Fast reorganization of connections in response to an event in a clustered computing system

• US 7,953,860 B2
• Filed: 08/12/2004
• Issued: 05/31/2011
• Est. Priority Date: 08/14/2003
• Status: Active Grant

• Current Assignee: Oracle International Corporation (Oracle Corporation)
• Original Assignee: Oracle International Corporation (Oracle Corporation)
• Inventors: Irudayaraj, Rajkumar, Colrain, Carol
• Primary Examiner(s): Etienne; Ario
• Assistant Examiner(s): Williams; Clayton R
• Application Number:
  US10/917,663
• Publication Number:
  US 20050038801A1
• Time in Patent Office: 2,483 Days
• Field of Search: None
• US Class Current: 709/227
• CPC Class Codes: G06F 11/1443 Transmit or communication errors
• CPC Class Codes: G06F 16/21 Design, administration or maintenance of databases

Abstract

Techniques for fast recovery and/or balancing of connections to a clustered computing system provide management of such connections by determining a number of connections to load balance across nodes and by triggering creation of such connections. In one aspect, a notification of an event regarding the clustered computing system is received by a connection pool manager, a pool of connections to the system is identified based on the notification, and one or more connections from the pool are processed in response to the event. According to an embodiment, the notification comprises the identities of the service, database, server instance and machine that correspond to the event.

Claims

1. A computer-implemented method comprising:

receiving a notification of an event regarding a clustered computing system;

based on said notification, identifying a pool of connections to said clustered computing system;

wherein said notification is received out-of-band of any connection in said pool of connections;

wherein each connection in said pool of connections is a connection to a database session of one or more database sessions; and

in response to said event, processing one or more connections from said pool, wherein said processing includes;

marking said one or more connections as invalid; and

aborting said one or more connections;

after receiving said notification of said event, receiving a second notification that identifies a start of a particular service that is hosted on a particular server instance of said clustered computing system; and

determining a second number of said one or more connections for allocating across a plurality of active server instances that host said particular service;

wherein the method is performed by one or more computers.

2. The method of claim 1, further comprising:

recording the location of sessions associated with connections from said pool of connections; and

identifying said one or more connections for said processing based on said location of sessions matching information in said notification.

3. The method of claim 1, wherein identifying includes identifying a pool that is associated with one or more services that are identified in said notification;

and

wherein said processing includes invoking a process to create one or more connections in association with said pool of connections.

4. The method of claim 3, wherein invoking a process to create one or more connections includes invoking a process on a clustered database server, and wherein said process allocates said one or more connections across said clustered database server based on state changes that are associated with said particular service that is hosted by one or more instances associated with said clustered database server.

5. The method of claim 3, wherein said particular service is hosted by one or more instances associated with a clustered database server.

6. The method of claim 3, wherein said particular service is hosted by one or more instances associated with a clustered database server configured in a plurality of clusters that are processing instructions as a computing grid.

7. The method of claim 1,

wherein said notification identifies termination of a particular service that was hosted on a particular server instance of said clustered computing system;

wherein said identifying further includes identifying said pool that is associated with (a) said particular service, (b) a database being used in said clustered computing system, (c) said particular server instance, and (d) a node of said clustered computing system, that are identified in said notification;

wherein said processing of connections further includes

identifying one or more connections, from said pool, that is associated with said database and said particular server instance.

8. The method of claim 1, wherein determining said second number further comprises:

identifying a first number of said one or more connections that are removed from said pool; and

determining said second number based on said first number.

9. The method of claim 8, further comprising:

requesting creation of at least a portion of said second number of said one or more connections, for immediate allocation across said plurality of active server instances that host said particular service.

10. The method of claim 9, wherein determining said second number includes determining said second number by dividing said first number by a number of said plurality of active server instances that host said particular service.

11. The method of claim 10, wherein

if said second number is not greater than zero, then

closing a subset of connections from said pool,

recreating said subset of connections in association with said pool for allocating said subset of connections to said plurality of active server instances that host said particular service.

12. The method of claim 11, wherein allocating said subset of connections is based on server performance metrics that are associated with said plurality of active server instances that host said particular service.

13. The method of claim 1,

wherein said notification identifies termination of a node that is hosting one or more server instance that is hosting at least a portion of a particular service that is provided by said clustered computing system;

wherein said identifying includes identifying a pool that is associated with said particular service that is identified in said notification; and

wherein said processing of connections includes

identifying one or more connections from said pool that are associated with said node;

marking said one or more connections as invalid, and

aborting said one or more connections so that one or more clients are interrupted.

14. The method of claim 13, wherein said aborting includes interrupting TCP/IP-based communications associated with said one or more connections.

15. The method of claim 13, wherein said aborting allows said one or more clients to quickly recover from said event.

16. A non-transitory computer-readable 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:

receiving a notification of an event regarding a clustered computing system;

based on said notification, identifying a pool of connections to said clustered computing system;

wherein said notification is received out-of-band of any connection in said pool of connections;

wherein each connection in the pool of connections is a connection to a database session of one or more database sessions; and

in response to said event, processing one or more connections from said pool, wherein said processing includes;

marking said one or more connections as invalid; and

aborting said one or more connections;

after receiving said notification of said event, receiving a second notification that identifies a start of a particular service that is hosted on a particular server instance of said clustered computing system; and

determining a second number of said one or more connections for allocating across a plurality of active server instances that host said particular service.

17. The non-transitory computer-readable storage medium of claim 16, further comprising instructions which, when executed by the one or more processors, causes the one or more processors to perform:

recording the location of sessions associated with connections from said pool of connections; and

identifying said one or more connections for said processing based on said location of sessions matching information in said notification.

18. The non-transitory computer-readable storage medium of claim 16, wherein identifying includes identifying a pool that is associated with one or more services that are identified in said notification;

and

wherein said processing includes invoking a process to create one or more connections in association with said pool of connections.

19. The non-transitory computer-readable storage medium of claim 18, wherein invoking a process to create one or more connections includes invoking a process on a clustered database server, and wherein said process allocates said one or more connections across said clustered database server based on state changes that are associated with said particular service that is hosted by one or more instances associated with said clustered database server.

20. The non-transitory computer-readable storage medium of claim 18, wherein said particular service is hosted by one or more instances associated with a clustered database server.

21. The non-transitory computer-readable storage medium of claim 18, wherein said particular service is hosted by one or more instances associated with a clustered database server configured in a plurality of clusters that are processing instructions as a computing grid.

22. The non-transitory computer-readable storage medium of claim 16, wherein said notification identifies termination of a particular service that was hosted on a particular server instance of said clustered computing system;

wherein said identifying further includes identifying said pool that is associated with (a) said particular service, (b) a database being used in said clustered computing system, (c) said particular server instance, and (d) a node of said clustered computing system, that are identified in said notification;

wherein said processing of connections further includes identifying one or more connections, from said pool, that is associated with said database and said particular server instance.

23. The non-transitory computer-readable storage medium of claim 16, wherein instructions to perform determining said second number further comprises instructions that cause the one or more processors to perform:

identifying a first number of said one or more connections that are removed from said pool; and

determining said second number based on said first number.

24. The non-transitory computer-readable storage medium of claim 23, further comprising one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform:

requesting creation of at least a portion of said second number of said one or more connections, for immediate allocation across said plurality of active server instances that host said particular service.

25. The non-transitory computer-readable storage medium of claim 24, wherein determining said second number includes determining said second number by dividing said first number by a number of said plurality of active server instances that host said particular service.

26. The non-transitory computer-readable storage medium of claim 25, wherein if said second number is not greater than zero, then closing a subset of connections from said pool,

recreating said subset of connections in association with said pool for allocating said subset of connections to said plurality of active server instances that host said particular service.

27. The non-transitory computer-readable storage medium of claim 26, wherein allocating said subset of connections is based on server performance metrics that are associated with said plurality of active server instances that host said particular service.

28. The non-transitory computer-readable storage medium of claim 16, wherein said notification identifies termination of a node that is hosting one or more server instance that is hosting at least a portion of a particular service that is provided by said clustered computing system;

wherein said identifying includes identifying a pool that is associated with said particular service that is identified in said notification; and

wherein said processing of connections includes identifying one or more connections from said pool that are associated with said node;

marking said one or more connections as invalid, and

aborting said one or more connections so that one or more clients are interrupted.

29. The non-transitory computer-readable storage medium of claim 28, wherein said aborting includes interrupting TCP/IP-based communications associated with said one or more connections.