Asynchronous interconnect protocol for a clustered DBMS

• US 7,801,997 B2
• Filed: 03/30/2006
• Issued: 09/21/2010
• Est. Priority Date: 03/30/2006
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Ng, Jack Hon Wai, Lam, Cheuk Lun, Pereyra, Herbert Walter, Kalmuk, David
• Primary Examiner(s): Backer; Firmin
• Assistant Examiner(s): Golabbakhsh; Ebrahim
• Application Number:
  US11/278,046
• Publication Number:
  US 20070260714A1
• Time in Patent Office: 1,636 Days
• Field of Search: 709/223, 709/227, 709/229
• US Class Current: 709/227
• CPC Class Codes: G06F 16/24554 Unary operations; Data partitioning operations
• CPC Class Codes: G06F 16/273 Asynchronous replication or reconciliation

Abstract

A method, system and computer program product for an asynchronous interconnection between nodes of a clustered database management system (DBMS). Node timestamps are provided when each of the nodes in the cluster are started. Two or more communication conduits are established between the nodes. Each communication conduit between a local node and a remote node has an associated session identifier. The session identifiers and the timestamp from the remote node are associated to each communication conduit and the associated local node in the cluster. A timestamp is received from the remote node at the local node when establishing communication to determine if the remote node corresponds to the remote node incarnation identified by the timestamp and if DBMS communication between nodes can be initiated.

Claims

1. For a clustered database management system (DBMS) operatively coupled to a data processing system having memory for storing a database, a method for asynchronous interconnection between a first and a second node of the clustered database management system, the first and the second node each having a processor and a memory, the method comprising the steps of:

a) associating a first node timestamp with the first node when the first node is started and a second node timestamp with the second node when the second node is started;

b) establishing a connection between the first node and second node by two or more communication conduits, wherein the two or more communication conduits each have an associated session identifier;

c) associating the session identifiers with the first node timestamp and the second node timestamp;

d) receiving, at the first node, another timestamp from the second node;

e) comparing the received another timestamp from the second node to the second node timestamp associated with the session identifiers; and

f) initiating, in response to the comparing, communications between the first and second nodes via the communication conduits.

2. The method of claim 1, wherein step f) further comprises commencing communication between the first node and second node when the received another timestamp matches the second node timestamp associated with one of the session identifiers.

3. The method of claim 1, wherein step b) further comprises associating respective status identifiers of the communication conduits with the respective session identifiers, wherein the status identifiers are defined as either up or down for each respective session identifier.

4. The method of claim 3, wherein step b) further comprises associating a communications socket identifier with one of the session identifiers when one of the communication conduits is established.

5. The method of claim 4, further comprising:

for each node, storing in a global node manager object a node timestamp and associated session identifiers, received timestamps from other nodes to which the node is connected in sessions corresponding to associated session identifiers, and status identifiers for identifying a status of each of the communication conduits associated with the session identifiers of the respective node; and

storing in a link table object, for each communication conduit, a session identifier, a socket identifier, and a status identifier of each of the communication conduits.

6. The method of claim 5, wherein step b) further comprises setting the status identifier of a communication conduit to up in the node manager object and the link table object when the connection is established.

7. The method of claim 6, further comprising setting the status identifier of the communications conduit which is associated with the session identifier of the communication conduit to down in the link table object when a drop or reset event is detected.

8. The method of claim 7, further comprising prior to step f) the steps of:

determining if the status identifier associated with the session identifier in the node manager object for the second node is equal to a value indicating that the communication conduit corresponding to the associated session identifier is up;

determining, if it is determined that the status identifier is up, if a session identifier of the communication conduit in the node manager object and a session identifier of the communication conduit in the link table object match; and

establishing a communication socket to initiate a connect call to the second node on the associated communication conduit when the status identifier is equal to up and the session identifiers of the communication conduit match.

9. The method of claim 8, further comprising storing a socket identifier for the established communication socket in the link table object of the communication conduit.

10. The method of claim 5, further comprising subsequent to step f) the steps of:

determining that the second node has failed;

setting the status identifier of the communication conduit associated with the second node to a value indicating that the communication conduit is down in the node manager object when communication between the first node and second node is not established due to the failed second node;

closing the communication conduits to the second node; and

incrementing the session identifiers associated with the closed communication conduits in the global node manager object.

11. The method of claim 1, wherein communication between the first node and the second node facilitates coordination of database transactions, wherein the database transactions comprise database queries and database updates.

12. For a clustered database management system (DBMS) operatively coupled to a data processing system, a computer programmed product comprising a non-transitory computer readable storage medium containing computer executable code for asynchronous interconnection between a first and a second node of the clustered database management system, the computer executable code on the computer programmed product comprising:

associating a first node timestamp with the first node when the first node is started and a second node timestamp with the second node when the second node is started;

establishing a connection between the first node and second node by two or more communication conduits, wherein the two or more communication conduits each have an associated session identifier;

associating the session identifiers with the first node timestamp and the second node timestamp;

receiving, at the first node, another timestamp from the second node;

comparing the received another timestamp from the second node to the second node timestamp associated with the session identifiers; and

initiating, in response to the comparing, communications between the first and second nodes via the communication conduits.

13. The computer programmed product of claim 12, wherein the initiating communications between the first and second nodes further comprises commencing communication between the first node and second node when the received another timestamp matches the second node timestamp associated with one of the session identifiers.

14. The computer programmed product of claim 12, wherein the establishing a connection between the first node and second node further comprises associating a communications socket identifier with one of the session identifiers when one of the communication conduits is established.

15. The computer programmed product of claim 14, wherein the computer executable code further comprises

for each node, storing in a global node manager object a node timestamp and session identifiers, received timestamps from other nodes to which the node is connected in sessions corresponding to associated session identifiers, and status identifiers for identifying a status of each of the communication conduits associated with the session identifiers of the respective node;

and

storing in a link table object, for each communication conduit, a session identifier, a socket identifier, and a status identifier of each of the communication conduits.

16. The computer programmed product of claim 15, wherein the computer executable code further comprises

determining if the status identifier associated with the session identifier in the node manager object for the second node is equal to a value indicating that the communication conduit corresponding to the associated session identifier is up;

determining, if it is determined that the status identifier is up, if a session identifier of the communication conduit in the node manager object and a session identifier of the communication conduit in the link table object match; and

establishing a communication socket to initiate a connect call to the second node on the associated communication conduit when the status identifier is equal to up and the session identifiers of the communication conduit match.

17. The computer programmed product of claim 15, wherein the computer executable code further comprises

determining that the second node has failed;

setting the status identifier of the communication conduit associated with the second node to a value indicating that the communication conduit is down in the node manager object when communication between the first node and second node is not established due to the failed second node;

closing the communication conduits to the second node; and

incrementing the session identifiers associated with the closed communication conduits in the global node manager object.

18. The computer programmed product of claim 12, wherein communication between the first node and the second node facilitates coordination of database transactions, wherein the database transactions comprise database queries and database updates.

19. For a clustered database management system (DBMS) operatively coupled to a data processing system, a system for asynchronous interconnection between a first and a second node of the clustered database management system, the system comprising:

a module, executed on a processor, for associating a first node timestamp with the first node when the first node is started and a second node timestamp with the second node when the second node is started;

a module for establishing a connection between the first node and second node by two or more communication conduits, wherein the two or more communication conduits each have an associated session identifier;

a module for associating the session identifiers with the first node timestamp and the second node timestamp;

a module for receiving, at the first node, another timestamp from the second node;

a module for comparing the received another timestamp from the second node to the second node timestamp associated with the session identifiers; and

a module for initiating, in response to the comparing, communications between the first and second nodes via the communication conduits.

20. The system of claim 19, wherein the module for initiating communications between the first and another further comprises commencing communication between the first node and second node when the received second node timestamp matches the second node timestamp associated with one of the session identifiers.

21. The system of claim 20, wherein the module for establishing a connection between the first node and second node further comprises associating a communications socket identifier with one of the session identifiers when one of the communication conduits is established.

22. The system of claim 21, further comprising:

a module for storing, for each node, in a global node manager object a node timestamp and session identifiers, received timestamps from other nodes to which the node is connected in sessions corresponding to associated session identifiers, and status identifiers for identifying a status of each of the communication conduits associated with the session identifiers of the respective node; and

a module for storing in a link table object, for each communication conduit, a session identifier, a socket identifier, and a status identifier of each of the communication conduits.

23. The system of claim 22, further comprising a module for:

determining if the status identifier associated with the session identifier in the node manager object for the second node is equal to a value indicating that the communication conduit corresponding to the associated session identifier is up;

determining, if it is determined that the status identifier is up, if a session identifier of the communication conduit in the node manager object and a session identifier of the communication conduit in the link table object match; and

establishing a communication socket to initiate a connect call to the second node on the associated communication conduit when the status identifier is equal to up and the session identifiers of the communication conduit match.

24. The system of claim 23, further comprising a module for:

determining that the second node has failed;

setting the status identifier of the communication conduit associated with the second node to a value indicating that the communication conduit is down in the node manager object when communication between the first node and second node is not established due to the failed second node;

closing the communication conduits to the second node; and

incrementing the session identifiers associated with the closed communication conduits in the global node manager object.

25. The system of claim 19, wherein communication between the first node and the second node facilitates coordination of database transactions, wherein the database transactions comprise database queries and database updates.