Block Reallocation Planning During Read-Ahead Processing

• US 20080270706A1
• Filed: 04/27/2007
• Published: 10/30/2008
• Est. Priority Date: 04/27/2007
• Status: Active Grant

• Current Assignee: Network Appliance Incorporated (NetApp, Inc.)
• Original Assignee: Network Appliance Incorporated (NetApp, Inc.)
• Inventors: Fair, Robert L., English, Robert M.
• Granted Patent: US 7,996,445 B2
• Time in Patent Office: Days
• US Class Current: 711/137
• CPC Class Codes: G06F 12/0862 with prefetch
• CPC Class Codes: G06F 12/0866 for peripheral storage systems, e.g. disk cache
• CPC Class Codes: G06F 3/061 Improving I/O performance
• CPC Class Codes: G06F 3/064 Management of blocks
• CPC Class Codes: G06F 3/0653 Monitoring storage devices or systems
• CPC Class Codes: G06F 3/0689 Disk arrays, e.g. RAID, JBOD

Abstract

A data storage system pre-fetches data blocks from a mass storage device, then determines whether reallocation of the pre-fetched blocks would improve access to them. If access would be improved, the pre-fetched blocks are written to different areas of the mass storage device. Several different implementations of such data storage systems are described.

Claims

1. A method comprising:

pre-fetching a plurality of data blocks from a mass storage device;

determining whether reallocation would improve access to the plurality of data blocks; and

if access would be improved, writing the plurality of data blocks to different locations of the mass storage device.

2. The method of claim 1, further comprising:

monitoring activity of a storage system; and

predicting the plurality of data blocks to be pre-fetched based on the activity.

3. The method of claim 1 wherein the plurality of data blocks is a pre-fetched plurality of data blocks, the method further comprising:

fetching a requested plurality of data blocks from the mass storage device in response to a request from a client, and wherein

said determining comprises determining whether reallocation would improve access to a combined plurality of data blocks including the requested plurality and the pre-fetched plurality; and

said writing comprises writing the combined plurality of data blocks.

4. The method of claim 1 wherein the plurality of data blocks is a subset of a larger set of related data blocks.

5. The method of claim 1 wherein the plurality of data blocks contains data from a file.

6. The method of claim 1 wherein the plurality of data blocks contains a portion of a filesystem.

7. The method of claim 1 wherein a data block of the plurality of data blocks is part of two different sequences of data blocks.

8. A system comprising:

a communication interface to receive requests from a client;

a processor to interpret the requests;

filesystem logic to locate data on a mass storage device, wherein the data is identified by the requests;

prediction logic to identify additional data on the mass storage device that may soon be requested; and

reallocation logic to determine whether defragmenting the additional data is likely to improve access to the additional data.

9. The system of claim 8, further comprising:

a mass storage device.

10. The system of claim 9 wherein the mass storage device is a plurality of storage devices operated as a Redundant Array of Independent Disks (“

RAID array”

).

11. The system of claim 8 wherein the communication interface is a network communication interface.

12. The system of claim 8 wherein the communication interface is a Small Computer Systems Interface (“

SCSI”

).

13. The system of claim 8 wherein the communication interface is a Fiber Channel (“

FC”

) interface.

14. A computer-readable medium containing data and instructions to cause a programmable processor to perform operations comprising:

maintaining a filesystem on a mass storage subsystem;

predicting a plurality of data blocks of the mass storage subsystem that are not required yet but are expected to be required soon;

pre-fetching the plurality of data blocks into a cache memory;

evaluating a physical location of the plurality of data blocks to decide whether access to the plurality of data blocks could be improved by moving the plurality of data blocks; and

if access could be improved, moving the plurality of data blocks.

15. The computer-readable medium of claim 14 wherein maintaining the filesystem comprises:

maintaining an inode to identify a sequence of data blocks that make up a file; and

maintaining a block map to distinguish between used data blocks and unused data blocks.

16. The computer-readable medium of claim 15 wherein the inode is a first inode and the sequence of data blocks is a first sequence, and wherein maintaining the filesystem comprises:

allocating a second inode to identify a second sequence of data blocks that make up a file, wherein a data block of the first sequence is also in the second sequence.

17. The computer-readable medium of claim 14 wherein moving the plurality of data blocks comprises writing the plurality of data blocks in sequential order.

18. The computer-readable medium of claim 14 wherein moving the plurality of data blocks comprises writing the plurality of data blocks contiguously.

19. The computer-readable medium of claim 14, containing additional data and instructions to cause the programmable processor to perform operations comprising:

monitoring activity affecting the filesystem, and wherein

the predicting operation refers to information collected by the monitoring operation.

20. A method comprising:

performing a read of a sequential set of data blocks from a mass storage device;

detecting at least two subsets of the sequential set of data blocks that are not contiguous with the sequential set and not contiguous with each other; and

reallocating the at least two subsets to a contiguous area of the mass storage device.

21. The method of claim 20, further comprising:

predicting a client interaction that will require data from the sequential set of data blocks, wherein

the read operation is performed before the client interaction.

22. The method of claim 20 further comprising:

placing data blocks of the at least two subsets in ascending order.

23. The method of claim 20, wherein the at least two subsets are contiguous but not sequential after reallocation.