Machine learning to enhance redundant array of independent disks rebuilds

Patent metadata

Field	Value
Publication number	US-10691543-B2
Application number	US-201715811822-A
Country	US
Kind code	B2
Filing date	Nov 14, 2017
Priority date	Nov 14, 2017
Publication date	Jun 23, 2020
Grant date	Jun 23, 2020

Abstract

Machine logic (for example, software) for storing: (i) a plurality of files across multiple disk drives of a RAID array and (ii) checksum data for the files. The machine logic rules assign priority values to each file based on one or more attributes of the files. These priority values are used to determine the order in which files are rebuilt using the checksum data so that the most important files are rebuilt first.

First claim

What is claimed is:

1. A computer-implemented method comprising:
storing a plurality of blocks of data in a striped redundant array of independent devices (RAID) storage system that includes a plurality of storage devices;
storing parity data for the plurality of blocks of data;
updating, by machine logic, a priority assignment algorithm for assigning the priority data values based, at least in part, on a review of relative weights assigned to a set of factors considered when assigning priority data values and what priority data values were assigned to each block of data of the plurality of blocks of data;
for each given block of data of the plurality of blocks of data, assigning a priority data value to the given block of data based on the priority assignment algorithm; and
responsive to a failure of a first storage device of the plurality of storage devices, rebuilding, using the parity data and data of the blocks of data stored on the plurality of storage devices other than the first storage device, blocks of data that were stored on the first storage device in an order determined by priority values of the blocks of data that were stored on the first storage device.

2. The method of claim 1 wherein the assigning a priority data value is based, at least in part, upon at least one of the following: block size, read frequency, read trend, data protection inclusion, and number of references to a deduped block.

3. The method of claim 1 wherein the assigning a priority data value includes clustering each block of data of the plurality of blocks of data with a k-means clustering algorithm.

4. The method of claim 1 wherein the rebuilding rebuilds the blocks of data onto a second storage device.

5. The method of claim 4 further comprising:
remapping the parity data associated with the data blocks of the first storage device to the data blocks of the second storage device.

6. The method of claim 1 wherein the assigning a priority data value for the given block also assigns an identical priority value to blocks of data containing metadata associated with the given block.

7. A computer program product comprising:
a non-transitory computer readable storage medium; and
computer code stored on the non-transitory computer readable storage medium, with the computer code including instructions for causing a processor(s) set to perform operations including the following:
storing a plurality of blocks of data in a striped redundant array of independent devices (RAID) storage system that includes a plurality of storage devices,
storing parity data for the plurality of blocks of data,
updating, by machine logic, a priority assignment algorithm for assigning the priority data values based, at least in part, on a review of relative weights assigned to a set of factors considered when assigning priority data values and what priority data values were assigned to each block of data of the plurality of blocks of data,
for each given block of data of the plurality of blocks of data, assigning a priority data value to the given block of data based on the priority assignment algorithm, and
responsive to a failure of a first storage device of the plurality of storage devices, rebuilding, using the parity data and data of the blocks of data stored on the plurality of storage devices other than the first storage device, blocks of data that were stored on the first storage device in an order determined by priority values of the blocks of data that were stored on the first storage device.

8. The product of claim 7 wherein the assigning a priority data value is based, at least in part, upon at least one of the following: block size, read frequency, read trend, data protection inclusion, and number of references to a deduped block.

9. The product of claim 7 wherein the assigning a priority data value includes clustering each block of data of the plurality of blocks of data with a k-means clustering algorithm.

10. The product of claim 7 wherein the rebuilding rebuilds the blocks of data onto a second storage device.

11. The product of claim 10 wherein the computer code further includes instructions for causing the processor(s) set to perform the following operation:
remapping the parity data associated with the data blocks of the first storage device to the data blocks of the second storage device.

12. The product of claim 7 wherein the assigning a priority data value for the given block also assigns an identical priority value to blocks of data containing metadata associated with the given block.

13. A computer system comprising:
a processor(s) set;
a machine readable storage device; and
computer code stored on the machine readable storage device, with the computer code including instructions for causing the processor(s) set to perform operations including the following:
storing a plurality of blocks of data in a striped redundant array of independent devices (RAID) storage system that includes a plurality of storage devices,
storing parity data for the plurality of blocks of data,
updating, by machine logic, a priority assignment algorithm for assigning the priority data values based, at least in part, on a review of relative weights assigned to a set of factors considered when assigning priority data values and what priority data values were assigned to each block of data of the plurality of blocks of data,
for each given block of data of the plurality of blocks of data, assigning a priority data value to the given block of data based on the priority assignment algorithm, and
responsive to a failure of a first storage device of the plurality of storage devices, rebuilding, using the parity data and data of the blocks of data stored on the plurality of storage devices other than the first storage device, blocks of data that were stored on the first storage device in an order determined by priority values of the blocks of data that were stored on the first storage device.

14. The system of claim 13 wherein the assigning a priority data value is based, at least in part, upon at least one of the following: block size, read frequency, read trend, data protection inclusion, and number of references to a deduped block.

15. The system of claim 13 wherein the assigning a priority data value includes clustering each block of data of the plurality of blocks of data with a k-means clustering algorithm.

16. The system of claim 13 wherein the rebuilding rebuilds the blocks of data onto a second storage device.

17. The system of claim 13 wherein the assigning a priority data value for the given block also assigns an identical priority value to blocks of data containing metadata associated with the given block.

Assignees

• IBM

Inventors

• Harrington Shaun E
• Barajas Gonzalez Emmanuel
• Watkins Jr Franklin D
• Anna Gary

Classifications

• G06N20/00

  Machine learning · CPC title

• G06F11/1092 Primary

  Rebuilding, e.g. when physically replacing a failing disk · CPC title

• G06F3/0689

  Disk arrays, e.g. RAID, JBOD · CPC title

• G06F3/0659

  Command handling arrangements, e.g. command buffers, queues, command scheduling · CPC title

• G06F3/0632

  by initialisation or re-initialisation of storage systems · CPC title