num="0082">In some embodiments, the method further includes the following operation. The expanded capacity of the each volume is obtained based on the total available capacity of the newly added disks and a quantity of volumes of the original RAID.

In some embodiments, the method further includes the following operation: the total available capacity of the newly added disks is calculated based on a RAID level that the original RAID belongs and the capacity of the newly added disk.

In some embodiments, the method further includes the following operation: the total newly added LBA range of the newly added RAID is recorded into the newly added LBA mapping table.

In some embodiments, the method further includes the following operation: a quantity of the newly added disks of the new RAID, LDNs of the newly added disks, and corresponding physical disk slot numbers are recorded into the newly added LBA mapping table.

Specifically, after physical disks are added to expand the capacity of the RAID, one new RAID Lba Map is mapped by a capacity expansion part, which is recorded as SubMap1 (that is, the newly added LBA mapping table). The newly added volume LBA range is obtained by converting the capacity of the newly added physical disks, and is also characterized by using LBA. Meanwhile, the information to be recorded by the SubMap1 includes an ID of an array that the SubMap1 belongs, the quantity of the newly added disks, and a correspondence between the LDN (0 to N−1) of the newly added disks and PDN (disk slot numbers).

Every time the capacity of the RAID array is expanded, a new SubMap (that is, an LBA mapping table) is newly added. Therefore, assuming that the capacity of the RAID array is expanded for N times, the array has one and only one initial Map0 and N SubMaps. Moreover, after the capacity of the RAID is expanded each time, stripe data chunks do not need to be re-arranged, so a data migration task of a disk does not need to be performed, the overhead and the complexity of a CPU of a system are greatly reduced.

In some embodiments, the operation that each newly added volume LBA range is recorded into a corresponding struct includes the following operation: start LBA value and an end LBA value of each newly added volume LBA range are recorded into the corresponding struct.

In some embodiments, the method further includes the following operation: a quantity of the newly added disks and a quantity of newly added chunks of each volume into the corresponding struct.

In some embodiments, the method further includes the following operation: an upper limit of a quantity of structs is set for the each volume.

Specifically, for a volume, a StripeSequence (that is, a struct) is designed for each volume to save a capacity LBA offset and capacity automatically allocated to the volume by expanding the capacity of the RAID to achieve a purpose of automatically expanding the capacity of the volume, thereby greatly simplifying operation steps of a user.

Each StripeSequence includes LBA information mapped by the disks added to the RAID array in the same batch. Each volume contains 128 (that is, an upper limit of a quantity) StripeSequences at most. Each StripeSequence contains the following information:

struct VolumeSS

{

UINT8 numDisks;//the number of physical disks contained in this StripeSequence

UINT16 stripeChunks;//the number of stripe chunks

UINT64 startLba;//a start LBA of the StripeSequence in a RAID

UINT64 maxLba;//a maximum LBA of the StripeSequence in a RAID

}

Therefore, when the capacity of the RAID is expanded each time, a StripeSequence will be newly added to the volume on the array to automatically allocate the expanded capacity of the RAID array to the volume, so that the capacity of the volume does not need to be expanded manually by a user.

When a host performs reading and writing on the volume, which StripeSequence of the volume the LBA belongs is determined first through an LBA value of the volume, information is obtained from the StripeSequence that the LBA belongs, and a corresponding array LBA corresponding to a volume LBA is calculated. A method for calculating the LDN and the PBA according to the array LBA includes the following operations.

- (1) A number of a chunk where the array LBA is located is calculated:

$Chunk_number = LBA / Chunk_size$

- (2) An offset in the chunk where the array LBA is located is calculated:

$Chunk_offset = LBA % Chunk_size$

- (3) A number of a stripe where the array LBA is located is calculated:

$Stripe_number = Chunk_number / numDisks$

- (4) An LDN where the array LBA is located is calculated:

$LDN_number = Chunk_number % numDisks$

- (5) A PBA is calculated:

$Disk_PBA = Stripe_number * Chunk_size + Chunk_offset$

“%” represents a remainder operation; and Chunk is a chunk, is a partition of a physical storage medium on a disk, and is used for the RAID array to calculate the granularity of data chunk movement in the RAID array, and is a unit that forms a stripe.

The embodiments of the present disclosure are applicable to expanding the capacity of the RAID array in a manner of adding hard disks and automatically allocating the expanded capacity to a volume. FIG. 4 is a schematic flowchart of automatically allocating expanded capacity to a volume after capacity of a RAID is expanded according to an embodiment of the present disclosure. With reference to FIG. 4, this process further includes the following embodiments.

In some embodiments, the method further includes the following operation: values of verification chunks of each stripe are recalculated based on the data chunks of the original RAID and the data chunks of the newly added disks to update the verification chunks.

In some embodiments, the operation that the newly added disks are chunked into chunks and the chunks are added to the original RAID to obtain a new RAID further includes the following operations: the newly added disks are chunked into chunks and the chunks are added to the original RAID, and whether a quantity and capacity of the newly added disks meet specifications of the original RAID is verified. In response to that the quantity and the capacity of the newly added disks meet the specifications of the original RAID, it is determined that the new RAID is obtained.

In some embodiments, the method further includes the following operation: in response to that the quantity and the capacity of the newly added disks do not meet the specifications of the original RAID, an error is returned to determine that a new RAID is incapable of being obtained.

In some embodiments, the method further includes the following operations: in response to that the target LBA value does not have a corresponding struct, it is determined that the target LBA value does not belong to the newly added volume LBA range of the current volume, and a corresponding physical storage area is directly found according to an LBA mapping relationship between the current volume and the original RAID and a mapping relationship between the LBA and a corresponding PBA of the original RAID to perform reading and writing.

In this embodiment, if the target LBA value does not have a corresponding struct, the target LBA value does not belong to the newly added volume LBA range of the current volume, and reading and writing are performed according to a reading and writing mode of an original RAID.

The method for expanding capacity of a RAID in the embodiments of the present disclosure can be implemented based on a specific system architecture. The system architecture is as shown in FIG. 5, wherein a RAID card, as a PCIe device, is hooked in a PCIE topology of a Root Complex (RC) in a host, and the host is mapped to a storage device (the RAID card device in the host as shown in FIG. 5). The host and the RAID card perform data interaction by using the PCIE high-speed bus protocol, and the application layer performs storage IO command interaction by using the NVMe protocol. The RAID card includes a RAID controller, which is used to cooperate with a host to complete relevant control of a PCIe device, PCIe DMA data transfer and NVMe, so as to realize the passage of the RAID card and the host; The interior of the RAID card may also have a plurality of modules, including a processor, an IO manager, an XOR calculation module, a cache module, etc., and the plurality of modules cooperatively complete data reading and writing and check calculation in a RAID process, so as to achieve RAID protection of user data and ensure the reliability and performance of the data; The RAID card can further include a physical disk manager, which is used for realizing the management of a physical disk to be connected, and organizing an independent physical disk into a logical capacity according to a RAID level so as to be used by a host.

A second aspect of the embodiments of the present disclosure further provides a system for expanding capacity of a RAID. FIG. 6 is a schematic diagram of an embodiment of a system for expanding capacity of a RAID according to the present disclosure. As shown in FIG. 6, a system for expanding capacity of a RAID includes: a new RAID module 10, configured to: in response to receiving a request for expanding capacity of an original RAID, chunk newly added disks into chunks and add the chunks to the original RAID to obtain a new RAID, and obtain a total newly added LBA range of the new RAID based on total available capacity of the newly added disks; a traversing module 20, configured to traverse all volumes in the new RAID, obtain, for each volume, a corresponding newly added volume LBA range based on the total newly added LBA range, and record each newly added volume LBA range into a corresponding struct; a determination module 30, configured to: in response to performing reading and writing on a current volume, determine whether a target LBA value of the current volume has a corresponding struct; and a reading and writing module 40, configured to: in response to that the target LBA value has a corresponding target struct, confirm, based on the target struct, an actual LBA value in the new RAID corresponding to the target LBA value, and perform reading and writing by using a physical storage area of a PBA corresponding to the actual LBA value.

A third aspect of the embodiments of the present disclosure further provides a non-transitory readable storage medium. FIG. 7 is a schematic diagram of a non-transitory readable storage medium for implementing a method for expanding capacity of a RAID according to an embodiment of the present disclosure. As shown in FIG. 7, the non-transitory readable storage medium 3 stores computer program instructions 31. The computer program instructions 31 implement the following steps when executed by a processor.

In response to receiving a request for expanding capacity of an original RAID, newly added disks are chunked into chunks and the chunks are added to the original RAID to obtain a new RAID, and a total newly added LBA range of the new RAID is obtained based on total available capacity of the newly added disks.

All volumes in the new RAID are traversed, a corresponding newly added volume LBA range is obtained for each volume based on the total newly added LBA range, and each newly added volume LBA range is recorded into a corresponding struct.

In response to performing reading and writing on a current volume, whether a target LBA value of the current volume has a corresponding struct is determined.

In response to that the target LBA value has a corresponding target struct, an actual LBA value in the new RAID corresponding to the target LBA value is confirmed based on the target struct, and reading and writing are performed by using a physical storage area of a PBA corresponding to the actual LBA value.

In some embodiments, the operations that newly added disks are chunked into chunks and the chunks are added to the original RAID to obtain a new RAID include the following operations: the newly added disks are chunked into the chunks according to chunk sizes of the original RAID, locations of data chunks and verification chunks in a stripe of the original RAID are remained unchanged, and the newly added disks after chunking are directly added to the original RAID to obtain the new RAID.

In some embodiments, the operation that a corresponding newly added volume LBA range is obtained for each volume based on the total newly added LBA range includes the following operation: a corresponding newly added volume LBA range is allocated for each volume based on expanded capacity and a stripe range of the each volume, and the total newly added LBA range.