Relocating data from an end of life storage drive based on storage drive loads in a data storage system using mapped RAID (redundant array of independent disks) technology

• US 10,289,336 B1
• Filed: 02/05/2018
• Issued: 05/14/2019
• Est. Priority Date: 01/31/2018
• Status: Active Grant

• Current Assignee: Emc IP Holding Company LLC (Dell Technologies Inc.)
• Original Assignee: Emc IP Holding Company LLC (Dell Technologies Inc.)
• Inventors: Liu, Qingyun, Kang, Jamin, Gong, Shaoqin, Sun, Ree, Gao, Jian
• Primary Examiner(s): Portka, Gary J
• Application Number:
  US15/888,647
• Time in Patent Office: 463 Days
• Field of Search: 711114, 714 622, 714 63
• CPC Class Codes: G06F 11/1076 Parity data used in redundant arrays of independent storages, e.g. in RAID systems
• CPC Class Codes: G06F 3/0616 in relation to life time, e.g. increasing Mean Time Between Failures [MTBF]
• CPC Class Codes: G06F 3/0619 in relation to data integrity, e.g. data losses, bit errors
• CPC Class Codes: G06F 3/0647 Migration mechanisms
• CPC Class Codes: G06F 3/067 Distributed or networked storage systems, e.g. storage area networks [SAN], network attached storage [NAS]
• CPC Class Codes: G06F 3/0689 Disk arrays, e.g. RAID, JBOD

Abstract

Technology for relocating data stored in an end of life source storage drive that identifies individual destination drive extents corresponding to drive extents located in the source storage drive by selecting individual storage drives within a set of healthy storage drives in a group of storage drives in a round robin manner and identifying an individual destination drive extent from each selected storage drive. The corresponding destination drive extents replace the drive extents located in the source storage drive, and are used to store data copied from the drive extents located in the source storage drive. Data is copied in parallel from a set of multiple drive extents located in the source storage drive that is selected based on the current total input/output load determined for each storage drive in the set of healthy storage drives, and on the storage drives on which the corresponding destination drive extents are located.

Claims

1. A method of relocating data stored in a storage drive that is one of a group of storage drives communicably coupled to a storage processor in a data storage system that provides RAID (Redundant Array of Independent Disks) data protection for at least one storage object, the method comprising:

generating a drive extent pool from the storage drives communicably coupled to the storage processor, wherein the drive extent pool indicates a plurality of drive extents located in the storage drives, wherein each drive extent indicated by the drive extent pool comprises a contiguous region of non-volatile data storage located in one of the storage drives, and wherein each drive extent has a size equal to a predetermined drive extent size;

generating a RAID mapping table, wherein the RAID mapping table contains a plurality of RAID extent entries, wherein each RAID extent entry indicates a unique plurality of drive extents that are allocated from the drive extent pool to the RAID extent entry and that are used by the storage processor to persistently store host data written to a portion of an address space of the storage object corresponding to the RAID extent entry and associated parity data, and wherein each RAID extent entry indicates the same total number of drive extents;

in response to receiving an end of life indication from a source storage drive within the group of storage drives, relocating the data stored in the source storage drive by;

i) for each drive extent located in the source storage drive, identifying a corresponding destination drive extent to replace the drive extent located in the source storage drive by selecting, in a round robin manner, a next storage drive from within a set of healthy storage drives in the group of storage drives and allocating a free drive extent that is located in the selected storage drive from the drive extent pool, and

ii) after a corresponding destination drive extent has been identified for each drive extent located in the source storage drive, generating a parallel copy list, wherein the parallel copy list indicates drive extents located in the source storage drive from which data is copied in parallel by the storage processor to the corresponding destination drive extents, wherein the parallel copy list is generated based on current total I/O loads of the storage drives in the set of healthy storage drives, and wherein each one of the drive extents located in the source storage drive indicated by the parallel copy list has a corresponding destination drive extent located on a healthy storage drive that has a lowest current total I/O load at a time at which the indication of that drive extent located on the source storage drive is added to the parallel copy list.

2. The method of claim 1, wherein generating the parallel copy list indicating drive extents located in the source storage drive from which data is copied by the storage processor in parallel to the corresponding destination drive extents further comprises:

detecting that the parallel copy list is not full; and

responsive to detecting that the parallel copy list is not full, i) determining a current total I/O load for each storage drive in the set of healthy storage drives, ii) identifying a storage drive in the set of healthy storage drives having a lowest current total I/O load, iii) identifying a drive extent located in the source storage drive corresponding to a destination drive extent located in the healthy storage drive having the lowest current total I/O load, and iv) adding, to the parallel copy list, an indication of the drive extent located in the source storage drive corresponding to the destination drive extent located in the healthy storage drive having the lowest current total I/O load.

3. The method of claim 2, wherein determining the current total I/O load for each storage drive in the set of healthy storage drives further comprises:

for each storage drive in the set of healthy storage drives, determining a current host I/O load equal to a total number of host I/O operations currently being performed on the storage drive;

for each storage drive in the set of healthy storage drives, determining a current background I/O load equal to a total number of background I/O operations currently being performed on the storage drive, wherein the background I/O operations are performed at a lower priority than the host I/O operations, and wherein the background I/O operations include copy operations performed in parallel by the storage processor from drive extents located in the source storage drive and indicated by the parallel copy list to corresponding destination drive extents; and

for each storage drive in the set of healthy storage drives, determining a current total I/O load that is equal to the sum of the current host I/O load for the storage drive and the current background I/O load for the storage drive.

4. The method of claim 1, wherein selecting, in a round robin manner, a next storage drive from within the set of healthy storage drives further comprises selecting a next storage drive in a predetermined sequential ordering of the set of healthy storage drives in following a storage drive on which is located the drive extent that was last allocated from the drive extent pool as a destination drive extent.

5. The method of claim 4, wherein the predetermined sequential ordering of the set of health storage drives comprises an order of slots into which the storage drives are connected within an array of storage drives that is communicably coupled to the storage processor.

6. The method of claim 5, wherein selecting, in a round robin manner, a next storage drive from within the set of healthy storage drives further comprises, before allocating a free drive extent that is located in an initially selected storage drive from the drive extent pool:

determining whether any drive extent located in the initially selected storage drive was allocated to a RAID extent entry to which the drive extent located in the source storage drive was also allocated; and

responsive to determining that a drive extent located in the initially selected storage drive was allocated to a RAID extent entry to which the drive extent located in the source storage drive was also allocated to the RAID extent entry in the RAID mapping table to which the drive extent located in the source storage drive was allocated, selecting a next storage drive in the predetermined sequential ordering of the set of healthy storage drives following the initially selected storage drive as the selected storage drive.

7. The method of claim 2, wherein detecting that the parallel copy list is not full comprises detecting that a background operation copying data from a drive extent located in the source storage drive indicated by the parallel copy list to the corresponding destination dive extent has completed.

8. The method of claim 7, wherein the end of life indication received from the source storage drive indicates that a threshold maximum number of reserved disk sectors in the source storage drive have been used to replace original disk sectors to which I/O operations were directed, in response to failures of the original disk sectors.

9. The method of claim 8, wherein the threshold maximum number of the reserved disk sectors comprises a total number of reserved disk sectors in the source storage drive.

10. A data storage system that provides RAID (Redundant Array of Independent Disks) data protection for at least one storage object, comprising:

at least one storage processor including processing circuitry and a memory storing program code executable on the processing circuitry;

at least one group of storage drives communicably coupled to the storage processor; and

wherein the storage processor is configured and arranged to;

generate a drive extent pool from the storage drives communicably coupled to the storage processor, wherein the drive extent pool indicates a plurality of drive extents located in the storage drives, wherein each drive extent indicated by the drive extent pool comprises a contiguous region of non-volatile data storage located in one of the storage drives, and wherein each drive extent has a size equal to a predetermined drive extent size;

generate a RAID mapping table, wherein the RAID mapping table contains a plurality of RAID extent entries, wherein each RAID extent entry indicates a unique plurality of drive extents that are allocated from the drive extent pool to the RAID extent entry and that are used by the storage processor to persistently store host data written to a portion of an address space of the storage object corresponding to the RAID extent entry and associated parity data, and wherein each RAID extent entry indicates the same total number of drive extents;

wherein to relocate data stored in a source storage drive within the group of storage drives in response to receiving an end of life indication from the source storage the storage processor is configured and arranged to;

i) for each drive extent located in the source storage drive, identify a corresponding destination drive extent to replace the drive extent located in the source storage drive by selecting, in a round robin manner, a next storage drive from within a set of healthy storage drives in the group of storage drives and allocating a free drive extent that is located in the selected storage drive from the drive extent pool, and

ii) after a corresponding destination drive extent has been identified for each drive extent located in the source storage drive, generate a parallel copy list, wherein the parallel copy list indicates drive extents located in the source storage drive from which data is copied in parallel by the storage processor to the corresponding destination drive extents, wherein the parallel copy list is generated based on current total I/O loads of the storage drives in the set of healthy storage drives, wherein each one of the drive extents located in the source storage drive indicated by the parallel copy list has a corresponding destination drive extent located on a storage drive that has a lowest current total I/O load at a time at which the indication of that drive extent located on the source storage drive is added to the parallel copy list.

11. The data storage system of claim 10, wherein to generate the parallel copy list indicating drive extents located in the source storage drive from which data is copied by the storage processor in parallel to the corresponding destination drive extents the storage processor is further configured and arranged to:

detect that the parallel copy list is not full; and

responsive to detecting that the parallel copy list is not full, i) determine a current total I/O load for each storage drive in the set of healthy storage drives, ii) identify a storage drive in the set of healthy storage drives having a lowest current total I/O load, iii) identify a drive extent located in the source storage drive corresponding to a destination drive extent located in the storage drive having the lowest current total I/O load, and iv) add, to the parallel copy list, an indication of the drive extent located in the source storage drive corresponding to the destination drive extent located in the storage drive having the lowest current total I/O load.

12. The data storage system of claim 11, wherein to determine the current total I/O load for each storage drive in the set of healthy storage drives the storage processor is further configured and arranged to:

for each storage drive in the set of healthy storage drives, determine a current host I/O load equal to a total number of host I/O operations currently being performed on the storage drive;

for each storage drive in the set of healthy storage drives, determine a current background I/O load equal to a total number of background I/O operations currently being performed on the storage drive, wherein the background I/O operations are performed at a lower priority than the host I/O operations, and wherein the background I/O operations include background copy operations performed in parallel by the storage processor from drive extents located in the source storage drive and indicated by the parallel copy list to corresponding destination drive extents; and

for each storage drive in the set of healthy storage drives, determine a current total I/O load that is equal to the sum of the current host I/O load for the storage drive and the current background I/O load for the storage drive.

13. The data storage system of claim 10, wherein to select, in a round robin manner, a next storage drive from within the set of healthy storage drives, the storage processor is configured and arranged to identify a next storage drive in a predetermined sequential ordering of the set of healthy storage drives following a storage drive on which is located the drive extent that was last allocated from the drive extent pool as a destination drive extent.

14. The data storage system of claim 13, wherein the predetermined sequential ordering of the set of healthy storage drives in the group of storage drives comprises an order of slots into which the storage drives are connected within an array of storage drives that is communicably coupled to the storage processor.

15. The data storage system of claim 14, wherein to select, in a round robin manner, a next storage drive from within the set of healthy storage drives the storage processor is further configured and arranged to, before allocating a free drive extent that is located in an initially selected storage drive from the drive extent pool:

determine whether any drive extent located in the initially selected storage drive was allocated to a RAID extent entry to which the drive extent located in the source storage drive was also allocated; and

responsive to a determination that a drive extent located in the initially selected storage drive was allocated to a RAID extent entry to which the drive extent located in the source storage drive was also allocated, select a next storage drive in the predetermined sequential ordering of the set of healthy storage drives in the group of storage drives following the initially selected storage drive as the selected storage drive.

16. The data storage system of claim 11, wherein to detect that the parallel copy list is not full the storage processor is further configured and arranged to detect that a background operation copying data from a drive extent located in the source storage drive indicated by the parallel copy list to the corresponding destination dive extent has completed.

17. The data storage system of claim 16, wherein the end of life indication received from the source storage drive indicates that a threshold maximum number of reserved disk sectors in the source storage drive have been used to replace original disk sectors to which I/O operations were directed, in response to failures of the original disk sectors.

18. The data storage system of claim 17, wherein the threshold maximum number of the reserved disk sectors comprises a total number of reserved disk sectors in the source storage drive.

19. A non-transitory computer readable medium for relocating data stored in a storage drive that is one of a group of storage drives communicably coupled to a storage processor in a data storage system that provides RAID (Redundant Array of Independent Disks) data protection for at least one storage object, comprising instructions stored thereon, that when executed on processing circuitry in a storage processor, perform the steps of:

generating a drive extent pool from the storage drives communicably coupled to the storage processor, wherein the drive extent pool indicates a plurality of drive extents located in the storage drives, wherein each drive extent indicated by the drive extent pool comprises a contiguous region of non-volatile data storage located in one of the storage drives, and wherein each drive extent has a size equal to a predetermined drive extent size;

generating a RAID mapping table, wherein the RAID mapping table contains a plurality of RAID extent entries, wherein each RAID extent entry indicates a unique plurality of drive extents that are allocated from the drive extent pool to the RAID extent entry and that are used by the storage processor to persistently store host data written to a portion of an address space of the storage object corresponding to the RAID extent entry and associated parity data, and wherein each RAID extent entry indicates the same total number of drive extents;

in response to receiving an end of life indication from a source storage drive within the group of storage drives, relocating the data stored in the source storage drive by;

i) for each drive extent located in the source storage drive, identifying a corresponding destination drive extent to replace the drive extent located in the source storage drive by selecting, in a round robin manner, a next storage drive from within a set of healthy storage drives in the group of storage drives and allocating a free drive extent that is located in the selected storage drive from the drive extent pool, and

ii) after a corresponding destination drive extent has been identified for each drive extent located in the source storage drive, generating a parallel copy list, wherein the parallel copy list indicates drive extents located in the source storage drive from which data is copied in parallel by the storage processor to the corresponding destination drive extents, wherein the parallel copy list is generated based on current total I/O loads of the storage drives in the set of healthy storage drives, wherein each one of the drive extents located in the source storage drive indicated by the parallel copy list has a corresponding destination drive extent located on a storage drive having a lowest current total I/O load at a time at which the indication of that drive extent located on the source storage drive is added to the parallel copy list.