Method for dynamically implementing N+K redundancy in a storage subsystem

• US 7,308,532 B1
• Filed: 11/12/2004
• Issued: 12/11/2007
• Est. Priority Date: 11/12/2004
• Status: Active Grant

• Current Assignee: Oracle America, Inc. (Oracle Corporation)
• Original Assignee: Sun Microsystems Incorporated (Oracle Corporation)
• Inventors: Wood, Robert B., Talagala, Nisha D.
• Primary Examiner(s): Shah, Sanjiv
• Assistant Examiner(s): Ayash, Marwan
• Application Number:
  US10/987,625
• Time in Patent Office: 1,124 Days
• Field of Search: None
• US Class Current: 711/112
• CPC Class Codes: G06F 11/1076 Parity data used in redundant arrays of independent storages, e.g. in RAID systems
• CPC Class Codes: G06F 11/1088 Reconstruction on already foreseen single or plurality of spare disks
• CPC Class Codes: G06F 11/2094 Redundant storage or storage space G06F11/2056 takes precedence

Abstract

In one embodiment, a storage subsystem includes a plurality of storage arrays each including a plurality of storage devices. The storage subsystem also includes a plurality of array controllers each coupled to one or more of the plurality of storage arrays. One or more of the arrays corresponds to a failure group. Each array controller may create a storage volume including storage devices belonging to one or more of plurality of storage arrays. In addition, the storage subsystem includes a redundancy controller that may be configured to implement N+K redundancy. The redundancy controller includes configuration functionality that may determine a number of redundant system data blocks to be stored on different storage devices for a given stripe of data that is dependent upon particular values of N and K and upon physical system configuration information.

Claims

1. A storage subsystem comprising:

a plurality of storage arrays each including a plurality of storage devices;

a plurality of array controllers each coupled to one or more of the plurality of storage arrays and configured to create a storage volume including storage devices belonging to one or more of the plurality of storage arrays, wherein one or more arrays of the plurality of storage arrays corresponds to a failure group; and

a redundancy controller coupled to the plurality of array controllers and configured to implement N+K redundancy, wherein the redundancy controller further includes configuration functionality configured to automatically determine a number of redundant system data blocks to be stored on different storage devices for a given stripe of data dependent upon particular values of N and K provided as system configuration parameters and dependent upon physical system configuration information;

wherein the configuration functionality is further configured to;

determine a total number of failure groups (G′

);

determine a maximum number of storage devices (M) in any one of the storage arrays in response to determining that N+K>

G′

; and

allocate M−

K additional storage devices to another portion of storage devices in response to determining M>

K and in response to determining there are M−

K failure groups that include fewer than M storage devices allocated to the storage volume.

2. The storage subsystem as recited in claim 1, wherein the particular values of N and K are initial values provided as system configuration parameters.

3. The storage subsystem as recited in claim 2, wherein the system configuration parameters include one or more service level agreements that specify a particular storage volume availability and storage utilization.

4. The storage subsystem as recited in claim 2, wherein the system configuration parameters include a particular storage utilization associated with a software application.

5. The storage subsystem as recited in claim 1, wherein in response to detecting a change in the physical system configuration information, the configuration functionality is further configured to dynamically and automatically reconfigure the storage volume and to determine the number of redundant system data blocks to be stored on different storage devices for a given stripe of data.

6. The storage subsystem as recited in claim 5, wherein the physical system configuration information includes a total number of the storage devices and a total number of failure groups.

7. The storage subsystem as recited in claim 1, wherein a failure of a storage device is a change in the physical system configuration information.

8. The storage subsystem as recited in claim 1, wherein the configuration functionality is further configured to encode redundant system data blocks with an N+K redundancy code for reconstructing the system data.

9. The storage subsystem as recited in claim 1, wherein the configuration functionality is further configured to:

determine the total number of available storage devices in all of the failure groups (T);

allocate one available storage device from each of the failure groups until N+K storage devices are allocated in response to determining that N+K≦

G′

; and

allocate from all of the failure groups until N+K storage devices are allocated in response to determining that N+K>

G′ and

N+K≦

T.

10. The storage subsystem as recited in claim 1, wherein the configuration functionality is further configured to provide a failure notification in response to determining there are no M−

K failure groups that include fewer than M storage devices allocated to the storage volume.

11. The storage subsystem as recited in claim 1, wherein in response to detecting a change in the system configuration parameters, the configuration functionality is further configured to dynamically and automatically reconfigure the storage volume and to determine a new number of redundant system data blocks to be stored on different storage devices for a given stripe of data.

12. The storage subsystem as recited in claim 10, wherein the configuration functionality is further configured to reallocate one or more storage devices belonging to existing storage volumes as spare storage devices by reducing a number of storage devices allocated as redundant storage devices in the existing volumes in response to determining that N+K>

G′ and

N+K>

T.

13. A method for dynamic configuration of a storage subsystem including a plurality of storage arrays each including a plurality of storage devices, said method comprising:

creating a storage volume including storage devices belonging to one or more of plurality of storage arrays, wherein one or more arrays of the plurality of storage arrays corresponds to a failure group;

implementing N+K redundancy, by automatically determining a number of redundant system data blocks to be stored on different storage devices for a given stripe of data dependent upon particular values of N and K provided as system configuration parameters and dependent upon physical system configuration information;

determining a total number of failure groups (G′

);

determining a maximum number of storage devices (M) in any one of the storage arrays in response to determining that N+K>

G′

; and

allocating M−

K additional storage devices to another portion of storage devices in response to determining M>

K and in response to determining there are M−

K failure groups that include fewer than M storage devices allocated to the storage volume.

14. The method as recited in claim 13, wherein the system configuration parameters include one or more service level agreements that specify a particular storage volume availability and storage utilization using the particular values of N and K as initial values.

15. The method as recited in claim 13 further comprising dynamically reconfiguring the storage volume and determining the number of redundant system data blocks to be stored on different storage devices for a given stripe of data in response to detecting a change in the physical system configuration information.

16. The method as recited in claim 15, wherein the physical system configuration information includes a total number of the storage devices and a total number of the failure groups.

17. The method as recited in claim 13 further comprising:

in response to detecting a failure associated with a storage device in the storage volume, allocating a spare storage device belonging to a same failure groups as the failing storage device and reconstructing the system data to the allocated storage device; and

allocating a spare storage device from another failure groups that does not include a storage device belonging to the storage volume in response to determining there are no spare storage devices available in the same failure groups as the failing storage device.

18. The method as recited in claim 13, further comprising dynamically and automatically reconfiguring the storage volume and determining a new number of redundant system data blocks to be stored on different storage devices for a given stripe of data in response to detecting a change in system configuration parameters.

19. A storage subsystem comprising:

a plurality of storage arrays each including a plurality of storage devices;

a plurality of array controllers each coupled to one or more of the plurality of storage arrays and configured to create a storage volume including storage devices belonging to one or more of the plurality of storage arrays, wherein one or more arrays of the plurality of storage arrays corresponds to a failure group; and

a redundancy controller coupled to the plurality of array controllers and configured to implement N+K redundancy, wherein the redundancy controller further includes configuration functionality configured to automatically determine a number of redundant system data blocks to be stored on different storage devices for a given stripe of data dependent upon particular values of N and K provided as system configuration parameters and dependent upon physical system configuration information;

wherein the configuration functionality is further configured to;

in response to detecting a failure associated with a storage device in the storage volume, allocate a spare storage device belonging to a same failure group as the failing storage device and reconstruct the system data to the allocated storage device;

allocate a spare storage device from another failure group that does not include a storage device belonging to the storage volume in response to determining there are no spare storage devices available in the same failure groups as the failing storage device;

determine a maximum number of storage devices (M) in any of the failure groups and allocate a spare storage device belonging to a storage array that is not already contributing M storage devices in response to determining there are no spare storage devices from another failure group that does not include a storage device belonging to the storage volume; and

allocate two storage devices from two different failure groups, using the first of the two storage devices to reconstruct the contents of the failed storage device and using the second of the two storage devices to create a new redundant storage device in response to determining there are no spare storage devices belonging to a failure group that is not already contributing M storage devices.