Oplogging for online recovery in direct connection client server systems

• US 20060085426A1
• Filed: 10/14/2004
• Published: 04/20/2006
• Est. Priority Date: 10/14/2004
• Status: Active Grant

• Current Assignee: Oracle International Corporation (Oracle Corporation)
• Original Assignee: Oracle International Corporation (Oracle Corporation)
• Inventors: Verma, Sanjay, Hoang, Chi-Kim, Edson, Kirk Meredith, McAuliffe, Mark Lawrence
• Granted Patent: US 7,739,244 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 11/1438 Restarting or rejuvenating

Abstract

Clients are permitted to directly attach to a client server system. A oplog management system tracks changes as clients make the changes to control structures of the direct connection client server system. The Oplogging system stores changes in oplogs. Then, if a client fails, the Oplogging system can repair the control structures, either by undoing the changes made by the client, or else completing the changes in place of the client; depending on principles defined.

Claims

1. A computer-based apparatus, comprising:

a system supporting simultaneous access by a first client and a second client, said first client and said second client directly attached to the system;

a oplog generator to generate an oplog storing information about a change in a control structure of the system by said first client; and

a oplog cleaner process to repair said control structure of the system if said first client fails before completing said change.

2. A computer-based apparatus according to claim 1, wherein said information stored in the oplog includes information enabling the oplog cleaner process to undo said change in said control structure.

3. A computer-based apparatus according to claim 2, wherein the oplog includes:

a transaction identifier identified with said first client;

a table identifier identifying a table being modified by said first client; and

a cursor identifier associated with the transaction identifier and the table identifier.

4. A computer-based apparatus according to claim 2, wherein the oplog includes:

a first list identifier;

a second list identifier;

an element identifier identifying an element to be moved from the first list to the second list;

a first next element identifier identifying a next element in the first list; and

a second next element identifier identifying a next element in the second list.

5. A computer-based apparatus according to claim 1, wherein said information stored in the oplog includes information enabling the oplog cleaner process to complete said change in said control structure.

6. A computer-based apparatus according to claim 5, wherein the oplog includes:

a list identifier;

a length for a list identified by the list identifier; and

an operation to be performed iteratively on said list.

7. A computer-based apparatus according to claim 1, further comprising a connection monitor daemon to monitor a connection between the system and said first client.

8. A computer-based apparatus according to claim 7, wherein the connection monitor daemon is operative to launch the oplog cleaner process if the connection monitor daemon detects a failure of said first client.

9. A computer-based apparatus according to claim 1, wherein the system is a database system.

10. A computer-based apparatus according to claim 9, wherein the database is an in-memory database system.

11. A computer-based apparatus according to claim 9, further comprising a connection point for said first client to connect to the database.

12. A computer-based apparatus according to claim 11, further comprising a second connection point for a second client to connect to the database.

13. A method for using a system, comprising:

enabling a client to directly attach to the system;

logging changes to be made by the client to a control structure of the system;

monitoring for a failure of the client; and

if the client fails, repairing the control structure of the system while maintaining an availability of the system.

14. A method according to claim 13, wherein:

the method further comprises enabling a second client to directly attach to the system; and

repairing the control structure of the system includes if the client fails, repairing the control structure of the system without disconnecting the second client from the system.

15. A method according to claim 13, wherein logging changes to be made by the client to the control structure of the system includes:

maintaining a list of open transactions;

associating each open transaction in the list of open transactions with a cursor; and

associating each cursor with a table in the system, the table to be modified as part of the open transaction associated with the cursor.

16. A method according to claim 15, wherein repairing the control structure of the system includes:

removing a first association between the cursor and the table, if the first association exists; and

removing a second association between the open transaction and the cursor, if the second association exists.

17. A method according to claim 13, wherein logging changes being made by the client to the control structure of the system includes:

identifying a list in the control structure of the system being changed by the client;

identifying an operation to be performed by the client on an element in the list; and

storing information to undo the operation to be performed by the client on the element.

18. A method according to claim 17, wherein identifying an operation includes identifying the element to be added to the list.

19. A method according to claim 17, wherein identifying an operation includes identifying the element to be removed from the list.

20. A method according to claim 19, wherein repairing the control structure of the system includes:

identifying the element to be removed as a hole in the list; and

iteratively moving the hole until the list is continuous.

21. A method according to claim 17, wherein identifying an operation includes identifying the element to be moved within the list.

22. A method according to claim 17, wherein repairing the control structure of the system includes restoring the element in the list to an original state.

23. A method according to claim 13, wherein logging changes being made by the client to the control structure of the system includes:

identifying a list in the control structure of the system being changed by the client;

identifying an operation to be performed by the client on elements in the list; and

storing information about the operation.

24. A method according to claim 23, wherein repairing the control structure of the system includes:

identifying elements to which the client has not performed the operation; and

iteratively performing the operation on the elements.

25. A method according to claim 24, wherein:

identifying an operation includes identifying a hole move operation to be performed by the client on elements in the list; and

iteratively performing the operation on the elements includes iteratively moving a hole past the elements in the list until the list is continuous.

26. A method according to claim 13, wherein:

enabling a client to directly attach to the system includes enabling a client to directly attach to a database system;

logging changes includes logging changes to be made by the client to a control structure of the database system; and

repairing the control structure of the system includes repairing the control structure of the database system while maintaining an availability of the database system.

27. A method according to claim 26, wherein:

enabling a client to directly attach to a database system includes enabling a client to directly attach to an in-memory database system;

logging changes includes logging changes to be made by the client to a control structure of the in-memory database system; and

repairing the control structure of the system includes repairing the control structure of the in-memory database system while maintaining an availability of the in-memory database system.

28. An article comprising a machine-accessible medium having associated data, wherein the data, when accessed, results in a machine performing:

enabling a client to directly attach to a system;

logging changes to be made by the client to a control structure of the system;

monitoring for a failure of the client; and

if the client fails, repairing the control structure of the system while maintaining an availability of the system.

29. An article according to claim 28, wherein:

the machine-accessible data further including associated data that, when accessed, results in, enabling a second client to directly attach to the system; and

repairing the control structure of the system includes if the client fails, repairing the control structure of the system without disconnecting the second client from the system.

30. An article according to claim 28, wherein logging changes to be made by the client to the control structure of the system includes:

maintaining a list of open transactions;

associating each open transaction in the list of open transactions with a cursor; and

associating each cursor with a table in the system, the table to be modified as part of the open transaction associated with the cursor.

31. An article according to claim 30, wherein repairing the control structure of the system includes:

removing a first association between the cursor and the table, if the first association exists; and

removing a second association between the open transaction and the cursor, if the second association exists.

32. An article according to claim 28;

wherein logging changes being made by the client to the control structure of the system includes;

identifying a list in the control structure of the system being changed by the client;

identifying an operation to be performed by the client on an element in the list; and

storing information to undo the operation to be performed by the client on the element.

33. An article according to claim 32, wherein identifying an operation includes identifying the element to be added to the list.

34. An article according to claim 32, wherein identifying an operation includes identifying the element to be removed from the list.

35. An article according to claim 34, wherein repairing the control structure of the system includes:

identifying the element to be removed as a hole in the list; and

iteratively moving the hole until the list is continuous.

36. An article according to claim 32, wherein identifying an operation includes identifying the element to be moved within the list.

37. An article according to claim 32, wherein repairing the control structure of the system includes restoring the element in the list to an original state.

38. An article according to claim 28, wherein logging changes being made by the client to the control structure of the system includes:

identifying a list in the control structure of the system being changed by the client;

identifying an operation to be performed by the client on elements in the list; and

storing information about the operation.

39. An article according to claim 38, wherein repairing the control structure of the system includes:

identifying elements to which the client has not performed the operation; and

iteratively performing the operation on the elements.

40. An article according to claim 39, wherein:

identifying an operation includes identifying a hole move operation to be performed by the client on elements in the list; and

iteratively performing the operation on the elements includes iteratively moving a hole past the elements in the list until the list is continuous.

41. An article according to claim 28, wherein:

enabling a client to directly attach to the system includes enabling a client to directly attach to a database system;

logging changes includes logging changes to be made by the client to a control structure of the database system; and

repairing the control structure of the system includes repairing the control structure of the database system while maintaining an availability of the database system.

42. An article according to claim 28, wherein:

enabling a client to directly attach to a database system includes enabling a client to directly attach to an in-memory database system;

logging changes includes logging changes to be made by the client to a control structure of the in-memory database system; and

repairing the control structure of the system includes repairing the control structure of the in-memory database system while maintaining an availability of the in-memory database system.