Transaction-lossless distributed deployment method and system

Abstract

The invention relates a transaction-lossless distributed deployment method and system. The method comprises the following steps: step 1, establishing connection between a front-end processor and a plurality of web clients; step 2, copying the configurations of the web clients in an application of the front-end processor, and loading the copied configurations into an internal memory for processing; step 3, setting a global variable, and circularly reading the processed configurations so that the front-end processor realizes equal resource distribution on every web application ends; step 4, enabling the front-end processor to receive a request sent by a client and orderly extracting the url of each web client; step 5, detecting the web clients to judge whether the obtained urls are available, and if so, going to step 6, otherwise, going to step 7; step 6, enabling the client to access the web clients via the front-end processor, and ending; and step 7, enabling the client to access a standby url in the front-end processor, and ending. Due to the addition of the front layer, the system stability and good user perception are improved.

Description

technical field [0001] The invention relates to a distributed deployment method and system for lossless transactions. Background technique [0002] With the development of the Internet era and the increase of network users, users have higher and higher requirements for the systems they use. No matter how high the performance of a single host or application is, it can no longer meet the requirements. Alibaba proposed the concept of de-IOE and realized de-IOE To reduce maintenance costs and achieve cost reduction, more and more service providers are adopting distributed cloud deployment, load balancing, multi-machine hot backup and other deployment solutions to avoid single point of failure and performance pressure. [0003] At present, most of the distributed deployments of large-scale systems are applications that are evenly distributed to multiple devices through load balancer configuration transactions. If the application is abnormal, the request will no longer be distrib...

AI Technical Summary

Problems solved by technology

[0003] At present, most of the distributed deployments of large-scale systems are applications that are evenly distributed to multiple devices through load balancer configuration transactions. If the application is abnormal, the request will no longer be distributed to the application. However, when the transaction volume is very large, such as dozens to hundreds of transactions per second, the period from the load balance judgment abnormality to the distribution block has already caused a large number of transaction failures , cannot be accepted normally

Images