Fault recovery method and system

The invention discloses a fault recovery method and system applied to a pre-established storage system. The storage system comprises at least one storage node group. Each storage node group comprises M first storage nodes executing storage work and N second storage nodes not executing the storage work, wherein M and N are both positive integers. The method comprises: when any first storage node has a fault, determining that the first storage node with the fault is a faulted storage node, wherein a storage node group corresponding to faulted storage nodes is a faulted storage node group; selecting one second storage node in the faulted storage node group as a substitutive storage node according to a preset principle; and replacing the faulted storage node with the substitutive storage node, and completing the storage work that the faulted storage node needs to execute. Therefore, the workload of operation and maintenance is reduced, the recovery time of system faults is shortened, the data volume of migration is reduced, and the risk caused by system recovery is lowered.