Robust end-of-log processing

• US 7,756,838 B2
• Filed: 12/12/2005
• Issued: 07/13/2010
• Est. Priority Date: 12/12/2005
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: East, Jeffrey A., Felipe, Luis, Lindell, Steven J., Kulesza, Jakub
• Primary Examiner(s): Jalil; Neveen Abel
• Assistant Examiner(s): BUCKINGHAM, KELLYE DEE
• Application Number:
  US11/301,168
• Publication Number:
  US 20070136380A1
• Time in Patent Office: 1,674 Days
• Field of Search: 707/200
• US Class Current: 707/682
• CPC Class Codes: G06F 11/1471 involving logging of persistent data for recovery
• CPC Class Codes: G06F 11/1474 in transactions G06F16/20 takes precedence
• CPC Class Codes: G06F 11/2094 Redundant storage or storage space G06F11/2056 takes precedence
• CPC Class Codes: G06F 2201/82 Solving problems relating to consistency

Abstract

A method of invalidating an unused log file in a database includes examining a discard log information storage area and determining if a discard log exists. If the discard log exists, the log is invalidated. The headers of each of the log files that have been fully invalidated are then invalidated. The discard log is then fully invalidated and a new log file may be generated on top of the discarded log. The method may be interrupted by a failure, a changeover, or crash and will be repeated until it completes.

Claims

1. A method for processing database transaction log files, the method comprising:

invalidating a discard log file of a transaction log of a first database, wherein the discard log file is an unusable transaction log file comprising one or more log blocks of the transaction log having unusable data content that is to be discarded before the first database can begin processing new log records of database transactions received from a second database,

wherein invalidating the discard log file comprises;

(a) reading discard log information from a storage area, the discard log information comprising an indicator that the discard log file exists, a starting point identifier of the discard log file, and an ending point identifier of the discard log file, the discard log file comprising at least one log block of the transaction log having unusable data content that is to be discarded;

(b) replacing the unusable data content of the at least one log block of the discard log file with an identifiably invalid data pattern that makes the log block content of the discard log file identifiable as invalid when scanning the transaction log, said identifiably invalid data pattern comprising a pattern of all zeros, a pattern of all ones or a pattern having a fixed combination of zeros and ones; and

invalidating header information of the at least one log block; and

(c) resetting the discard log information, the reset clearing the indicator that the discard log file exists; and

overlaying identifiably invalid content of a discarded log block of the discard log file with new log records received from the second database, and

using existing identifiably invalid content of the discard log file to detect an end of the transaction log, when scanning the transaction log during a recovery process from failure.

2. The method of claim 1, further comprising resuming the method from step (a) if a failover occurs in any of steps (a), (b), and (c).

3. The method of claim 1, wherein replacing the unusable data content of the at least one log block of the transaction log that is to be discarded with an identifiably invalid data pattern comprises writing the identifiably invalid data pattern into the at least one log block in a physical file order within a logical file beginning at a beginning logical sequence number and progressing to an end logical sequence number.

4. The method of claim 1, wherein replacing the unusable data content of the at least one log block of the transaction log that is to be discarded with an identifiably invalid data pattern comprises writing an identifiably invalid data pattern into header information of log blocks that are invalidated.

5. The method of claim 4, wherein header information of a last log block is invalidated before header information of a preceding log block, wherein if a failover occurs, header information is retained in preceding log blocks to allow repeating the invalidating.

6. The method of claim 1, wherein resetting the discard log information further comprises resetting the starting point identifier of the discard log file and the ending point identifier of the discard log file.

7. The method of claim 1, wherein the invalidating is performed in response to initiating a changeover from the first database operating as a principal to the second database operating as a mirror, wherein said changeover causes the second database to operate as the principal and the first database to operate as the mirror.

8. A method for processing database transaction log files, the method comprising:

invalidating a discard log file of a transaction log of a first database, wherein the discard log file is an unusable transaction log file comprising one or more log blocks of the transaction log having unusable data content that is to be discarded before the first database can begin processing new log records of database transactions received from a second database,

wherein invalidating the discard log file comprises;

(a) reading discard log information from a storage area, the discard log information comprising an indicator that the discard log file exists, a starting point identifier of the discard log file, and an ending point identifier of the discard log file, the discard log file comprising at least a first log block of the transaction log having unusable data content that is to be discarded;

(b) replacing the unusable data content of the first log block of the transaction log that is to be discarded with an identifiably invalid data pattern that makes the log block content of the discard log file identifiable as invalid when scanning the transaction log, said identifiably invalid data pattern comprising a pattern of all zeros, a pattern of all ones or a pattern having a fixed combination of zeros and ones; and

invalidating header information of the at least one log block;

(c) updating the starting point identifier of the discard log file to correspond to a next sequential log block having unusable data content following the first log block;

asynchronously invalidating a remaining portion of the discard log file starting with the next sequential log block while the first database is receiving and processing new log records from the second database;

using existing identifiably invalid content of the discard log file to detect an end of the transaction log, when scanning the transaction log during a recovery process from failure.

9. The method of claim 8, wherein the invalidating is performed in response to initiating a changeover from the first database operating as a principal to the second database operating as a mirror, wherein said changeover causes the second database to operate as the principal and the first database to operate as the mirror.

10. The method of claim 8, wherein asynchronously invalidating a remaining portion of the discard log file comprises:

writing an identifiably invalid data pattern into log blocks of the discarded log file; and

writing an invalid data pattern into headers of the log blocks of the discarded log files.

11. The method of claim 8, further comprising resetting the discard log information, the reset clearing the indicator that the discard log file exists.

12. The method of claim 8, further comprising resuming the method from step (a) if a failover occurs in any of steps (a), (b), and (c).

13. The method of claim 8, wherein reading discard log information from a storage area further comprises reading a starting point identifier that is past the header of the at least one block.

14. The method of claim 8, wherein replacing the unusable data content comprises writing an identifiably invalid data pattern into the at least one log block in a physical file order within a logical file beginning at a beginning logical sequence number and progressing to an end logical sequence number.

15. A computer-readable storage medium having computer-executable instructions to perform a method for processing database transaction log files, the method comprising:

invalidating a discard log file of a transaction log of a first database, wherein the discard log file is an unusable transaction log file comprising one or more log blocks at the end of the transaction log having unusable data content that is to be discarded before the first database can begin processing new log records of database transactions received from a second database,

wherein invalidating the discard log file comprises;

(a) reading discard log information from a storage area, the discard log information comprising an indicator that the discard log file exists, a starting point identifier of the discard log file, and an ending point identifier of the discard log file, the discard log file comprising at least one log block of the transaction log having unusable data content that is to be discarded;

(b) replacing the unusable data content of the at least one log block of the discard log file with an identifiably invalid data pattern that makes the log block content of the discard log file identifiable as invalid when scanning the transaction log, said identifiably invalid data pattern comprising a pattern of all zeros, a pattern of all ones or a pattern having a fixed combination of zeros and ones; and

invalidating header information of the at least one log block; and

(c) writing a pattern of identifiably invalid data into header information of the at least one log block; and

(d) resetting the discard log information, the reset clearing the indicator that the discard log file exists; and

overlaying identifiably invalid content of a discarded log block of the discard log file with new log records received from the second database, and

using existing identifiably invalid content of the discard log file to detect an end of the transaction log, when scanning the transaction log during a recovery process from failure.

16. The computer-readable storage medium of claim 15, wherein replacing the unusable data content of the at least one log block of the transaction log that is to be discarded with an identifiably invalid data pattern comprises writing an identifiably invalid data pattern into the at least one log block in a physical file order within a logical file beginning at a beginning logical sequence number and progressing to an end logical sequence number.

17. The computer-readable storage medium of claim 15, wherein replacing the unusable data content of the at least one log block of the transaction log that is to be discarded with an identifiably invalid data pattern comprises writing an identifiably invalid data pattern into header information of log blocks that are invalidated.

18. The computer-readable storage medium of claim 17, wherein header information of a last log block is invalidated before header information of a preceding log block, wherein if a failover occurs, header information is retained in preceding log blocks to allow repeating the invalidating.

19. The computer-readable storage medium of claim 15, further comprising computer-executable instructions for resuming the method from step (a) if a failover occurs in any of steps (a), (b), (c), and (d).