公开(公告)号：US4761785B1

公开(公告)日：1996-03-12

申请号：US87324986

申请日：1986-06-12

Applicant: IBM

Inventor： LAWLOR FRANCIS D ,  STEWART TERRENCE J ,  CLARK BRIAN E ,  SCHMIDT-STUMPF E ,  TIMMS JR GEORGE D

IPC: G06F3/06 ,  G06F11/10 ,  G06F11/14 ,  G06F11/20 ,  G06F12/16 ,  G11B20/18 ,  G11C29/00

CPC classification number: G06F11/1076 ,  G06F11/14 ,  G06F2211/104 ,  G11B20/1833

公开(公告)号：CA2044521A1

公开(公告)日：1991-12-22

申请号：CA2044521

申请日：1991-06-13

Applicant: IBM

Inventor： BOND MILTON F ,  CLARK BRIAN E ,  MCROBERTS RAYMOND S

IPC: G06F3/06 ,  G06F11/10 ,  G11B20/18 ,  G06F12/00

Abstract: A storage management mechanism resident on a storage controller maintains parity records on the storage units it services. The storage management mechanism includes a status map indicating, for each data block, the location of the corresponding parity block, and the status of the data block. If a single storage unit fails, the system continues to operate, and the storage management mechanism is placed in a failure operating mode. While in failure operating mode, the storage management mechanism checks the status map before accessing data on the failed storage unit. If the data has not yet been reconstructed, storage management first reconstructs the data in that block of storage by successively reading and accumulating an Exclusive-OR of the same blocks on all storage units in the parity group, including the parity block. The block of reconstructed data is stored in the location of the parity block, and the status map is updated to indicate that the block has been reconstructed. Once the data has been reconstructed, it is only necessary to read from or write to the former parity block directly. In the same manner, storage management will reconstruct a block of storage on the failed unit before writing to any corresponding block on a non-failed unit, if the block has not yet been reconstructed. In an alternate embodiment, spare areas of storage in the non-failing storage units are allocated to the reconstructed data. The total of these spare areas constitute a virtual spare storage unit. As data is reconstructed, it is placed in the virtual spare unit, and parity is maintained in the normal fashion.

公开(公告)号：CA1270333A

公开(公告)日：1990-06-12

申请号：CA535598

申请日：1987-04-27

Applicant: IBM

Inventor： CLARK BRIAN E ,  LAWLOR FRANCIS D ,  SCHMIDT-STUMPF WERNER E ,  STEWART TERRENCE J ,  TIMMS GEORGE D JR

IPC: G06F3/06 ,  G06F11/10 ,  G06F11/14 ,  G06F11/20 ,  G06F12/16 ,  G11B20/18

Abstract: A storage management mechanism distributes parity blocks corresponding to multiple data blocks substantially equally among a set of storage devices. N storage units in a set are divided into a multiple of equally sized address blocks, each containing a plurality of records. Blocks from each storage unit having the same address ranges form a stripe of blocks. Each stripe has a block on one storage device containing parity for the remaining blocks of the stripe. Further stripes also have parity blocks, which are distributed on different storage units. Parity updating activity associated with every change to a data record is therefore distributed over the different storage units, enhancing access characteristics of the set of storage devices. The parity updating activity also includes the use of an independent version number stored with each data record and corresponding version numbers stored with the parity record. Each time a data record is changed, its version number is incremented and the corresponding version number in the parity record is incremented with the parity record update.