Checkpoint-Free In Log Mining For Distributed Information Sharing

• US 20110010392A1
• Filed: 08/10/2010
• Published: 01/13/2011
• Est. Priority Date: 07/31/2007
• Status: Active Grant

• Current Assignee: Cristina Schmidt, Lei Gao, Lik Wong, Nimar S. Arora, Thuyan Hoang
• Original Assignee: Cristina Schmidt, Lei Gao, Lik Wong, Nimar S. Arora, Thuyan Hoang
• Inventors: Arora, Nimar S., Schmidt, Cristina, Hoang, Thuyan, Gao, Lei, Wong, Lik
• Granted Patent: US 9,009,104 B2
• Time in Patent Office: Days
• US Class Current: 707/776
• CPC Class Codes: G06F 16/2358 Change logging, detection, and notification replication G06F16/27

Abstract

Techniques for replicating data between database systems without taking checkpoints are provided. In an embodiment, a capture process restarts. Upon restarting, the capture process reestablishes an association with an apply process. A particular logical time maintained by the apply process is then communicated to the capture process. Upon receiving the particular logical time, the capture process restarts mining from this particular logical time.

Claims

1. A method comprising:

a capture process establishing an association with an apply process at a destination node in a session, wherein said apply process at said destination node, during said session, maintains a record of a logical time of a source node, and wherein logical change records of transactions mined by said capture process from transaction logs of said source node, up to said logical time, have been applied by said apply process at said destination node;

said capture process restarting;

said capture process reestablishing said association with said apply process in a new session;

said capture process receiving said logical time from said apply process;

said capture process restarting mining said transaction logs of said source node using said logical time received from said apply process;

wherein said method is performed by one or more computing devices.

2. A method of claim 1, wherein said source node assigns system change numbers to said transactions at said source node, and wherein said logical time is one of said system change numbers.

3. A method of claim 1, wherein said source node assigns system change numbers to said logical change records at said source node, and wherein said logical time is one of said system change numbers.

4. A method of claim 1, wherein said capture process does not save any checkpoint information during said session and wherein said capture process does not retrieve any checkpoint information during said new session.

5. A method of claim 1, further comprising:

said capture process reestablishing said association with said apply process at said destination node in a second new session after said apply process restarts, wherein said apply process at said destination node, during said new session, maintains a record of a new logical time of said source node, and wherein said apply process at said destination node, in said second new session, retrieves said record of said new logical time of said source node;

said capture process receiving said new logical time from said apply process;

said capture process restarting mining said transaction logs of said source node using said new logical time received from said apply process.

6. A method of claim 1, further comprising said capture process preventing logical change records of inflight transactions at said source node at said logical time of said source node from being sent to said apply process in said new session.

8. A non-transitory machine-readable storage medium carrying one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform the method recited in claim 1.

9. A non-transitory machine-readable storage medium carrying one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform the method recited in claim 2.

10. A non-transitory machine-readable storage medium carrying one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform the method recited in claim 3.

11. A non-transitory machine-readable storage medium carrying one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform the method recited in claim 4.

12. A non-transitory machine-readable storage medium carrying one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform the method recited in claim 5.

13. A non-transitory machine-readable storage medium carrying one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform the method recited in claim 6.

7. A method comprising:

reestablishing an association with an apply process, wherein the apply process stores a particular system change number as the lowest value for all subsequent system change numbers;

receiving the particular system change number from the apply process;

establishing a plurality of active transactions at a particular time, wherein the particular time corresponding to a beginning of a particular transaction that is associated with the particular system change number, and wherein the plurality of active transactions excludes the particular transaction whose start transaction records is assigned with the particular system change number;

receiving one or more records that are related to the plurality of active transactions;

preventing sending the one or more records to the apply process.

14. A non-transitory machine-readable storage medium carrying one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform the method recited in claim 7.