A data writing and reading method, device and distributed object storage cluster

Abstract

The embodiment of the present invention discloses a data writing and reading method, a device and a distributed object storage cluster. The data writing method includes: a storage server receives data to be written sent by a client, and divides the to-be-written data into multiple sub-systems. data; write each sub-data into the corresponding network hard disk, and for each sub-data, send the corresponding relationship between the serial number of the sub-data and the identification information of the network hard disk written by the sub-data to the distributed object storage Metadata server in the cluster. Using the embodiment shown in the present invention, each sub-data divided by the data to be written is written into the corresponding network hard disk. The storage characteristics of the network hard disk are that the space occupied by the stored data is the same as the size of the data itself. Multiple sub-data divided into input data are written to multiple network hard disks correspondingly, which does not occupy extra space, avoids wasting resources, and does not need to wait for merging with other data, which improves the efficiency of reading and writing data.

Description

technical field [0001] The present invention relates to the technical field of cloud storage, in particular to a method and device for writing and reading data, and a distributed object storage cluster. Background technique [0002] Usually, the storage space in the storage device is pre-divided into multiple sub-storage spaces of a specific size, for example, the 100M storage space is divided into 30M, 30M and 40M sub-storage spaces. In this case, the data is written to When stored in a storage device, the data can be written into a certain storage space of the storage device, for example, 10M of data is written into a 30M sub-storage space of the storage device. However, during the writing process, the 10M data will occupy the 30M storage space, and other data cannot be written again, resulting in a waste of resources. [0003] In view of the above defects, when writing data currently, multiple data are usually combined into a larger data before writing, for example, thre...

AI Technical Summary

Problems solved by technology

But using this method, when writing a 10M file into a 30M sub-storage space, you need to wait for other files that can be merged. If there are no other files to be written, or the data size of other files to be written is not suitable, write The efficiency of entering data will be very low

Images