Method and apparatus for providing write recovery of faulty data in a non-redundant raid system

• US 6,854,071 B2
• Filed: 05/14/2001
• Issued: 02/08/2005
• Est. Priority Date: 05/14/2001
• Status: Expired due to Fees

• Current Assignee: GlobalFoundries, Inc.
• Original Assignee: International Business Machines Corporation
• Inventors: King, Allen, Lynch, Jr., Michael Attmer
• Primary Examiner(s): Beausoliel, Robert
• Assistant Examiner(s): DUNCAN, MARC M
• Application Number:
  US09/854,543
• Publication Number:
  US 20020169996A1
• Time in Patent Office: 1,366 Days
• Field of Search: 714/7, 714/8, 714/54, 369/53.36, 369/53.35, 369/47.14, 369/53.15
• US Class Current: 714/6.13
• CPC Class Codes: G06F 11/0727 in a storage system, e.g. in a DASD or network based storage system drivers for digital recording or reproducing units G06F3/06; circuits for error detection or correction within digital recording or reproducing units G11B20/18; for distributed storage of data in networks, e.g. transport arrangements for network file system [NFS], storage area networks [SAN] or network attached storage [NAS], H04L67/1097
• CPC Class Codes: G06F 11/0775 Content or structure details of the error report, e.g. specific table structure, specific error fields
• CPC Class Codes: G06F 11/0793 Remedial or corrective actions recovery from an exception in an instruction pipeline G06F9/3861; by retry G06F11/1402; for recovering from a failure of a protocol instance or entity H04L69/40

Abstract

A table for identifying potentially bad location addresses based on prior knowledge and performing specific operations using the table data to ensure write recovery is disclosed. A Write Recovery Table (WRT) is provided, which consists of a list of LBAs requiring such special write recovery action so that a WRITE AND VERIFY command is issued instead of a WRITE command. If the WRITE AND VERIFY command fails, the RAID controller can issue a REASSIGN BLOCKS command for the bad block and re-issue the WRITE AND VERIFY command. If WRITE AND VERIFY commands are not supported, then the system can use a READ command with the “Force Unit Access” flag bit being set to “1” to verify the written data. Further, the WRT may be combined with Bad Data Table (BDT) to reject a Read Request from the host for a matched LBA (that is listed therein) because the data in the addressed block is known incorrect when a write recovery action is not warranted.

Claims

1. A method for handling write errors in a non-redundant disk array storage system, comprising:

issuing a command to write and verify data transfer when requested to write data to a specified location address that is stored in an error table; and

reassigning the data to a new location address when the command to write and verify the data transfer is unsuccessful; and

wherein the error table comprises a combined write recovery table and bad data table.

2. The method of claim 1 wherein the error table records error locations on a disk drive.

3. The method of claim 1 further comprising transferring the data to the specified location address when a good status is returned in response to the command to write and verify the data transfer.

4. The method of claim 1 wherein the command to write and verify the data transfer comprises a write/verify command.

5. The method of claim 1 wherein the command to write and verify the data transfer comprises a write command followed by a read command with the force unit access bit being set to 1 to force the data to be read from the disk drive rather than from disk drive cache.

6. The method of claim 1 wherein the command to write and verify the data transfer is unsuccessful only when a predetermined number of write and verify retries is exhausted.

8. The method of claim 1 further comprising deleting the specified location address that is stored in the error table when the reassign of data is successful.

9. The method of claim 1 further comprising deleting the specified location address that is stored in the error table when the command to write and verify the data transfer is successful.

7. A method for handling write errors in a non-redundant disk array storage system, comprising:

issuing a command to write and verify data transfer when requested to write data to a specified location address that is stored in an error table;

reassigning the data to a new location address when the command to write and verify the data transfer is unsuccessful;

verifying the reassigning of data to a new location address;

detecting an error for the reassigning;

retrying the reassigning for a predetermined number of times; and

returning a fatal error status and marking the disk offline when the predetermined number of times is exhausted.

10. A storage system, comprising:

an array of storage devices; and

a storage controller, coupled to the array of storage devices, the storage controller including a memory for maintaining an error table, the storage controller further being configured to issue a command to write and verify data transfer when requested to write data to a specified location address that is stored in an error table and reassign the data to a new location address when the command to write and verify the data transfer is unsuccessful; and

wherein the error table comprises a combined write recovery table and bad data table.

11. The storage system of claim 10 wherein the error table records data error locations on a disk drive.

12. The storage system of claim 10 wherein the controller transfers the data to the specified location address when a good status is returned in response to the command to write and verify the data transfer.

13. The storage system of claim 10 wherein the command to write and verify the data transfer comprises a write/verify command.

14. The storage system of claim 10 wherein the command to write and verify the data transfer comprises a write command followed by a read command with the force unit access bit being set to 1 to force the data to be read from the storage device rather than from cache of the storage controller.

15. The storage system of claim 10 wherein the command to write and verify the data transfer is unsuccessful only when a predetermined number of write and verify retries is exhausted.

16. The storage system of claim 10 wherein the storage controller deletes the specified location address that is stored in the error table when the reassign of data is successful.

17. The storage system of claim 10 wherein the storage controller deletes the specified location address that is stored in the error table when the command to write and verify the data transfer is successful.

18. A storage system, comprising:

an array of storage devices; and

a storage controller, coupled to the array of storage controller including a memory for maintaining an error table, the storage controller further being configured to issue a command to write and verify data transfer when requested to write data to a specified location address that is stored in an error table and reassign the data to a new location address when the command to write and verify the data transfer is unsuccessful; and

wherein the storage controller further verifies the reassigning of data to new location address, detects an error for the reassigning, retries the reassigning for a predetermined number of times, returns a fatal error status and marks the storage device offline when the predetermined number of times is exhausted.

19. An error table disposed in a storage controller, the error table configured with addresses for data error locations on a storage device so that for a write request, the storage controller knows to perform a write and verify command to transfer the data to the storage device and to verify the successful transfer of the data to the storage device;

said error table further comprising a write recovery bit to indicate when an address associated therewith requires a write recovery during a write; and

wherein the write recovery bit allows a bad data table to be combined with the write recovery table.

20. The error table of claim 19 wherein on a read the error table is searched only for an address specified in a read request together with an associated write recovery bit being set to 0.

21. A computer program product for handling write errors in a non-redundant disk array storage system, said program product comprising:

a computer readable medium;

first program instructions to issue a command to write and verify data transfer when requested to write data to a specified location address that is stored in an error table;

second program instructions to reassign the data to a new location address when the command to write and verify the data transfer is unsuccessful;

third program instructions to verify the reassignment of data to a new location address;

fourth program instructions to detect an error for said reassignment;

fifth program instructions to retry said reassignment for a predetermined number of times; and

sixth program instructions to return a fatal error status and mark the disk offline when the predetermined number of times is exhausted; and

wherein

said first, second, third, fourth, fifth and sixth program instructions are recorded on said medium.