Data reading/writing method, system and device, and computer readable storage medium

Abstract

The invention provides a data reading / writing method, system and device, and a computer readable storage medium. The method comprises the steps of pre-creating a mapping relationship between an identifier of memcg and an identifier of blkcg; in a data write-back process, obtaining the identifier of the memcg according to an index node of a dirty page; according to the identifier of the memcg, creating a write-back structure; according to the write-back structure and the mapping relationship, obtaining blkcg speed limitation information; and according to the blkcg speed limitation information,performing IO operation on data write-back. In the way, different processes have higher equity degree in IO use; mutual influence among the processes is reduced; especially for IO intensive applications and low-speed disks, a better resource control capability is achieved; realization is easy; corresponding speed limitation policies can be set without modifying an interface of any application layer; and application limitation of CGroups is eliminated, so that the IO speed limitation application of the CGroups is expanded.

Description

technical field [0001] The present application relates to the field of computer technology, in particular to a method, system, device and computer-readable storage medium for reading and writing data. Background technique [0002] In computers such as Linux systems, IO operations on files, such as read operations and / or write operations, are often involved. Among them, the data is first written to the memory (Page Cache) for caching, and then written back to the disk The way to improve IO performance is buffered IO (buffered IO). Currently, buffered IO is used. Because most applications are read and write operations, when the same storage device is shared by multiple processes, there will be mutual influence, and even a large amount of IO generated by a certain process will cause other processes to block for a long time. For this kind of problem, IO speed limit operation is usually performed. [0003] An existing IO speed limit method is to use CGroups V2 (Control Groups, ...

AI Technical Summary

Problems solved by technology

[0003] An existing IO speed limit method is to use CGroups V2 (Control Groups, control group), which can limit the speed of writeback (write-back). This function can identify the process of generating dirty data in the memory, so that the The blkcg information (speed limit information) is passed to the block device layer to achieve the purpose of speed limit. However, CGroups V2 has a necessary premise that memcg and blkcg must use the same hierarchical structure, otherwise the blkcg information cannot be obtained according to the memcg logo, and currently The widely used Linux system background process management and most applications are multi-level CGroups user interface. This premise is difficult to meet at present. Therefore, the application of this IO speed limit is very limited, and the resource control ability is unbalanced.

Images