe includes free stripe units other than the target stripe unit, then it is determined that the data to be written includes the target stripe unit to be written with the second data.

In a possible implementation manner, the sending unit 1202 may be specifically configured to: carry the delayed write indication in the data write instruction of the first data sent to the target disk, and send it to the target disk; or, send the instruction to the target disk Sending a first operation instruction including a delayed write instruction, where the first operation instruction is an instruction other than the first data write instruction. Wherein, the delayed writing indication includes a delayed writing identifier and a target storage address of the first data.

In another embodiment, as shown in FIG. 13 , the data access device 1300 provided in the embodiment of the present application may include a prediction unit 1301 , an indication sending unit 1302 and a data receiving unit 1303 . The data access device is used to implement the functions in the method embodiment shown in FIG. 9 above. When the data access device is used to realize the function of the method embodiment shown in FIG. 10: the predicting unit 1301 can be used to perform S1001, the instruction sending unit 1302 can be used to perform S1002 and S1003, and the data receiving unit 1303 can be used to perform S1004 . For example: the prediction unit 1301 is used to determine the target storage address and the target disk of the prefetched data; the target disk includes a cache unit; the instruction sending unit 1302 is used to send the prefetch data indication generated based on the target storage address to the target disk; prefetch The data indication is used to instruct the target disk to read the prefetch data from the target storage address into the cache unit; and, send the first data read instruction for the prefetch data to the target disk; the data receiving unit 1303 is used to receive the target disk Prefetched data fetched and sent from the cache unit.

In a possible implementation manner, the predicting unit 1301 can be specifically configured to: determine whether the data reading mode is sequential reading according to the logical addresses in multiple disk read instructions received continuously, and if the data reading mode is sequential For reading, determine the target storage address and target disk of the prefetched data according to the logical address in the currently received disk read instruction.

In a possible implementation manner, the predicting unit 1301 may be specifically configured to: during the process of repairing the disk to be repaired, determine the target storage address and The target disk; the stripe unit to be repaired is a stripe unit in the disk to be repaired; the target stripe includes multiple stripe units located on different disks.

In a possible implementation manner, the instruction sending unit 1302 can be specifically configured to: carry the prefetch data instruction in the second data read instruction sent to the target disk, and send it to the target disk; the second data read instruction It is used to instruct the target disk to read data from a specified storage address; or, to send a second operation instruction including a prefetch data instruction to the target disk; the second operation instruction is an instruction other than the second data read instruction.

Based on the same inventive concept as the above method embodiment, the embodiment of the present application also provides a data access device, which can be applied to a magnetic disk. The data access device can be used to implement the functions of the embodiments of the data access method performed by the above-mentioned disk, and thus can realize the beneficial effects of the embodiments of the above-mentioned data access method.

In one embodiment, as shown in FIG. 14 , the data access device 1400 provided in the embodiment of the present application may include a receiving unit 1401 and an executing unit 1402 . The data access device is used to implement the functions in the method embodiment shown in FIG. 9 above. When the data access device is used to implement the functions of the method embodiment shown in FIG. 9: the receiving unit 1401 may be used to perform S901, and the execution unit 1402 may be used to perform S902. For example: the receiving unit 1401 is configured to receive the first data sent by the disk controller, and receive the delayed writing indication of the first data sent by the disk controller; the delayed writing indication is used to instruct the target disk to temporarily store the first data in the target In the cache unit of the disk; the execution unit 1402 is configured to temporarily store the first data in the cache unit of the disk, wherein the first data satisfies the following condition: the first data is the data to be written into the first stripe unit of the target disk , the disk controller predicts that after the first data, the to-be-written data written to the disk includes the second data to be written to the second stripe unit, wherein the second stripe unit and the first stripe unit belong to For the same target stripe, the second stripe unit and the first stripe unit belong to different disks.

In a possible implementation manner, the receiving unit 1401 may be specifically configured to: receive a data write instruction of the first data sent by the disk controller, and acquire a delayed write instruction carried in the data write instruction of the first data; The data writing instruction of the first data includes the first data; or, receiving a first operation instruction including a delayed writing instruction sent by the disk controller, the first operation instruction is an instruction other than the data writing instruction of the first data. Wherein, the delayed writing indication includes a delayed writing identifier and a target storage address of the first data.

In another embodiment, as shown in FIG. 15 , the data access device 1500 provided in the embodiment of the present application may include an instruction receiving unit 1501 , an instruction executing unit 1502 and a data transmission unit 1503 . The data access device is used to implement the functions in the method embodiment shown in FIG. 11 above. When the data access device is used to realize the function of the method embodiment shown in FIG. 11: the instruction receiving unit 1501 can be used to execute S1101, the instruction execution unit 1502 can be used to execute S1102, and the data transmission unit 1503 can be used to execute S1103 and S1104. For example: the instruction receiving unit 1501 is used to receive the prefetch data instruction sent by the disk controller; the prefetch data instruction is generated by the disk controller based on the target storage address of the prefetch data; the instruction execution unit 1502 is used to transfer the prefetch data Indicate that the prefetch data stored in the target storage address carried is read into the cache unit of the disk; the data transmission unit 1503 is configured to receive the first data read instruction for the prefetch data sent by the disk controller, and read the instruction from the cache unit Get the prefetched data and send the prefetched data to the disk controller.

In a possible implementation manner, the instruction receiving unit 1501 may be specifically configured to: receive the second data read instruction sent by the disk controller, obtain the prefetch data indication carried in the data read instruction; The command is used to instruct the target disk to read data from a specified storage address; or, to receive a separate operation command sent by the disk controller including the prefetch data instruction.

Based on the same inventive concept as the above-mentioned method embodiments, the embodiments of the present application also provide a disk controller, which can be used to realize the functions of the above-mentioned method embodiments, so that the beneficial features of the above-mentioned method embodiments can be realized. Effect. The disk controller may be the above-mentioned RAID controller, a host, a computer, or a server.

In some embodiments, the structure of the disk controller 1600 may be as shown in FIG. 16 , including a processor 1601 and a memory 1602 connected to the processor 1601 .

Optionally, the processor 1601 and the memory 1602 may be connected to each other through a bus, and the processor 1601 may be a general processor, such as a microprocessor, or other conventional processors. The bus may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus or the like. The bus can be divided into address bus, data bus, control bus and so on.

The memory 1602 is used to store instructions or programs executed by the processor 1601, or store input data required by the processor 1601 to run the instructions or programs, or store data generated after the processor 1601 runs the instructions or programs. The processor 1601 may include one or more processing units, and different processing units may be independent devices or integrated into one or more processors. The processor 1601 may further include a controller, and the controller may generate an operation control signal according to the instruction operation code and the timing signal, and complete the control of fetching and executing instructions.

Exemplarily, the processor 1601 in the disk controller 1600 is configured to execute instructions or programs stored in the memory 1602 to execute the functions in the method embodiment shown in FIG. 8 or FIG. 10 . For example, when the disk controller 1600 is used to implement the method shown in FIG. 8 , the processor 1601 is used to execute the functions of the above data acquisition unit 1201 and the sending unit 1202 . When the disk controller 1600 is used to implement the method shown in FIG. 10 , the processor 1601 is used to execute the functions of the prediction unit 1301 , the indication sending unit 1302 and the data receiving unit 1303 .

Based on the same inventive concept as the above-mentioned method embodiments, the embodiments of the present application further provide a disk, which can be used to implement the functions of the above-mentioned method embodiments, and thus can realize the beneficial effects of the above-mentioned method embodiments. The disk can be a solid state drive or a mechanical hard drive.

In some embodiments, the structure of the disk 1700 may be as shown in FIG. 17 , including a processor 1710 and a memory 1720 connected to the processor 1710 , where the memory 1720 may include a cache unit 1721 and a non-volatile memory 1722 .

Optionally, the processor 1710 and the memory 1720 may be connected to each other through a bus, and the processor 1710 may be a general processor, such as a microprocessor, or other conventional processors. The bus can be a PCI bus or an EISA bus or the like.

The cache unit 1721 in the memory 1720 may be used to temporarily store data to be written into the non-volatile memory 1722 or data to be read from the non-volatile memory 1722 . The non-volatile memory 1722 can be used to provide data storage services and store data used in various business scenarios.

The memory 1720 may also be used to store instructions or programs executed by the processor 1710, or store input data required by the processor 1710 to run the instructions or programs, or store data generated after the processor 1710 runs the instructions or programs. The instructions or programs executed by the processor 1710 can be stored in the non-volatile memory 1722, the processor 1710 can include one or more processing units, different processing units can be independent devices, and can also be integrated in one or more processing units. device.

In one embodiment, the processor 1710 in the disk 1700 can run instructions or programs stored in the memory 1720 to execute the functions in the method embodiment shown in FIG. 9 or FIG. 11 . For example, when the disk 1700 is used to implement the method shown in FIG. 9 , the processor 1701 is used to execute the functions of the above-mentioned receiving unit 1401 and executing unit 1402 . When the disk 1700 is used to implement the method shown in FIG. 11 , the processor 1701 is used to execute the functions of the above instruction receiving unit 1501 , instruction executing unit 1502 and data transmitting unit 1503 .

It can be understood that, the structure illustrated in the embodiment of the present application does not constitute a specific limitation on the computing device. In other embodiments of the present application, the computing device may include more or fewer components than shown, or combine some components, or separate some components, or arrange different components. The illustrated components can be realized in hardware, software or a combination of software and hardware.

The method steps in the embodiments of the present application may be implemented by means of hardware, or may be implemented by means of a processor executing computer programs or instructions. A computer program or instructions may constitute a computer program product. The embodiment of the present application also provides a computer program product including computer executable instructions. In one embodiment, the computer-executable instructions are used to cause a computer to execute the functions in the method embodiment shown in FIG. 8 or FIG. 10 . In another embodiment, the computer-executable instructions are used to cause the computer to execute the functions in the method embodiment shown in FIG. 9 or FIG. 11 .

Computer-executable instructions may be stored in a computer-readable storage medium, and the embodiment of the present application further provides a computer-readable storage medium, and the computer-readable storage medium stores executable instructions. In one embodiment, the computer-executable instructions are used to cause a computer to execute the functions in the method embodiment shown in FIG. 8 or FIG. 10 . In another embodiment, the computer-executable instructions are used to cause the computer to execute the functions in the method embodiment shown in FIG. 9 or FIG. 11 .

The computer-readable storage medium provided by the embodiment of the present application may be random access memory (random access memory, RAM), flash memory, read-only memory (read-only memory, ROM), programmable read-only memory (programmable ROM, PROM), Erasable programmable read-only memory (erasable PROM, EPROM), electrically erasable programmable read-only memory (electrically ePROM, EEPROM), registers, hard disk, removable hard disk, CD-ROM or any other form known in the art computer readable storage medium.

Computer-executable instructions may be stored in or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, the computer programs or instructions may be transmitted from a website, computer, server or A data center transmits to another website site, computer, server, or data center via wired or wireless means. The computer-readable storage medium may be any available medium that can be accessed by a computer, or a data storage device such as a server or a data center integrating one or more available media. The available medium may be a magnetic medium, such as a floppy disk, a hard disk, or a magnetic tape; it may also be an optical medium, such as a digital video disc (digital video disc, DVD); it may also be a semiconductor medium, such as a solid-state hard disk.

In each embodiment of the present application, if there is no special explanation and logical conflict, the terms and/or descriptions between different embodiments are consistent and can be referred to each other, and the technical features in different embodiments are based on their inherent Logical relationships can be combined to form new embodiments. Furthermore, the terms "comprising" and "having", as well as any variations thereof, are intended to cover a non-exclusive inclusion, for example, of a sequence of steps or elements. A method, system, product or device is not necessarily limited to those steps or elements explicitly listed, but may include other steps or elements not explicitly listed or inherent to the process, method, product or device.

Although the application has been described in conjunction with specific features and embodiments thereof, it will be apparent that various modifications and combinations can be made thereto without departing from the spirit and scope of the application. Accordingly, the specification and drawings are merely illustrative of the solutions defined by the appended claims, and are deemed to cover any and all modifications, changes, combinations or equivalents within the scope of the application.

Apparently, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. In this way, if the modifications and variations of the embodiments of the present application fall within the scope of the claims of the present application and their equivalent technologies, the present application is also intended to include these modifications and variations.

Claims

A data access method, characterized in that it is applied to a disk controller, the method comprising:

Obtain the first data to be written into the target disk;

When it is determined that the first data satisfies the set condition, send the first data and a delayed write indication of the first data to the target disk; the delayed write indication is used to indicate the target The disk temporarily stores the first data in the cache unit of the target disk.
The method according to claim 1, wherein the setting conditions include: the first data is the data to be written into the first stripe unit of the target disk, and the data to be written after the first data The input data includes the second data to be written into the second stripe unit; the second stripe unit and the first stripe unit belong to the same target stripe, and the second stripe unit and the first stripe unit The stripe units belong to different disks.
method according to claim 1, is characterized in that, described setting condition comprises:

If the first data is the data to be written into the first stripe unit of the target disk, and it is predicted that the data to be written after the first data includes the second data to be written into the second stripe unit, then determining that the first data satisfies the set condition; the second stripe unit and the first stripe unit belong to the same target stripe, and the second stripe unit and the first stripe unit belong to different disks; or,

The set conditions include

If the first data is the verification data to be written to the third stripe unit of the target disk, and it is predicted that the data to be written after the first data includes the second data to be written to the second stripe unit , it is determined that the first data satisfies the set condition; the second stripe unit and the third stripe unit belong to the same target stripe, and the second stripe unit and the third stripe unit Cells belong to different disks.
The method according to claim 2 or 3, wherein the first data is the data carried in the received write disk instruction; the method further comprises:

If it is determined that the data writing mode is sequential writing, and the target stripe includes an idle stripe unit other than the target stripe unit, then it is determined that the data to be written includes the second data.
The method according to claim 4, characterized in that, by the following manner, determine whether the data writing mode is sequential writing:

Determine whether the data writing mode is sequential writing according to the logical addresses in the multiple disk writing instructions received continuously.
The method according to any one of claims 1 to 5, wherein the delayed writing instruction includes a delayed writing identifier and a target storage address of the first data.
The method according to any one of claims 1-6, wherein the sending the delayed writing instruction of the first data to the target disk comprises:

Carrying the delayed write instruction in a first data write instruction and sending it to the target disk; the first data write instruction includes the first data; or,

Sending a first operation instruction including the delayed write instruction to the target disk; the first operation instruction is an instruction other than the first data write instruction.
The method according to any one of claims 1-7, wherein the first data is verification data, and after sending the delayed writing instruction of the first data to the target disk, the The method also includes:

If the second data is acquired, update data is determined based on the second data; the update data is updated verification data;

sending a second data write instruction to the target disk, the second data write instruction including the update data; the second data write instruction is used to instruct the target disk to write the update data into the in the non-volatile memory of the target disk described above.
A data access method, characterized in that it is applied to a disk controller, the method comprising:

Determining a target storage address and a target disk for prefetching data; the target disk includes a cache unit;

sending a prefetch data indication generated based on the target storage address to the target disk; the prefetch data indication is used to instruct the target disk to read the prefetch data from the target storage address to the cache in the unit;

sending a first data read instruction for the prefetched data to the target disk;

receiving the prefetch data obtained and sent by the target disk from the cache unit.
The method according to claim 9, wherein said determining the target storage address and target disk of the prefetched data comprises:

If the data read mode is determined to be sequential reading according to the logical addresses in multiple read disk instructions received continuously, then determine the target storage address and address of the prefetched data according to the logical addresses in the currently received disk read instructions The target disk.
The method according to claim 9, wherein said determining the target storage address and target disk of the prefetched data comprises:

In the process of repairing the disk to be repaired, the target storage address and the target disk of the prefetched data are determined according to the target stripe to which the stripe unit to be repaired belongs; the stripe unit to be repaired is the A stripe unit in a disk; the target stripe includes a plurality of stripe units located on different disks.
The method according to any one of claims 9-11, wherein the sending the prefetch data instruction generated based on the target storage address to the target disk comprises:

Carrying the prefetch data instruction in a second data read instruction sent to the target disk, and sending it to the target disk; the second data read instruction is used to instruct the target disk to read from a specified storage address read data; or,

Sending a second operation instruction including the prefetch data indication to the target disk; the second operation instruction is an instruction other than the second data read instruction.
A data access method, characterized in that it is applied to a disk, the method comprising:

receiving the first data sent by the disk controller, and receiving a delayed write indication of the first data sent by the disk controller; the delayed write indication is used to instruct the disk to temporarily store the first data In the cache unit of the disk;

Temporarily storing the first data in a cache unit of the disk.
The method according to claim 13, wherein the delayed writing indication includes a delayed writing identifier and a target storage address of the first data.
The method according to claim 13 or 14, wherein the receiving the delayed writing indication of the first data sent by the disk controller comprises:

receiving a first data write instruction sent by the disk controller, and acquiring the delayed write instruction carried in the first data write instruction; the first data write instruction includes the first data; or ,

receiving a first operation instruction sent by the disk controller and including the delayed write instruction; the first operation instruction is an instruction other than the first data write instruction.
The method according to any one of claims 13-15, wherein the first data is verification data, and after the first data is temporarily stored in the cache unit of the disk, the method Also includes:

receiving a second data write instruction for update data sent by the disk controller; the update data is updated check data; the second data write instruction instructs the disk to write the update data into a target storage address, the target storage address is the same as the target storage address of the verification data;

According to the target storage address, the update data is written into the non-volatile memory of the magnetic disk, and the verification data is deleted in the cache unit.
A data access method, characterized in that it is applied to a disk, the method comprising:

receiving a prefetch data indication sent by the disk controller; the prefetch data indication carries a target storage address of the prefetch data;

reading the prefetched data stored in the storage space corresponding to the target storage address into the cache unit of the disk;

receiving a first data read instruction for the prefetched data sent by the disk controller;

Acquire the prefetch data from the cache unit, and send the prefetch data to the disk controller.
The method according to claim 17, wherein the receiving the prefetch data indication sent by the disk controller comprises:

receiving a second data read instruction sent by the disk controller, and obtaining the prefetch data instruction carried in the second data read instruction; the second data read instruction is used to instruct the disk to read from a specified store address to read data; or,

receiving a second operation instruction sent by the disk controller and including the instruction to prefetch data; the second operation instruction is an instruction other than the second data read instruction.
A data access device is characterized in that it is applied to a disk controller, and the device includes:

a data acquisition unit, configured to acquire the first data to be written into the target disk;

A sending unit, configured to send the first data and a delayed write indication of the first data to the target disk when it is determined that the first data satisfies a set condition; the delayed write indication is used and instructing the target disk to temporarily store the first data in a cache unit of the target disk.
A data access device is characterized in that it is applied to a disk controller, and the device includes:

A prediction unit, configured to determine a target storage address and a target disk for prefetching data; the target disk includes a cache unit;

an instruction sending unit, configured to send to the target disk a prefetch data instruction generated based on the target storage address; the prefetch data instruction is used to instruct the target disk to transfer the prefetch data from the target storage address read into the cache unit; and, send a first data read instruction for the prefetched data to the target disk;

A data receiving unit, configured to receive the prefetch data obtained and sent by the target disk from the cache unit.
A data access device is characterized in that it is applied to a magnetic disk, and the device includes:

a receiving unit, configured to receive the first data sent by the disk controller, and receive a delayed writing indication of the first data sent by the disk controller; the delayed writing indication is used to instruct the disk to write the temporarily storing the first data in the cache unit of the disk;

The execution unit is configured to temporarily store the first data in the cache unit of the disk.
A data access device is characterized in that it is applied to a magnetic disk, and the device includes:

An instruction receiving unit, configured to receive a prefetch data indication sent by the disk controller; the prefetch data indication carries a target storage address of the prefetch data;

an instruction execution unit, configured to read the prefetched data stored in the storage space corresponding to the target storage address into the cache unit of the disk;

a data transmission unit, configured to receive a first data read instruction for the prefetched data sent by the disk controller, obtain the prefetched data from the cache unit, and send the prefetched data to The disk controller.
A disk controller, characterized by comprising a memory and a processor, the memory stores a computer program that can run on the processor, and when the computer program is executed by the processor, the The processor implements the method described in any one of claims 1-8, or the method described in any one of claims 9-12.
A magnetic disk, characterized by comprising a memory and a processor, the memory stores a computer program that can run on the processor, and when the computer program is executed by the processor, the processor Implement the method according to any one of claims 13-16, or the method according to any one of claims 17-18.
A data storage system, characterized in that it comprises:

A disk controller as claimed in claim 23, and a disk as claimed in claim 24.
A computer-readable storage medium, characterized in that computer-executable instructions are stored, and the computer-executable instructions are used to cause a computer to execute the method according to any one of claims 1 to 12, or to execute the method according to any one of claims The method described in any one of 13 to 18.
A computer program product, characterized in that it contains computer-executable instructions, the computer-executable instructions are used to make a computer execute the method according to any one of claims 1 to 12, or execute the method according to any one of claims 13 to 12 The method of any one of 18.