Recovery objects in an object oriented computing environment

• US 5,151,987 A
• Filed: 10/23/1990
• Issued: 09/29/1992
• Est. Priority Date: 10/23/1990
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Priven, Michael P., Abraham, Robert L., Moorman, Thomas P.
• Primary Examiner(s): Canney, Vincent P.
• Application Number:
  US07/602,443
• Time in Patent Office: 707 Days
• Field of Search: 371/12, 364/285.2, 395/575
• US Class Current: 714/20
• CPC Class Codes: G06F 11/1435 using file system or storage system metadata

Abstract

The present invention is directed to systems and methods for recovering from unplanned failures in object oriented computing environments. The systems and methods of the present invention recover from unplanned failures in an efficient manner by storing recovery information in recovery objects. During recovery operations from an unplanned failure, object instance methods (which were abnormally terminated by the unplanned failure) use the recovery information to identify committable actions which were executed prior to the unplanned failure. The object instance methods then execute committable actions which were not executed prior to the unplanned failure. Thus, according to the present invention, the committable actions which were executed prior to the unplanned failure are not re-executed, and the committable actions which were not executed prior to the planned failure are executed.

Claims

1. A method for recovering from computer failures in an object oriented computing environment, said method adapted to operate on one or more computer platforms and comprising the steps of:

(a) creating persistent recovery objects associated with object instances having object instance methods, each of said object instance methods comprising;

(1) multiple committable actions which cannot safely be re-executed; and

(2) multiple commit messages;

(b) storing recovery information sufficient to recover from the failures in said recovery objects;

(c) committing committable data, including said recovery information, to a non-volatile storage device.

2. In a computer-based system for performing work, a method adapted to operate in a local and in one or more remote object oriented computing environments, for recovering from computer failures, said method adapted to operate on one or more computer platforms and comprising the steps of:

(a) receiving a user command which specifies the work to be performed;

(b) creating a queued message request according to said user command, wherein said queue message request comprises a run method including;

(1) multiple committable actions for performing the work and which cannot safely be re-executed;

(2) multiple commit messages;

(c) creating a recovery object comprising recovery information for said queued message request;

(d) initializing said recovery information, such that said recovery information refers to a first of said committable actions;

(e) identifying object instances which are referenced by said run method;

(f) committing committable data including said recovery information to a non-volatile storage device;

(g) placing said queued message request on a queue;

(h) sending a run message to said run method; and

(i) in response to said run message, executing said run method according to said recovery information;

wherein said recovery information is sufficient to recover from the computer failures.

3. The method of claim 2, wherein step (i) comprises the steps of:

(1) executing said committable action which is referenced by said recovery information;

(2) updating said recovery information such that said recovery information refers to a next of said committable actions that has not been executed;

(3) sending one of said commit messages to

said recovery information to said non-volatile storage device;

(4) repeating steps (1) - (3) until said committable actions have all been executed.

4. The method of claim 3, wherein step (i) further comprises the steps of:

(5) restoring said queue;

(6) sending said run message to said run method; and

(7) performing steps (1) - (4);

wherein steps (5) - (7) are performed if said run method has not fully executed when one of said computer failures occurs.

5. The method of claim 2, wherein said queue exists in said local object oriented computing environment.

6. The method of claim 2, wherein said queue exists in one of said remote object oriented computing environments.

7. The method of claim 6, wherein step (g) comprises the step of sending said queued message request, said recovery object, and said referenced object instances to said remote object oriented computing environment.

8. The method of claim 2, wherein said recovery object is a persistent object instance.

9. The method of claim 2, wherein said committable data in step (f) also includes said queued message request and said referenced object instances.

10. The method of claim 2, wherein step (a) comprises the steps of:

(1) displaying a user option screen on a terminal;

(2) waiting for said user command;

(3) receiving said user command;

(4) using said user command to select an object instance and an action message, wherein said object instance contains an action method associated with said action message, and wherein said action method is adapted to create said queued message request; and

(5) sending said action message to said action method.

11. The method of claim 2, wherein step (h) comprises the step of using a selection policy to select said queued message request to execute.

12. In a computer-based system for performing work, a method adapted to operate in one or more object oriented computing environments, for recovering from computer failures, said method adapted to operate on one or more computer platforms and comprising the steps of:

(a) receiving a user command which specifies the work to be performed;

(b) using said user command to select an object instance, wherein said object instance comprises a run method including;

(1) multiple committable actions for performing the work and which cannot safely be re-executed; and

(2) multiple commit messages;

(c) sending a run message to said run method; and

(d) in response to said run message, executing said run method according to a recovery information contained in a recovery object associated with said object instance;

wherein said recovery information is sufficient to recover from the computer failures.

13. The method of claim 12, wherein said recovery information refers to a first of said committable actions, and wherein said recovery information is committed in a non-volatile storage device.

14. The method of claim 13, wherein step (d) comprises the steps of:

(1) executing said committable action which is referenced by said recovery information;

(2) updating said recovery information such that said recovery information refers to a next of said committable actions that has not been executed;

(3) sending one of said commit messages to commit said recovery information to said non-volatile storage device;

(4) repeating steps (1) - (3) until said committable actions have all been executed.

15. The method of claim 14, wherein step (d) further comprising the steps of:

(5) receiving said user command;

(6) using said user command to select said object instance;

(7) sending said run message to said run method; and

(8) performing steps (1) - (4);

wherein steps (5) - (8) are performed if said run method has not fully executed when one of said computer failures occurs.

16. A computer-based system adapted to operate on one or more computer platforms and in one or more object oriented computing environments, for recovering from computer failures, said system comprising:

(1) one or more object instances comprising object instance methods, each of said object instance methods comprising;

(a) multiple committable actions which cannot safely be re-executed;

(b) multiple commit messages;

(2) persistent recovery objects associated with said object instances;

(3) means for storing recovery information sufficient to recover from the failures in said recovery objects;

(4) one or more non-volatile storage devices;

(5) means for committing committable data including said recovery information to said nonvolatile storage devices

17. In a computer-based system for performing work, a recovery system adapted to operate on one or more computer platforms and in a local and in one or more remote object oriented computing environments, for recovering from computer failures, said recovery system comprising:

(1) means for receiving a user command which specifies the work to be performed;

(2) means for creating a queued message request according to said user command, wherein said queue message request comprises a run method including;

(a) multiple committable actions for performing the work and which cannot safely be re-executed;

(b) multiple commit messages;

(3) means for creating a recovery object comprising recovery information for said queued message request;

(4) means for initializing said recovery information, such that said recovery information refers to a first of said committable actions;

(5) means for identifying object instances which are referenced by said run method;

(6) one or more non-volatile storage devices;

(7) means for committing committable data including said recovery information to said non-volatile storage devices;

(8) means for placing said queued message request on a queue;

(9) means for sending a run message to said run method; and

(10) in response to said run message, means for executing said run method according to said recovery information;

wherein said recovery information is sufficient to recover from the computer failures.

18. The recovery system of claim 17, wherein said means for executing comprises:

(a) means for executing said committable action which is referenced by said recovery information;

(b) means for updating said recovery information such that said recovery information refers to a next of said committable actions that has not been executed; and

(c) means for sending one of said commit messages to commit said recovery information to said non-volatile storage devices.

19. The recovery system of claim 17, further comprising means for restoring said queue if one of the computer failures occurs.

20. The recovery system of claim 17, wherein said queue exists in said local object oriented computing environment.

21. The recovery system of claim 17, wherein said queue exists in one of said remote object oriented computing environments.

22. The recovery system of claim 21, wherein said means for placing comprises means for sending said queued message request, said recovery object, and said referenced object instances to said remote object oriented computing environment.

23. The recovery system of claim 17, wherein said recovery object is a persistent object instance.

24. The recovery system of claim 17, wherein said committable data also includes said queued message request and said referenced object instances.

25. The recovery system of claim 17, wherein said means for receiving a user command comprises:

(a) means for displaying a user option screen on a terminal;

(b) means for receiving said user command;

(c) means for using said user command to select an object instance and an action message, wherein said object instance contains a action method associated with said action message, and wherein said action method is adapted to create said queued message request; and

(5) sending said action message to said action method.

26. The recovery system of claim 17, wherein said means for sending a run message to said run method comprises means for using a selection policy to select said queued message request to execute.

27. In a computer-based system for performing work, a recovery system adapted to operate on one or more computer platforms and in one or more object oriented computing environments, for recovering from computer failures, said recovery system comprising:

(1) means for receiving a user command which specifies the work to be performed;

(2) means for using said user command to select an object instance, wherein said object instance comprises a run method including;

(a) multiple committable actions for performing the work and which cannot safely be re-executed; and

(b) multiple commit messages;

(3) one or more non-volatile storage devices;

(4) means for sending a run message to said run method; and

(5) in response to said run message, means for executing said run method according to a recovery information contained in a recovery object associated with said object instance;

wherein said recovery information is sufficient to recover from the computer failures.

28. The recovery system of claim 27, wherein said recovery information refers to a first of said committable actions, and wherein said recovery information is committed in said non-volatile storage devices.

29. The recovery system of claim 28, wherein said means for executing comprises:

(a) means for executing said committable action which is referenced by said recovery information;

(b) means for updating said recovery information such that said recovery information refers to a next of said committable actions that has not been executed; and

(c) means for sending one o.f said commit messages to commit said recovery information to said non-volatile storage device.