Research And Implementation Of Dockers Container Group Automated Deployment System

Traditional software organizations set development,testing,operation,and quality assurance as separate departments,which to a certain extent has brought great development and progress to the software industry.But at the same time,multi-departmental collaboration has also ushered in many challenges,such as resource independence,personnel management,and exchanges.In addition,with more and more software customers and more complex software requirements,the requirements for software development and testing are getting higher and higher.Development and testing require many servers to work in parallel to ensure progress.However,most small and medium-sized enterprises do not have enough physical server resources to satisfy the complicated testing work.This topic focuses on the pain points of the enterprise team and develops an automated deployment system for Docker container clusters,and introduces the entire process of customized automatic deployment of enterprise-level docker combined with product systems,and optimization after deployment.This topic uses DevOps theory,combined with Docker technology,uses Docker Swarm tools and System Imager tools to build an automatic deployment system for Docker clusters,which realizes multi-host one-click automatic deployment,which not only avoids human errors,but also ensures the diversity of container versions after deployment.The consistency with the environment of the same version also improves the efficiency of the team.It can be achieved within a few minutes to complete a few hours or even a week of environmental deployment in the traditional way.After this automated deployment is completed,developers can use the platform for integration testing,and testers can use this deployment system for system testing to more effectively ensure the quality of software product release.In the subject research,first analyze the environment segmentation characteristics of Docker containers,and combine our actual project environment on this basis.According to different versions of system files,write them into the file pool according to Dockerfile,so that different hosts can obtain different information from the file pool.The target file,then the base image and the file are combined to produce the target image,and then different versions of the container group environment are automatically deployed according to the target image.Secondly,after the deployment of the container cluster environment,there will be bottlenecks in the machine performance,so it is necessary to set performance benchmarks and visualization graphs,measure them based on visualization,and perform subsequent optimization solutions.Finally,optimize the scheduling strategy of the deployed Docker Swarm node,use the binary comparison method,add a certain subjective tone operator,and make more reasonable scheduling decisions when the Swarm node weights are the same,to maximize the single Load balancing of cluster nodes.The research results of this subject have been put into practical application within the company.The solution of automatically deploying container groups in different versions is a unique innovation,which greatly improves the work efficiency of team members.In addition,the binary comparison method optimizes the scheduling strategy of Docker Swarm and also ensures the load of the host machine and reduces the company's cost.Finally,through Portainer,a visualized operation and maintenance solution was provided for the operation and maintenance,and the research on the subject was truly put into production.