Message processing method, server cluster and message processing system

Abstract

The application provides a message processing method, a server cluster and a message processing system. The server cluster comprises multiple queue servers. When writing a request message sent by a first client into a target log file, a target queue server triggers backup queue servers in the server cluster to synchronize the target log file. When the backup queue servers push the synchronization request message to corresponding backup message queues, the backup queue servers feed backup completion information back to the target queue server. After receiving a first preset number of pieces of backup completion information, the target queue server pushes the request message to a target message queue for storage, and feeds success response information for the request message back to the first client. Thus, the production reliability and storage reliability of the request message are ensured. Moreover, as a multi-copy mechanism is adopted, the reliability and availability of the system are guaranteed under the condition that the target queue server is in abnormal operation.

Description

Technical field [0001] This application mainly relates to the field of message processing applications, and more specifically to a message processing method, server cluster and system. Background technique [0002] In applications in various fields, message transmission between clients has become the basic process of realizing application services. Referring to Figure 1(a), traditional message transmission is usually one-to-one transmission between message producers and message consumers. It is easy to cause message loss, and it is often impossible to achieve message reuse, which affects work efficiency. [0003] In order to solve the above problems, as shown in Figure 1(b), currently open source message queue middleware is usually set up between the message producer and the message consumer to store the message generated by the message producer and forward it to the message as needed For consumers, message reuse is realized, and problems such as message loss caused by the sudden ...

AI Technical Summary

Problems solved by technology

[0004] However, the existing open source message queue middleware mainly adopts single-click or master-slave structure. Once one of the nodes fails, the entire system will be unavailable. If there are lost messages during this period, the messages will not be traceable , thus affecting the reliable and stable operation of the system; moreover, when the number of nodes in the open source message queue middleware is adjusted, it will trigger global data rebalancing, affecting system availability

Images