The Preliminary Exploration Of G/S Distributed Cluster Monitoring

After the information age, it is hoped that a better spatial information services. For example, in high-definition resolution, massive resources. The emergence of G/S model (Geo-Information Browser/Spatial Information Server) create new opportunities for the Spatial information technology.G/S model is distributed cluster systems based internet network. The Three-dimensional geo-information browser is the realization of dynamic spatial information access, three-dimensional performance, spatial information query and spatial analysis functions. Spatial Information in Geographic Data Geography Markup Language HGML, constitute the performance of spatial information exchange, organization and management of standards and formats. Distributed clustering approach to spatial information to build server-side, responsible for data storage, processing and transmission.Distributed cluster system as a result of the vast and complex subject and we need to monitor its operation, the timely detection of failures, performance issues, and server group to do a unified configuration management. This article analyzes the current mainstream of the server monitoring technology to study the working principle of distributed clusters, build a Xen virtual experiment environment, the "application-based, customer experience and business-oriented data" means the monitoring, On this basis, further analysis of the monitoring system requirement specification for the system design process and follow the rules of software engineering, the development of G/S distributed cluster monitoring system. After testing in the experimental environment, the system achieved the desired results.Monitoring system using Linux Bash Shell language, combined with awk language, and deployed under a variety of Linux open source software. The system monitoring consists of four parts:monitor module, analysis module, alarm module, modules, and a unified log configuration management module Distributed system in accordance with the provision of a list of all the cluster nodes, one by one into the cycle of detection. Therefore, apart from an overall monitoring, for all real servers in order to do independent testing of the cycle. At the same time, the contents of each response with a correct copy of the content to check, to determine the completeness and correctness of the content. For the dynamic pages by the proposed language from the database to a number of key business data, but also access to local, and to analyze, organize, with a view of the average volume of business expected to determine some of the hidden potential, it is difficult to use technology in ways to monitor the issue. Detection of cases found in error, he immediately called the police function, to remind the staff on duty at the scene. At the same time, the problems will be recorded in the log file, at the same time, a summary of error conditions to the way e-mail to inform the relevant managers.Subject reached the following results:1. analysis of the current mainstream of the monitoring technology, a "firmly rooted in the user experience and business analysis-oriented" concept of the new monitor.2. in-depth study of the working principle of distributed clusters proposed for G/S distributed cluster monitoring the basis of the theory.3. set up a Xen virtual experiment platform, the two virtual sub-network, 20 Linux servers, as G/S distributed server cluster environment for the experiment.4. summed up the G/S cluster of distributed monitoring and control system requirements specification, design a workflow monitoring system, follow the rules of software engineering, has developed an initial G/S distributed cluster monitoring system.5. Pass the test to verify the monitoring system of G/S system is the actual role.