High-availability and high-consistency database cluster system and command processing method thereof

Abstract

The invention discloses a high-availability and high-consistency database cluster system and a command processing method thereof. The system comprises a command processing rule setting unit, an initial configuration unit and a plurality of nodes, wherein the nodes are connected through a private network; each node can provide complete data storage, access and command execution management services; the initial configuration unit is used for providing a configuration interface and determining a primary node and secondary nodes in the nodes in a manual configuration manner; and the command processing rule setting unit is used for setting a command processing rule and ensuring data among the nodes of the database cluster system in the command processing process to be synchronous and consistent. According to the system and the method, each node in the database cluster system can provide complete storage and access services; in the command processing process, the data of the nodes in the database cluster system is synchronous and consistent; and a single node fault does not influence normal work of the database cluster system and also does not cause data loss.

Description

technical field [0001] The invention relates to a database cluster system with high availability and strong consistency and a command processing method thereof. Background technique [0002] RAC (Real Application Cluster, real application cluster) is a parallel cluster of Oracle. Oracle instances located on different nodes access the same Oracle database at the same time. Nodes communicate through a private network. All control files, online logs and data files are stored in On a shared storage device, it can be read and written by all nodes in the cluster; this cluster method has certain limitations: [0003] 1) Data synchronization between instances [0004] RAC provides multiple instances. RAC will first cache the written data in the memory of the instance, and then synchronize it to the shared storage. Sometimes it takes a long time to synchronize the data. For example: when multiple instances write data at the same time and immediately query the newly written data, th...

AI Technical Summary

Problems solved by technology

[0002] RAC (Real Application Cluster, real application cluster) is a parallel cluster of Oracle. Oracle instances located on different nodes access the same Oracle database at the same time. Nodes communicate through a private network. All control files, online logs and data files are stored in On a shared storage device, it can be read and written by all nodes in the cluster; this cluster method has certain limitations:

When the amount of data is small, the time required is acceptable; but when the amount of data is large, the time it takes to maintain consistency may not be acceptable

[0005] On the other hand, when an instance fails, if the cached data in the instance has not been synchronized to the shared storage, this part of the data will be lost and cannot be recovered

[0006] 2) Database single point of failure

[0007] There is only one data file in RAC, when the shared storage fails, the service will not be provided

Images