Method for achieving file lock and metadata server

Abstract

The invention discloses a method for achieving a file lock and a metadata server. The method comprises the steps of enabling the metadata server to receive a request message sent by a request party, wherein the request message comprises a lock applying message, and a serial number of a file requested to be locked and identification information of the request party are carried in the lock applying message; enabling the metadata server to obtain file description information of the file according to the serial number, wherein file lock holding attributes and file lock update time attributes are contained in the file description information in an extended mode; enabling the metadata server to judge whether the file lock is added in the file or not according to the file lock holding attributes, if the file lock is already added and a holding party of the file lock is not the request party, judging whether the file lock added on the file is overtime or not according to the file lock update time attributes, and updating an attribute value of the file lock holding attributes to be the identification information of the request party and updating an attribute value of the file lock update time attributes to be the current time if the file lock added on the file is overtime.

Description

technical field [0001] The invention relates to the field of distributed file systems, in particular to a method for realizing file locks and a metadata server. Background technique [0002] Distributed file systems generally adopt a client / server architecture (refer to figure 1 ), the data is stored on the server, and the client application can access files located on the remote server as if they were accessing the local file system. Taking time as a clue, the development of distributed file systems can be divided into three stages: 1) network file system (1980s), 2) shared storage (SAN) cluster file system (1990s), 3) object-oriented parallel file system system (2000s). At present, the more influential distributed file systems include Google File System (GFS), Luster, and HDFS. [0003] The important role of the distributed file system in the era of cloud computing is increasingly prominent. In a cloud computing environment, the distributed file system can be applied i...

AI Technical Summary

Problems solved by technology

[0005] However, in a distributed environment, the management of file locks is much more complicated than that of a stand-alone file system. The difficulty lies not in the creation of file locks, but in the reasonable and normal release of file locks.

The distributed file system is faced with multiple processes across nodes. Each process or node may suddenly crash and cannot complete the normal release of file locks, resulting in the corresponding data / files being unable to be accessed by other processes / nodes and continuing to provide Serve

Images