Data access method and storage array

Patent metadata

Field	Value
Publication number	US-11301157-B2
Application number	US-202016859234-A
Country	US
Kind code	B2
Filing date	Apr 27, 2020
Priority date	Dec 29, 2017
Publication date	Apr 12, 2022
Grant date	Apr 12, 2022

Abstract

A disk array includes SSD of different page sizes and a virtual disk is constructed based on the SSDs of different page sizes. The constructed virtual disk formed by a hybrid-page chunk which includes a large-page chunk, a small-page chunk, and the small-page chunk, is not provided for the host for use, and is only used to migrate data corresponding to the large-page chunk. Based on the virtual disk formed by the hybrid-page chunk, the data with a large granularity can be stored in the SSD with a large page size, and data with a small granularity is migrated to the SSD with a small page size.

First claim

What is claimed is:

1. A data access method implemented by a processor of a storage array, the data access method comprising:
receiving, from a host, a first data access request for accessing first data stored in the storage array, wherein the first data access request carries a first logical address of the first data, wherein the storage array comprises first solid-state disks (SSDs) and second SSDs, wherein a page size of each first SSD is greater than a page size of each second SSD, wherein a logical address space formed by the first SSDs is divided into large-page logical blocks, wherein each large-page logical block is divided into large-page logical pages based on the page size of the first SSD, wherein a logical address space constituted by the second SSDs is divided into small-page logical blocks, wherein each small-page logical block is divided into small-page logical pages based on the page size of the second SSD, wherein a large-page logical block and a small-page logical block constitute a hybrid-page logical block, wherein a logical unit provided for the host for use is constructed based on the large-page logical block in the hybrid-page logical block, wherein a conversion table is set for each hybrid-page logical block, and wherein the conversion table records a mapping relationship between a large-page logical page, a data annotation, and a small-page-index table address;
searching for a first large-page logical page corresponding to the first logical address and a large-page logical page address of the first large-page logical page, wherein the first large-page logical page belongs to a large-page logical block of a first hybrid-page logical block;
searching a conversion table corresponding to the first hybrid-page logical block for a first data annotation corresponding to the first large-page logical page;
obtaining a small-page-index table based on a small-page-index table address that is recorded in the conversion table and that corresponds to the first large-page logical page when the first data annotation is a small-page annotation, wherein the small-page annotation indicates that data corresponding to the first large-page logical page is stored in a small-page SSD corresponding to a small-page chunk;
determining, in the small-page-index table based on the large-page logical page address, a small-page logical page corresponding to the large-page logical page address and a small-page logical page address of the small-page logical page, wherein a storage space that corresponds to the small-page logical page address and that is of a second SSD is for storing data migrated from a storage space that corresponds to the large-page logical page address and that is of a first SSD; and
obtaining the first data from the second SSD based on the small-page logical page address.

2. The data access method of claim 1 , further comprising:
receiving, from the host, a second data access request that is for accessing second data stored in the storage array, wherein the second data access request carries a logical address of the second data;
searching for a second large-page logical page corresponding to the logical address of the second data and a large-page logical page address in the second large-page logical page, wherein the second large-page logical page belongs to a large-page logical block of a second hybrid-page logical block;
searching a conversion table corresponding to the second hybrid-page logical block for a data annotation corresponding to the second large-page logical page; and
obtaining the second data from the first SSD based on the large-page logical page address in the second large-page logical page when the data annotation corresponding to the second large-page logical page is a large-page annotation.

3. The data access method of claim 2 , wherein the conversion table further records a quantity of access times of each large-page logical page, and wherein the data access method further comprises increasing by 1 a quantity of access times of the second large-page logical page in which an amount of the data obtained from the first SSD is less than a preset value when the second data access request is a rewrite request.

4. The data access method of claim 2 , wherein a large-page logical block in each hybrid-page logical block has a large-page logical page table, wherein the large-page logical page table records a correspondence between a large-page logical page comprised in each large-page logical block and a logical page in the first SSD; and wherein the data access method further comprises:
obtaining, from the large-page logical page table, a logical page that corresponds to the second large-page logical page and that is in the first SSD;
determining a logical page address that corresponds to the large-page logical page address and that is in the first SSD;
generating an access sub-request, wherein the access sub-request carries the logical page and the logical page address in the first SSD; and
sending the access sub-request to the first SSD to instruct the first SSD to obtain, based on the logical page and the logical page address in the first SSD, the data requested in the access sub-request.

5. The data access method of claim 1 , further comprising:
periodically detecting in the conversion table a large-page logical page whose quantity of access times is greater than a preset value;
creating a small-page-index table for the large-page logical page whose quantity of access times is greater than the preset value;
recording an address of the small-page-index table in the conversion table, wherein the small-page-index table records N blank small-page logical pages, N is a quantity of small-page logical pages comprised in each large-page logical page, and N≥1;
migrating, to a storage space that corresponds to the N blank small-page logical pages and that is of the second SSD, data that is stored in the first SSD and that corresponds to the large-page logical page whose quantity of access times is greater than the preset value;
modifying in the small-page annotation a data annotation of the large-page logical page whose quantity of access times is greater than the preset value; and
setting to zero the quantity of access times of the large-page logical page whose quantity of access times is greater than the preset value.

6. The data access method of claim 5 , wherein the small-page-index table records a quantity of access times of each small-page logical page, and wherein the data access method further comprises:
calculating, at an interval of a preset time, a variance and a mean value of quantities of access times corresponding to all the N blank small-page logical pages in the small-page-index table; and
migrating, to the first SSD corresponding to a large-page logical page corresponding to the small-page-index table when the variance is less than a threshold T 1 and the mean value is greater than a threshold T 2 , data that corresponds to the N blank small-page logical pages and that is in the second SSD.

7. The data access method of claim 1 , wherein a small-page logical block in each hybrid-page logical block has a small-page logical page table, wherein the small-page logical page table records a correspondence between a small-page logical page comprised in each small-page logical block and a logical page in the second SSD and wherein the data access method further comprises:
obtaining, from the small-page logical page table, a logical page and a logical page address of the logical page in the second SSD that correspond to each small-page logical page and a small-page logical page address of the small-page logical page;
generating an access sub-request for each logical page and a logical page address of the logical page in the second SS

Assignees

• Huawei Tech Co Ltd

Inventors

• Hu Pan

Classifications

• G06F3/067

  Distributed or networked storage systems, e.g. storage area networks [SAN], network attached storage [NAS] · CPC title

• G06F3/064

  Management of blocks · CPC title

• G06F3/0644

  Management of space entities, e.g. partitions, extents, pools · CPC title

• G06F3/0647 Primary

  Migration mechanisms · CPC title

• G06F3/0679

  Non-volatile semiconductor memory device, e.g. flash memory, one time programmable memory [OTP] · CPC title