Design And Implementation Of Server Cluster Monitoring System For Webjoin

With the development of Internet Technology, especially the rapid rise of mobile internet, the number of users who access internet has been made a sharp increase, which causes a huge performance pressure to internet central servers. To solve this problem the Cluster Technology begins to gain widespread use and development, through a server cluster system which consisting of multiple application servers to play a single application server unmatched performance, availability and scalability. However, due to the independence and loose among servers in the cluster server system, server cluster monitoring system is required to monitor, maintain and handle exceptions.However, with the continuous evolution and development of server cluster system, the number of server node is increasing rapidly and the type of it also varies into many kinds. In addition, monitoring range of server cluster extends from traditional hardware-lever to the operating system, applications, services and other software levels. And the duty of cluster monitoring system also extends from data acquisition and display to abnormal discovery and processing. In order to meet the new demands, this paper designed and implemented a new server cluster monitoring system.In this paper, the server cluster monitoring system is divided into three data layers from the bottom to top, which are the raw data layer, the structured data layer and data application layer. The raw data layer where monitoring data is collected uses asynchronous message method for distributed data aggregated collection. The structured data layer where the initial processing of the raw data is complete organizes raw data into structured data according to the categories of data, and realizes the storage of heterogeneous data. The data application layer whose feature is the use of monitoring data completes two functions, namely the deal with abnormal situations and visual display monitoring data. In addition, this paper designed and implemented the server cluster monitoring system for the actual server cluster system-WebJoin.