基于新型中间件的大型望远镜控制计算机集群资源监控系统

As the telescope of the highest spectrum acquiring rate, GuoShouJing(also named LAMOST) Telescope combines a large aperture with a wide field of view. It has been already successfully observed and distributed more than 4.1 million spectra, since October 2001. . LAMOST has very complex and sophisticated hardware and software systems, and contains 8 subsystems. Each sybsystem has many computers, and these computers cooperate with each other to jointly complete the observation task. All these computers have different types, different operation systems, and are placed in different positions. For such a complex computers cluster, wes can not be testing and monitoring fully by manual. Thus, the cluster resource monitoring system( especially in monitoring critical node) plays an important supporting role for running stability and efficiency of the large telescopes such as LAMOST.. The project will design and develope a resource monitoring system for computer cluster of the large telescope. We prepare using the new lightweight message middleware technology(ZMQ), and develope real-time resource message acquisition program, collecting computers' message(CPU usage, memory usage, hard disk space, NIC uplink and downlink transmission rate, etc.) on each node, and transfer these message to the monitoring server by ZMQ. Monitoring Server is responsible for the storage of classified information, real-time display and alarm functions in accordance with the conditions set. Thereby improving the observing efficiency of large telescopes.

郭守敬(LAMOST)望远镜兼顾大口径与大视场，是光谱获取率最高的天文望远镜之一。自2011年以来已获得目标光谱超过410万条。. LAMOST望远镜设计结构复杂，由八个子系统组成。各子系统都由大量计算机集群来完成工作。这些计算机种类繁多，位置分散，操作系统各异。对于如此复杂的计算机集群，不可能完全由人工实现检测和监控。因此，集群资源监控尤其是对于关键节点计算机的监控对于LAMOST这类大型望远镜的稳定运行和观测效率提高起到了重要的辅助作用。. 本项目设计开发计算机集群硬件资源监控系统。拟定采用新型消息中间件技术ZMQ，各节点上开发采集程序，采集计算机的实时资源(CPU使用率、内存占用率、硬盘剩余空间、网卡上下行传输速率等），并通过ZMQ传输给监控服务器。监控服务器负责实现对信息分类存储、实时显示以及根据设置条件预警等功能。从而提高大型望远镜运行和维护效率。

现代大型天文望远镜的控制一般由多台独立（或组成集群）的计算机来完成。这些计算机的性能、效率与可靠性直接影响整个望远镜的稳定运行。因此，需要研制一套实时对各控制计算机（节点）硬件资源信息进行采集、存储、监控的软件系统，并提供一定的预警功能。这样的系统可以有效排除隐患进而提高望远镜整体观测运行效率。本项目在深入分析郭守敬（LAMOST）望远镜观测运行需求的基础上，采用异步协程技术，设计和开发了一套基于Python语言的硬件资源监控系统软件，系统可以高效稳定地采集获取计算机各种状态，也可以获得相应部件的实时信息，并提供了多种人机交互方式，为后续开发提供了扩展接口。该系统部署于LAMOST环境中，实际运行表明整个系统取得了较好的效果。