Disk activity balancing in mapped raid storage arrays by data remapping

Patent metadata

Field	Value
Publication number	US-10146459-B1
Application number	US-201715581944-A
Country	US
Kind code	B1
Filing date	Apr 28, 2017
Priority date	Apr 28, 2017
Publication date	Dec 4, 2018
Grant date	Dec 4, 2018

Abstract

Techniques are disclosed which monitor the distribution of I/O activity across the storage drives of a data storage system and detect non-uniform distributions of I/O operations across the storage drives by monitoring I/O operations directed to disk extents supporting mapped RAID extents. In response to detecting non-uniform distributions of I/O operations across the storage drives, these techniques identify LUNs with non-uniform I/O patterns and redistribute data stored in those LUNs across RAID extents supporting those LUNs.

First claim

What is claimed is:

1. A method of storing data in a data storage system using mapped RAID, the method comprising:
providing, by processing circuitry:
a set of RAID extents for storing data in the data storage system, each RAID extent having a respective set of disk extents allocated thereto, each disk extent supported by an extent of storage on a physical storage drive of a set of physical storage drives in the data storage system; and
a set of logical units (LUNs), each LUN storing LUN data in a set of RAID extents belonging to that LUN;
calculating, by the processing circuitry, a respective drive temperature for each physical storage drive of the set of physical storage drives in the data storage system, the respective drive temperature for each physical storage drive indicating a quantity of I/O requests directed to that physical storage drive by the processing circuitry over an interval of time;
generating an excessive variation result indicating that a level of variation amongst the respective drive temperatures for each physical storage drive of the set of physical storage drives exceeds a threshold; and
in response to generating the excessive variation result, performing a LUN shuffling operation to redistribute LUN data of a LUN across the set of RAID extents belonging to that LUN.

2. The method of claim 1 , wherein calculating, by the processing circuitry, the respective drive temperature for each physical storage drive of the set of physical storage drives includes measuring a respective extent temperature for each disk extent of each physical storage drive, the respective extent temperature for disk extent indicating a quantity of I/O requests directed to that disk extent by the processing circuitry over the interval of time.

3. The method of claim 2 , wherein the method further comprises selecting, a particular LUN as the LUN, by:
comparing the respective disk extent temperatures of disk extents associated with a candidate LUN to an average disk extent temperature for the candidate LUN;
generating a LUN extent comparison result indicating that at least one disk extent associated with the candidate LUN has a respective disk extent temperature differing from the average disk extent temperature for the candidate LUN; and
in response to generating the LUN extent comparison result, providing the candidate LUN as the selected LUN.

4. The method of claim 1 ,
wherein each LUN is represented as a set of chunks, each chunk being assigned to a RAID extent according to a current mapping of chunks to RAID extents; and
wherein performing the LUN shuffling operation includes transforming the current mapping of chunks to RAID extents into a shuffled mapping of chunks to RAID extents in a cyclical manner with respect to the RAID extents.

5. The method of claim 4 , wherein each chunk of each LUN corresponds to contiguous range of logical block addresses belonging to that LUN, the chunk having a data capacity in a range between 1 and 128 megabytes.

6. The method of claim 1 , wherein the respective drive temperature for each physical storage drive is calculated based exclusively on a quantity of write requests directed to that physical storage drive.

7. The method of claim 1 , wherein the method further comprises performing at subsequent time intervals:
recalculating the respective drive temperature for each physical storage drive of the set of physical storage drives over each subsequent time interval;
generating additional excessive variation results indicating that levels of variation amongst the respective drive temperatures for each physical storage drive exceed the threshold; and
in response to generating the excessive variation results, performing additional LUN shuffling operations to redistribute LUN data of one or more LUNs across the set of RAID extents belonging to those LUNs.

8. The method of claim 7 , wherein the subsequent time intervals occur once each day.

9. A data storage system comprising:
a set of physical storage drives;
memory, storing a set of instructions; and
processing circuitry coupled to the memory and the set of physical storage drives, the processing circuitry constructed and arranged to execute the instructions stored by the memory, causing the processing circuitry to:
provide a set of RAID extents for storing data in the data storage system, each RAID extent having a respective set of disk extents allocated thereto, each disk extent supported by an extent of storage on a physical storage drive of the set of physical storage drives in the data storage system;
provide a set of logical units (LUNs), each LUN storing LUN data in a set of RAID extents belonging to that LUN;
calculate a respective drive temperature for each physical storage drive of the set of physical storage drives in the data storage system, the respective drive temperature for each physical storage drive indicating a quantity of I/O requests directed to that physical storage drive by the processing circuitry over an interval of time;
generate an excessive variation result indicating that a level of variation amongst the respective drive temperatures for each physical storage drive of the set of physical storage drives exceeds a threshold; and
perform, in response to generating the excessive variation result, a LUN shuffling operation to redistribute LUN data of a LUN across the set of RAID extents belonging to that LUN.

10. The data storage system of claim 9 , wherein the processing circuitry, when calculating the respective drive temperature for each physical storage drive of the set of physical storage drives is constructed and arranged to measure a respective extent temperature for each disk extent of each physical storage drive, the respective extent temperature for disk extent indicating a quantity of I/O requests directed to that disk extent by the processing circuitry over the interval of time.

11. The data storage system of claim 10 , wherein the processing circuitry is further constructed arranged to:
compare the respective disk extent temperatures of disk extents associated with a candidate LUN to an average disk extent temperature for the candidate LUN;
generate a LUN extent comparison result indicating that at least one disk extent associated with the candidate LUN has a respective disk extent temperature differing from the average disk extent temperature for the candidate LUN; and
provide the candidate LUN as the selected LUN in response to the LUN extent comparison result.

12. The data storage system of claim 9 ,
wherein each LUN is represented as a set of chunks, each chunk being assigned to a RAID extent according to a current mapping of chunks to RAID extents; and
wherein performing the LUN shuffling operation includes transforming the current mapping of chunks to RAID extents into a shuffled mapping of chunks to RAID extents in a cyclical manner with respect to the RAID extents.

13. The data storage system of claim 9 wherein the processing circuitry is further constructed and arranged to:
recalculate the respective drive temperature for each physical storage drive of the set of physical storage drives over each subsequent time interval; generate additional excessive variation results indicating that levels of variation amongst the respective drive temperatures for each physical storage drive exceed the threshold; and
perform additional LUN shuffling operations to redistribute LUN data of one or more LUNs across the set of RAID extents belonging to those LUNs, in response to generating the excessive variation results.

14. A computer program product comprising a non-transitory, computer-readable

Assignees

• Emc Ip Holding Co Llc

Inventors

• Gao Jian
• Wahl Michael
• Han Geng
• Gao Hongpo
• Yang Lifeng
• Kang Jamin

Classifications

• G06F3/0619 Primary

  in relation to data integrity, e.g. data losses, bit errors · CPC title

• G06F3/0665

  at area level, e.g. provisioning of virtual or logical volumes · CPC title

• G06F3/0689

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

• G06F3/0616

  in relation to life time, e.g. increasing Mean Time Between Failures [MTBF] · CPC title

• G06F3/065

  Replication mechanisms · CPC title