Maintaining Data Integrity in Data Servers Across Data Centers

• US 20100036885A1
• Filed: 08/05/2008
• Published: 02/11/2010
• Est. Priority Date: 08/05/2008
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Wang, Hao, Shen, Jinmei
• Granted Patent: US 8,676,760 B2
• Time in Patent Office: Days
• US Class Current: N/A
• CPC Class Codes: G06F 11/2028 eliminating a faulty processor or activating a spare
• CPC Class Codes: G06F 11/2035 without idle spare hardware
• CPC Class Codes: G06F 11/2097 maintaining the standby controller/processing unit updated initialisation or re-synchronisation thereof G06F11/1658 and subgroups

Abstract

A method, computer program product and system for maintaining data integrity of multiple copies of a piece of data in data servers of multiple data centers, includes maintaining a cluster of catalog servers, where one of the catalog servers is selected as the master catalog server, storing the piece of data in a primary data server chosen from the data servers, duplicating the piece of data to one or more backup data servers chosen from the data servers, recording a data state information to the master catalog server, duplicating the data state information to the cluster of catalog servers, and updating the data servers and the cluster of catalog servers when the primary data server fails.

Claims

1. A method for maintaining data integrity of a plurality of copies of a piece of data, the plurality of copies being stored in a plurality of data servers across multiple data centers, comprising:

selecting a primary data server from the plurality of data servers to store the piece of data, and duplicating the piece of data to other data servers in the plurality of data servers;

maintaining a cluster of catalog servers, wherein one of the catalog servers is selected as the master catalog server;

recording a data state information of the piece of data to the master catalog server, and duplicating the data state information to other catalog servers in the cluster of catalog servers; and

responsive to a failure of the primary data server, selecting a new primary server from the plurality of data servers, and updating the cluster of catalog servers.

2. The method of claim 1, wherein the updating comprises:

registering a new data state information containing the new primary data server to the master catalog server; and

duplicating the new data state information to other catalog servers in the cluster of catalog servers.

3. The method of claim 2, further comprising, before selecting the new primary data server,

managing a majority group and a minority group of data servers disconnected by the failing primary data server;

controlling functions of data servers in the minority group;

compiling dirty records of data in the minority group of data servers;

removing the failing primary data server by connecting each server in the plurality of data servers to a headmaster controller; and

regrouping the data servers connected to the headmaster controller.

4. The method of claim 1, further comprising routing a service request from a user to a closest data server in the plurality of data servers.

5. A computer program product for maintaining data integrity of a plurality of copies of a piece of data, the plurality of copies being stored in a plurality of data servers across multiple data centers, the computer program product comprising:

a computer usable medium having computer usable program code embodied therewith, the computer usable program code comprising;

instructions to select a primary data server from the plurality of data servers to store the piece of data, and to duplicate the piece of data to other data servers in the plurality of data servers;

instructions to maintain a cluster of catalog servers, wherein one of the catalog servers is selected as the master catalog server;

instructions to record a data state information of the piece of data to the master catalog server, and to duplicate the data state information to other catalog servers in the cluster of catalog servers; and

responsive to a failure of the primary data server, instructions to select a new primary server from the plurality of data servers, and to update the cluster of catalog servers.

6. The computer program product of claim 5, wherein the instructions to update comprise:

instructions to register a new data state information containing the new primary data server to the master catalog server; and

instructions to duplicate the new data state information to other catalog servers in the cluster of catalog servers.

7. The computer program product of claim 6, further comprising, before instructions to select the new primary data server,

instructions to manage a majority group and a minority group of data servers disconnected by the failing primary data server;

instructions to control functions of data servers in the minority group;

instructions to compile dirty records of data in the minority group of data servers;

instructions to remove the failing primary data server by connecting each server in the plurality of data servers to a headmaster controller; and

instructions to regroup the data servers connected to the headmaster controller.

8. The computer program product of claim 5, further comprising instructions to route a service request from a user to a closest data server in the plurality of data servers.

9. A computer system comprising:

a processor;

a memory operatively coupled with the processor;

a storage device operatively coupled with the processor and the memory; and

a computer program product for maintaining data integrity of a plurality of copies of a piece of data, the plurality of copies being stored in a plurality of data servers across multiple data centers, the computer program product comprising;

a computer usable medium having computer usable program code embodied therewith, the computer usable program code comprising;

instructions to select a primary data server from the plurality of data servers to store the piece of data, and to duplicate the piece of data to other data servers in the plurality of data servers;

instructions to maintain a cluster of catalog servers, wherein one of the catalog servers is selected as the master catalog server;

instructions to record a data state information of the piece of data to the master catalog server, and to duplicate the data state information to other catalog servers in the cluster of catalog servers; and

responsive to a failure of the primary data server, instructions to select a new primary server from the plurality of data servers, and to update the cluster of catalog servers.

10. The computer system of claim 9, wherein the instructions to update comprise:

instructions to register a new data state information containing the new primary data server to the master catalog server; and

instructions to duplicate the new data state information to other catalog servers in the cluster of catalog servers.

11. The computer system of claim 10, further comprising, before instructions to select the new primary data server,

instructions to manage a majority group and a minority group of data servers disconnected by the failing primary data server;

instructions to control functions of data servers in the minority group;

instructions to compile dirty records of data in the minority group of data servers;

instructions to remove the failing primary data server by connecting each server in the plurality of data servers to a headmaster controller; and

instructions to regroup the data servers connected to the headmaster controller.

12. The computer system of claim 9, further comprising instructions to route a service request from a user to a closest data server in the plurality of data servers.

13. A computer system for maintaining data integrity of a plurality of copies of a piece of data, the plurality of copies being stored in a plurality of data servers across multiple data centers, comprising:

a network partition manager configured to detect and manage first and second network partitions created when one of the plurality of data centers fails;

a headmaster controller configured to control functions of data centers in at least one of the first and second network partitions;

a router configured to redirect all write transactions to the second network partition to the first network partition;

a repair manager configured to reconnect a select data server in the second network partition to the headmaster controller, so that the headmaster controller can remove function limitations on the select data server;

a recovery manager configured to recover status and privilege information of the select data server after the repair manager establishes a connection to the headmaster controller;

a temporal regrouping manager configured to remove a failing data server in the failing data center, and to regroup other data servers in the plurality of data servers;

a load and traffic controller configured to redirect traffic to the other data servers according to the service capability of each of the other data servers; and

an inter-data-center connection manager configured to monitor connections between the data centers.

14. The computer system of claim 13, wherein the first network partition comprises more data centers than the second network partition.

15. The computer system of claim 13, wherein the functions comprise removing a transaction commit privilege of data centers in a select network partition, and keeping non-dirty reads to the select network partition.

16. The computer system of claim 15, wherein the select network partition is one of the first and second network partitions that has a smaller number of data centers than the other one.

17. The computer system of claim 13, wherein the router is also configured to compile dirty records.

18. The computer system of claim 13, wherein the repair manager consistently attempts to connect to the partitioned headmaster controller according to a preset schedule.