Design And Implementation Of High Availability Service Fault Recovery System For OpenStack

With the popularity of cloud computing as a business computing model,Enterprises no longer need to pay attention to the underlying IT infrastructure,but only need to take corresponding services from cloud computing to meet their business requirements.As the standard of open source cloud computing,more and more enterprises are transferring services to OpenStack cloud platform.OpenStack provides convenience for cloud computing,but there are also some problems.The control node has intensive logic processing.As the control and management hub of the whole system,its performance is limited for a single control node,which can not meet the large-scale computing cluster and intensive traffic.OpenStack's control layer and computing layer run a large number of service components,each component may fail,and face the risk of single node failure.Whether node or service component failure,it will reduce the user's sense of experience.Therefore,improving the scalability and high availability of OpenStack is one of the key research issues.Aiming at the above problems,a high availability fault recovery system for OpenStack is designed and implemented.The main work is as follows:1.Combining high availability cluster with OpenStack distributed system,the scheme of high availability cluster and fault recovery system is given.The scheme transforms singlepoint control node into distributed multi-control node cluster system,manages system components by cluster,and applies high-availability load balancing to OpenStack components.It solves the bottleneck and single-point failure.For node failure,it designs automatic recovery and migration function of fault,to improve system scalability and high-availability;2.Based on the above scheme,designed a variety of cluster high-availability cooperation mechanisms,which are high-availability control layer cluster cooperation system,including cluster management system,multi-node dual-link heartbeat;high-availability computing layer cluster cooperation system,multi-service multi-virtual IP load balancing agent system and high-availability fault recovery and self-migration system.3.Completed the implementation of the high availability fault recovery system function module,built the experimental cloud platform and carried a variety of functional validation and performance tests for the system.The results show that the functions of the high availability fault recovery system cooperate with each other,realizing cluster high availability and fast fault transfer and recovery,and ensuring the stable provision of cloud services.