Per file dirty region logging

• US 20080010515A1
• Filed: 05/15/2006
• Published: 01/10/2008
• Est. Priority Date: 05/15/2006
• Status: Active Grant

• Current Assignee: Oracle America, Inc. (Oracle Corporation)
• Original Assignee: Sun Microsystems Incorporated (Oracle Corporation)
• Inventors: Wong, Brian, McDougall, Richard, Shepler, Spencer, Robinson, David
• Granted Patent: US 7,730,351 B2
• Time in Patent Office: Days
• US Class Current: 714/015
• CPC Class Codes: G06F 11/2082 Data synchronisation

Abstract

A method for dirty region logging of a file that includes receiving a request to open the file, determining each of a plurality of component files associated with the file, opening each of the plurality of component files, writing to a region of the at least one of the plurality component files, and updating a dirty region log (DRL) associated with the one of the plurality of component files to reflect the write to the region.

Claims

1. A method for dirty region logging of a file, comprising:

receiving a request to open the file;

determining each of a plurality of component files associated with the file;

opening each of the plurality of component files;

receiving a request to write to a region of the at least one of the plurality component files;

updating a dirty region log (DRL) associated with at least one of the plurality of component files to reflect the write to the region in response to the request; and

writing to the region of the at least one of the plurality component files in response to the request.

2. The method of claim 1, further comprising

marking the file as a dirty in response to the request.

3. The method of claim 1, wherein determining each of the plurality of component files associated with the file comprises determining a location of the plurality of component files.

4. The method of claim 1, wherein the DRL corresponds to a bitmap.

5. The method of claim 1, wherein the file is stored on a server implementing Network File System version 4 (NFSv4).

6. The method of claim 1, wherein a server that received the request updates the DRL in response to the request.

7. The method of claim 1, wherein the DRL is updated by a client that sent the request to a server.

8. The method of claim 7, wherein the client uses a compound request to update the DRL, wherein the DRL is stored on the server.

9. The method of claim 1, wherein the DRL for the at least one of the plurality component files is stored in an extended attribute of the at least one of the plurality component files.

10. A method for recovering a file on a failed server comprising:

determining whether the file is dirty;

determining a first component file and a second component file associated with the file, if the file is dirty;

obtaining a first dirty region log (DRL) associated with the first component file and a second DRL associated with the second component file;

merging the first DRL and the second DRL to obtain a merged DRL, wherein the merged DRL comprises at least one region selected from the group consisting of the first component file and second component file; and

synchronizing contents of the at least one region in the first component file and the second component file using the merged DRL.

11. The method of claim 10, wherein merging the first DRL and the second DRL comprises performing a union operation between dirty regions in the first DRL and dirty regions in the second DRL.

12. The method of claim 10, wherein the failed server implements Network File System version 4 (NFSv4).

13. The method of claim 10, wherein the first DRL is located on a client and the second DRL is located on a server, wherein the file is located on the server.

14. A system, comprising:

a plurality of servers each configured to store one of a plurality of component files associated with a file and to store a dirty region log (DRL) associated with each of the plurality of component files; and

a metadata server operatively connected to the plurality of servers configured to store a location of each of the plurality of component files of the file.

15. The system of claim 14, wherein metadata server is further configured to:

receive a request to perform a write operation on the file; and

determine each of the plurality of component files associated with the file; and

record that the file is dirty in response to write request.

16. The system of claim 14, wherein each of the plurality of servers is configured to:

determine whether the file is dirty using the metadata server;

determine a first component file and a second component file associated with the file, if the file is dirty;

obtain a first dirty region log (DRL) associated with the first component file and a second DRL associated with the second component file;

merge the first DRL and the second DRL to obtain a merged DRL, wherein the merged DRL comprises at least one region selected from the group consisting of the first component file and second component file; and

synchronize contents of the at least one region in the first component file and the second component file using the merged DRL,

wherein the first component file is located a first one of the plurality of servers and the second component file located on second one of the plurality of servers.

17. The system of claim 14, wherein merging the first DRL and the second DRL comprises performing a union operation between dirty regions in the first DRL and dirty regions in the second DRL.