Distributed file system file pre-reading method and system

Abstract

The invention discloses a distributed file system file pre-reading method and system. The pre-reading method comprises the steps that 1, whether file pre-reading needs to be conducted in a current cache or not is judged; if yes, 2, a current pre-reading position is obtained, and the current pre-reading position is the initial position of the initial file needing to be pre-read after the text segment is pre-read in the file needing to be read; and 3, a continuous part in a predetermined size range is extracted from the current pre-reading position in the to-be-read file into the cache to serveas a pre-reading text segment. The method comprises the steps of judging whether file pre-reading needs to be performed in a current cache or not; the current pre-reading position is acquired after the pre-reading is required, and the current pre-reading position is the initial position of the initial to-be-pre-read file after the pre-read text segment in the to-be-read file is pre-read, so that normal and accurate pre-reading can be realized regardless of a pre-reading mechanism of a file continuous scene or a file discontinuous scene, and the pre-reading efficiency is improved.

Description

technical field [0001] The invention relates to the technical field of distributed file systems, in particular to a file pre-reading method and system of a distributed file system. Background technique [0002] In the era of cloud computing and big data, massive data is stored in storage servers. When reading data, file reading performance is very important. [0003] At present, the commonly used method in the distributed file system is pre-reading, that is, while reading a piece of data in the file, the data at one end after reading is extracted and put into the buffer. When the file is read continuously, this method can effectively improve The reading efficiency of the file is improved to improve the reading performance, but when the file is read non-sequentially, the performance improvement of this method is minimal. [0004] Therefore, how to read files discontinuously and improve the performance of reading files is the direction of efforts of workers in this field. C...

AI Technical Summary

Problems solved by technology

[0003] At present, the commonly used method in the distributed file system is pre-reading, that is, while reading a piece of data in the file, the data at one end after reading is extracted and put into the buffer. When the file is read continuously, this method can effectively improve The read efficiency of the file, thereby improving the read performance, but when the file is read non-sequentially, the performance improvement of this method will maintain its slight

Images