Fast application notification in a clustered computing system

• US 7,747,717 B2
• Filed: 08/12/2004
• Issued: 06/29/2010
• Est. Priority Date: 08/14/2003
• Status: Active Grant

• Current Assignee: Oracle International Corporation (Oracle Corporation)
• Original Assignee: Oracle International Corporation (Oracle Corporation)
• Inventors: Colrain, Carol
• Primary Examiner(s): Meky; Moustafa M
• Assistant Examiner(s): SHIU, HO T
• Application Number:
  US10/917,660
• Publication Number:
  US 20050038772A1
• Time in Patent Office: 2,147 Days
• Field of Search: None
• US Class Current: 709/223
• CPC Class Codes: G06F 2209/5016 Session
• CPC Class Codes: G06F 2209/544 Remote
• CPC Class Codes: G06F 9/5061 Partitioning or combining of resources
• CPC Class Codes: G06F 9/542 Event management; Broadcasting; Multicasting; Notifications

Abstract

With fast notification of changes to a clustered computing system, through which a number of events are published for system state changes, applications can quickly recover and sessions can quickly be rebalanced. When a resource associated with a service experiences a change in status, such as a termination or a start/restart, a notification event is immediately published. Notification events contain information to enable subscribers to identify, based on matching a session signature, the particular sessions that are affected by the change in status, and to respond accordingly. This allows sessions to be quickly aborted and ongoing processing to be quickly terminated when a resource fails, and allows fast rebalancing of work when a resource is restarted.

Claims

1. A method for communicating changes about a clustered computing environment that comprises a plurality of interconnected nodes that host clustered server instances, comprising the computer-implemented steps of:

an application client establishing one or more sessions with a server instance performing work within a cluster, said server instance being one of said clustered server instances;

wherein said application client resides outside of said cluster;

wherein said server instance is an allocation of resources of a particular node for performing said work, and said session requests performing said work;

in response to establishing a session with said server instance, said application client generating session identification information that identifies a resource and a session for which said resource performs work;

in response to a status change to said resource, generating first data that identifies said resource and second data that indicates a status of said resource;

publishing said first data and said second data to a set of one or more subscribers, including said application client, wherein said publishing transmits on an out-of-band connection that is different from a connection used by the one or more sessions; and

in response to receiving said first data and said second data over said out-of-band connection, said application client using said first data, said second data, and said session identification information to identify one or more sessions affected by said status change;

wherein the steps are performed by a plurality of computers.

2. The method of claim 1, wherein said status of said resource that is identified in said second data is one from a group consisting of (a) termination of said resource, (b) starting of said resource, and (c) unable to restart said resource.

3. The method of claim 1, wherein said cluster is a database cluster, and wherein said resource is identified in said first data by identifying said database cluster that is affected by said status change.

4. The method of claim 3, wherein said work is associated with a service, and wherein said resource is further identified in said first data by identifying said service that is affected by said status change.

5. The method of claim 4, wherein a location of said resource is further identified in said first data by identifying an instance and a node that are affected by said status change.

6. The method of claim 3, further comprising the computer-implemented step of:

identifying one or more sessions with said database cluster that are affected by said status change, based on matching said session identification information with said first data.

7. The method of claim 6, further comprising the computer-implemented step of:

interrupting said one or more sessions.

8. The method of claim 1, wherein said resource is identified in said first data by identifying a node that is affected by said status change.

9. The method of claim 1, wherein said work is associated with a service, and wherein said resource is identified in said first data by identifying that said service has terminated at a particular instance and by identifying said particular instance at which said service has terminated.

10. The method of claim 1, wherein said work is associated with a service, and wherein said resource is identified in said first data by identifying that an entire service has terminated and by identifying said service that has terminated.

11. The method of claim 1, wherein said resource is identified in said first data by identifying that a particular instance has terminated and by identifying said particular instance that has terminated.

12. The method of claim 1, wherein said resource is identified in said first data by identifying that all said instances have terminated and that identifies said cluster with which said instances are associated.

13. The method of claim 1, wherein said work is associated with a service, and wherein said resource is identified in said first data by identifying that said service has started at a particular instance and by identifying said particular instance on which said service has started.

14. The method of claim 13, wherein said resource is identified in said first data by identifying a number of instances that support said service that has started.

15. The method of claim 1, wherein said work is associated with a service, and wherein said resource is identified in said first data by identifying that said service has started on any instance and by identifying said service that has started.

16. The method of claim 15, wherein said resource is identified in said first data by identifying a number of instances that support said service that has started.

17. The method of claim 1, wherein said resource is identified in said first data by identifying that a particular instance has started and by identifying said particular instance that has started.

18. The method of claim 1, wherein said resource is identified in said first data by identifying that an instance has started and by identifying said cluster with which said instance is associated.

19. The method of claim 1, wherein said resource is identified in said first data by identifying that a node has terminated and by identifying said node that has terminated.

20. The method of claim 1, wherein the publishing comprises publishing said first data and said second data through processes that are not part of clusterware that manages said cluster.

21. The method of claim 1, wherein a subscriber of the set of one or more subscribers to said first data and said second data is a connection pool manager that responds to said status change by redistributing, based on said first data and said second data, connections to said cluster.

22. The method of claim 1, wherein a subscriber of the set of one or more subscribers to said first data and said second data is a client application that responds to said status change by requesting, based on said first data and said second data, redistribution within said cluster of said work that is affected by said status change.

23. The method of claim 1, wherein a subscriber of the set of one or more subscribers to said first data and said second data is a batch job that responds to said first data and said second data by calling for execution of a routine within said cluster based on said status change.

24. The method of claim 1, wherein said work is associated with a service, and wherein said resource is identified in said first data by identifying that said service is not restarting so that subscribing applications are interrupted from retrying to use said service.

25. A computer-readable volatile or non-volatile storage medium storing one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform:

an application client establishing one or more sessions with a server instance performing work within a cluster, said server instance being one of said clustered server instances;

wherein said application client resides outside of said cluster;

wherein said server instance is an allocation of resources of a particular node for performing said work, and said session requests performing said work;

in response to establishing a session with said server instance, said application client generating session identification information that identifies a resource and a session for which said resource performs work;

in response to a status change to said resource, generating first data that identifies said resource and second data that indicates a status of said resource;

a processor publishing said first data and said second data to a set of one or more subscribers, including said application client, wherein said publishing transmits on an out-of-band connection that is different from a connection used by the one or more sessions; and

in response to receiving said first data and said second data over said out-of-band connection, said application client using said first data, said second data, and said session identification information to identify one or more sessions affected by said status change.

26. The computer-readable volatile or non-volatile storage medium of claim 25, wherein

said status of said resource that is identified in said second data is one from a group consisting of (a) termination of said resource, (b) starting of said resource, and (c) unable to restart said resource.

27. The computer-readable volatile or non-volatile storage medium of claim 25,

wherein said cluster is a database cluster, and wherein said resource is identified in said first data by identifying said database cluster that is affected by said status change.

28. The computer-readable volatile or non-volatile storage medium of claim 27,

wherein said work is associated with a service, and wherein said resource is further identified in said first data by identifying said service that is affected by said status change.

29. The computer-readable volatile or non-volatile storage medium of claim 28,

wherein a location of said resource is further identified in said first data by identifying an instance and a node that are affected by said status change.

30. The computer-readable volatile or non-volatile storage medium of claim 27,

the instructions further comprising instructions that cause the one or more processor to perform:

identifying one or more sessions with said database cluster that are affected by said status change, based on matching said session identification information with said first data, wherein said session identification information is associated with said session and said first data identifies said resource.

31. The computer-readable volatile or non-volatile storage medium of claim 30,

further comprising instructions which cause the one or more processors to perform:

interrupting said one or more sessions.

32. The computer-readable volatile or non-volatile storage medium of claim 25,

wherein said resource is identified in said first data by identifying a node that is affected by said status change.

33. The computer-readable volatile or non-volatile storage medium of claim 25,

wherein said work is associated with a service, and wherein said resource is identified in said first data by identifying that said service has terminated at a particular instance and by identifying said particular instance at which said service has terminated.

34. The computer-readable volatile or non-volatile storage medium of claim 25,

wherein said work is associated with a service, and wherein said resource is identified in said first data by identifying that an entire service has terminated and by identifying said service that has terminated.

35. The computer-readable volatile or non-volatile storage medium of claim 25,

wherein said resource is identified in said first data by identifying that a particular instance has terminated and by identifying said particular instance that has terminated.

36. The computer-readable volatile or non-volatile storage medium of claim 25,

wherein said resource is identified in said first data by identifying that all said instances have terminated and that identifies said cluster with which said instances are associated.

37. The computer-readable volatile or non-volatile storage medium of claim 25,

wherein said work is associated with a service, and wherein said resource is identified in said first data by identifying that said service has started at a particular instance and by identifying said particular instance on which said service has started.

38. The computer-readable volatile or non-volatile storage medium of claim 37,

wherein said resource is identified in said first data by identifying a number of instances that support said service that has started.

39. The computer-readable volatile or non-volatile storage medium of claim 25,

wherein said work is associated with a service, and wherein said resource is identified in said first data by identifying that said service has started on any instance and by identifying said service that has started.

40. The computer-readable volatile or non-volatile storage medium of claim 39,

wherein said resource is identified in said first data by identifying a number of instances that support said service that has started.

41. The computer-readable volatile or non-volatile storage medium of claim 25,

wherein said resource is identified in said first data by identifying that a particular instance has started and by identifying said particular instance that has started.

42. The computer-readable volatile or non-volatile storage medium of claim 25,

wherein said resource is identified in said first data by identifying that an instance has started and by identifying said cluster with which said instance is associated.

43. The computer-readable volatile or non-volatile storage medium of claim 25,

wherein said resource is identified in said first data by identifying that a node has terminated and by identifying said node that has terminated.

44. The computer-readable volatile or non-volatile storage medium of claim 25,

wherein publishing comprises publishing said first data and said second data through processes that are not part of clusterware that manages said cluster.

45. The computer-readable volatile or non-volatile storage medium of claim 25,

wherein a subscriber of the set of one or more subscribers to said first data and said second data is a connection pool manager that responds to said status change by redistributing, based on said first data and said second data, connections to said cluster.

46. The computer-readable volatile or non-volatile storage medium of claim 25,

wherein a subscriber of the set of one or more subscribers to said first data and said second data is a client application that responds to said status change by requesting, based on said first data and said second data, redistribution within said cluster of said work that is affected by said status change.

47. The computer-readable volatile or non-volatile storage medium of claim 25,

wherein a subscriber of the set of one or more subscribers to said first data and said second data is a batch job that responds to said first data and said second data by calling for execution of a routine within said cluster based on said status change.

48. The computer-readable volatile or non-volatile storage medium of claim 25,

wherein said work is associated with a service, and wherein said resource is identified in said first data by identifying that said service is not restarting so that subscribing applications are interrupted from retrying to use said service.