Raid rebuild algorithm with low I/O impact

• US 10,120,769 B2
• Filed: 04/13/2016
• Issued: 11/06/2018
• Est. Priority Date: 04/13/2016
• Status: Active Grant

• Current Assignee: Dell Products LP (Dell Technologies Inc.)
• Original Assignee: Dell Products LP (Dell Technologies Inc.)
• Inventors: Sreedhar M, Deepu Syam, Agarwal, Sandeep, Kumar P. K., Krishna
• Primary Examiner(s): Guyton, Philip
• Application Number:
  US15/097,795
• Publication Number:
  US 20170300393A1
• Time in Patent Office: 937 Days
• CPC Class Codes: G06F 11/1076 Parity data used in redundant arrays of independent storages, e.g. in RAID systems
• CPC Class Codes: G06F 11/1658 Data re-synchronization of a redundant component, or initial sync of replacement, additional or spare unit
• CPC Class Codes: G06F 2201/805 Real-time

Abstract

A disclosed storage management method includes detecting an unrecoverable failure associated with a logical block of a first physical storage device that is one of a plurality of storage devices within a redundant virtual drive that also includes a hot spare drive. Data for the unrecoverable block may be rebuilt from data in the remaining storage devices and stored in a logical block of the hot spare drive. One or more logical block maps may be maintained to identify unrecoverable logical blocks and to indicate the logical blocks and storage devices to which each of the unrecoverable logical blocks is relocated. I/O operations that access “good” logical blocks are normally while accesses to unrecoverable logical blocks are rerouted according to the logical block map. One or more unrecoverable thresholds may be supported to initiate operations to replace storage devices containing unrecoverable blocks exceeding an applicable threshold.

Claims

1. A storage method, comprising:

detecting an unrecoverable error associated with a particular logical block of a virtual drive comprising a plurality of physical storage devices and a hot spare drive;

identifying a particular source disk and a particular stripe associated with the particular logical block address;

determining whether the unrecoverable error is a first error associated with the particular stripe;

responsive to determining the unrecoverable error is the first error associated with the particular stripe, performing operations comprising;

regenerating data corresponding to the particular logical block from remaining source disks in the virtual drive;

storing the regenerated data to a first stripe of the hot spare drive; and

creating an entry in a mapping structure, the entry associating the particular source disk and the particular stripe of the particular logical block address with the first stripe of the hot spare drive; and

responsive to determining the unrecoverable error is not the first unrecoverable error associated with the particular stripe, performing operations comprising;

copying a different logical block, stored in a different stripe of the particular source disk, to a second stripe of the hot spare drive;

regenerating the logical block from;

the first stripe of the hot spare drive; and

unmapped portions of the particular stripe on other source disks;

storing the regenerated logical block to the different stripe;

creating an entry mapping the particular logical block to the different stripe; and

creating an entry mapping the different logical block to the second stripe of the hot spare drive.

2. The method of claim 1, further comprising:

detecting an unrecoverable error associated with the first stripe of the hot spare drive;

regenerating data for the first stripe of the hot spare drive;

storing the regenerated data to a second stripe of the hot spare drive.

3. The method of claim 2, wherein regenerating the first stripe of the distributed hot spare drive includes regenerating from at least one remapped stripe.

4. The method of claim 3, wherein the remapped stripe is remapped to a source disk.

5. The method of claim 3, wherein the remapped stripe is remapped to the hot spare drive.

6. The method of claim 1, further comprising:

detecting a number of logical block errors on a first source disk exceeding a single disk threshold; and

performing replacement operations comprising;

copying remaining good blocks of the first source disk to a replacement disk;

copying remapped portions of the first source disk to the replacement disk; and

removing from the mapping structure, entries associated with the first source disk.

7. The method of claim 1, further comprising:

responsive to detecting a total number of logical block errors for the virtual drive exceeding a virtual drive threshold, performing operations comprising;

identifying a candidate source disk based on a number of logical block errors associated with the candidate source disk;

copying remaining good blocks of the candidate source disk to a replacement disk;

copying remapped portions of the first source disk to the replacement disk; and

removing from the mapping structure, entries associated with the first source disk.

8. A storage controller comprising:

a processor; and

a computer readable medium including processor executable instructions that, when executed by the processor, cause the processor to perform operations comprising;

detecting an unrecoverable failure associated with a logical block of a first physical storage device, wherein the first physical storage device comprises a first physical storage device of a redundant array of inexpensive drives (RAID) 5 virtual drive, the logical block comprises a first logical block of a data stripe, and wherein rebuilding data for the logical block comprises performing an exclusive or (EXOR) operation on remaining portions of the data stripe in the remaining physical storage devices;

rebuilding data for the logical block from data in remaining physical storage devices of the virtual drive;

storing rebuilt data resulting from the rebuilding of data in a first logical block of a hot spare drive;

maintaining an address map for the first physical storage device, wherein the address map;

identifies logical blocks of the first physical storage device containing unrecoverable failures associated with the first physical storage device; and

indicates relocated logical blocks for each of the logical blocks identified;

detecting a second unrecoverable error associated with a second logical block on a second physical storage device of the virtual drive;

determining whether a data stripe corresponding to the second logical block includes any other logical blocks associated with a unrecoverable error; and

responsive to determining that the second logical block is of the same data stripe as the first logical block, performing same-stripe rebuilding operations comprising;

copying data for a different logical block of the second physical storage device to a distributed hot spare drive; and

performing an exclusive OR (EXOR) of;

the relocated first logical block on the distributed hot spare drive; and

a third logical block, associated with a third physical storage device, of the virtual drive.

9. The storage controller of claim 8, wherein the operations include maintaining a respective address map for each of the plurality of physical storage devices.

10. The storage controller of claim 8, wherein the operations include:

detecting a third unrecoverable error associated with a first logical block of the distributed hot spare drive;

rebuilding the data from;

the relocated logical block of the second physical storage device; and

the third logical block associated with the third physical storage device.

11. The storage controller of claim 8, wherein the operations include:

detecting a DEVICE COUNT event triggered by a quantity of unrecoverable errors associated with a particular physical storage device exceeding a DEVICE COUNT threshold; and

performing replacement operations comprising;

copying remaining good blocks of the first source disk to a replacement disk;

copying remapped portions of the first source disk to the replacement disk; and

removing from the mapping structure, entries associated with the first source disk.

12. The storage controller of claim 11, wherein the operations include:

detecting a SPARE COUNT event triggered by a quantity of rebuilt logical blocks stored on the hot spare drive associated with a particular physical storage device exceeding a DEVICE COUNT threshold;

identifying a candidate source disk based on a number of logical block errors associated with the candidate source disk;

copying remaining good blocks of the candidate source disk to a replacement disk;

copying remapped portions of the first source disk to the replacement disk; and

removing from the mapping structure, entries associated with the first source disk.

13. An information handling system, comprising:

a processor; and

computer readable storage accessible to the processor and including processor executable program instructions that, when executed by the processor, cause the processor to perform operations comprising;

detecting an unrecoverable failure associated with a logical block of a first physical storage device, wherein the first physical storage device comprises a first physical storage device of a redundant array of inexpensive drives (RAID) 5 virtual drive and the logical block comprises a first logical block of a data stripe, and;

rebuilding data for the logical block from data in remaining physical storage devices of the virtual drive wherein rebuilding data for the logical block comprises performing an exclusive or (EXOR) operation on remaining portions the data stripe in the remaining physical storage devices;

storing rebuilt data resulting from the rebuilding of data in a first logical block of a hot spare drive; and

maintaining an address map for the first physical storage device, wherein the address map;

identifies logical blocks of the first physical storage device containing unrecoverable failures associated with the first physical storage device; and

indicates relocated logical blocks for each of the logical blocks identified;

detecting a second unrecoverable error associated with a second logical block on a second physical storage device of the virtual drive;

determining whether a data stripe corresponding to the second logical block includes any other logical blocks associated with a unrecoverable error; and

responsive to determining that the second logical block is of the same data stripe as the first logical block, performing same-stripe rebuilding operations comprising;

copying data for a different logical block of the second physical storage device to the distributed hot spare; and

performing an exclusive OR (EXOR) of;

the relocated first logical block on the hot spare drive; and

a third logical block, associated with a third physical storage device, of the virtual drive.