System and method for dynamic configuration of replicated database servers

• US 7,805,407 B1
• Filed: 06/16/2004
• Issued: 09/28/2010
• Est. Priority Date: 06/16/2004
• Status: Active Grant

• Current Assignee: Oracle America, Inc. (Oracle Corporation)
• Original Assignee: Oracle America, Inc. (Oracle Corporation)
• Inventors: Verbeke, Jerome M., Nadgir, Neelakanth M.
• Primary Examiner(s): Jalil; Neveen Abel
• Assistant Examiner(s): Syed; Farhan M
• Application Number:
  US10/869,664
• Time in Patent Office: 2,295 Days
• Field of Search: 707/101, 707/102, 707/103
• US Class Current: 707/634
• CPC Class Codes: G06F 11/1662 the resynchronized component or unit being a persistent storage device re-synchronization of failed mirror storage G06F11/2082; rebuild or reconstruction of parity RAID storage G06F11/1008
• CPC Class Codes: G06F 11/2097 maintaining the standby controller/processing unit updated initialisation or re-synchronisation thereof G06F11/1658 and subgroups
• 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

Abstract

System and method for the dynamic configuration of replicated database servers. Embodiments may provide a mechanism to dynamically and automatically replace lost server nodes with a pool node to maintain a necessary or desired level of replication. Some embodiments may provide a mechanism or mechanisms to dynamically and automatically increase the number of server nodes to meet an increase in demand, and to dynamically and automatically decrease the number of server nodes during periods of decreased demand. To maintain consistency among the database replicas, some embodiments may provide a mechanism or mechanisms for synchronizing the database replicas on the server nodes. In one embodiment, the nodes may be peer nodes in a peer-to-peer network, and the pool nodes may be a peer group.

Claims

1. A replicated database system, comprising:

a plurality of server nodes each comprising a replica of a database, wherein the replicas of the database on the server nodes are concurrently accessible to service requests to the database; and

one or more pool nodes, wherein each of the one or more pool nodes is capable of becoming a server node in the replicated database system;

wherein the plurality of server nodes is configured to;

determine whether a new server node is needed in the plurality of server nodes; and

recruit one of the pool nodes as the new server node in the plurality of server nodes, wherein, to recruit one of the pool nodes as the new server node, at least one of the plurality of server nodes is configured to broadcast a message to the pool nodes requesting the pool nodes to volunteer to become the new server node.

2. The replicated database system as recited in claim 1, wherein, to determine whether a new server node is needed, the plurality of server nodes is further configured to detect that one of the server nodes has gone down, wherein the server node that has gone down is replaced with the new server node.

3. The replicated database system as recited in claim 2, wherein each of the plurality of server nodes is further configured to communicate with the other server nodes at periodic or aperiodic intervals, and wherein, to detect that one of the server nodes has gone down, the plurality of server nodes is further configured to detect that the one of the server nodes that has gone down has not communicated with the other server nodes for a specified number of intervals.

4. The replicated database system as recited in claim 1, wherein the pool node recruited as the new server node is configured to:

receive the broadcast message;

stop execution of a current process on the recruited pool node in response to the received broadcast message; and

start execution of a replicated database server process on the recruited pool node in response to the received broadcast message.

5. The replicated database system as recited in claim 1, wherein, to recruit one of the pool nodes as the new server node, the at least one of the plurality of server nodes is further configured to:

receive a response message to the broadcast message from the recruited pool node indicating that the recruited pool node is available to become a server node; and

configure the recruited pool node as the new server node in response to the response message from the recruited pool node.

6. The replicated database system as recited in claim 1, wherein, to recruit one of the pool nodes as the new server node, the at least one of the plurality of server nodes is further configured to replicate the database to the recruited pool node.

7. The replicated database system as recited in claim 1, wherein, to recruit one of the pool nodes as the new server node, at least one of the plurality of server nodes is configured to provide software and configuration information to the recruited pool node to enable the recruited pool node to operate as a server node.

8. The replicated database system as recited in claim 1, wherein, to determine that a new server node is needed, the plurality of server nodes is further configured to determine that demand on the replicated database has increased above a demand threshold.

9. The replicated database system as recited in claim 1, wherein the plurality of server nodes is further configured to:

determine that demand on the replicated database has decreased below a demand threshold;

select one of the server nodes to be released; and

release the selected server node, wherein the selected server node becomes a pool node.

10. The replicated database system as recited in claim 1, wherein each of the plurality of server nodes is further configured to:

periodically or aperiodically broadcast a current status of the server node'"'"'s replica of the database to the other server nodes;

receive from one or more of the other server nodes differences between the broadcast current status of the server node'"'"'s replica of the database and a current status of the replica of the database on the one or more other server nodes; and

update the server node'"'"'s replica of the database according to the received differences.

11. The replicated database system as recited in claim 1, wherein the pool nodes and server nodes are configured to participate as peer nodes in a peer group on a peer-to-peer network.

12. The replicated database system as recited in claim 10, wherein the peer-to-peer network is implemented according to JXTA platform protocols.

13. The replicated database system as recited in claim 1, wherein the pool nodes and server nodes are configured to participate as nodes of a grid computing framework.

14. The replicated database system as recited in claim 1, wherein the replicated database is used to track tasks distributed to computing nodes of a distributed computing system.

15. A server node, comprising:

a processor; and

a memory comprising program instructions and a replica of a database, wherein the program instructions are executable by the processor to implement a replicated database server process configured to communicate with one or more other server nodes to;

determine whether a new server node is needed, wherein, to determine that a new server node is needed, the replicated database server process is further configured to communicate with the one or more other server nodes to determine that demand on the replicated database has increased above a demand threshold; and

recruit a pool node from one or more pool nodes each capable of becoming a server node as the new server node;

wherein each server node comprises a replica of the database, and wherein the replicas of the database on the server nodes are concurrently accessible to service requests to the database.

16. The server node as recited in claim 15, wherein, to determine that a new server node is needed, the replicated database server process is further configured to communicate with the other server nodes to detect that one of the server nodes has gone down, wherein the server node that has gone down is replaced with the new server node.

17. The server node as recited in claim 16, wherein the replicated database server process is further configured to communicate with the other server nodes at periodic or aperiodic intervals, and wherein, to detect that one of the server nodes has gone down, the replicated database server process is further configured to detect that the server node that has gone down has not communicated with the other server nodes for a specified number of intervals.

18. The server node as recited in claim 15, wherein, to recruit a pool node as the new server node, the replicated database server process is further configured to broadcast a message to the one or more pool nodes requesting a volunteer to become the new server node.

19. The server node as recited in claim 18, wherein, to recruit a pool node as the new server node, the replicated database server process is further configured to:

receive a response message to the broadcast message from the recruited pool node indicating that the recruited pool node is available to become a server node; and

configure the recruited pool node as the new server node in response to the response message from the recruited pool node.

20. The server node as recited in claim 15, wherein, to recruit a pool node as the new server node, the replicated database server process is further configured to replicate the database to the recruited pool node.

21. The server node as recited in claim 15, wherein, to recruit a pool node as the new server node, the replicated database server process is further configured to provide software and configuration information to the recruited pool node to enable the recruited pool node to operate as a server node.

22. The server node as recited in claim 15, wherein the replicated database server process is further configured to communicate with the other server nodes to:

determine that demand on the replicated database has decreased below a demand threshold;

select one of the server nodes to be released; and

release the selected server node, wherein the selected server node becomes a pool node.

23. The server node as recited in claim 15, wherein the replicated database server process is further configured to:

periodically or aperiodically broadcast a current status of the server node'"'"'s replica of the database to the other server nodes;

receive from one or more of the other server nodes differences between the broadcast current status of the server node'"'"'s replica of the database and a current status of the replica of the database on the one or more other server nodes; and

update the server node'"'"'s replica of the database according to the received differences.

24. A system, comprising:

means for replicating a database on a plurality of server nodes, wherein the replicas of the database on the server nodes are concurrently accessible to service requests to the database;

means for the server nodes determining whether a new server node is needed in the plurality of server nodes;

means for the server nodes recruiting a pool node from one or more pool nodes as the new server node, wherein each of the one or more pool nodes is capable of becoming a server node in the replicated database system;

means for the server nodes determining that one of the server nodes is no longer needed; and

means for the one of the server nodes to be released, wherein the released server node becomes a pool node.

25. The system as recited in claim 24, further comprising means for synchronizing the replicas of the database on the two or more server nodes.

26. A method for dynamic configuration of a replicated database system, comprising:

a plurality of server nodes determining that a new server node is needed in the plurality of server nodes, wherein each of the plurality of server nodes comprises a replica of a database, and wherein the replicas of the database on the server nodes are concurrently accessible to service requests to the database, wherein said determining that a new server node is needed further comprises determining that demand on the replicated database system has increased above a demand threshold;

the plurality of server nodes recruiting a pool node from one or more pool nodes to become the new server node in the plurality of server nodes, wherein each of the one or more pool nodes is capable of becoming a server node in the replicated database system; and

configuring the recruited pool node to operate as the new server node in the plurality of server nodes.

27. The method as recited in claim 26, wherein said determining that a new server node is needed comprises detecting that one of the server nodes has gone down, wherein the server node that has gone down is replaced with the new server node.

28. The method as recited in claim 27, wherein said detecting that one of the server nodes has gone down comprises:

each of the plurality of server nodes communicating with the other server nodes at periodic or aperiodic intervals; and

one or more of the plurality of server nodes detecting that the one of the server nodes has not communicated with the other server nodes for a specified number of intervals.

29. The method as recited in claim 26, wherein said recruiting a pool node to become the new server node comprises at least one of the plurality of server nodes broadcasting a message to the pool nodes requesting one or more of the pool nodes to volunteer to become the new server node.

30. The method as recited in claim 29,

wherein said recruiting a pool node to become the new server node further comprises one of the plurality of server nodes receiving a response message to the broadcast message from the recruited pool node indicating that the recruited pool node is available to become a server node;

and

wherein said configuring the recruited pool node to operate as the new server node comprises the one of the plurality of server nodes configuring the recruited pool node as the new server node in response to the response message.

31. The method as recited in claim 26,

wherein said recruiting a pool node to become the new server node comprises the recruited pool node receiving a broadcast message requesting a pool node to volunteer to become the new server node;

and

wherein said configuring the recruited pool node to operate as the new server node comprises;

stopping execution of a current process on the recruited pool node in response to the received broadcast message; and

starting execution of a replicated database server process on the recruited pool node in response to the received broadcast message.

32. The method as recited in claim 26, wherein said configuring the recruited pool node to operate as the new server node comprises one of the plurality of server nodes replicating the database to the recruited pool node.

33. The method as recited in claim 26, wherein said configuring the recruited pool node to operate as the new server node comprises one of the plurality of server nodes providing software and configuration information to the recruited pool node to enable the recruited pool node to operate as a server node.

34. The method as recited in claim 26, further comprising:

the plurality of server nodes determining that demand on the replicated database has decreased below a demand threshold;

the plurality of server nodes selecting one of the server nodes to be released; and

releasing the selected server node, wherein the selected server node becomes a pool node.

35. The method as recited in claim 26, further comprising:

each of the server nodes periodically or aperiodically broadcasting a current status of the server node'"'"'s replica of the database to the other server nodes;

the server node receiving from one or more of the other server nodes differences between the broadcast current status of the server node'"'"'s replica of the database and a current status of the replica of the database on the one or more other server nodes; and

the server node updating the server node'"'"'s replica of the database according to the received differences.

36. The method as recited in claim 26, wherein the pool nodes and server nodes are peer nodes in a peer group on a peer-to-peer network.

37. The method as recited in claim 26, wherein the replicated database is used to track tasks distributed to computing nodes of a distributed computing system.

38. A computer-accessible storage medium, comprising program instructions, wherein the program instructions are computer-executable to implement:

a plurality of server nodes determining that a new server node is needed in the plurality of server nodes, wherein each of the plurality of server nodes comprises a replica of a database, and wherein the replicas of the database on the server nodes are concurrently accessible to service requests to the database, wherein, in said determining that a new server node is needed, the program instructions are further computer-executable to implement determining that demand on the replicated database has increased above a demand threshold;

the plurality of server nodes recruiting a pool node from one or more pool nodes to become the new server node in the plurality of server nodes, wherein each of the one or more pool nodes is capable of becoming a server node in the replicated database system; and

configuring the recruited pool node to operate as the new server node in the plurality of server nodes.

39. The computer-accessible storage medium as recited in claim 38, wherein, in said determining that a new server node is needed, the program instructions are further computer-executable to implement detecting that one of the server nodes has gone down, wherein the server node that has gone down is replaced with the new server node.

40. The computer-accessible storage medium as recited in claim 39, wherein, in said detecting that one of the server nodes has gone down, the program instructions are further computer-executable to implement:

each of the plurality of server nodes communicating with the other server nodes at periodic or aperiodic intervals; and

one or more of the plurality of server nodes detecting that the one of the server nodes has not communicated with the other server nodes for a specified number of intervals.

41. The computer-accessible storage medium as recited in claim 38, wherein, in said recruiting a pool node to become the new server node, the program instructions are further computer-executable to implement at least one of the plurality of server nodes broadcasting a message to the pool nodes requesting one or more of the pool nodes to volunteer to become the new server node.

42. The computer-accessible storage medium as recited in claim 41,

wherein, in said recruiting a pool node to become the new server node, the program instructions are further computer-executable to implement one of the plurality of server nodes receiving a response message to the broadcast message from the recruited pool node indicating that the recruited pool node is available to become a server node;

and

wherein, in said configuring the recruited pool node to operate as the new server node, the program instructions are further computer-executable to implement the one of the plurality of server nodes configuring the recruited pool node as the new server node in response to the response message.

43. The computer-accessible storage medium as recited in claim 38,

wherein, in said recruiting a pool node to become the new server node, the program instructions are further computer-executable to implement the recruited pool node receiving a broadcast message requesting a pool node to volunteer to become the new server node;

and

wherein, in said configuring the recruited pool node to operate as the new server node, the program instructions are further computer-executable to implement;

stopping execution of a current process on the recruited pool node in response to the received broadcast message; and

starting execution of a replicated database server process on the recruited pool node in response to the received broadcast message.

44. The computer-accessible storage medium as recited in claim 38, wherein, in said configuring the recruited pool node to operate as the new server node, the program instructions are further computer-executable to implement one of the plurality of server nodes replicating the database to the recruited pool node.

45. The computer-accessible storage medium as recited in claim 38, wherein, in said configuring the recruited pool node to operate as the new server node, the program instructions are further computer-executable to implement one of the plurality of server nodes providing software and configuration information to the recruited pool node to enable the recruited pool node to operate as a server node.

46. The computer-accessible storage medium as recited in claim 38, wherein the program instructions are further computer-executable to implement:

the plurality of server nodes determining that demand on the replicated database has decreased below a demand threshold;

the plurality of server nodes selecting one of the server nodes to be released; and

releasing the selected server node, wherein the selected server node becomes a pool node.

47. The computer-accessible storage medium as recited in claim 38, wherein the program instructions are further computer-executable to implement:

each of the server nodes periodically or aperiodically broadcasting a current status of the server node'"'"'s replica of the database to the other server nodes;

the server node receiving from one or more of the other server nodes differences between the broadcast current status of the server node'"'"'s replica of the database and a current status of the replica of the database on the one or more other server nodes; and

the server node updating the server node'"'"'s replica of the database according to the received differences.

48. The computer-accessible storage medium as recited in claim 38, wherein the pool nodes and server nodes are peer nodes in a peer group on a peer-to-peer network.

49. The computer-accessible storage medium as recited in claim 38, wherein the replicated database is used to track tasks distributed to computing nodes of a distributed computing system.