UNIX服务器集中监控系统的设计与实现

发布时间：2018-07-05 06:31

本文选题：集中监控 + 守护进程　；参考：《中南林业科技大学》2013年硕士论文

【摘要】：随着信息技术的高速发展,为了提高业务处理效率,各行业数据集中的进程越来越快,业务种类也越来越多,服务器承载的工作就显得非常重要。UNIX操作系统具有独特的稳定、安全特性,成为大、中型服务器的首选操作系统。服务器承载的业务范围越大、数据越多,对系统的稳定运行率要求就越高；服务器的全天24小时运行,仅依靠人工的监控和管理已不能达到要求,企业需要一个自动地、准确的、实时的监控系统,来管理UNIX服务器群,以代替系统管理复杂的人工操作。目前流行的代理监控软件是大公司集成在服务器上的插件,修改难度大,模式单一,无法满足客户的特殊需求；企业自行开发的无代理式监控软件,设计虽然简单,但服务器资源占用率比较高,可移植性不强,精确度不高。针对以上的问题,论文结合代理和无代理监控方式的特点,提出一套新的UNIX服务器集中监控设计方案,通过把监控需求进行详细分类,利用CRON守护进程原理,实现了监控数据的高精度实时采集；通过在监控主机设置定时器,主动抓取数据,实现了数据传输和数据判断工作,这样减轻了UNIX服务器群的负担,节约了系统资源。部署在服务器上的脚本使用常用的UNIX系统操作命令编写,简单易懂,便于管理员维护和修改,增强了集中监控系统的可移植性；在监控主机上对监控数据进行二次开发,不影响UNIX服务器的正常运行,还提高了监控系统的可扩展性。最后,开发实现了集中监控系统,能准确监控服务器运行情况,且资源占用率低、可移植性好。经试运行一段时间后,基本上满足了大部分企业客户的需求,并能为系统管理员对系统运行评估提供了准确的信息。
[Abstract]:Along with the high - speed development of information technology , in order to improve the efficiency of business processing , the process of data concentration in various industries is faster and more , and the service type is more and more important . The UNIX operating system has unique stability and security characteristics , becoming the preferred operating system of large and medium servers . The larger the operating range of the server , the more data the more data is , the higher the requirements for the stable operation rate of the system ;
The whole day 24 - hour operation of the server , relying only on manual monitoring and management has not met the requirement , the enterprise needs an automatic , accurate and real - time monitoring system to manage the UNIX server farm instead of the complicated manual operation of system management . The current popular agent monitoring software is the plug - in which is integrated on the server by the large company , the modification difficulty is large , the mode is single , and the special needs of the customer cannot be met ;
The enterprise self - developed agentless monitoring software , the design is simple , but the server resource occupancy rate is relatively high , the portability is not strong , the precision is not high .

Aiming at the above problems , the paper puts forward a new set of centralized monitoring and monitoring schemes for UNIX servers based on the characteristics of agent and agentless monitoring , and realizes the high - precision real - time acquisition of monitoring data by using CRON daemon principle .
By setting the timer on the monitoring host , actively capturing the data , the data transmission and the data judging work are realized , the burden of the UNIX server group is reduced , the system resources are saved . The scripts deployed on the server are written by the usual UNIX system operation commands , are simple and easy to understand , facilitate the maintenance and modification of the administrator , and enhance the portability of the centralized monitoring system ;
secondary development of the monitoring data on the monitoring host does not affect the normal operation of the UNIX server , and also improves the expandability of the monitoring system .

Finally , a centralized monitoring system is developed , which can accurately monitor the operation of the server , and the resource occupancy rate is low and the portability is good . After the trial operation for a period of time , the requirements of most enterprise customers are basically met , and accurate information can be provided for the system administrator to run the evaluation system .
【学位授予单位】：中南林业科技大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP277;TP368.5

【参考文献】