Reducing reconstruct input/output operations in storage systems

• US 7,647,526 B1
• Filed: 12/06/2006
• Issued: 01/12/2010
• Est. Priority Date: 12/06/2006
• Status: Active Grant

• Current Assignee: NetApp, Inc.
• Original Assignee: NetApp, Inc.
• Inventors: Taylor, James A.
• Primary Examiner(s): DUNCAN, MARC M
• Application Number:
  US11/634,497
• Time in Patent Office: 1,133 Days
• Field of Search: 714/6, 714/7
• US Class Current: 714/6.32
• CPC Class Codes: G06F 11/1092 Rebuilding, e.g. when physically replacing a failing disk
• CPC Class Codes: G06F 11/1662 the resynchronized component or unit being a persistent storage device re-synchronization of failed mirror storage G06F11/2082; rebuild or reconstruction of parity RAID storage G06F11/1008
• CPC Class Codes: G06F 11/2094 Redundant storage or storage space G06F11/2056 takes precedence

Abstract

The present invention provides a system, method, and computer program product for partially pre-writing data stored in an array of mass storage devices to a spare disk dedicated to the array prior to detecting a failure in the array. The system then reconstructs data from the failed disk using data partially pre-written to the spare disk. Since some of the data has already been pre-written to the spare disk, embodiments of the present invention reduce the number of I/O operations performed by a storage system during a process of reconstructing data from the failed disk to the spare disk.

Claims

1. A method for reducing a number of I/O operations performed by a storage system during a process of reconstructing data from a failed device to a spare device, the method comprising:

partially pre-writing data stored in an array of mass storage devices to a spare device prior to detecting a failed device in the array of mass storage devices; and

reconstructing data from the failed device using the partially pre-written data without reading data from all the devices in the array.

2. The method of claim 1, further comprising:

replacing the failed device with the spare device upon identifying the failed device;

determining whether data from the failed device are pre-written to the spare device and the data are valid; and

responsive to a positive determination, reconstructing the data without reading the data from any device in the array or writing data to the spare device.

3. The method for claim 1, wherein the spare device is dedicated to the array.

4. The method of claim 1, further comprising:

identifying a device in the array from which data to be pre-written to a spare device for a stripe as the stripe number modulo total number of devices in the array.

5. The method of claim 1, wherein the data is partially pre-written to the spare device during an idle state.

6. The method of claim 1, wherein the data is partially pre-written to the spare device during I/O operations initiated by a client system requesting services of the storage system.

7. The method of claim 1, wherein pre-writing data to a spare device during a scrub operation further comprises writing, to the spare device, a result of a logical operation performed on data stored on two or more devices in the array of mass storage devices.

8. The method of claim 7, wherein the logical operation is an exclusive-OR (XOR) operation.

9. The method of claim 7, wherein reconstructing data on the failed device further comprises:

computing data from the failed device using data partially pre-written to the spare device and data received from other devices in the array.

10. The method of claim 9, further comprising:

responsive to data from the failed device being included in a result of the logical operation pre-written to the spare disk, reading into a memory, by the command executing module, data stored on the spare device;

receiving, by the command executing module, from the storage system, the XOR of data from the devices that were included in the result of logical operation pre-written to the spare device, other than data from the failed device;

XOR'"'"'ing the data being read; and

writing a result of the XOR operation to the spare device.

11. The method of claim 9, further comprising:

responsive to data from the failed device not being included in a result of the logical operation pre-written to the spare disk, reading into a memory, by the command executing module, data stored on the spare device;

receiving, by the command executing module, from the storage system the XOR of data from the devices that were not included in the result of logical operation pre-written to the spare device, other than data from the failed device;

XOR'"'"'ing the data being read; and

writing a result of the XOR operation to the spare device.

12. The method of claim 1, further comprising alternating writing, to the spare device, a result of logical operation performed on data stored in a first part of the array in a stripe and a result of a logical operation on data stored in a second part of the array in the stripe.

13. The method of claim 12, wherein the first part of the array is a first half of the array and the second part of the array is the second half of the array.

14. The method of claim 12, wherein reconstructing data on the failed device further comprises:

determining whether data from the failed device is included in the result of the logical operation pre-written to the spare device;

responsive to a positive determination, reconstructing data from the failed device by;

reading data from the spare device;

reading data from the devices that were included in the result of the logical operation pre-written to the spare disk other than data from the failed device, without reading data from those device that were not included in the result of the logical operation;

performing a logical operation on the data being read; and

writing a result of the logical operation to the spare device.

15. The method of claim 14, wherein the logical operation is an exclusive-OR (XOR) operation.

16. The method of claim 15, further comprising:

responsive to a negative determination, reconstructing data on the failed device by reading data from the spare device,

reading data from the devices that were not included in the result of the logical operation pre-written to the spare device, other than data from the failed device, without reading data from those device that were included in the result of the logical operation pre-written to the spare device,

performing a logical operation on the data being read, and

writing a result of the logical operation to the spare device.

17. The method of claim 16, wherein the logical operation is an exclusive-OR (XOR) operation.

18. A storage system for reducing the number of I/O operations performed during a process of reconstructing data from a failed device to a spare device, the system comprising:

a scrub module executed by a processor, the scrub module configured to partially pre-write data stored in an array of mass storage devices to a spare device prior to detecting a failed device in the array of mass storage devices; and

a reconstruct module executed by the processor, the reconstruct module configured to reconstruct data from the failed device using the partially pre-written data without reading data from all the devices in the array.

19. The storage system of claim 18, further comprising:

a RAID controller module configured to identify a failed device in the array and to replace the failed device with a spare device.

20. The system of claim 18, wherein the reconstruct module is further configured to:

reconstruct the data without reading the data from all devices in the array, responsive to the data from the failed device being pre-written to the spare device and the data being valid.

21. The system of claim 18, wherein the scrub module is further configured to:

identify a device in the array from which data to be pre-written to a spare device for a stripe as the stripe number modulo total number of devices in the array.

22. The system of claim 18, wherein the scrub module is further configured to pre-write, to the spare device, a result of a logical operation performed on data stored in two or more devices in a stripe in the array.

23. The system of claim 18, further comprising:

a command executing module implemented at the spare device, configured to reconstruct data from the failed device using data partially pre-written to the spare device and data received, from the storage system, from other devices in the array.

24. The system of claim 23, wherein the command executing module comprises program code executed on one or more controllers at the spare device.

25. The system of claim 18, further comprising a data structure for indicating whether data stored on the spare device are valid.

26. The system of claim 19, wherein the RAID controller module, the scrub module, and the reconstruct module each comprising program code executed on one or more processors of the storage system.

27. The system of claim 18, wherein the scrub module is further configured to pre-write data to the spare device during an idle state.

28. The system of claim 18, wherein the scrub module is further configured to pre-write data to the spare device during I/O operations initiated by a client system requesting services of the storage system.

29. A method for recovering data from a device that has a media error in an array of mass storage devices having a failed device, wherein the data from the failed device is pre-written to a spare device prior to a device failure in the array, the method comprising:

reading data pre-written to the spare device;

reading data from other devices in the array other than the failed device and other than the device with the media error; and

recovering the data stored on the device that has a media error by performing a logical operation on the read data.

30. A method for recovering data from a device that has a media error in an array of mass storage devices having a failed device and a spare device storing a result of an application of a logical operation on data stored in the array, the method comprising:

determining whether the result of the application of the logical operation includes both data stored on a failed device and data stored on a device that has a media error;

responsive to a negative determination, determining whether the result of the application of the logical operation includes data stored on a device that has a media error;

responsive to a positive determination, reading data from those devices in the array that are included in the result of the logical operation pre-written to the spare device in the stripe except for the device with the media error;

reading the result of the logical operation from the spare device; and

recovering the data stored on the device that has a media error by performing a logical operation on the data being read and responsive to a negative determination

reading data from those devices in the array that are not included in the result of the logical operation pre-written to the spare device in the stripe except for the device with a media error;

reading the result of the logical operation from the spare device; and

recovering the data stored on the device that has a media error by performing a logical operation on the data being read.

31. A computer-program product comprising:

a computer-readable medium having computer program code embodied thereon for reducing a number of I/O operations performed by a storage system during a process of reconstructing data from a failed device to a spare device, the computer program code adapted to;

partially pre-write data stored in an array of mass storage devices to a spare device prior to detecting a failed device in the array of mass storage devices; and

reconstruct data from the failed device using the partially pre-written data without reading data from all the devices in the array.

32. The computer program code of claim 31, further adapted to:

replace the failed device with the spare device upon identifying the failed device;

determine whether data from the failed device are pre-written to the spare device and the data are valid; and

reconstruct the data without reading the data from any device in the array or writing data to the spare device responsive to a positive determination.