System and method dynamic cluster membership in a distributed data system

• US 7,035,858 B2
• Filed: 04/29/2002
• Issued: 04/25/2006
• Est. Priority Date: 04/29/2002
• Status: Active Grant

• Current Assignee: Oracle America, Inc. (Oracle Corporation)
• Original Assignee: Sun Microsystems Incorporated (Oracle Corporation)
• Inventors: Kannan, Mahesh, Dinker, Darpan, Gopinath, Pramod
• Primary Examiner(s): Gaffin, Jeffrey
• Assistant Examiner(s): Wu, Yicun
• Application Number:
  US10/134,961
• Publication Number:
  US 20030204509A1
• Time in Patent Office: 1,457 Days
• Field of Search: 707 9- 10, 707/104.1, 707/1, 709/221, 709/242, 709/220, 711/148
• US Class Current: 1/1
• CPC Class Codes: G06F 16/27 Replication, distribution or synchronisation of data between databases or within a distributed database system; Distributed database system architectures therefor
• CPC Class Codes: H04L 41/0681 Configuration of triggering conditions
• CPC Class Codes: H04L 41/0893 Assignment of logical groups to network elements
• CPC Class Codes: H04L 41/12 Discovery or management of network topologies
• CPC Class Codes: H04L 43/0811 by checking connectivity
• CPC Class Codes: H04L 67/1097 for distributed storage of data in networks, e.g. transport arrangements for network file system [NFS], storage area networks [SAN] or network attached storage [NAS]
• CPC Class Codes: Y10S 707/99945 Object-oriented database structure processing

Abstract

A distributed system provides for separate management of dynamic cluster membership and distributed data. Nodes of the distributed system may include a state manager and a topology manager. A state manager handles data access from the cluster. A topology manager handles changes to the dynamic cluster topology. The topology manager enables operation of the state manager by handling topology changes, such as new nodes to join the cluster and node members to exit the cluster. A topology manager may follow a static topology description when handling cluster topology changes. Data replication and recovery functions may be implemented, for example to provide high availability.

Claims

1. A computer implemented node for a clustered distributed data system, comprising:

a state manager configured to replicate application data in the node to a first other node in the clustered distributed data system; and

a cluster topology manager configured to maintain information regarding the membership of nodes in the clustered distributed data system;

wherein the cluster topology manager is configured to detect changes in the node membership of the clustered distributed data system, wherein the cluster topology manager is configured to send an event notification to the state manager in response to a change in node membership in the clustered distributed data system; and

wherein the state manager is configured to replicate the application data to a second other node in the clustered distributed data system in response to the event notification.

2. The node as recited in claim 1, wherein the event notification indicates another node joined the clustered distributed data system.

3. The node as recited in claim 1, wherein the event notification indicates a node has left the clustered distributed data system.

4. The node as recited in claim 1, wherein the cluster topology manager is configured to send and receive messages as it transitions between finite states to maintain information regarding the membership of nodes in the clustered distributed data system.

5. The node as recited in claim 4, wherein the cluster topology manager is configured to send a message to verify the failure of a failed node, in the cluster to another node in the clustered distributed data system before sending an event notification to the state manager indicating a node failure in the clustered distributed data system.

6. The node as recited in claim 4, wherein the cluster topology manager is configured to attempt, according to a predefined order, to connect to a cluster node of the clustered distributed data system and send a message to the cluster node to request membership in the clustered distributed data system before sending the event notification indicating it has joined the clustered distributed data system.

7. The node as recited in claim 4, wherein the cluster topology manager is configured to send to its next node a message indicating a new node joined the clustered distributed data system in response to receiving a message requesting membership in the clustered distributed data system from the new node and wait for a message indicating the addition to the clustered distributed data system is complete before sending an event notification to the state manager indicating the new node has joined the clustered distributed data system.

8. The node as recited in claim 4, wherein the cluster topology manager is configured to send to its next node a message indicating a new node joined the clustered distributed data system in response to receiving a message indicating a new node joined the clustered distributed data system.

9. The node as recited in claim 1, wherein the cluster topology manager is configured to participate in a node membership management process according to a finite state automata to maintain a node topology matching a predetermined static topology for active nodes of the clustered distributed data system.

10. A computer implemented method for a node in a clustered distributed data system managing cluster membership, the method comprising:

a state manager replicating application data stored in the node to a first other node in the clustered distributed data system;

a cluster topology manager maintaining information regarding the membership of nodes in the clustered distributed data system;

the cluster topology manager detecting changes in the node membership of the clustered distributed data system;

the cluster topology manager sending an event notification to the state manager in response to a change in node membership in the clustered distributed data system; and

the state manager replicating the application data to a second other node in the clustered distributed data system in response to the event notification.

11. The method as recited in claim 10, wherein the event notification indicates another node joined the clustered distributed data system.

12. The method as recited in claim 10, wherein the event notification indicates a node left the clustered distributed data system.

13. The method as recited in claim 10, further comprising the cluster topology manager sending and receiving messages as it transitions between finite states to maintain information regarding the membership of nodes in the clustered distributed data system.

14. The method as recited in claim 13, the cluster topology manager sending a message to verify a failure of a failed node in the cluster to another node in the clustered distributed data system before sending the event notification indicating a node failure in the clustered distributed data system.

15. The method as recited in claim 13, further comprising:

the cluster topology manager connecting, according to a predefined order, to a cluster node of the clustered distributed data system; and

after said connecting and before sending an event notification indicating the node has joined the clustered distributed data system, the cluster topology manager sending a message to the cluster node to request membership in the clustered distributed data system.

16. The method as recited in claim 13, further comprising:

the cluster topology manager receiving a message from a new node requesting membership in the clustered distributed data system;

after said receiving the cluster topology manager sending to its next node a message indicating the new node joined the clustered distributed data system; and

after said sending the cluster topology manager waiting for a message indicating the addition to the clustered distributed data system is complete before sending an event notification indicating the new node has joined the clustered distributed data system.

17. The method as recited in claim 13, further comprising:

the cluster topology manager receiving a message indicating a new node joined the clustered distributed data system from its previous node; and

after said receiving, the cluster topology manager sending to its next node a message indicating the new node joined the clustered distributed data system and sending the event notification indicating a node joined the clustered distributed data system.

18. The method as recited in claim 10, wherein the cluster topology manager is configured to participate in a node membership management process according to a finite state automata to maintain a node topology matching a predetermined static topology for active nodes of the clustered distributed data system.

19. A computer system comprising a process and memory including instructions executable by the processor for:

a state manager replicating application data stored in the node to a first other node in the clustered distributed data system;

a cluster topology manager maintaining information regarding the membership of nodes in the clustered distributed data system;

the cluster topology manager detecting changes in the node membership of the clustered distributed data system;

the cluster topology manager sending an event notification to the state manager in response to a change in node membership in the clustered distributed data system; and

the state manager replicating the application data to a second other node in the clustered distributed data system in response to the event notification.

20. The computer system as recited in claim 19, wherein the event notification indicates another node joined the clustered distributed data system.

21. The computer system as recited in claim 19, wherein the event notification indicates a node has left the clustered distributed data system.

22. The computer system as recited in claim 19, further comprising the cluster topology manager sending and receiving messages as it transitions between finite states to maintain information regarding the membership of nodes in the clustered distributed data system.

23. The computer system as recited in claim 22, the cluster topology manager sending a message to verify a failure of a failed node in the cluster to another node in the clustered distributed data system before sending an event notification indicating a node failure in the clustered distributed data system.

24. The computer system as recited in claim 22, further comprising:

the cluster topology manager connecting, according to a predefined order, to a cluster node of the clustered distributed data system; and

after said connecting, the cluster topology manager sending a message to the cluster node to request membership in the clustered distributed data system before sending an event notification indicating the node has joined the clustered distributed data system.

25. The computer system as recited in claim 22, further comprising:

the cluster topology manager receiving a message from a new node requesting membership in the clustered distributed data system;

after said receiving, the cluster topology manager sending to its next node a message indicating the new node joined the clustered distributed data system; and

after said sending, the cluster topology manager waiting for a message indicating the addition to the clustered distributed data system is complete before sending the event notification indicating the new node has joined the clustered distributed data system.

26. The computer system as recited in claim 22, further comprising:

the cluster topology manager receiving a message indicating a new node joined the clustered distributed data system from its previous node; and

after said receiving, the cluster topology manager sending to its next node a message indicating the new node joined the clustered distributed data system and sending the event notification indicating the new node has joined the clustered distributed data system.

27. The computer system as recited in claim 19, wherein the cluster topology manager is configured to participate in a node membership management process according to a finite state automata to maintain a node topology matching a predetermined static topology for active nodes of the clustered distributed data system.