Method and apparatus for path selection of storage systems

• US 10,416,914 B2
• Filed: 12/16/2016
• Issued: 09/17/2019
• Est. Priority Date: 12/22/2015
• 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, Bing, Lv, Man, Ni, James Lei, Mei, Martin Chaojun
• Primary Examiner(s): Grullon, Francisco A
• Application Number:
  US15/381,428
• Publication Number:
  US 20170177264A1
• Time in Patent Office: 1,005 Days
• CPC Class Codes: G06F 3/0613 in relation to throughput
• CPC Class Codes: G06F 3/0635 by changing the path, e.g. traffic rerouting, path reconfiguration
• CPC Class Codes: G06F 3/0653 Monitoring storage devices or systems
• CPC Class Codes: G06F 3/0689 Disk arrays, e.g. RAID, JBOD

Abstract

Embodiments of the present disclosure provide methods and apparatus for path selection in a storage system. The storage system includes a plurality of storage devices, each storage device being connected to a plurality of different HBA ports of a server via a plurality of paths; the method comprises: monitoring a system-wide performance metric of the storage system to obtain dynamic performance statistics; detecting a performance-related event based on the dynamic performance statistics; and selecting, from the plurality of paths, an active path for a storage device of the plurality of storage devices based on the dynamic performance statistics and a result of the detecting. With the methods or apparatus according to the embodiments of the present disclosure, performance bottleneck can be avoided or timely eliminated, load balance can be achieved, and system resources can be utilized more effectively.

Claims

1. A method for a storage system, the storage system including a plurality of storage devices, each of the plurality of storage devices being connected to a plurality of different host bus adapter (HBA) ports of a server via a plurality of paths, the method comprising:

monitoring a system-wide performance metric of the storage system to obtain dynamic performance statistics, wherein the system wide performance metric includes at least one of delay and resource utilization of the plurality of different HBA ports;

detecting a performance related event based on the dynamic performance statistics; and

selecting, from the plurality of paths, an active path for a storage device of the plurality of storage devices based on the dynamic performance statistics and a result of the detecting, wherein selecting an active path for the storage device of the plurality of storage devices includes;

dividing, based on the dynamic performance statistics and the result of the detecting, the plurality of storage devices into a plurality of subsets respectively corresponding to the plurality of different HBA ports; and

routing storage devices in each of the plurality of subsets to their corresponding HBA ports.

2. The method of claim 1, wherein monitoring a system-wide performance metric of the storage system comprises:

monitoring a performance metric of the plurality of different HBA ports.

3. The method of claim 2, wherein the performance metric of the plurality of different HBA ports includes throughput.

4. The method of claim 2, wherein monitoring a system-wide performance metric of the storage system further comprises monitoring at least one of:

local performance metrics of the plurality of storage devices, and

local performance metrics of the plurality of paths.

5. The method of claim 1, wherein the performance related event includes at least one of:

a performance bottleneck at a HBA port of the plurality of different HBA ports,

imbalance of performance of the plurality of HBA ports, and

performance of a storage device of the plurality of storage devices being lower than a threshold.

6. The method of claim 1, wherein selecting, from the plurality of paths, an active path for a storage device of the plurality of storage devices based on the dynamic performance statistics and a result of the detecting comprises:

in response to detecting a performance bottleneck or a heavy workload at a first HBA port of the plurality of different HBA ports, performing dynamic path selection for a storage device of the plurality of storage devices so as to re-route at least part of a workload of the first HBA port to a further HBA port of the plurality of HBA ports.

7. The method of claim 6, wherein performing dynamic path selection for a storage device of the plurality of storage devices so as to re-route at least part of a workload of the first HBA port to a further HBA port of the plurality of HBA ports comprises:

computing an expected workload required to be offloaded from the first HBA port;

selecting a group of storage devices from the plurality of storage devices based on the computing; and

switching an active path for the group of storage devices to a path of a further HBA port other than the first HBA port.

8. The method of claim 7, wherein computing an expected workload required to be offloaded from the first HBA port comprises:

computing the expected workload required to be offloaded at least based on a performance metric of the first HBA port and a predetermined threshold.

9. The method of claim 7, wherein selecting a group of storage devices from the plurality of storage devices based on the computing comprises:

selecting a group of storage devices from the plurality of storage devices, such that a sum of current performance metrics of the group of storage devices matches the computed expected workload.

10. The method of claim 7, wherein selecting a group of storage devices from the plurality of storage devices based on the computing comprises:

selecting a plurality of groups of storage devices from the plurality of storage devices, such that a sum of current performance metrics of each of the plurality of groups of storage devices matches the computed expected workload; and

selecting a group of storage devices from the plurality of groups of storage devices based on a predetermined criterion.

11. The method of claim 10, wherein selecting a group of storage devices from the plurality of groups of storage devices based on a predetermined criterion comprises:

selecting, from the plurality of groups of storage devices, a group of storage devices with a maximum sum of priority weights.

12. An apparatus for a storage system, the storage system including a plurality of storage devices, each of the plurality of storage devices being connected to a plurality of different host bus adapter (HBA) ports of a server via a plurality of paths, the apparatus comprising:

a performance monitor configured to monitor a system-wide performance metric of the storage system to obtain dynamic performance statistics, wherein the system wide performance metric includes at least one of delay and resource utilization of the plurality of different HBA ports;

an event detector configured to detect a performance-related event based on the dynamic performance statistics; and

a path selector configured to select, from the plurality of paths, an active path for a storage device of the plurality of storage devices based on the dynamic performance statistics and a result of the detecting, wherein the path selector is configured to select, from the plurality of paths, an active path for a storage device of the plurality of storage devices by;

dividing, based on the dynamic performance statistics and the result of the detecting, the plurality of storage devices into a plurality of subsets respectively corresponding to the plurality of different HBA ports; and

routing storage devices of each of the plurality of subsets to their corresponding HBA port.

13. The apparatus of claim 12, wherein the performance monitor comprises:

an HBA port performance monitor configured to monitor a system-wide performance metric of the storage system by monitoring a performance metric of the plurality of different HBA ports.

14. The apparatus of claim 13, wherein the performance metric of the plurality of different HBA ports include throughput.

15. The apparatus of claim 13, wherein the performance monitor further includes a local performance monitor configured to monitor at least one of:

local performance metrics of the plurality of storage devices, and

local performance metrics of the plurality of paths.

16. The apparatus of claim 12, wherein the performance related event includes at least one of:

a performance bottleneck at a HBA port of the plurality of different HBA ports,

imbalance of performance of the plurality of HBA ports, and

performance of a storage device of the plurality of storage devices being lower than a threshold.

17. The apparatus of claim 12, wherein the path selector further comprises:

a load balancer configured to, in response to detecting a performance bottleneck or a heavy workload at a first HBA port of the plurality of different HBA ports, re-route at least part of a workload of the first HBA port to a further HBA port of the plurality of HBA ports.

18. The apparatus of claim 17, wherein the load balancer comprises:

an offloaded workload computing device configured to, in response to detecting a performance bottleneck or a heavy workload at the first HBA port of the plurality of different HBA ports, compute an expected workload required to be offloaded from the first HBA port;

a storage device selector configured to select a group of storage devices from the plurality of storage devices based on the computing; and

a path switching device configured to switch an active path for the group of storage devices to a path of a further HBA port other than the first HBA port.