Method and system for promoting metadata service reliability

Abstract

A method for improving metadata service reliability and a relative system belongs to computer memory technical field, for fully utilizing node resource to improve the reliability of metadata service. The inventive method comprises normal operation, main metadata server accident, backup metadata server lift, sub metadata server accident and second backup metadata server lift. The inventive system comprises a client node, a memory node, a main metadata server and a sub metadata server, which respectively comprises an object metadata service module, a memory resource manage module, and a high available module, wherein the high available module comprises a metadata copy module, an accident shield module, a detect module and a select module, the select module can select the memory node with minimum load as the backup metadata server, and the accident shield server is switched to next metadata server, and selects the backup metadata server to form a master-salve backup chain service mode, to improve the reliability of metadata service.

Description

technical field [0001] The invention belongs to the technical field of computer storage, and in particular relates to a method and a system for improving the reliability of metadata service. Background technique [0002] Metadata is often referred to as "data of data". In order to reduce server overhead and increase transmission rate, network storage systems often separate metadata from data, and use a dedicated metadata server to manage metadata, so that data transmission does not need to go through the server. For transit, you only need to obtain enough metadata information in advance. The metadata server thus becomes a hub connecting client nodes and storage nodes, and it is easy to form a single point of failure. Therefore, it is necessary to seek methods to improve the reliability of metadata services. [0003] Traditional network storage systems mostly use master-slave standby mode to provide reliable metadata services, such as Pang Liping, He Feiyue, Yue Jianhui, etc...

AI Technical Summary

Problems solved by technology

[0004] However, due to its own characteristics, the master-slave mutual backup method fails to fully utilize the resources of other nodes in the storage system, resulting in the inability to make further breakthroughs in reliability; when both the master metadata server and the slave metadata server fail, Although other parts of the storage system are still operating normally, the system cannot provide storage services due to the role of the server as the central hub.

Images