Metadata organization and management method and system of distributed file system

Abstract

The invention discloses a meta data organization and management method of a distributed file system. The method includes the steps: receiving meta data request message packets, parsing content of the meta data request message packets, putting the meta data request message packets into a corresponding message queue according to parsed types of the meta data request message packets, extracting meta data request message packets from message queues, parsing the meta data request message packets so as to obtain request types and request content of the meta data request message packets, acquiring meta data information stored in a tree B+ according to the meta data request message packets, performing corresponding read-write operation on the meta data information, putting results obtained after read-write operation in a meta data request response message packet, and returning the meta data request response message packet to a client side. The method can solve the problem of poor meta data processing performance of an MDS (meta data server) caused by existing meta data read-write high-concurrency operation, and processing capacity of the MDS is improved. .

Description

technical field [0001] The invention belongs to the technical field of computer storage, and more particularly relates to a metadata organization and management method and system of a distributed file system. Background technique [0002] With the continuous expansion of data information and the wide application of unstructured data in various industries, the distributed file system provides an ideal solution for the storage management of unstructured data. Most current distributed file systems usually process metadata and data separately, that is, separate control flow from data flow, so as to obtain better system scalability and IO concurrency. The system under this strategy is mainly composed of a tripartite architecture of metadata server MDS (Meta Data Server), data storage server, and client. MDS is mainly responsible for storing file namespaces, responding to service requests, and providing metadata information such as file storage locations. In the file system, the ...

AI Technical Summary

Problems solved by technology

[0004] Aiming at the above defects or improvement needs of the prior art, the present invention provides a metadata organization that distinguishes metadata read and write requests and adopts copy-on-write B+ tree organization to manage metadata in a distributed file system metadata server The purpose of the management method is to provide directory-based operations for the client and separate the client’s read and write operations on the B+ tree, thereby solving the technical problem of low metadata processing performance caused by the existing metadata read and write high concurrent operations on the MDS , to improve the processing capacity of the metadata server

Images