Method and apparatus for the reliability of host data stored on fibre channel attached storage subsystems

• US 20070038913A1
• Filed: 07/26/2005
• Published: 02/15/2007
• Est. Priority Date: 07/26/2005
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Allen, James, Mathews, Thomas, Shankar, Ravi, Wightwick, Glenn, Sharma, Satya
• Granted Patent: US 7,472,332 B2
• Time in Patent Office: Days
• US Class Current: 714/758
• CPC Class Codes: G11B 20/1813 by adding special bits or symbols to the coded information G11B20/1809 takes precedence

Abstract

A method for improving the reliability of host data stored on Fibre Channel attached storage subsystems by performing end-to-end data integrity checks. When a read or write operation is initiated, an initial checksum for data in the read/write operation is generated and associated with the data, wherein the association exists through a plurality of layers of software and attached storage subsystems. The initial checksum is passed with the data in the read/write path. When a layer of software in the read/write path receives the initial checksum and data, the layer performs an integrity check of the data, which includes generating another checksum and comparing it to the initial checksum. If the checksums do not match, the read/write operation fails and the error is logged. If the checksums match, the integrity check is repeated through each layer in the read/write path to enable detecting data corruption at the point of source.

Claims

1. A method in a data processing system for providing end-to-end data integrity checks, the method comprising:

performing a read/write operation, wherein the read/write operation comprises generating an initial checksum for data in the read/write operation and associating the initial checksum with the data, wherein the association exists through a plurality of layers of software and attached storage subsystems in the data processing system;

sending the initial checksum with the data in a read/write path;

responsive to receiving the initial checksum and data at a layer of software in the read/write path, performing an integrity check of the data, wherein the integrity check comprises generating a comparison checksum and determining whether the comparison checksum matches the initial checksum; and

if the initial and comparison checksums do not match, failing the read/write operation.

2. The method of claim 1, further comprising:

if the initial and comparison checksums match, repeating the integrity check in each layer of the plurality of layers of software and attached storage subsystems in the read/write path.

3. The method of claim 1, wherein determining whether the comparison checksum matches the initial checksum at each layer in the plurality of layers of software and attached storage subsystems allows for detecting data corruption at a point of source.

4. The method of claim 1, wherein failing the read/write operation includes logging an error and sending a read/write failure notification to a user.

5. The method of claim 1, wherein the initial checksum is sent in a header of a communication medium.

6. The method of claim 5, wherein the communications medium is one of Fibre Channel and iSCSI.

7. The method of claim 1, wherein the integrity check is performed without imposing a restriction on structure, form, and content of the data.

8. The method of claim 1, wherein performing the integrity check in the plurality of layers of software and attached storage subsystems in the read/write path prevents corrupted data from being stored on disk.

9. A data processing system for providing end-to-end data integrity checks, comprising:

a resource manager, wherein the resource manager initiates a read/write operation, wherein the read/write operation comprises generating an initial checksum for data in the read/write operation and associating the initial checksum with the data, wherein the association exists through a plurality of layers of software and attached storage subsystems in the data processing system, and wherein the resource manager sends the initial checksum with the data in a read/write path; and

layers of software and storage subsystems connected to the resource manager, wherein a layer in the read/write path receives the initial checksum and data and performs an integrity check of the data, wherein the integrity check comprises generating a comparison checksum and determining whether the comparison checksum matches the initial checksum, and wherein the layer fails the read/write operation if the initial and comparison checksums do not match.

10. The data processing system of claim 9, further comprising:

wherein the integrity check is repeated in each layer of the plurality of layers of software and attached storage subsystems in the read/write path if the initial and comparison checksums match.

11. The data processing system of claim 9, wherein determining whether the comparison checksum matches the initial checksum at each layer in the plurality of layers of software and attached storage subsystems allows for detecting data corruption at a point of source.

12. The data processing system of claim 9, wherein failing the read/write operation includes logging an error and sending a read/write failure notification to a user.

13. The data processing system of claim 9, wherein the initial checksum is sent in a header of a communication medium.

14. The data processing system of claim 13, wherein the communications medium is one of Fibre Channel and iSCSI.

15. The data processing system of claim 9, wherein the integrity check is performed without imposing a restriction on structure, form, and content of the data.

16. The data processing system of claim 9, wherein performing the integrity check in the plurality of layers of software and attached storage subsystems in the read/write path prevents corrupted data from being stored on disk.

17. A computer program product in a computer readable medium for providing end-to-end data integrity checks, comprising:

first instructions for performing a read/write operation, wherein the read/write operation comprises generating an initial checksum for data in the read/write operation and associating the initial checksum with the data, wherein the association exists through a plurality of layers of software and attached storage subsystems in the data processing system;

second instructions for sending the initial checksum with the data in a read/write path;

third instructions for performing an integrity check of the data in response to receiving the initial checksum and data at a layer of software in the read/write path, wherein the integrity check comprises generating a comparison checksum and determining whether the comparison checksum matches the initial checksum; and

fourth instructions for failing the read/write operation if the initial and comparison checksums do not match.

18. The computer program product of claim 17, further comprising:

fifth instructions for repeating the integrity check in each layer of the plurality of layers of software and attached storage subsystems in the read/write path if the initial and comparison checksums match.

19. The computer program product of claim 17, wherein determining whether the comparison checksum matches the initial checksum at each layer of the plurality of layers of software and attached storage subsystems allows for detecting data corruption at a point of source.

20. The computer program product of claim 17, wherein failing the read/write operation includes logging an error and sending a read/write failure notification to a user.