Research On OpenStack Based Clustered Cloud Platform With High Availability

In recent decades,cloud computing technology has developed rapidly and has been applied to every corner of life.With cloud computing technology,people can get IT services in a more convenient and cheaper way.At present,all major enterprises in the society rely on cloud computing technology to complete their own cloud business,to ensure that their business runs flexibly and efficiently.Most enterprises use a cloud platform provided by a third-party cloud computing service provider.However,the controllability of the third-party cloud platform is too low,and the user has limited operation and management rights to the cloud platform.Therefore,more and more users start to build their own cloud platforms.Among the many open source cloud computing technologies,OpenStack has been widely recognized for its open source and stable advantages.However,the cloud platform built with OpenStack cannot meet the high availability of servers and the high concurrency of server clusters.In view of the inability of OpenStack to achieve server high availability and clustering defects,this thesis proposes a high availability clustering solution based on Keepalived+LVS(short for Linux Virtual Server).Keeping the status of the load balancing server through Keepalived,when the load balancing host fails,quickly switch to the load balancing slave,and the load balancing slave continues to provide services,thus ensuring the high availability of the load balancing server;balancing the load through LVS the user request sent by the host is sent to the real server of the backend according to a certain rule,and the real server processes the user's request together,thereby realizing the high availability cluster of the real server.Finally,this thesis has conducted a lot of testing on the cloud platform and improved the cloud platform based on the test results.Through the high availability test of the cloud platform,it is found that when the MySQL service failure occurs in the real server,the availability of the cloud platform drops sharply.This thesis proposes an improved scheme: real-time monitoring of the running status of the real server MySQL service,the network service is automatically shut down when the failure of MySQL service is detected,and finally the purpose of improving the availability of the cloud platform is achieved.Through the cluster performance test on the cloud platform,it is found that the performance bottleneck of the cloud platform is the load balancing server.By improving the performance of the load balancing server running memory,the experiment proves that the cloud platform that improves the running memory of the load balancing server has been greatly improved on the concurrent processing capability and the response speed of the request.In order to verify that the high-availability cluster cloud platform can be put into production and use normally,this thesis applies the cloud platform to the road condition monitoring system to achieve the purpose of real-time monitoring of road surface conditions.