Control of sparing in storage systems

• US 8,201,018 B2
• Filed: 09/18/2007
• Issued: 06/12/2012
• Est. Priority Date: 09/18/2007
• Status: Active Grant

• Current Assignee: Hewlett Packard Enterprise Development LP (Hewlett-Packard Enterprise Company)
• Original Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Inventors: Navarro, Guillermo, Manic, Milos, Umberger, David K.
• Primary Examiner(s): Beausoliel, Jr., Robert
• Assistant Examiner(s): Arcos, Jeison C
• Application Number:
  US11/857,335
• Publication Number:
  US 20090077418A1
• Time in Patent Office: 1,729 Days
• Field of Search: 714/16, 714/54, 708/103
• US Class Current: 714/6.22
• 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 11/3419 by assessing time
• CPC Class Codes: G06F 11/3485 for I/O devices

Abstract

Embodiments include methods, apparatus, and systems for controlling of sparing in a storage system. In one embodiment, a method compares a first amount of time to complete sparing of data from a failed disk in a storage system with a second amount of time to complete a user request to the storage system in order to determine when to create a copy of the data from the failed disk.

Claims

1. A method, comprising:

sparing data to reconstruct a copy of the data after a disk failure in a storage system;

inputting, to a controller and while continuing to spare the data, a first parameter that includes a response time to perform an input/output (I/O) request from an application to a disk in a disk array of the storage system;

inputting, to the controller and while continuing to spare the data, a second parameter that includes an amount of data spared to the disk; and

controlling, while continuing to spare the data and using the first and second parameters, both executing I/O requests from the application to the disk and a rate of sparing the data to the disk to reconstruct the copy of the data after the disk failure in the storage system, wherein the sparing to the disk continues when the application has I/O requests pending to the disk, and the controller uses the first and second parameters to determine when to continue to spare the data and when to discontinue to spare the data.

2. The method of claim 1 further comprising:

inputting, to the controller, a third parameter that includes an amount of time spent on sparing data to the disk;

controlling, using the first, second, and third parameters, both executing I/O requests from the application to the disk and the rate of sparing to the disk.

3. The method of claim 1 further comprising:

inputting, to the controller, a third parameter that includes a number of I/Os that arrive at the disk array during a given time period;

controlling, using the first, second, and third parameters, both executing I/O requests from the application to the disk and sparing to the disk.

4. The method of claim 1, further comprising:

inputting, to the controller, a third parameter that includes a response time for the disk array to perform reads and writes from user requests;

controlling, using the first, second, and third parameters, both executing I/O requests from the application to the disk and sparing to the disk.

5. The method of claim 1 further comprising, deciding, based on the first and second parameters, whether to execute an I/O request pending in a queue from the application or continue with the sparing.

6. The method of claim 1 further comprising, determining the rate at which to spare the data to the disk based on balancing a time needed to complete the sparing and a latency of user requests for I/O operations.

7. A method, comprising:

sparing data to reconstruct a copy of the data after a disk failure in a storage system;

inputting, to a controller and while continuing to spare the data, a first parameter that includes a response time to perform an input/output (I/O) request from an application to a disk in a disk array of the storage system;

inputting, to the controller and while continuing to spare the data, a second parameter that includes an amount of data spared to the disk;

fuzzifying the first parameter and second parameter; and

controlling, while continuing to spare the data and using the first and second parameters, both executing I/O requests from the application to the disk and a rate of sparing the data to the disk to reconstruct the copy of the data after the disk failure in the storage system, wherein the sparing to the disk continues when the application has I/O requests pending to the disk, and the controller uses the first and second parameters to determine when to continue to spare the data and when to temporarily discontinue to spare the data.

8. A computer readable medium storing instructions for causing a computer to execute a method, comprising:

commencing a sparing operation to spare data to a disk after a disk failure in a storage system;

measuring input parameters that include a response time to perform an input/output (I/O) request from an application to the storage system and an amount of data spared to the disk;

fuzzifying the input parameters to determine when to continue the sparing operation to spare the data the disk and when to temporarily discontinue the sparing operation to spare the data to the disk; and

continuing the sparing operation to spare the data to the disk while user requests are pending to the disk.

9. The computer readable medium of claim 8, wherein the input parameters are fuzzified to values of low, medium, and high.

10. The computer readable medium of claim 8 further comprising, utilizing three parameters to select one of continuing the sparing operation of sparing the data to the disk and stopping the sparing operation of sparing the data to the disk, wherein the three parameters include the response time to perform the I/O request, the amount of data spared, and an amount of time spent on sparing data to the disk.

11. The computer readable medium of claim 8 further comprising:

normalizing a first parameter that includes a response time to perform a user request;

determining, based on the first parameter, whether to continue the sparing operation to the disk or temporarily hold the sparing operation.

12. The computer readable medium of claim 8 further comprising:

determining a maximum latency for responding to a user request to the storage system;

determining, based on the maximum latency, whether to continue the sparing operation or temporarily hold the sparing operation.

13. The computer readable medium of claim 8 further comprising:

determining a time elapsed since the disk failure;

calculating, based on the time elapsed, when in time to commence sparing the data to the disk.

14. The computer readable medium of claim 8 further comprising:

temporarily holding the sparing operation;

determining an amount of data spared to the disk during the sparing operation;

calculating, based on the amount of data spared to the disk, when in time to resume the sparing operation of sparing to the disk.

15. A storage system, comprising:

a memory for storing an algorithm;

a processor for executing the algorithm to;

spare data to a disk after a disk failure in a disk array of a storage system; and

analyze plural parameters to determine when to temporarily stop reconstructing the data to the disk in order to execute an input/output (I/O) request from an application to the disk array, wherein the plural parameters include a response time to perform the I/O request and an amount of data spared, and wherein sparing to the disk continues while I/O requests from a host are pending to the disk in a queue in the storage system.

16. The storage system of claim 15, wherein the processor further executes the algorithm to determine, based on the plural parameters, between executing a pending I/O request to the disk from the host and continuing the sparing to the disk while the I/O request remains in the queue.

17. The storage system of claim 15, wherein the processor further executes the algorithm to determine a rate at which to reconstruct the data to the disk, the rate based on a balance between a time needed to complete reconstructing the data to the disk and a latency of user requests for I/O operations.

18. The storage system of claim 15, wherein the plural parameters are normalized and input into a fuzzy neural net that includes a first section for outputting first fuzzy values and a second section that receives the fuzzy values to determine whether to continue sparing or discontinue sparing.

19. The storage system of claim 15, wherein the processor further executes the algorithm to determine a fraction of data already spared in order to determine when to favor sparing over latency to perform user requests.

20. A storage system, comprising:

a memory for storing an algorithm;

a processor for executing the algorithm to;

reconstruct data to a disk after a disk failure in a disk array of a storage system; and

fuzzify plural parameters to determine when to continue to reconstruct the data to the disk and when to temporarily stop reconstructing the data to the disk in order to execute an input/output (I/O) request from an application to the disk array, wherein the plural parameters include a response time to perform the I/O request and an amount of data reconstructed, and wherein reconstructing the data to the disk continues while I/O requests from a host are pending to the disk in a queue in the storage system.