基于Kubernetes的云原生数据库集群部署方案  

Cluster deployment scheme of Cloud Native Database based on Kubernetes

作　　者：马晋 赵思亮 赵芳 MA Jin;ZHAO Siliang;ZHAO Fang(Chongqing Meteorological Information and Technical Support Center,Chongqing 401147,China)

机构地区：[1]重庆市气象信息与技术保障中心,重庆401147

出　　处：《自动化与仪器仪表》2022年第10期55-59,共5页Automation & Instrumentation

基　　金：重庆市气象部门智慧气象技术创新团队项目(ZHCXTD-202018)。

摘　　要：针对气象结构化数据在单机数据库环境下存在缺乏灵活扩展和资源利用率低下的问题,提出了一种基于Kubernetes的云原生数据库部署方案,将MySQL数据库运行在Kubernetes+容器的环境,通过集群部署调优,形成分布式云原生数据库服务能力。该方案利用Shared Everything+Shared Storage的存储计算分离架构实现资源池化高效管理;利用Shared Nothing的分布式架构,实现数据水平分片、水平扩展。测试结果证明,该方案具有较好的横向扩展和数据查询性能。In view of the lack of flexible expansion and low resource utilization of meteorological structured data in the stand-alone database environment,a deployment scheme of cloud native database based on Kubernetes is proposed,which runs MySQL database in the environment of Kubernetes+container and optimizes through cluster deployment to form the serv-ice capability of distributed cloud native database.The scheme uses the storage and computing separation architecture of Shared Everything+Shared storage to achieve efficient resource pooling management,and uses the distributed architecture of Shared Nothing to achieve horizontal data fragmentation and horizontal expansion.The test results show that the scheme has good performance of horizontal expansion and data query.

关 键 词：Kubernetes 容器编排 云原生 分布式数据库 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]