云环境 MMM 数据库集群双主多从架构模型构建磁

摘要

As the running of Hunan rural integrated services platform of national rural and agricultural information demonstration province and the fundamental platform of the Hunan agriculture IOT ,new requirements are proposed by big data operation and safety for agriculture .For the requirement ,an optimization model is proposed of master‐master replication manager for MySQL on Linux cluster for RedHat and MySQL .Three design models are mainly components of the model , including ,the front layer to realize the main switch from MySQL‐Proxy ,the data service layer to integrate the service node , monitoring and single image ,the share data layer to realize remote data synchronization ,backup and disaster recovery .The tests have proved that not only the problem of the performance bottleneck of the whole database cluster can be solved by in ‐creasing servers by the way ,but also the effective service can be achieved through a large of reading and writing under not in ‐creasing server hardware performance .The model can enhance the agricultural data security ,robustness and reliability under the cloud .%随着湖南国家农村农业信息化示范省综合服务平台与农村物联网基础平台的运行，对农业大数据的操作与安全提出了新的要求。为了适应这一要求，针对 RedHat 与 MySQL 系统，提出一个基于 Linux 系统可扩展的 MySQL 数据库集群双主多从架构优化模型。该模型主要由三个设计模块组成，即前端代理层实现主备热切换与读写分离；数据服务层对服务节点进行整合、监控与单一镜像；数据共享层实现远程数据同步备份容灾。测试证明在不提高服务器硬件性能的条件下，仅通过增加服务器数量就能解决整个数据库集群的性能瓶颈问题，而且能够在大批量读写数据的环境下进行有效的服务，加强云平台下农业数据的安全性、健壮性及可靠性。