OpenStack Fault Diagnosis Based On Machine Learning

With the rapid development of cloud computing technology,more and more enterprises and individuals put their services and systems on the cloud platform.The cloud platform can quickly build the development environment and allocate computing resources according to the specific needs of different users.OpenStack is the technology foundation of most public and private cloud platforms.With OpenStack,service providers and enterprises can implement cloud services internally.For cloud platforms,ensuring system reliability is very important.For many enterprise-scale large-scale cloud computing services,there are tens of thousands of nodes,and so many nodes are very likely to cause system failure.Coupled with the complexity of the OpenStack system itself,the complexity of each node task makes it difficult to detect faults in time,and it is difficult to locate faults.This paper proposes an algorithm for fault diagnosis.This algorithm only needs the original log and can be implemented without any marked data.First,the log format is processed,the log format is unified,and the log information is stored in the database for visual analysis.Then a time window is established,and features are extracted in the time window.After parameter selection,cluster analysis is performed according to the extracted features to find the fault time period.At last,we extract the log of the fault period and perform text analysis to find the component and cause of the fault.The experimental results show that this algorithm has very good effect and has high practical value.