一种高可用性集群系统的架构  

Framework of High Availability Cluster System

作　　者：钱勇[1] 吴坚[1] 

机构地区：[1]西南科技大学计算机学院,四川绵阳621010

出　　处：《兵工自动化》2008年第5期64-66,73,共4页Ordnance Industry Automation

摘　　要：基于J2EE技术的B/S架构的客户控制端实现方案,其NodeN为集群的节点,Monitor和HA为一个节点上的2个"精灵"进程。HA进程为当前节点的活跃的精灵进程,负责与其管理的服务进程的代理进程交互。节点与节点间的关于成员籍和服务的状态等信息的交互则通过Heartbeat进行。Monitor进程与HA进程互为备份,时刻监视HA进程的状态,避免由于HA进程意外终止而给系统造成SPOF。A B/S mode framework of console based on J2EE for server console control, whose nodeN is the node of cluster, while Monitor process and Ha process are two daemons in one node. Ha process is the active daemon on a node, which is responsible for communications with services＇ agents related to it. The inter-node communication about memberships and statuses of services are through heartbeat. Monitor and Ha process are backup for each other, and Monitor always checks Ha process＇s status, avoiding introducing SPOF （Single Point of Failure） to cluster systems for accidentally killing of Ha.

关 键 词：集群 控制端 J2EE HEARTBEAT 

分 类 号：TP393.02[自动化与计算机技术—计算机应用技术]