Data reading method and terminal

Abstract

The invention discloses a data reading method and a terminal, and the method comprises the steps: building a database agent layer, changing data in a main node when a data change operation is received, and storing the changed update data in the agent layer; and when a data reading operation is received, judging whether the master-slave synchronization is delayed or not, if so, calculating a delayed data volume according to the delay time and the number of transactions per second of the database, comparing data contained in the delayed data volume with the updated data cached in the proxy layer to obtain unsynchronized data and synchronized data. Through interaction between the agent layer and the database, synchronized data can be read from the slave node, unsynchronized data can be read from the master node, and the phenomenon that a client cannot read data or dirty data from the slave node to cause business abnormity is avoided, so that the data reading accuracy is ensured, meanwhile, and the reading amount of the master node is reduced. And the pressure of the main node is reduced to a certain extent.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to a data reading method and a terminal. Background technique [0002] Today, with the vigorous development of the mobile Internet, there are various databases, and due to the use of a large number of Internet users, a large amount of data will be brought, which will put a lot of pressure on the database. In order to alleviate the pressure on the database, the general architecture of the database is distributed deployment to improve the QPS (Query Per Second, query rate per second) data of the database. [0003] However, as the number of users continues to increase, there are higher requirements for the QPS of the database. At this time, the read-write separation architecture of the master-slave database is generally adopted, and the data is written on the master node, and the read data is on the slave node. on the node. Using this read-write separation architecture can further ...

AI Technical Summary

Problems solved by technology

However, under such a read-write separation architecture, there will be a certain delay due to data synchronization from the master node to each slave node. If network jitter or other abnormal conditions cause master-slave delays, at this time, the client slave node The data will not be read, or the read data will be dirty, resulting in abnormal business

Images