IO data sequential pre-reading method and device, equipment and storage medium

Abstract

The invention discloses an IO data sequential pre-reading method and device, equipment and a storage medium. The method comprises the steps: recording a reading track for IO data in a logical volume in each step, updating the reading track into a track tree corresponding to the logical volume, wherein each track node in the track tree records an initial logical address and a continuous data lengthof the IO data; traversing all track nodes in the track tree according to a sequence from low to high of the initial logic addresses at intervals of first preset time, and positioning a plurality ofcontinuous adjacent track nodes meeting a sampling condition, a sampling condition being that the discontinuous data length between two adjacent track nodes is lower than a first threshold value; andsampling the plurality of continuous adjacent track nodes to obtain a corresponding total data length, and if the total data length exceeds a second threshold, triggering sequential pre-reading of IOdata corresponding to the plurality of continuous adjacent track nodes. According to the invention, multi-channel concurrent sequential reading and skip sequential reading can be effectively identified, so that the reading performance is improved.

Description

Technical field [0001] The present invention relates to the field of storage technology, and in particular to a sequential pre-reading method, device, equipment and storage medium for IO data. Background technique [0002] Read ahead means that the file system reads more file contents than expected for the application at one time and caches them in the cache (CACHE), so that when the next read request comes, it can be read directly from the CACHE. [0003] At present, CACHE pre-reading is generally based on logical (LV) volumes. Only when the IO stream on the logical volume is a continuous sequential stream, pre-reading will be triggered. In this way, when the front-end business IO command comes down, the file will be read directly from the cache. There is no need to read the file from the disk, which can improve the performance of sequential reading of the volume. [0004] In the process of realizing the present invention, the inventor found that there are at least the fol...

AI Technical Summary

Problems solved by technology

[0005] Usually, the business end reads and accesses the volume concurrently, that is, multiple threads read the data of a volume at the same time. At this time, although each thread reads the data sequentially, when multiple threads read at the same time, each The thread does not set a specific access priority, so that the sequential read operations of each thread will be mixed together, resulting in the inability to trigger read-ahead; in addition, in actual operation, for discontinuous sequential reads, that is, skip sequential reads, may be Read-ahead cannot be triggered due to failure to identify continuous sequential reads, resulting in reduced data read performance

Images