Method and system for striping data to accommodate integrity metadata

• US 7,051,155 B2
• Filed: 08/05/2002
• Issued: 05/23/2006
• Est. Priority Date: 08/05/2002
• Status: Active Grant

• Current Assignee: Oracle America, Inc. (Oracle Corporation)
• Original Assignee: Sun Microsystems Incorporated (Oracle Corporation)
• Inventors: Wong, Brian, Talagala, Nisha
• Primary Examiner(s): MOAZZAMI, NASSER G
• Application Number:
  US10/212,861
• Publication Number:
  US 20040024963A1
• Time in Patent Office: 1,387 Days
• Field of Search: 711/114, 714 1- 7, 714/52, 714/770
• US Class Current: 711/114
• CPC Class Codes: G06F 11/1076 Parity data used in redundant arrays of independent storages, e.g. in RAID systems
• CPC Class Codes: G06F 2211/1007 Addressing errors, i.e. silent errors in RAID, e.g. sector slipping and addressing errors
• CPC Class Codes: G06F 2211/104 Metadata, i.e. metadata associated with RAID systems with parity

Abstract

A data layout mechanism is described for allocating metadata within a storage system employing data striping. The data layout mechanism includes a number of storage devices, each of the storage devices having storage spaces allocated to store individual data stripe units associated with a number of stripes. The data layout mechanism further includes a plurality of metadata chunks allocated within the storage devices such that (1) metadata associated with at least two data stripe units of the same stripe is stored within a single metadata chunk, and (2) the metadata chunks are evenly distributed across the storage devices.

Claims

1. A data layout mechanism comprising:

a plurality of storage devices, each of said storage devices having storage spaces allocated to store individual data stripe units in a plurality of stripes;

a plurality of metadata chunks, each metadata chunk including integrity metadata for each individual data stripe unit of a respective one of the plurality of stripes, and wherein said metadata chunks are allocated within the storage devices such that (1) said integrity metadata associated with at least two individual data stripe units of a common data stripe is stored within a single metadata chunk, and (2) said plurality of metadata chunks are distributed across the storage devices.

2. The data layout mechanism of claim 1, wherein the integrity metadata for each individual data stripe units of an entire stripe is stored within a single metadata chunk.

3. The data layout mechanism of claim 1, wherein the integrity metadata is selected from the group consisting of checksum data, address data or combination thereof.

4. The data layout mechanism of claim 1, wherein storage space for each respective metadata chunk is allocated adjacent to one of the data stripe units associated with the corresponding stripe.

5. The data layout mechanism of claim 1, wherein storage space for each respective metadata chunk is allocated within one of the storage devices that is not configured to store data associated with the respective stripe.

6. The data layout mechanism of claim 1, further comprising a controller to receive data to be stored on the storage devices and to divide the data into individual data stripe units and distribute the individual data stripe units across the storage devices.

7. The data layout mechanism of claim 6, wherein said controller is configured to generate the integrity metadata for each data stripe unit in the respective stripe and to store the integrity metadata for each individual data stripe unit of an entire stripe within a single metadata chunk.

8. The data layout mechanism of claim 6, wherein said controller is configured to copy the integrity metadata for each individual data stripe unit of an entire stripe, and to store the copy of the integrity metadata on one of said plurality of storage devices.

9. A method comprising:

allocating storage spaces within a plurality of storage devices to store individual data stripe units in a plurality of stripes;

generating integrity metadata for each individual data stripe unit in each individual data stripe unit of a respective one of the plurality of stripes; and

allocating storage spaces within said storage devices to store the integrity metadata such that (1) said integrity metadata associated with at least two individual data stripe units of a common data stripe is stored within a single metadata chunk, and (2) said metadata chunks associated with said plurality of stripes are distributed across the storage devices.

10. The method of claim 9, wherein the integrity metadata for each individual data stripe unit of an entire stripe is stored within a single metadata chunk.

11. The method of claim 9, further comprising:

receiving data to be stored on the storage devices;

dividing the data into individual data stripe units and storing the individual data stripe units by distributing the individual data stripe units across the storage devices; and

storing the integrity metadata for each individual data stripe unit in the respective stripe in one of the storage devices adjacent to the data associated with the respective stripe, wherein the stored integrity metadata associated with different stripes is distributed across the storage devices.

12. The method of claim 11, further comprising:

copying the integrity metadata for each individual data stripe unit in the respective stripe; and

storing the copy of the integrity metadata on another one of said storage devices adjacent to the data associated with the respective stripe.

13. The method of claim 9, further comprising:

receiving data to be stored on the storage devices;

dividing the data into individual data stripe units and storing the individual data stripe units by distributing the individual data stripe units across the storage devices;

storing the integrity metadata for each individual data stripe units in the respective stripe on one of the storage devices that is not configured to store the data associated with the respective stripe, wherein the stored integrity metadata associated with different stripes is distributed across the storage devices.

14. The method of claim 13, further comprising:

copying the integrity metadata for each individual data stripe units in the respective stripe; and

storing the copy of the integrity metadata on another one of said storage devices that is not configured to store data associated with the respective stripe.

15. The method of claim 9, wherein the integrity metadata is selected from the group consisting of checksum data, address data or combination thereof.

16. A machine-readable medium containing instructions, which when executed by a processor cause said processor to perform operations comprising:

allocating storage spaces within a plurality of storage devices to store individual data stripe units in a plurality of stripes;

generating an integrity metadata for each individual data stripe unit of a respective one of the plurality of stripes; and

allocating storage spaces within said storage devices to store the integrity metadata for each individual data stripe units of the respective one of the plurality of stripes such that (1) said integrity metadata associated with at least two individual data stripe units of a common data stripe is stored within a single metadata chunk, and (2) said metadata chunks associated with said plurality of stripes are distributed across the storage devices.

17. The machine-readable medium of claim 16, wherein the integrity metadata for each individual data stripe units in the entire stripe is stored within a single metadata chunk.

18. The machine-readable medium of claim 16, wherein the operations further comprise:

receiving data to be stored on the storage devices;

dividing the data into individual data stripe units and storing the individual data stripe units by distributing the individual data stripe units across the storage devices; and

storing the integrity metadata for each individual data stripe units in the respective stripe in one of the storage devices adjacent to the data associated with the respective stripe, wherein the stored integrity metadata associated with different stripes is distributed across the storage devices.

19. The machine-readable medium of claim 18, wherein the operations further comprise:

copying the integrity metadata for each individual data stripe units in the respective stripe; and

storing the copy of the integrity metadata on another one of said storage devices adjacent to the data associated with the respective stripe.

20. The machine-readable medium of claim 16, wherein the operations further comprise:

receiving data to be stored on the storage devices;

dividing the data into individual data stripe units and storing the individual data stripe units by distributing the individual data stripe units across the storage devices;

combining the integrity metadata for each individual data stripe units in the respective stripe; and

storing the combined integrity metadata on one of the storage devices that is not configured to store the data associated with the respective stripe, wherein the stored combined integrity metadata associated with different stripes is distributed across the storage devices.

21. The machine-readable medium of claim 20, wherein the operations further comprise:

copying the integrity metadata for each individual data stripe units in the respective stripe; and

storing the copy of the integrity metadata on another one of said storage devices that is not configured to store data associated with the respective stripe.

22. The machine-readable medium of claim 16, wherein the integrity metadata is selected from the group consisting of checksum data, address data or combination thereof.

23. An apparatus comprising:

means for allocating storage spaces within a plurality of storage devices to store individual data stripe units in a plurality of stripes;

means for allocating storage space within said storage devices to store an integrity metadata for each individual data stripe unit of a respective one of the plurality of stripes such that (1) integrity metadata associated with at least two data stripe units of a common data stripe is stored within a single metadata chunk, and (2) metadata chunks associated with said plurality of stripes are distributed across the storage devices.

24. The apparatus of claim 23, wherein the integrity metadata for each of the individual data stripe units in an entire stripe is stored within a single metadata chunk.

25. The apparatus of claim 23, wherein the integrity metadata is selected from the group consisting of checksum data, address data or combination thereof.

26. The apparatus of claim 23, wherein the storage space for each respective metadata chunk is allocated adjacent to data associated with the corresponding stripe.

27. The apparatus of claim 23, further comprising:

means for receiving data to be stored on the storage devices;

means for dividing the data into individual data stripe units and storing the individual data stripe units by distributing the individual data stripe units across the storage devices; and

means for storing the integrity metadata for each of the individual data stripe units in the respective stripe in one of the storage devices adjacent to the data associated with the respective stripe, wherein the stored metadata chunks associated with different stripes are distributed across the storage devices.

28. The apparatus of claim 27, further comprising:

means for copying the integrity metadata for each of the individual data stripe units in the respective stripe; and

means for storing the copy of the integrity metadata on another one of said storage devices adjacent to the data associated with the respective stripe.

29. The apparatus of claim 23, wherein the storage space for integrity metadata is allocated on one of the storage devices that is not configured to store data associated with the respective stripe.

30. The apparatus of claim 23, further comprising:

means for receiving data to be stored on the storage devices;

means for dividing the data into individual data stripe units and storing the individual data stripe units by distributing the individual data stripe units across the storage devices;

means for generating the integrity metadata for each individual data stripe unit in the respective stripe; and

means for storing the integrity metadata for each individual data stripe unit in the respective stripe on one of the storage devices that is not configured to store the data associated with the respective stripe, wherein the stored integrity metadata associated with different stripes are evenly distributed across the storage devices.

31. The apparatus of claim 30, further comprising:

means for copying the combined integrity metadata associated with the respective stripe; and

means for storing the copy of the combined integrity metadata on another one of said storage devices that is not configured to store data associated with the respective stripe.