Selfreparing method of nodes in cloud computing cluster

The invention provides a selfreparing method of nodes in a cloud computing cluster, and relates to the technical field of cloud computing. The selfreparing method comprises the steps that first, a cloud management system monitors the operating state of each node; second, when breakdown happens to one node, a system starts a failover process, and all cloud virtual machines on the node are automatically started on other nodes which are sufficient in resource and normal in operation; third, a management system controls the node to be started again through an IPMI system, and the state is detected; fourth, if the state is normal after starting, the node is continuously added into the cluster for being used; fifth, if restarting repairing fails, the IPMI system is used for controlling the node to achieve network starting, then the automatic finding and installing function of the cloud management system is used for reinstalling systems and settings to the node, and the node is added in the cluster after setting is