自OpenStack开源项目诞生以来, 云计算技术与企业解决方案得到了飞速发展, 云平台更是在企业内部大面积部署, 随着规模化与集成化的高度发展, 也带来一系列问题.其中单节点的稳定性与高可用性尤为重要, 是保障企业持续性、高容错、高并发服务业务的关键.针对OpenStack各相关服务组件在长时间运行期间容易导致的单点故障 (SPoF) 问题, 结合多种负载均衡、IP漂移、网络路由冗余 (VRRP) 、数据库集群 (galera) 、分布式集群 (Ceph) 等技术, 提出一种基于OpenStack高可用平台的分布式存储方案.该方案能在单节点发生故障、网络拥堵、数据高爆发等情况下通过业务的负载均衡、服务实例故障转移、热备份和热迁移、软硬件冗余等方式实现整个分布式平台的高可用, 具有测试环境和生产环境下的可行性.%Since the birth of OpenStack open source project, cloud computing technology and enterprise solutions have developed rapidly, and the cloud platform have been deployed largely within enterprises.With the high development of scale and integration, a series of problems have been brought out.The stability and high availability of single node is particularly important, which is the key to ensure enterprise persistence, high fault tolerance and high concurrency service business.For a single point of failure (SPoF) easily caused by various OpenStack related components during a long period of operation, combined with a variety of technologies such as load balancing, IP drift, network routing redundancy (VRRP), database cluster (galera) and Ceph, we propose a distributed storage scheme based on OpenStack high availability platform.The scheme can realize the high availability of the whole distributed platform in the case of single node fault, network congestion, high data outbreak by means of load balancing, service instance failover, hot backup and hot transfer, software and hardware redundancy, which is feasible in both the test environment and the production environment.
展开▼
