Database error processing method and device and database cluster access system

Abstract

The invention discloses a database error processing method and device and a database cluster access system. Whether cluster switching is carried out or not is determined according to the statistical result of the access errors detected by all the data access nodes, so the situation that the database cluster access errors are inaccurately counted due to the fact that the data access nodes independently record the access errors of the data access nodes is avoided, thereby avoiding the misjudgement of clustering switching; Therefore, the accuracy of database cluster access error statistics and cluster switching is improved. Besides, when cluster switching needs to be carried out, the database configuration center is controlled to push update configuration information to all the data access nodes in a unified mode, all the data access nodes complete switching of the database clusters at the same time, it is avoided that the main database cluster and the standby database cluster have access of the data access nodes at the same time, and data consistency is guaranteed. Meanwhile, the whole switching process is automatically completed by the system, so that the delay of manual switching of the database clusters is avoided, the timeliness of switching of the database clusters is improved, and the availability of access of the database clusters is ensured.

Description

Technical field [0001] The present application relates to the field of computer technology, and more particularly to a database error handling method, a device and a database cluster access system. Background technique [0002] With the large-scale application of the distributed storage system, in order to improve the data reliability and stability of the distributed storage system, the settings each contain a plurality of storage nodes, and maintain data consistency between the active / standby cluster. When the primary cluster is influenced by disk, network, or other business bursts, it can continue to provide services through the 备. [0003] In the prior art, the switching of the active / standby cluster requires the data access node to complete by updating the cluster address, however, if all data access nodes cannot complete the host-standive cluster switching, the hosting cluster will appear simultaneously by the data access node access, affecting The consistency of the sta...

AI Technical Summary

Problems solved by technology

[0003] In the existing technology, the switchover of the active and standby clusters requires the data access nodes to update the cluster addresses. However, if all the data access nodes cannot complete the switchover of the active and standby clusters at the same time, the active and standby clusters will be accessed by the data access nodes at the same time, affecting The consistency of the read and write data of the active and standby clusters and the availability of the distributed storage system lead to the inconsistency of the data in the active and standby clusters of the distributed storage system, which in turn makes the distributed storage system unusable

Images