Design And Development Of Resource Management Platform For Life Cycles Of Servers In A Cluster

Like human beings,servers are also alive.There are many states in the whole life cycle of the server from the release to the elimination,which is similar to the birth,the aging and the death of human beings.At present,the Internet industry has been developing rapidly.Large Internet companies own millions of servers,and many small-and medium-sized enterprises usually have tens of thousands of servers.That is to say,the cost on serversposes a great burden to an enterprise.This paper aims to manage the entire life cycle of servers.Based on the complete data of machine resource in the life cycle,an efficient resource application,delivery,utilization and recycling process is established to make the machine operate quicklythrough standardized construction,data collection and platform-based solutions,thus satisfyingdata capacity expansion,rapid business iteration and other purposes.The research focuses on resource view,resource flow,empty machine recycling,automatic maintenance and resource management platform,which have involved the full life cycle of machines,aiming to reduce enterprises' cost on servers and improve the efficiency of resource flow.Resource view provides multi-dimensional statistical analysis and auxiliary decision-making of machine resource metadata.Resource flow dominates machine life cycle management,including initialization,standby pool management,allocation,maintenance,elimination,etc.,to solve the problem of inefficient resource management.A mechanism for empty machine recycling is set up,in which,a white list is added to cold standby machine,and empty machines that are not in white list will be automatically recycled after a certain period of time.According to the specifications of accessory and fault detection of server manufacturers,the fault machine with one machine and many parts should be maintained automatically on demand.The resource view takes puppet to collect machine information,Python to process the reported information,Python's Django framework to implement enterprise CMDB,and mysql to store data.The API of resource flow,machine maintenance,empty machine recycling and automatic maintenance is developed using PHP,and some libraries of Python are adopted for processing.Based on the abundance of Python libraries,processing the libraries with Python data can minimize the workload;at the same time,taking advantage of the high efficiency of PHP script language,PHP is used to develop API.The core value of this paper is that the construction of data capacity and the improvement of resource flow efficiency can ultimately reduce enterprises' cost on servers.In the aspect of data capacity building,the good state of machines in their life cycle is maintained,the analysis tools for data query is provided.In the aspect of efficiency improvement,the efficiency of resource delivery,expenditure,audit,recycling and other circulation processes is optimized by means of automation and platform approaches.