Alibaba Group Holding Limited
System and method for facilitating storage system operation with global mapping to provide maintenance without a service interrupt

Abstract:

The system receives a request to write data with an associated LBA. The system stores, in a global mapping table, a mapping of the LBA to a PBA assigned by a master node. The PBA is associated with a first storage drive in a first collection of storage drives, which includes a first set of storage drives and a plurality of redundant storage drives. The system writes, based on the PBA, the data to the first storage drive. In response to detecting that the first storage drive is defective, the system replaces the first storage drive with a redundant storage drive by reconstructing data stored in the defective storage drive based on a copy of the data. The system thus allows the first collection to remain online while replacing the defective storage drive.