Design And Implementation Of Container Cloud Platform Based On Kubernetes And Docker

With the vigorous development of high technology information technology,the dividends brought by the achievements of scientific and technological progress have penetrated into every household,and everyone can feel the influence and change of the power of science and technology on our lives.The momentum of scientific and technological services has surpassed the momentum of scientific and technological research,and the number and types of various application services are also increasing with the advancement of technology,and the functions that can be provided are constantly improving.The user's satisfaction with various application services continues to increase,and at the same time they also put forward more requirements for service extensions.Many current services are developed using the microservice architecture,which contains many mutually coupled components.This kind of service architecture leads to the adjustment of many modules involved in each service upgrade and maintenance,and each module is dependent on each other,which will eventually lead to slow service upgrades and unable to meet rapid iterative upgrades.At the same time,the speed and success rate of automatic deployment of the business will be very low,which makes the service stability poor,which requires increase maintenance costs.The use of cloud platform container virtualization and automatic deployment technology is a very good choice.This thesis uses Docker container virtualization technology to accomplish the low-overhead runtime environment isolation and multiple environment operation of applications,and designs and implements an application-centric private container cloud platform with Kubernetes container orchestration engine as the core.This paper firstly introduces the background and significance of the subject research,describes the related frameworks and technologies involved,analyzes the requirements for building a container cloud platform based on Kubernetes and Docker in many aspects,and makes a detailed design and implementation.The cloud platform itself is installed through containers,and the components are also provided as containers to the outside world,reducing external pollution and intrusion into the cluster host.The container cloud platform uses Nginx reverse proxy server externally,and through the design of multiple service discovery and exposure methods,the container applications running on the cloud platform can provide services externally independently,and the load balancing of network access is designed to improve the availability and throughput of the cloud platform.The container cloud platform uses Harbor private image repository to store,upload and distribute container images internally,and achieves restricted access to image resources by external users through permission control,which solves the security problem of image resources.The cloud platform uses the Volume container data volume to perform persistent operations on logs and other information generated when the container is running,which solves the problem of data loss caused by container downtime.This article has made improvements in security monitoring based on the native Kubernetes,using the Prometheus tool to monitor the cluster nodes and the container data running on the nodes in real time.In terms of cloud platform resource scheduling,a preemptive scheduling strategy is designed and implemented in conjunction with production application scenarios.The policy allows users to customize the Pod priority,and in the scheduling process,the cluster will be scheduled in order according to the Pod priority.When the cluster resources are insufficient,the policy supports hanging low-priority Pods to release resources,and high-priority Pods will seize resources,which realizes the function of efficient utilization of cloud platform resources.In order to demonstrate the effect of the implementation of the cloud platform-related functions,the code project was completed through Spring Boot,Thymeleaf and Mybatis frameworks to interact with My SQL database for data persistence operations.This paper designs and implements an automated deployment cloud platform based on Kubernetes and Docker virtualization technology,which provides a unified runtime environment for container virtualized application services,effectively solves the problem of inter-module coupling,increases the stability of service operation,and can more perfectly solve the automatic deployment and maintenance of large-scale projects or services in production environments.It provides an example for the purpose and process of automated container deployment.This thesis concludes with a test of the functionality of the container cloud platform based on Kubernetes and Docker,and the test results show that the container cloud platform can provide stable and effective containerized application management services.