Containerized Design And Practice Of Sino Probe Data Management System

This thesis researches the cloud computing technology and container technology,analyzes the existing problems of the Sino Probe data management and sharing system,and proposes a solution to address these problems using cloud computing and container technologies.Using this solution,this thesis designs and implements a new container cluster management system,and containerizes the Sino Probe data management and sharing system.The existing Sino Probe data management and sharing service system is a classic traditional web application architecture that applications are directly deployed in the operating system of each server.But compared with containerized way,this mode has many disadvantages as follows:1)Applications are directly deployed in the operating system.When the function of the program is updated,the code file of the program needs to be replaced.It is easy to make mistakes when someone deploy it but not the developer.2)The storage mode is NFS dedicated storage device.When the data volume reaches the upper limit of storage,expanded this kind of dedicated storage device needs to buy the special device,which always be high price.Besides,when the device is damaged,the data is difficult to recover,so the disaster recovery rate is low.3)No suitable monitoring mechanism,it is difficult to observe the running state of clusters and applications.Aiming at the existing problems of Sino Probe data management and sharing service system,this thesis designs a new containerized cluster management system by using cloud computing technology,docker container technology and rancher container cluster management platform.Realized the containerization of Sino Probe data management sharing service system.The main achievements of this paper are as bellow:1)Designed and developed Dockerfile both of ‘Sino Probe data management and sharing service application' and database,built their images.Built Rancher container cluster management platform by using three servers in the lab,and use it deploy theimages to achieve containerization of Sino Probe data management and sharing service application.2)Integrated the Longhorn distributed block storage management system and Minio distributed object storage technology into he rancher container cluster management platform,realizing the distributed storage management system of container cluster.This makes storage resource management simpler,storage resource capacity easier to expand or shrink,and storage resource more secure.3)Integrated Nexus3 image warehouse,gitlab code warehouse and Jenkins continuous integrated management technology into the rancher container cluster management platform,realizing deployment and update of the application workflow pipeline,simplified the application management mode.From applications development to deployment are all operated by developers,which is more effective and efficient.4)Integrated Prometheus monitoring technology and grafana data visualization technology into the rancher container cluster management platform to realize the monitoring of the container's running state,and give the visual chart display.The main achievement of this thesis is to introduce the newly cloud computing and container technologies including Rancher container cluster management technology and Kubernetes cluster management technology into the Sino Probe data management and sharing service system,and to build a container cluster management platform including the unified resource scheduling module,continuous integration module and monitoring module.Compared with the existing Sino Probe data management system,the new one reduces the complexity of system management and improves the flexibility,disaster recovery rate and scalability of the system.The new system also gives a simple Rancher container cluster management web application,which is more convenient and efficient than directly managing the application at the operating system level.This thesis makes a significant exploration and demonstration for the application and promotion of cloud computing container technology in the field of geoscience data management and data analytics.