| 摘要 |
Embodiments of the present invention provide a data recovery manner, where a currently processed dataset is divided into n data blocks, and m redundant code blocks of the dataset are calculated, where m is greater than or equal to 1. The method further includes: when one or more bad blocks exist in the n data blocks, analyzing a position in which damaged data of each bad block in the n data blocks is located in each bad block; dividing, according to the analyzed position, each block of the n data blocks and each block of the m code blocks into h sub-blocks according to a same division rule, and in all sub-blocks included in the n data blocks and the m code blocks, using, according to the analyzed position, sub-blocks located in a same position in each data block and each code block as one sub-set; and for one sub-set with a bad sub-block, recovering damaged data of a bad sub-block in the sub-set by using data of an undamaged sub-block belonging to the sub-set with a bad sub-block. Correspondingly, the embodiments of the present invention further provide a data recovery device. The data recovery method and device according to the embodiments of the present invention can improve data reliability. |
