Reconciliation of geo-replicated database clusters

• US 9,405,816 B2
• Filed: 03/05/2013
• Issued: 08/02/2016
• Est. Priority Date: 03/05/2013
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Inventors: Verbitski, Alexandre, Xia, Hao, Voutilainen, Santeri, Talius, Tomas, Nelson, Michael, Kuchibhotla, Sree, Shah, Vishrut
• Primary Examiner(s): DWIVEDI, MAHESH H
• Application Number:
  US13/784,825
• Publication Number:
  US 20140258229A1
• Time in Patent Office: 1,246 Days
• Field of Search: 707/625, 707/634
• US Class Current: 1/1
• CPC Class Codes: G06F 16/273 Asynchronous replication or reconciliation

Abstract

A database associated with a primary database cluster may be replicated in a backup database cluster located in a different location in order to provide a highly-available fault tolerant database service. The databases are reconciled through a cluster management module distributed in each database cluster. The cluster management module uses a set of reconciliation data structures to track locally the reconciled states of each database in each database cluster, the operations made locally to the databases in each database cluster, and the author of the operations. The cluster management module in each database cluster engages in a stateless messaging protocol using the set of reconciliation data structures to determine whether or not the databases may be reconciled.

Claims

1. A method for reconciling data stored on a primary database with a secondary database, the method comprising:

tracking a current state identifier of the primary database that identifies a last reconciled state with the secondary database, a last system change made in the primary database, and a last system change made in the secondary database;

transmit a request to reconcile changes in the primary database with the secondary database, the request including the current state identifier and a payload; and

receiving confirmation that the primary database has been synchronized with the secondary database, when the current state identifier of the primary database matches a current state identifier of the secondary database and no system operations have been performed on the primary database or the secondary database since a last reconciliation,

wherein the current state identifier of the secondary database identifies a last reconciled state with the primary database, a last system change made in the primary database, and a last system change made in the secondary database by a system.

2. The method of claim 1, further comprising:

generating a new state identifier when a user performs an operation on the primary database that is not reconciled with the secondary database; and

wherein the transmission of the current state identifier includes the new state identifier,

wherein the new state identifier is used to update the current state identifier of the secondary database.

3. The method of claim 1, further comprising:

updating a previous state identifier of the primary database with the current state identifier of the primary database, the previous state identifier of the primary database representing a state prior to the current state identifier; and

updating the current state identifier of the primary database with the new state identifier of the primary database.

4. The method of claim 1, further comprising:

delaying reconciliation of the primary database with the secondary database when a current state identifier of the primary database matches the current state identifier of the secondary database data and there is at least one system operation performed on the primary database or the secondary database since the last reconciliation.

5. The method of claim 1, wherein the payload contains changes made on the primary database since a last reconciliation state.

6. The method of claim 1, further comprising:

receiving a new state identifier of the secondary database, a current state identifier of the secondary database, and a payload not reconciled in the primary database;

when the received current state identifier of the secondary database matches a previous state identifier of the primary database, updating the primary database with the payload.

7. The method of claim 6, further comprising:

updating a previous state identifier of the primary database with the current state identifier of the secondary database and the current state identifier of the primary database with the new state identifier of the secondary database.

8. The method of claim 1 wherein the primary database is hosted on a primary database cluster and the secondary database is hosted on a secondary database cluster, the primary database cluster located in a same geographic location as the secondary database cluster.

9. The of claim 1 wherein the primary database is hosted on primary database cluster and the secondary database is hosted on a secondary database cluster, the primary database cluster located in a different geographic location as the secondary database cluster.

10. A system, comprising:

a primary database cluster and a secondary database cluster; and

the primary database cluster having at least one processor and a memory;

the at least one processor configured to;

establish a relationship between a primary database in the primary database cluster and the secondary database in the secondary database cluster that allows a user to perform one or more operations on the primary database and on the secondary database concurrently and independently;

track a current state identifier and a previous state identifier for the primary database, the current state identifier represents a last reconciled state with the secondary database and a last system change made in the primary database, and a last system change made in the secondary database;

reconcile operations performed on the primary database with the secondary database based on a current state identifier of the primary database and the current state identifier of the secondary database; and

reconcile operations performed on the secondary database with the primary database based on a current state identifier of the secondary database and a previous state identifier of the primary database.

11. The system of claim 10, wherein the at least one processor is further configured to:

transmit a first message to the secondary database requesting reconciliation with the primary database after the primary database has been unavailable;

receive a second message from the secondary database including the current state identifier of the secondary database and a payload; and

reconcile the primary database with the payload when the current state identifier of the secondary database matches the previous state identifier of the primary database.

12. The system of claim 10, wherein the at least one processor is further configured to:

transmit a first message to the secondary database requesting reconciliation, the first message including the current state identifier of the primary database; and

receive a second message from the secondary database rejecting the reconciliation request when a previous state identifier of the secondary database does not match the current state identifier of the primary database.

13. The system of claim 12, wherein the at least one processor is further configured to:

suspend the relationship between the primary database and the secondary database.

14. The system of claim 10,

wherein the current state identifier of the primary database includes a statename, a first sequence value and a second sequence value, the statename identifying a current reconciled state of the primary database, the first sequence value identifying a last system change made in the primary database, and the second sequence value identifying a last system change made in the secondary database;

and

wherein the at least one processor is further configured to;

transmit a first message to the secondary database requesting reconciliation, the first message including the current state identifier of the primary database; and

receive a second message rejecting the reconciliation request when the statename of the current state identifier of the primary database matches the statement of the current state identifier of the secondary database, the first sequence value of the current state identifier of the primary database does not match the first sequence value of the current state identifier of the secondary database, and second sequence value of the current state identifier of the primary database does not match the second sequence value of the current state identifier of the secondary database.

15. The system of claim 10 wherein the primary database cluster and the secondary database cluster are in different geographic regions.

16. A device, comprising:

at least one processor and a memory;

the at least one processor configured to;

establish a relationship with a first database so that the first database and the second database are synchronized;

track operations performed on the second database using a current state identifier, the current state identifier associated with a last reconciled state with the first database, a last system change made in the first database, and a last system change made in the second database;

receive a first message requesting reconciliation between a first database and a second database, the first message including a current state identifier of the first database and a payload; and

incorporating the payload into the second database when the current state identifier of the first database matches the current state identifier of the second database.

17. The device of claim 16,

wherein a current state identifier includes a statename, a first sequence value and a second sequence value, the statename identifying a current reconciled state, the first sequence value identifying a last system change made in the first database, and the second sequence value identifying a last system change made in the secondary database, and

wherein the at least one processor is further configured to:

receive a second message requesting reconciliation, the second message including the current state identifier of the first database; and

transmit a response to retry the reconciliation at a later time when the statename of the current state identifier of the first database matches the statename of the current state identifier of the second database, the first sequence value of the current state identifier of the first database does not match the first sequence value of the current state identifier of the second database, and the second sequence value of the current state identifier of the first database does not match the second sequence value of the current state identifier of the second database.

18. The device of claim 16, wherein the at least one processor is further configured to:

receive a second message including a current state identifier of the first database; and

transmit a response including the current state identifier of the second database and a payload,

wherein the first database is updated with the payload when the current state identifier of the second database matches the previous state identifier of the first database.

19. The device of claim 16, wherein the at least one processor is further configured to:

receive a second message requesting reconciliation including a current state identifier of the first database; and

transmit a response rejecting the reconciliation request when a previous state identifier of the second database does not match the current state identifier of the first database.

20. The device of claim 16, wherein the first database and the second database are located in different geographic locations.