The utility model discloses a cluster storage system capable of automatically repairing a fault node, and aims at providing a cloud storage system capable of self-repair. The cluster storage system comprises a storage cluster, a monitoring and management server and multiple standby servers, wherein the storage cluster comprises multiple storage servers, the monitoring and management server is respectively connected with all the standby servers and all the storage servers in the storage cluster, the monitoring and management server is equipped with an input/output interface communicated with the external, and a user stores data in the storage servers or reads data from the storage servers via the input/output interface of the monitoring and management server. At the same time, the monitoring and management server monitors healthy states of each storage server, and if fault occurs a certain storage server, one standby server is used to replace the storage server to ensure normal operation of the cloud storage system. The cluster storage system capable of automatically repairing the fault node is suitable for all cloud storage configurations.