System and method for detecting write errors in a storage device

• US 20050005191A1
• Filed: 05/05/2004
• Published: 01/06/2005
• Est. Priority Date: 06/11/2003
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Judd, Ian David
• Granted Patent: US 7,380,198 B2
• Time in Patent Office: Days
• US Class Current: 714/5
• CPC Class Codes: G06F 11/1076 Parity data used in redundant arrays of independent storages, e.g. in RAID systems
• CPC Class Codes: G06F 2211/1059 Parity-single bit-RAID5, i.e. RAID 5 implementations
• CPC Class Codes: G11B 20/18 Error detection or correction; Testing , e.g. of drop-outs

Abstract

A system for detecting write errors in a storage device is disclosed. The system comprises a storage device; within the storage device, means for storing one or more data blocks in a storage group, the storage group comprising the one or more data blocks and a check block, wherein the check block comprises one of the group of: a combination of the one or more data blocks of the storage group, a combination of one or more bits of a logical block address associated with the storage group, and a combination of one or more bits of a phase field that is updated each time the storage group is written; means for updating the check block each time the storage group is written; and means for detecting write errors by checking the check block.

Claims

1. A system for detecting write errors in a storage device, said system comprising:

a storage device;

within said storage device, means for storing one or more data blocks in a storage group, said storage group including said one or more data blocks and a check block, wherein said check block includes one of the group of;

a combination of said one or more data blocks of said storage group,

a combination of one or more bits of a logical block address associated with said storage group, and

a combination of one or more bits of a phase field that is updated each time said group is written;

means for updating said check block each time said group is written; and

means for detecting write errors by checking said check block when a group is read.

2. The system of claim 1, wherein said combination is a logical Exclusive-OR combination.

3. The system of claim 1, wherein said phase field comprises a single bit value, which single-bit value is inverted each time the group is written.

4. The system of claim 1, wherein said phase field comprises a multi-bit value, which multi-bit value is updated each time said group is written.

5. The system of claim 1, further comprising:

a non-volatile table for the expected phase field values, wherein said table comprises a reserved disk area within said storage device; and

a working copy of said table cached in a controller of said system.

6. The system of claim 5, further comprising a non-volatile log, said log arranged to record an entry before a write operation.

7. The system of claim 6, wherein said log is arranged to retain updates to a working copy of said table in said controller, which updates have not yet been stored in said non-volatile table.

8. The system of claim 1, further comprising a non-volatile log, said log arranged to record an entry before a disk write operation starts, said entry being arranged for an operation selected from the group of (1) invalidation or (2) deletion on completion of the write operation.

9. A method for detecting write errors in a storage device, said method comprising:

within a storage device, storing one or more data blocks in a storage group;

creating a check block including one of a set including;

a combination of said one or more data blocks of said storage group,

a combination of one or more bits of a logical block address associated with said storage group, and

a combination of one or more bits of a phase field that is updated each time said group is written;

including said check block in said storage group;

updating said check block each time said storage group is written; and

detecting write errors in said data blocks by checking said check block when a group is read.

10. The method of claim 9, wherein said creating step further comprises creating said check block with a combination that is a logical Exclusive-OR combination.

11. The method of claim 9, wherein said creating step further comprises creating said check block with a phase field that comprises a single bit value, and inverting said single-bit value each time said group is written.

12. The method of claim 9, wherein said creating step further comprises creating said check block with a phase field that comprises a multi-bit value, which multi-bit value is updated each time said group is written.

13. The method of claim 9, further comprising:

generating a non-volatile table for phase field values, wherein said table comprises a reserved disk area within said storage device; and

caching a working copy of said table in a controller of said system.

14. The method of claim 13, further comprising recording a non-volatile log, said log arranged to record an entry before a disk write operation starts.

15. The method of claim 14, further comprising arranging said log to retain updates to a working copy of said table in said controller, which updates have not yet been stored in said non-volatile table.

16. The method of claim 9, further comprising establishing a non-volatile log, said log arranged to record an entry before a disk write operation starts, said entry being arranged for an operation selected from the group of (1) invalidation or (2) deletion on completion of the write operation.

17. A computer program product in a computer-readable medium for detecting write errors in a storage device, said computer program product comprising:

a computer-readable medium;

instructions on the computer-readable medium for storing one or more data blocks in a storage group, said storage group comprising said one or more data blocks and a check block, wherein said check block includes one of the group of;

a combination of said one or more data blocks of said storage group,

a combination of one or more bits of a logical block address associated with said storage group, and

a combination of one or more bits of a phase field that is updated each time said storage group is written;

instructions on the computer-readable medium for updating said check block each time said storage group is written; and

instructions on the computer-readable medium for detecting write errors by checking said check block when a group is read.

18. The computer program product of claim 17, wherein said storing instructions further comprise instructions for creating said check block to include a combination that is a logical Exclusive-OR combination.

19. The computer program product of claim 17, wherein said storing instructions further comprise instructions for creating said check block to include a phase field that comprises a single bit value, which single-bit value is inverted each time the group is written.

20. The computer program product of claim 17, wherein said storing instructions further comprise instructions for creating said check block to include a phase field, which comprises a multi-bit value, which multi-bit value is updated each time said group is written.

21. The computer program product of claim 17, further comprising instructions on the computer-readable medium for:

creating a non-volatile table for the expected phase field values, wherein said table comprises a reserved disk area within said storage device; and

generating a working copy of said table cached in a controller of said system.

22. The computer program product of claim 21, further comprising instructions on the computer-readable medium for recording a non-volatile log, said log arranged to record an entry before a disk write operation starts.

23. The computer program product of claim 22, wherein said instructions for recording said log further comprise instructions for arranging said log to retain updates to a working copy of said table in said controller, which updates have not yet been stored in said non-volatile table.

24. The computer program product of claim 17, further comprising instructions on the computer-readable medium for generating a non-volatile log, wherein said log is arranged to record an entry before a write operation, said entry being arranged for an operation selected from the group of (1) invalidation or (2) deletion on completion of the write operation.