Universal multi-path driver for storage systems including an external boot device with failover and failback capabilities

• US 7,406,617 B1
• Filed: 11/22/2004
• Issued: 07/29/2008
• Est. Priority Date: 11/22/2004
• Status: Active Grant

• Current Assignee: Unisys Corporation
• Original Assignee: Unisys Corporation
• Inventors: Athreya, Giridhar, Ortiz, Juan Carlos
• Primary Examiner(s): Lohn; Joshua A
• Application Number:
  US10/996,271
• Time in Patent Office: 1,345 Days
• Field of Search: 714/4, 714/5, 714/7, 714/8, 714/18, 714/43, 709/226, 709/239
• US Class Current: 714/4.11
• CPC Class Codes: G06F 11/1443 Transmit or communication errors
• CPC Class Codes: G06F 11/201 between storage system components

Abstract

An embodiment of the invention is a technique for monitoring via a universal multipath driver (UMD) a first path to one of a plurality of virtual device objects. The virtual device objects are created by the UMD and correspond to a plurality of functional device objects. The functional device objects are created by a plurality of lower level drivers and correspond to a plurality of real physical devices having M device types including an external boot device. The UMD is a functional driver of each of the functional device objects. The lower level drivers control the real physical devices. The UMD detects a failure of the first path to a first virtual device object corresponding to a first real physical device and performs one of failover and failback for the first real physical device.

Claims

1. A method comprising:

monitoring via a universal multipath driver (UMD) a first path to one of a plurality of virtual device objects, the virtual device objects being created by the UMD and corresponding to a plurality of functional device objects, the functional device objects being created by a plurality of lower level drivers and corresponding to a plurality of real physical devices having M device types including an external boot device, the UMD being a functional driver of each of the functional device objects, the lower level drivers controlling the real physical devices;

detecting via the UMD a failure of the first path to a first virtual device object corresponding to a first real physical device; and

performing via the UMD one of failover and failback for the first real physical device.

2. The method of claim 1 wherein the first real physical device is a disk device and wherein performing failover comprises:

determining if an I/O request has failed on the first path;

if the I/O request has failed,

determining if the first path has failed,

if the first path has failed,

determining if the first teal physical device is a cluster disk;

determining if there is an alternate path to the real physical device;

if there is an alternate path, resending the I/O request through the alternate path.

3. The method of claim 2 wherein the first real physical device is a disk device and wherein performing failover further comprises:

if the first path has failed and if there is no alternate path,

returning an error status.

4. The method of claim 1 wherein the first real physical device is a disk device and wherein performing failover further comprises:

if the TI/O request has not failed,

making a buffer queue available for a next T/O request; and

returning a success status.

5. The method of claim 1 wherein the first real physical device is a disk device and wherein performing failover further comprises:

if the I/O request has failed and if the first path has not failed,

resending the I/O request through the first path; and

returning a success status.

6. The method of claim 1 wherein the first real physical device is a disk device and wherein performing failback comprises:

determining if a bad path list is empty;

if the bad path list is not empty,

determining if the first path has been completely removed;

if the first path has not been completely removed,

determining if the first path is back on line;

if the first path is back on line,

removing the first path from the bad path list.

7. The method of claim 1 wherein the first real physical device is a tape device and wherein performing failover comprises:

determining if there is a reservation conflict for a current path;

if there is a reservation conflict for the current path

determining if the reservation conflict is due to a failure on a previous path;

if the reservation conflict is due to a failure on a previous path,

breaking a reservation on the current path; and

resending I/O requests through the current path.

8. The method of claim 7 further comprising:

if there is a reservation conflict for the current path and if the reservation conflict is not due to a failure on a previous path,

determining if there is an alternate path to the real physical device;

if there is no alternate path,

breaking a reservation on the current path; and

resending I/O requests through the current path;

if there is an alternate path,

disabling the current path; and

resending I/O requests through the alternate path.

9. The method of claim 7 further comprising:

if there is no reservation conflict for a current path,

determining if the current path has failed;

if the current path has failed,

determining if there is an alternate path to the real physical device;

if there is an alternate path to the real physical device,

adding the current path to a bad path list; and

resending I/O requests on the alternate path;

if there is no alternate path to the real physical device,

returning an error status.

10. The method of claim 1 wherein the first real physical device is a tape device and wherein performing failback comprises:

determining if a bad path list is empty;

if the bad path list is not empty,

determining if the first path has been completely removed;

if the first path has not been completely removed,

determining if the first path is back on line;

if the first path is back on line,

removing the first path from the bad path list.

11. The method of claim 10 wherein performing failback further comprises:

if the bad path list is not empty,

determining if the first path is reserved;

if the first path is reserved, adding the first path to an active path list.

12. The method of claim 1 wherein the M device types include at least two of a disk device, a tape device, a redundant army of inexpensive disks (RAID) subsystem, and a tape library.

13. An article of manufacture comprising:

a machine-accessible storage medium including data that, when accessed by a machine, cause the machine to perform operations comprising;

monitoring via a universal multipath driver (UMD) a first path to one of a plurality of virtual device objects, the virtual device objects being created by the UMD and corresponding to a plurality of functional device objects, the functional device objects being created by a plurality of lower level drivers and corresponding to a plurality of real physical devices having M device types including an external boot device, the UMD being a functional driver of each of the functional device objects, the lower level drivers controlling the real physical devices;

detecting via the UMD a failure of the first path to a first virtual device object corresponding to a first real physical device; and

performing via the UMD one of failover and failback for the first real physical device.

14. The article of manufacture of claim 13 wherein the first real physical device is a disk device and wherein the data causing the machine to perform the operation of performing failover comprise data that, when accessed by the machine, cause the machine to perform operations comprising:

determining if an I/O request has failed on the first path;

if the I/O request has failed,

determining if the first path has failed,

if the first path has failed,

determining if the first real physical device is a cluster disk;

determining if there is an alternate path to the real physical device;

if there is an alternate path, resending the T/O request through the alternate path.

15. The article of manufacture of claim 14 wherein the first real physical device is a disk device and wherein the data causing the machine to perform the operation of performing failover further comprise data that, when accessed by the machine, cause the machine to perform operations comprising:

if the first path has failed and if there is no alternate path,

returning an error status.

16. The article of manufacture of claim 13 wherein the first real physical device is a disk device and wherein the data causing the machine to perform the operation of performing failover further comprise data that, when accessed by the machine, cause the machine to perform operations comprising:

if the I/O request has not failed,

making a buffer queue available for a next I/O request; and

returning a success status.

17. The article of manufacture of claim 13 wherein the first real physical device is a disk device and wherein the data causing the machine to perform the operation of performing failover further comprise data that, when accessed by the machine, cause the machine to perform operations comprising:

if the I/O request has failed and if the first path has not failed,

resending the I/O request through the first path; and

returning a success status.

18. The article of manufacture of claim 13 wherein the first real physical device is a disk device and wherein the data causing the machine to perform the operation of performing failback comprise data that when accessed by the machine, cause the machine to perform operations comprising:

determining if a bad path list is empty;

if the bad path list is not empty,

determining if the first path has been completely removed;

if the first path has not been completely removed,

determining if the first path is back on line;

if the first path is back on line,

removing the first path from the bad path list.

19. The article of manufacture of claim 13 wherein the first real physical device is a tape device and wherein the data causing the machine to perform the operation of performing failover comprise data that, when accessed by the machine, cause the machine to perform operations comprising:

determining if there is a reservation conflict for a current path;

if there is a reservation conflict for the current path,

determining if the reservation conflict is due to a failure on a previous path;

if the reservation conflict is due to a failure on a previous path,

breaking a reservation on the current path; and

resending I/O requests through the current path.

20. The article of manufacture of claim 19 wherein the data causing the machine to perform the operation of performing failover comprise data that, when accessed by the machine, cause the machine to perform operations comprising:

if there is a reservation conflict for the current path and if the reservation conflict is not due to a failure on a previous path,

determining if there is an alternate path to the real physical device;

if there is no alternate path,

breaking a reservation on the current path; and

resending I/O requests through the current path;

if there is an alternate path,

disabling the current path; and

resending I/O requests through the alternate path.

21. The article of manufacture of claim 19 wherein the data causing the machine to perform the operation of performing failover comprise data that, when accessed by the machine, cause the machine to perform operations comprising:

if there is no reservation conflict for a current path,

determining if the current path has failed;

if the current path has failed;

determining if there is an alternate path to the real physical device;

if there is an alternate path to the real physical device,

adding the current path to a bad path list; and

resending I/O requests on the alternate path;

if there is no alternate path to the real physical device,

returning an error status.

22. The article of manufacture of claim 13 wherein the first real physical device is a tape device and wherein the data causing the machine to perform the operation of performing failback comprise data that, when accessed by the machine, cause the machine to perform operations comprising:

determining if a bad path list is empty;

if the bad path list is not empty,

determining if the first path has been completely removed;

if the first path has not been completely removed,

determining if the first path is back on line;

if the first path is back on line,

removing the first path from the bad path list.

23. The article of manufacture of claim 22 wherein the data causing the machine to perform the operation of performing failback further comprise data that, when accessed by the machine, cause the machine to perform operations comprising:

if the bad path list is not empty,

determining if the first path is reserved; and

if the first path is reserved, adding the first path to an active path list.

24. The article of manufacture of claim 13 wherein the M device types include at least two of a disk device, a tape device, a redundant array of inexpensive disks (RAID) subsystem, and a tape library.

25. A system comprising:

a processor;

a plurality of devices having M device types coupled to the processor via a plurality of adapters; and

a memory coupled to the processor, the memory containing program code that, when executed by the processor, causes the processor to;

monitor via a universal multipath driver (UMD) a first path to one of a plurality of virtual device objects, the virtual device objects being created by the UMD and corresponding to a plurality of functional device objects, the functional device objects being created by a plurality of lower level drivers and corresponding to a plurality of real physical devices having M device types including an external boot device, the UMD being a functional driver of each of the functional device objects, the lower level drivers controlling the real physical devices;

detect via the UMD a failure of the first path to a first virtual device object corresponding to a first real physical device; and

perform via the UMD one of failover and failback for the first real physical device.

26. The system of claim 25 wherein the first real physical device is a disk device and wherein the program code causing the processor to perform failover comprises program code that, when executed by the processor, causes the processor to:

determine if an I/O request has failed on the first path;

if the I/O request has failed,

determine if the first path has failed,

if the first path has failed,

determine if the first real physical device is a cluster disk;

determine if there is an alternate path to the real physical device;

if there is an alternate path, resend the I/O request through the alternate path.

27. The system of claim 26 wherein the first real physical device is a disk device and wherein the program code causing the processor to perform failover further comprises program code that, when executed by the processor, causes the processor to:

if the first path has failed and if there is no alternate path,

return an error status.

28. The system of claim 25 wherein the first real physical device is a disk device and wherein the program code causing the processor to perform failover further comprises program code that, when executed by the processor, causes the processor to:

if the I/O request has not failed,

make a buffer queue available for a next I/O request; and

return a success status.

29. The system of claim 25 wherein the first real physical device is a disk device and wherein the program code causing the processor to perform failover further comprises program code that, when executed by the processor, causes the processor to:

if the I/O request has failed and if the first path has not failed,

resend the I/O request trough the first path; and

return a success status.

30. The system of claim 25 wherein the first real physical device is a disk device and wherein the program code causing the processor to perform failback comprises program code that, when executed by a processor, causes the processor to:

determine if a bad path list is empty;

if the bad path list is not empty,

determine if the first path has been completely removed;

if the first path has not been completely removed,

determine if the first path is back on line;

if the first path is back on line,

remove the first path from the bad path list.

31. The system of claim 25 wherein the first real physical device is a tape device and wherein the program code causing the processor to perform failover comprises program code that, when executed by the processor, causes the processor to:

determine if there is a reservation conflict for a current path;

if there is a reservation conflict for the current path,

determine if the reservation conflict is due to a failure on a previous path;

if the reservation conflict is due to a failure on a previous path,

break a reservation on the current path; and

resend I/O requests through the current path.

32. The system of claim 31 wherein the program code causing the processor to perform failover comprises program code that, when executed by the processor, causes the processor to:

if there is a reservation conflict for the current path and if the reservation conflict is not due to a failure on a previous path,

determine if there is an alternate path to the real physical device;

if there is no alternate path,

break a reservation on the current path; and

resend I/O requests through the current path;

if there is an alternate path,

disable the current path; and

resend I/O requests through the alternate path.

33. The system of claim 31 wherein the program code causing the processor to perform failover comprises program code that, when executed by a processor, causes the processor to:

if there is no reservation conflict for a current path,

determine if the current path has failed;

if the current path has failed,

determine if there is an alternate path to the real physical device;

if there is an alternate path to the real physical device,

add the current path to a bad path list; and

resend I/O requests on the alternate path;

if there is no alternate path to the real physical device,

return an error status.

34. The system of claim 25 wherein the first real physical device is a tape device and wherein the program code causing the processor to perform failback comprises program code that, when executed by a processor, causes the processor to:

determine if a bad path list is empty;

if the bad path list is not empty,

determine if the first path has been completely removed;

if the first path has not been completely removed,

determine if the first path is back on line;

if the first path is back on line,

remove the first path from the bad path list.

35. The system of claim 34 wherein the program code causing the processor to perform failback further comprises program code that, when executed by a processor, causes the processor to:

if the bad path list is not empty,

determine if the first path is reserved; and

if the first path is reserved, add the first path to an active path list.

36. The system of claim 25 wherein the M device types include at least two of a disk device, a tape device, a redundant array of inexpensive disks (RAID) subsystem, and a tape library.