Support distributed lock method of distributed database and distributed database system

Abstract

The invention provides a distributed lock method for supporting a distributed database. The distributed lock method comprises the following steps of taking lock resources required by the distributed database as a whole to form a virtual distributed lock manager, setting a local image in each node of the distributed database by the virtual distributed lock manager and enabling all local images to be kept synchronous; after receiving a request of the system by the local image of any node of the virtual distributed lock manager, determining a distributed lock object corresponding to each request and synchronizing the distributed lock object to all local images of other modes in the distributed lock manager; performing resource locking operation of the database by the nodes of the distributed database through the local images of the virtual distributed lock manager. The distributed lock method disclosed by the invention has the advantages that a transaction resource monopoly mode can be provided for the distributed database, so the access conflict of the resources under the distributed system is avoided; meanwhile, on the premise of high availability and high performance, the consistency of the resources in the distributed data system can be guaranteed.

Description

Technical field [0001] The present invention relates to the field of distributed databases, in particular to a distributed lock method supporting distributed databases and a distributed database system. Background technique [0002] With the rapid development of information technology, large database systems need to process and store more and more data, calculations are becoming more and more complex, and the challenges to performance are also increasing. Performance, reliability, and scalability are required. It will become more and more intense. At this time, a centralized database obviously cannot meet the demand. In order to meet the development needs of application services, distributed databases distribute data on different nodes of a computer network, and these data logically belong to the same system. [0003] In traditional database systems, when applications concurrently access database data resources, in order to ensure data consistency, resource locks are usually used ...

AI Technical Summary

Problems solved by technology

When the second transaction accesses the resource, it executes the application lock operation. Since the first transaction has obtained the exclusive lock, the application lock operation will be blocked, and the second transaction cannot perform subsequent resource access operations.

The problem with this method is that each operation of the application transaction needs to access the lock manager node, resulting in a decrease in the performance of the application transaction. At the same time, due to the use of an independent node as the lock manager, once the node fails, the entire distributed The database cannot provide external services, which reduces the high availability of the system

[0007] To sum up, the existing distributed database locks cannot guarantee the consistency of resources in the distributed data system under the premise of ensuring high availability and high performance, nor can they solve the conflict problem of concurrent access to the same resource on different nodes

Images