Method and apparatus for monitoring computer system objects to improve system reliability

• US 6,594,774 B1
• Filed: 09/07/1999
• Issued: 07/15/2003
• Est. Priority Date: 09/07/1999
• Status: Expired due to Term

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Moeller, Mark M., Chapman, Craig
• Primary Examiner(s): Beausoliel, Robert
• Assistant Examiner(s): Chu, Gabriel L.
• Application Number:
  US09/390,941
• Time in Patent Office: 1,407 Days
• Field of Search: 714/15, 714/48, 714/50, 714/21, 714/2, 714/38, 714/26
• US Class Current: 714/2
• CPC Class Codes: G06F 11/366 using diagnostics G06F11/0703 takes precedence

Abstract

Computer system reliability is improved using various techniques to monitor objects (e.g., processes, threads, DLLs, etc.) executing on the system. Such techniques include active techniques, in which information is continually communicated from the object to the monitor, and passive techniques, in which the object does not need to repeatedly provide information to the monitor. The monitor determines when an object in the computer system has failed, and initiates appropriate recovery action when such a failure is detected.

Claims

1. A system comprising:

a bus;

a processor coupled to the bus; and

a memory, coupled to the bus, including a plurality of instructions to be executed by the processor, the plurality of instructions including,

an object to generate a registration request, including a registration type identifying a way failure of the object is to be determined and a recovery type identifying a recovery action to initiate in the event failure of the object is detected, and

a monitor to receive the registration request, to monitor the object for failure in accordance with the registration type, and if failure of the object is detected then to initiate the recovery action in accordance with the recovery type.

2. A system as recited in claim 1, wherein the system comprises an automobile computer.

3. A system as recited in claim 1, further comprising a registration database, coupled to the monitor, to store registration requests from a plurality of objects.

4. A system as recited in claim 1, wherein the object is further to generate notification messages and repeatedly send the notification messages to the monitor within a specified amount of time after the previous notification message was sent.

5. A system as recited in claim 4, further comprising a notification log, coupled to the monitor, to store a record of the notification messages.

6. A system as recited in claim 1, wherein the monitor is to detect a failure of the object if a notification message is not received from the object within a specified amount of time after a previous notification message was received.

7. A system as recited in claim 1, wherein the monitor is to detect a failure if the object is no longer executing in the system.

8. A method in a system, the method comprising:

receiving a registration request from an object, the registration request including an indication of a way failure of the object is to be determined and a type of recovery to be attempted in the event of a detected failure of the object;

detecting, in accordance with the indication of the way failure of the object is to be determined, whether the object has failed; and

initiating recovery of the object in accordance with the indication of the type of recovery to be attempted in response to detecting that the object has failed.

9. A method as recited in claim 8, wherein the object comprises an application process.

10. A method as recited in claim 8, wherein the object comprises an application thread.

11. A method as recited in claim 8, wherein the detecting comprises detecting the object has failed if a notification is not received from the object within a specified amount of time.

12. A method as recited in claim 11, wherein the specified amount of time comprises an amount of time provided by the object plus an additional buffer amount of time.

13. A method as recited in claim 11, wherein:

the receiving comprises receiving the registration request from a first thread associated with the object;

the detecting comprises detecting the object has failed if a notification is not received from a second thread associated within the object within the specified amount of time; and

the second thread is different from the first thread.

14. A method as recited in claim 8, wherein the detecting comprises:

repeatedly checking whether the object is still running in the system; and

determining the object has failed if the object is no longer present running in the system.

15. A method as recited in claim 8, wherein the initiating recovery of the object comprises logging the detected failure.

16. A method as recited in claim 8, wherein the initiating recovery of the object comprises terminating the object and restarting the object.

17. A method as recited in claim 16, wherein the restarting the object comprises restarting a process corresponding to the object.

18. A method as recited in claim 16, wherein the registration request further includes a command line, and wherein the restarting the object comprises executing the command line.

19. A method as recited in claim 8, wherein the initiating recovery of the object comprises rebooting the system.

20. A method as recited in claim 8, wherein the initiating recovery of the object comprises rebooting the system after a subsequent event occurs.

21. A method as recited in claim 20, wherein the subsequent event comprises an ignition coupled to the system being turned off.

22. A method as recited in claim 20, wherein the subsequent event comprises activation of a power off button.

23. A method as recited in claim 8, further comprising sending, prior to the initiating recovery of the object, a message to the object to verify that the object is to be recovered.

24. A method as recited in claim 23, further comprising aborting the recovery in response to an indication being received from the object that the object is not to be recovered.

25. At least one computer-readable memory containing a computer program that is executable by a processor to perform the method recited in claim 8.

26. An automobile computer programmed to perform the method as recited in claim 8.

27. A method in an object, the method comprising

generating a registration request, the registration request including an indication of a way failure of the object is to be determined and a type of recovery to be attempted in the event of a detected failure of the object;

and

transmitting the registration request to a monitor.

28. A method as recited in claim 27, wherein the object comprises an application process.

29. A method as recited in claim 27, wherein the object comprises an application thread.

30. A method as recited in claim 27, further comprising:

receiving an indication from the monitor that the object was detected as having failed; and

providing a response to the monitor indicating whether recovery of the object is refused.

31. A method as recited in claim 27, further comprising generating notification messages and repeatedly sending the notification messages to the monitor within a specified amount of time after the previous notification message was sent.

32. At least one computer-readable memory containing a computer program that is executable by a processor to perform the method recited in claim 27.

33. A system comprising:

a plurality of objects running in the system; and

a monitor to,

monitor the plurality of objects,

detect when at least one of the plurality of objects has failed, and

initiate an appropriate recovery action when an object fails, wherein a type of the recovery action is previously identified by the object.

34. A system as recited in claim 33, wherein the monitor is to passively monitor the plurality of objects.

36. A system as recited in claim 34, wherein the monitor is to passively monitor the plurality of objects by repeatedly comparing an initial heap allocation for the object and a current heap allocation for the object.

37. A system as recited in claim 33, wherein the monitor is to actively monitor the plurality of objects.

38. A system as recited in claim 37, wherein the monitor is to actively monitor the plurality of objects by repeatedly receiving feedback from at least one of the plurality of objects running in the system.

39. A system as recited in claim 33, wherein at least one of the plurality of objects comprises a process.

40. A system as recited in claim 33, wherein at least one of the plurality of objects comprises a thread.

41. At least one computer-readable memory containing a computer program that is executable by a processor to perform the method recited in claim 33.

42. A system as recited in claim 33, implemented in an automobile computer.

35. A system comprising:

a plurality of objects running in the system; and

a monitor to,

passively monitor the plurality of objects by repeatedly comparing a list of currently running objects to a list of objects that should be running,

detect when at least one of the plurality of objects has failed, and

initiate an appropriate recovery action when an object fails.

43. A method comprising:

determining that an object in a system has failed in response to not receiving a notification from the object within a specified amount of time;

checking whether a test object is being scheduled to execute by a system scheduler; and

initiating recovery of the object if the test object is being scheduled, otherwise reversing the determination that the object has failed if the test object is not being scheduled.

44. A method as recited in claim 43, wherein the test object comprises a test thread.

45. A method as recited in claim 43, wherein the checking comprises comparing an interval at which the test object should be executed to a difference between a current time and a last time that the test object was scheduled to execute.

46. At least one computer-readable memory containing a computer program that is executable by a processor to perform the method recited in claim 43.

47. An automobile computer programmed to perform the method as recited in claim 43.

48. A computer-readable storage medium comprising computer-executable instructions that implement interface methods, the interface methods performing a function comprising:

recording a way failure of an object is to be determined and a type of recovery to be attempted in the event of a detected failure of the object, wherein the way failure of the object is to be determined is identified by the object.

49. A computer-readable storage medium of claim 48, the interface methods further performing a function comprising removing, from a database, the way failure of the object is to be determined and the type of recovery to be attempted.

50. A computer-readable storage medium of claim 48, the interface methods further performing a function comprising updating a notification log with a new notification time for the object.

51. A computer-readable storage medium of claim 48, wherein the type of recovery to be attempted is identified by the object.

52. A computer-readable storage medium comprising computer-executable instructions that implement interface methods, the interface methods performing a function comprising:

recording a way failure of an object is to be determined and a type of recovery to be attempted in the event of a detected failure of the object; and

further performing a function comprising informing the object that a recovery process has been initiated for the object and providing the object with an opportunity to abort the recovery process.

53. A method comprising:

recording a way failure of an object is to be determined and a type of recovery to be attempted in the event of a detected failure of the object, wherein the type of recovery to be attempted is identified by the object.

54. A method as recited in claim 53, wherein the object comprises an application process.

55. A method as recited in claim 53, wherein the object comprises an application thread.

56. A method as recited in claim 53, wherein the way failure of the object is to be determined is identified by the object.

57. A system comprising:

a database; and

a monitor configured to record, in the database, a way failure of an object is to be determined and a type of recovery to be attempted in the event of a detected failure of the object, wherein the way failure of the object is to be determined and the type of recovery to be attempted are both identified by the object.

58. A system as recited in claim 57, wherein the object comprises an application process.

59. A system as recited in claim 57, wherein the object comprises an application thread.

60. A system comprising:

a processor;

a memory, coupled to the processor, including a plurality of instructions to be executed by the processor, the plurality of instructions including,

a monitor to determine that an object in a system has failed in response to not receiving a notification from the object within a specified amount of time, to check whether a test object is being scheduled to execute by a system scheduler, to initiate recovery of the object if the test object is being scheduled, and to reverse the determination that the object has failed if the test object is not being scheduled.

61. A system as recited in claim 60, wherein the test object comprises a test thread.

62. A system as recited in claim 60, wherein the instructions to check whether the test object is being scheduled to execute by the system scheduler comprises instructions to compare an interval at which the test object should be executed to a difference between a current time and a last time that the test object was scheduled to execute.

63. A system as recited in claim 62, wherein the system comprises an automobile computer.