Metadata group design method based on real-time application group

Abstract

The invention provides a metadata group design method based on a real-time application group. The particular design process comprises the step of initial deployment, wherein initial configuration is carried out on a metadata group, the metadata group is composed of a plurality of metadata nodes, a configuration file is generated and synchronized to each metadata node, and the initialization of metadata storage is finished on each metadata node; the step of sequentially starting each metadata node when the group is started, wherein the metadata node with the smallest number in the configuration file automatically obtains the hegemony and guides the whole metadata group to build an overall consistency buffering and sharing lock; the step of fault recovery; and the step of adding new metadata nodes. Compared with the prior art, the metadata group design method based on the real-time application group achieves transverse expansion of the metadata group in a group file system, effectively resolves the performance bottleneck of metadata in a large-scale group file system, is high in practicability and is easy to popularize.

Description

technical field [0001] The invention relates to the technical field of computer clusters, in particular to a metadata cluster design method based on real-time application clusters. Background technique [0002] With the rapid development of network applications, the amount of network information data is increasing, and PB-level mass data storage is becoming more and more important. The traditional file system can no longer meet the requirements of large capacity, high reliability, and high performance of existing applications. In order to meet these new requirements, the application of new cluster storage systems has been widely valued. [0003] In traditional cluster file systems, metadata performance often becomes the performance bottleneck of the system. Early cluster file systems have one and only one metadata node responsible for metadata storage, access and processing. In a more advanced cluster file system, there are multiple metadata nodes providing services at the...

AI Technical Summary

Problems solved by technology

Its response to metadata access hotspots often lags behind, and cannot adapt to scenarios where metadata access hotspots often change, and there is currently no feasible engineering implementation

[0006] For scenarios where the absolute amount of metadata is not particularly large and the popularity of metadata access changes frequently, none of the above solutions can achieve good results, and there are problems that reduce system availability at the high availability level

Images