DISK ARRAY HAVING MIRROR CONFIGURATION AND REBUILDING METHOD THEREFOR,申请号US201615046931-传众专利搜索

DISK ARRAY HAVING MIRROR CONFIGURATION AND REBUILDING METHOD THEREFOR

申请公布号：US2016162360(A1)

申请号：US201615046931

申请日期：2016.02.18

申请公布日期：2016.06.09

申请人：

International Business Machines Corporation

发明人：Murakami Yoshikuni

分类号：G06F11/14;G06F3/06

主分类号：G06F11/14

地址：Armonk NY US

摘要：A disk array includes a buffer to temporarily store data (data X) from a host, a redundant first and second disk having a mirror configuration, and an auxiliary disk. A controller controls reading and writing of data X in response to a request. The controller is configured to: (a) copy to the auxiliary disk, data X stored in the buffer for rebuilding, in parallel with reading and writing data X, when the second disk fails; (b) log copied disk locations of data X copied in (a); (c) return a response to the host after completion; (d) reference a redundancy management table before copying in (a) to determine whether data X stored in the buffer has been copied; and (e) bypass copying in (a) and return to the host a response when current data X has been copied in (d).

主权项：1. A disk array having a mirror configuration to rebuild, the disk array comprising: a buffer to temporarily store data (data X) in a read or write request from a host (read/write request); a redundant first disk and second disk having a mirror configuration; an auxiliary disk used to perform rebuilding when the second disk fails; a redundancy management table to log disk locations of data stored in the first disk and copied to the auxiliary disk; and a controller to control reading and writing of the data X between the buffer and two redundant disks in response to a read or write request, the controller configured to: (a) copy to the auxiliary disk, the data X stored in the buffer for rebuilding, in parallel with reading and writing the data X, when the second disk fails; (b) log, in the redundancy management table, copied disk locations of the data X copied in (a); (c) return a ‘read/write request complete’ response to the host after completing the copying in (a); (d) reference the redundancy management table before the copying in (a) to determine whether the data X stored in the buffer has been copied; and (e) bypass the copying in (a) and return to the host a ‘read or write request complete’ response when a current data X has been copied in (d).