MECHANISMS FOR MOVING DATA IN A HYBRID AGGREGATE

• US 20100281230A1
• Filed: 04/29/2009
• Published: 11/04/2010
• Est. Priority Date: 04/29/2009
• Status: Active Grant

• Current Assignee: NetApp, Inc.
• Original Assignee: NetApp, Inc.
• Inventors: Strunk, John, Kimmel, Jeffrey S., Rabii, Faramarz
• Granted Patent: US 8,321,645 B2
• Time in Patent Office: Days
• US Class Current: 711/165
• CPC Class Codes: G06F 12/0866 for peripheral storage systems, e.g. disk cache
• CPC Class Codes: G06F 12/121 using replacement algorithms
• CPC Class Codes: G06F 2212/214 Solid state disk
• CPC Class Codes: G06F 2212/222 Non-volatile memory
• CPC Class Codes: G06F 3/0605 by facilitating the interaction with a user or administrator
• CPC Class Codes: G06F 3/0613 in relation to throughput
• CPC Class Codes: G06F 3/0647 Migration mechanisms
• CPC Class Codes: G06F 3/0649 Lifecycle management
• CPC Class Codes: G06F 3/0685 Hybrid storage combining heterogeneous device types, e.g. hierarchical storage, hybrid arrays

Abstract

At least certain embodiments disclose a method, system and apparatus for relocating data between tiers of storage media in a hybrid storage aggregate encompassing multiple tiers of heterogeneous physical storage media including a file system to automatically relocate the data between tiers. The hybrid storage aggregate includes one or more volumes, each volume including a volume block number space spanning at least a first-tier of storage media and a second tier of storage media of the multiple tiers of heterogeneous physical storage media and the hybrid storage aggregate further includes a control module to cooperatively manage the tiers of the multiple tiers of heterogeneous physical storage media and a file system coupled with the control module, the file system including a policy module configured to make policy decisions based on a set of one or more policies and configured to automatically relocate data between different tiers of the multiple tiers of heterogeneous physical storage media based on the set of policies.

Claims

1. A method comprising:

examining data in a hybrid storage aggregate encompassing multiple tiers of heterogeneous physical storage media using a policy module based on a set of policies, the hybrid storage aggregate including a volume having a contiguous volume block number space spanning a first-tier and a second-tier of the multiple tiers of heterogeneous physical storage media;

determining that the data is to be relocated to a different tier of physical storage media based on the set of policies; and

relocating the data to the different tier of physical storage media.

2. The method of claim 1, wherein the first-tier of physical storage media and the second-tier of physical storage media include writable, persistent storage media, and wherein the first-tier of storage media is a lower latency storage media than the second-tier of storage media.

3. The method of claim 2, wherein the first-tier of storage media comprises an electronic storage medium and the second-tier of storage media comprises a magnetic storage medium.

4. The method of claim 3, wherein the electronic storage medium comprises a solid-state device (SSD) and the magnetic storage medium comprises a hard disk device (HDD).

5. The method of claim 1, further comprising dynamically adjusting the set of policies by the policy module to adapt to changes at run-time to improve performance characteristics of the hybrid storage aggregate.

6. The method of claim 5, wherein the examining of the data is performed during read transactions and the changes at run-time include changes in read traffic patterns within the hybrid storage aggregate.

7. The method of claim 6, further comprising determining the data is to be relocated to a lower-latency tier of physical storage media when the data has been one or more of frequently and randomly accessed in read transactions.

8. The method of claim 6, further comprising determining that the data is to be relocated to a higher-latency tier of physical storage media when the data has been either infrequently or sequentially accessed in read transactions.

9. The method of claim 6, wherein the read transactions are write-after-read (WAR) transactions.

10. The method of claim 5, wherein the examining of the data is performed by a segment cleaner module running transparently to a user of the hybrid storage aggregate and the changes at run-time include changes in capacity of storage space among the multiple tiers of heterogeneous physical storage media.

11. The method of claim 10, further comprising:

monitoring the capacity of storage space of each tier of the multiple tiers of heterogeneous physical storage media; and

relocating data from the first-tier of storage media to the second-tier of storage media when the first-tier of storage media approaches full-capacity and the second-tier of storage media has sufficient storage capacity.

12. A method comprising:

dynamically adjusting a set of policies at run-time using a policy module within a storage server to improve performance characteristics of a hybrid storage aggregate, wherein the hybrid storage aggregate includes multiple tiers of heterogeneous storage media, each tier comprising a different storage media technology, the hybrid storage aggregate including a volume having a contiguous volume block number space spanning a first-tier and a second-tier of the multiple tiers of heterogeneous storage media;

examining data in the hybrid storage aggregate using the policy module based on a set of policies;

determining that the data is to be relocated to a different tier of storage media based on the set of policies; and

automatically relocating the data to the different tier of storage media in response to the determining.

13. The method of claim 12, wherein the first-tier of storage media and the second-tier of storage media include writable, persistent storage media, and wherein the first-tier of storage media is a lower latency storage media than the second-tier of storage media.

14. The method of claim 12, wherein examining the data is performed during a read transaction, and wherein the set of policies includes an access-based policy derived from read traffic patterns within the hybrid storage aggregate.

15. The method of claim 14, further comprising:

determining the data is to be relocated to a lower-latency tier of storage media when the data has been one or more of frequently and randomly accessed in the read transactions; and

determining the data is to be relocated to a higher-latency tier of storage media when the data has been either infrequently or sequentially accessed in the read transactions.

16. The method of claim 15, further comprising:

marking a header associated with the block of data in the read transaction as dirty; and

marking the header associated with the block of data as destined to the different tier of storage media.

17. The method of claim 16, further comprising:

marking a header associated with the block of data in the read transaction as fake dirty to update only a physical volume block number associated with the block of data without requiring an update to the virtual volume block number associated with the block of data; and

marking the header associated with the block of data as destined to the different tier of storage media.

18. The method of claim 17, further comprising:

checking the header associated with the block of data to determine if it is marked dirty or fake dirty; and

moving the block of data to the different tier of storage media when the header is marked dirty or fake dirty.

19. The method of claim 12, wherein the examining of the data is performed by a module running transparently to a user of the hybrid storage aggregate and wherein the set of policies includes a capacity-based policy derived from capacity of storage space of the multiple tiers of heterogeneous storage media.

20. The method of claim 19, further comprising:

monitoring the capacity of storage space of the multiple tiers of heterogeneous storage media; and

relocating data from the first-tier of storage media to the second-tier of storage media when the first-tier of storage media approaches full-capacity and the second-tier of storage media has sufficient storage capacity.

21. A storage system comprising:

a processor;

a hybrid storage aggregate encompassing multiple tiers of heterogeneous physical storage media and including a volume having a contiguous volume block number space spanning a first tier and a second tier of the multiple tiers of heterogeneous physical storage media; and

a memory coupled with the processor having a storage manager stored therein, the storage manager operable by the processor to automatically relocate data between tiers of the multiple tiers of heterogeneous physical storage media within the hybrid storage aggregate.

22. The storage system of claim 21, wherein the storage manager includes:

a write allocation module configured to perform access-based relocation of data within the hybrid storage aggregate during read operations;

a segment cleaner module running transparently to a user of the hybrid storage aggregate, the segment cleaner module configured to perform capacity-based relocation of data within the hybrid storage aggregate; and

a policy module configured to make policy decisions based on a set of policies stored in the memory and to control operation of the write allocation and segment cleaner modules based on the policy decisions.

23. The storage system of claim 22, wherein the first tier of physical storage media is to store data that is one or more of frequently and randomly accessed and the second-tier of physical storage media is to store data that is either infrequently or sequentially accessed.

24. The storage system of claim 23, wherein the first tier of storage media is a low-latency storage media having performance characteristics superior to the performance characteristics of the second tier of storage media.

25. The storage system of claim 24, wherein the first tier of storage media comprises a writeable, persistent electronic storage media and the second tier of storage media comprises a writeable, persistent magnetic storage media.

26. An apparatus comprising:

a hybrid storage aggregate including;

multiple tiers of heterogeneous storage media, each tier including a different storage media technology, wherein the hybrid storage aggregate includes a volume having a volume block number space spanning a first tier of storage media and a second tier of storage media of the multiple tiers of heterogeneous storage media;

a control module coupled with the hybrid storage aggregate to cooperatively manage the multiple tiers of heterogeneous storage media; and

a policy module coupled with the control module, the policy module configured to make policy decisions based on a set of policies and to automatically relocate data between different tiers of the multiple tiers of heterogeneous storage media based on the policy decisions.

27. The apparatus of claim 26, wherein the first tier of storage media is a lower-latency storage media than the second tier of storage media.

28. The apparatus of claim 27, wherein the control module is a RAID module.

29. The apparatus of claim 27, further comprising a write allocator, wherein the policy module directs the write allocator to automatically relocate data between tiers of the multiple tiers of heterogeneous storage media during read transactions based on access characteristics of the data.

30. The apparatus of claim 29, wherein the policy module directs the write allocator to automatically relocate data from the second-tier of storage media to the first-tier of storage media when the data is determined to be one or more of frequently and randomly accessed, and to automatically relocate data from the first-tier of storage media to the second-tier of storage media when the data is determined to be either infrequently or sequentially accessed.

31. The apparatus of claim 28, further comprising a segment cleaner, wherein the policy module directs the segment cleaner to automatically relocate data between tiers of the multiple tiers of heterogeneous storage media based on capacity of storage space among the tiers of storage media.

32. The apparatus of claim 32, wherein the policy module directs the segment cleaner to automatically relocate data from the first-tier of storage media to the second-tier of storage media when the first-tier of storage media approaches full-capacity and the second-tier of storage media has sufficient storage capacity.

33. A storage system comprising:

a processor;

a storage manager; and

a cache memory coupled with the processor, the cache memory including;

a higher-level cache memory formed of a random access memory (RAM); and

a lower-level cache memory formed of a first tier of non-volatile storage media in a hybrid storage aggregate having multiple tiers of heterogeneous storage media.

34. The storage system of claim 33, wherein the hybrid storage aggregate includes a volume having a contiguous volume block number space spanning the first tier of non-volatile storage media and a second tier of non-volatile storage media of the multiple tiers of heterogeneous storage media, wherein the first tier and second tier of non-volatile storage media include storage media of different technologies.

35. The storage system of claim 34, wherein the first tier of non-volatile storage media is a lower-latency tier of storage media than the second tier of non-volatile storage media.

36. The storage system of claim 35, wherein the processor is configured to request a block of data from a storage location in the cache memory.

37. The storage system of claim 36, wherein the storage manager is configured to:

return the block of data to the processor from the higher-level cache memory when the copy of the block of data is located within the higher-level cache memory; and

return the block of data to the processor from the lower-level cache memory when the copy of the block of data is located in the lower-level cache memory and not located in the higher-level cache memory.

38. The storage system of claim 37, wherein the storage manager is further configured to return the block of data to the processor from the second tier of non-volatile storage media when the copy of the block of data is located in the second tier of non-volatile storage media only.

39. The storage system of claim 38, wherein the copy of the block of data is stored in both the first tier and second tier of non-volatile storage media during a period of time after the block of data is relocated from the first tier to the second tier of non-volatile storage media and before metadata pointers associated with the block of data have been updated to reference the copy of the block of data stored in the second tier of non-volatile storage media.

40. The storage system of claim 39, wherein the block of data stored in the lower-level cache memory comprising the first tier of non-volatile storage media is accessible to a failover partner storage system in the event of a power outage or other system failure.

41. A method of caching blocks of data in storage system having a hybrid storage aggregate with multiple tiers of heterogeneous storage media, the method comprising:

incorporating a lower-latency tier of non-volatile storage media and a higher-latency tier of non-volatile storage media into the hybrid storage aggregate;

receiving a request for a block of data stored in the hybrid storage aggregate, wherein a copy of the block of data is stored in both the lower-latency tier of non-volatile storage media and the higher-latency tier of non-volatile storage media; and

returning the block of data from the lower-latency tier of non-volatile storage media in response to the request without having to access the higher-latency tier of non-volatile storage media.

42. The method of claim 41, wherein the copy of the block of data is stored in both the lower-latency tier and higher-latency tier of non-volatile storage media during a period of time after the block of data is relocated from the lower-latency tier to the higher-latency tier of non-volatile storage media and before metadata pointers associated with the block of data have been updated to point to the copy of the block of data stored in the higher-latency tier of non-volatile storage media.

43. The method of claim 42, wherein the block of data stored in the lower-latency tier of storage media is accessible to a failover partner storage system in the event of a power outage or other system failure.