Design And Monitoring Of Server System And Network System In CAFe

CAFe(The Chinese ADS front-end demo)superconducting linac is supported by Strategic Priority Research Program ADS(Accelerator Driven Sub-critical System)of the Chinese Academy of Sciences.It is also the critical part of twelve-five large scientific facilities.Control system of CAFe,as a core component of it,plays an important role in ensuring stability and reliability of the whole facility,which are the server system and the network system take the control software running and the data interactive support platform.In this paper,based on the cluster technology,optical network technology and Zabbix system monitoring technology,the technical improvement and function optimization of the server and network system are made from the aspects of system design,scheme implementation and running state monitoring,which improves the stability and maintainability of the server and network system of the CAFe control system.The main work of this paper is as follows:Firstly,In order to solve the problem of single server system running in the field,this paper proposes a technical proposal to build Open Stack,a distributed private cloud platform,and replace the traditional single-server system operation mode.Compared with traditional server system,the Open Stack platform has the following three advantages:1.it integrates hardware resources and disk arrays of multiple servers to realize online allocation of computing resources;2.Graphically display the usage status of server hardware resources and all virtual machine instances;3.Using the online hot migration technology of virtual machines,the operating stability and data security of the server platform have been greatly improved.Secondly,redesigned and transformed the existing network system.In response to frequent network disconnection and maintenance difficulties,CAFe’s network system was redesigned.Compared with the previous network system,the current network system has two new features: 1.Adopt a three-layer structure,use two redundant core switches,and use optical fiber transmission for the main data channel;2.Use SNMP protocol to automatically obtain the operating status data of all network switches.Finally,aiming at the failure of server software and network disconnection in the operation site of the device,a monitoring system based on Zabbix is designed and deployed.Distributed remote monitoring functions are implemented for the running status of the designed server cluster and network system,which provides technical guarantee for the smooth operation of the CAFe control system.At present,the relevant on-site reconstruction projects of the private cloud platform and network system have been completed,and the Zabbix monitoring system has completed functional tests in the actual operating environment and has been put into field use for one year.The Open Stack platform and network system showed good expected results in terms of stability and reliability,laying a solid foundation for the stable operation of the CAFe control system.