In a method for identifying a sub-health OSD in a data storage system, a first OSD receives a write request and replicates that write request to a second OSD. Subsequently, the first OSD sends a report regarding the write request to a management node in the system. The report includes an identifier of the first OSD, an identifier of the second OSD, and health status information of the second OSD. The management node determines, based on the received report, whether the second OSD is in a sub-health state.