Multi-path management

• US 9,535,631 B2
• Filed: 08/26/2014
• Issued: 01/03/2017
• Est. Priority Date: 08/26/2013
• Status: Active Grant

• Current Assignee: GlobalFoundries, Inc.
• Original Assignee: GlobalFoundries, Inc.
• Inventors: Fu, Xiaofei, Hu, Jiale, Shi, Peng, Yu, Zhihua, Zhang, Yaodong
• Primary Examiner(s): Farrokh, Hashem
• Application Number:
  US14/468,544
• Publication Number:
  US 20150058558A1
• Time in Patent Office: 861 Days
• US Class Current: 1/1
• CPC Class Codes: G06F 3/061 Improving I/O performance
• CPC Class Codes: G06F 3/0635 by changing the path, e.g. traffic rerouting, path reconfiguration
• 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

The invention discloses a multi-path management method and system for managing paths over which an IO is routed from a host to multiple logical volumes (LUNs) of a storage system, wherein the multiple LUNs of the storage system are mapped to the host, and the method comprises: grouping LUNs with the same paths between the host and the storage system into a group of LUNs; and using one path selection thread for all of LUNs in the group of LUNs to manage the same paths. The method and system lower the amount of work for path management and the amount of required resources.

Claims

1. A multi-path management method of managing paths over which an IO is routed from a host to multiple logical volumes (LUNs) of a storage system, wherein the multiple LUNs of the storage system are mapped to the host, the method comprising:

grouping LUNs with the same paths between the host and the storage system into a group of LUNs;

using one path selection thread for all LUNs in the group of LUNs to manage the same paths and;

identifying LUNs with the same paths between the host and the storage system by scanning at least some of the LUNs mapped to the host side, the identifying comprising;

sending a command to fiber channel switches through respective bus adapter HBA ports on the host respectively, to inquiry for port IDs of respective Fiber Channel (FC) adapter FC ports of the storage system accessible to the HBA ports;

receiving the port IDs of the respective FC ports accessible to the respective HBA ports returned from the fiber channel switches respectively;

sending a command to the respective accessible FC ports through the respective HBA ports to identify LUNs on different paths, if the respective HBA ports have established connections with the respective FC ports corresponding to the returned port IDs, wherein a path of an LUN is from a specific HBA port ID to a specific FC port ID; and

identifying the LUNs with the same paths between the host and the storage system according to the identified at least some of the LUNs mapped to the host side and LUNs on the different paths.

2. The method according to claim 1, wherein using one path selection thread for all LUNs in the group of LUNs to manage the same paths comprises:

maintaining for each path a path state of the path, wherein a path state is one of;

available, indicating that the path does not fail and can have an IO routed;

fail, indicating that the path fails and cannot have any IO routed; and

partly fail, indicating that the path can have an IO routed for some of the LUNs.

3. The method according to claim 2, wherein using one path selection thread for all LUNs in a group of LUNs to manage the same paths comprises:

receiving a request for routing an IO to an LUN;

selecting a path with a path status of available or partly fail among the same paths by a path selection thread used for a group of LUNs which the LUN belongs to; and

routing the IO to the LUN over the selected path.

4. The method according to claim 3, further comprising:

determining whether the IO is routed successfully;

if the IO is routed unsuccessfully;

determining whether the path is a last path with a path state of available for the group of LUNs which the LUN belongs to;

if the path is the last path with a path state of available for the group of LUNs which the LUN belongs to;

setting the state of the path on the group of LUNs to Partly Fail; and

setting a flag on the LUN to indicate that the LUN cannot have the IO routed over the path.

5. The method according to claim 4, wherein if the IO is routed unsuccessfully further comprises:

if the path is not the last path with a path state of available for the group of LUNs which the LUN belongs to;

setting the path state of the path on the group of LUNs to Fail, so that all the LUNs in the group of LUN will not have any IO routed over the path; and

reselecting another available path among the same paths by the path selection thread used for the group of LUNs which the LUN belongs to.

6. The method according to claim 5, wherein setting the path state of the path on the group of LUNs to fail further comprises:

determining whether other LUNs in the group of LUNs is accessible over the failed path;

if none of the other LUNs in the group of LUNs is accessible over the failed path;

keeping the state of the path on the group of LUN as Fail;

if some of the other LUNs in the group of LUNs are accessible over the failed path;

setting the path state of the path on the group of LUNs to Partly Fail; and

setting a flag on all the LUNs inaccessible over the path in the group of LUNs to indicate that these LUNs cannot have any IO routed over the path.

7. The method according to claim 2, wherein the method further comprises:

determining for a path in a path state of path fails or path partly fails whether all the LUNs in the group of LUNs are accessible by reusing the path;

if all the LUNs in the group of LUNs are accessible over the path;

setting the path state of the path on the groups of LUNs to available; and

if the original path state is Partly Fail, removing the flag on the LUNs which indicates the path is unavailable;

if some of the LUNs are accessible and some of the LUNs are inaccessible;

setting the path state of the path on the group of LUNs to Partly Fail; and

setting a flag on those inaccessible LUNs in the group of LUNs to indicate that

these LUNs cannot have any IO routed over the path; and

if none of the LUNs in the group of LUNs is accessible over the path;

continuing with a check.

8. A multi-path management system to manage paths over which an IO is routed from a host to multiple logical volumes (LUNs) of a storage system, wherein the multiple LUNs of the storage system are mapped to the host, comprising:

a grouping device, configured to group LUNs with the same paths between the host and the storage system into a group of LUNs;

a path selecting device, configured to use one path selection thread for all LUNs in a group of LUNs to manage the same paths; and

an identifying device, configured to identify the LUNs with the same paths between the host and the storage system, wherein the identifying device further comprises;

an inquiring device, configured to send a command to fiber channel switches through respective bus adapter HBA ports on the host respectively, to inquire for port IDs of respective Fiber Channel (FC) adapter FC ports of the storage system accessible to the HBA ports;

a receiving device, configured to receive the port IDs of the respective FC ports accessible to the respective HBA ports returned from the fiber channel switches respectively;

an LUN on different path identifying device, configured to send a command to the respective accessible FC ports through the respective HBA ports to identify LUNs on different paths, if the respective HBA ports have established connections with the respective FC ports corresponding to the returned port IDs, wherein a path of an LUN is from a specific HBA port ID to a specific FC port ID; and

an LUN on the same path identifying device, configured to identify the LUNs with the same paths between the host and the storage system according to the identified at least some of the LUNs mapped to the host side and LUNs on the different paths,

wherein the identifying device comprises a scanning device, configured to scan at least some of the LUNs mapped to the host side.

9. The system according to claim 8, wherein the system further comprises a path recovering device configured to:

determine for a path in a path state of path fails or path partly fails whether all the LUNs in the group of LUNs are accessible by reusing the path;

if all the LUNs in the group of LUNs are accessible over the path;

set the path state of the path on the groups of LUNs to available; and

if the original path state is Partly Fail, remove the flag on the LUNs which indicates the path is unavailable;

if some of the LUNs are accessible and some of the LUNs are inaccessible;

set the path state of the path on the group of LUNs to Partly Fail, and set a flag on those inaccessible LUNs in the group of LUNs to indicate that these LUNs cannot have any IO routed over the path; and

if none of the LUNs in the group of LUNs is accessible over the path, continue with a check.

10. The system according to claim 8, wherein the path selecting device comprises a state maintaining device configured to maintain for each path a path state of the path, wherein a path state is one of:

available, indicating that the path does not fail and can have an IO routed;

fail,

indicating that the path fails and cannot have any IO routed; and

partly fail, indicating that the path can have an IO routed for some of the LUNs.

11. The system according to claim 10, wherein the path selecting device is further configured to:

receive a request for routing an IO to an LUN;

select a path with a path status of Available or Partly Fail among the same paths by a path selection thread used for a group of LUNs which the LUN belongs to; and

route the IO to the LUN over the selected path.

12. The system according to claim 11, wherein the path selecting device is further configured to:

determine whether the IO is routed successfully;

if the IO is routed unsuccessfully;

determine whether the path is a last path with a path state of Available for the group of LUNs which the LUN belongs to;

if the path is the last path with a path state of Available for the group of LUNs which the LUN belongs to;

set the state of the path on the group of LUNs to Partly Fail, and

set a flag on the LUN to indicate that the LUN cannot have the IO routed over the path.

13. The system according to claim 12, wherein if the IO is routed unsuccessfully further comprises:

if the path is not the last path with a path state of available for the group of LUNs which the LUN belongs to, the path selecting device is further configured to;

set the path state of the path on the group of LUNs to Fail, so that all the LUNs in the group of LUN will not have any IO routed over the path; and

reselect another available path among the same paths by the path selection thread used for the group of LUNs which the LUN belongs to.

14. The system according to claim 13, wherein the route selecting device configured to set the path state of the path on the group of LUNs to Fail is further configured to:

determine whether other LUNs in the group of LUNs is accessible over the failed path;

if none of the other LUNs in the group of LUNs is accessible over the failed path;

keep the state of the path on the group of LUN as fail;

if some of the other LUNs in the group of LUNs are accessible over the failed path;

set the path state of the path on the group of LUNs to Partly Fail; and

set a flag on all the LUNs inaccessible over the path in the group of LUNs to indicate that these LUNs cannot have any IO routed over the path.

15. A multi-path management method of managing paths over which an IO is routed from a host to multiple logical volumes (LUNs) of a storage system, wherein the multiple LUNs of the storage system are mapped to the host, the method comprising:

grouping LUNs with the same paths between the host and the storage system into a group of LUNs;

using one path selection thread for all LUNs in the group of LUNs to manage the same paths comprising;

maintaining for each path a path state of the path, wherein the path state is one of;

available, indicating that the path does not fail and can have an IO routed;

fail, indicating that the path fails and cannot have any IO routed; and

partly fail, indicating that the path can have an IO routed for some of the LUNs;

receiving a request for routing an IO to an LUN;

selecting a path with a path status of available or partly fail among the same paths by a path selection thread used for a group of LUNs which the LUN belongs to; and

routing the IO to the LUN over the selected path; and

determining whether the IO is routed successfully;

if the IO is routed unsuccessfully;

determining whether the path is a last path with a path state of available for the group of LUNs which the LUN belongs to;

if the path is the last path with a path state of available for the group of LUNs which the LUN belongs to;

setting the state of the path on the group of LUNs to Partly Fail; and

setting a flag on the LUN to indicate that the LUN cannot have the IO routed over the path.