Method, system, and database archive for enhancing database archiving

• US 20070288490A1
• Filed: 06/12/2006
• Published: 12/13/2007
• Est. Priority Date: 06/12/2006
• Status: Active Grant

• Current Assignee: Teradata US, Inc. (Teradata Corporation)
• Original Assignee: RainStor Ltd. (Teradata Corporation)
• Inventors: Longshaw, Tom Benjamin
• Granted Patent: US 8,271,452 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 11/1464 for networked environments
• CPC Class Codes: G06F 11/1471 involving logging of persistent data for recovery
• CPC Class Codes: G06F 16/217 Database tuning G06F16/2282 takes precedence; database performance monitoring G06F11/3409
• CPC Class Codes: G06F 2201/80 Database-specific techniques
• CPC Class Codes: G06F 2201/84 Using snapshots, i.e. a logical point-in-time copy of the data

Abstract

A database archive, and a system and method for generating database archives is disclosed. A processor may store in a database archive an index that includes record entries that each references one or more record fields of a database, and may store the record fields independently of the database archive.

Claims

1. A database archiving method, comprising:

storing in a first database archive an index that includes at least one record entry, wherein each of the at least one record entry references at least one record field of a database; and

storing each of the at least one record field independently of the database archive.

2. The method of claim 1, wherein at least one of the at least one record field is a field of more than one record.

3. The method of claim 1, further comprising:

storing a plurality of record fields, the plurality including the at least one record field, wherein;

each of the plurality of record fields is a field of a record of the database; and

each field included in more than one record of any single table of the database is stored for the more than one record as a single record field referenced by a plurality of record entries, each of the plurality of record entries corresponding to a different one of the more than one record.

4. The method of claim 1, wherein each of the at least one record field is referenceable by (a) at least one of a plurality of database archives and (b) a current database index.

5. The method of claim 4, wherein:

the at least one record field is referenced by the each of the at least one record entry one of directly and indirectly; and

a record field is referenced indirectly by a reference of a record entry to a record data element that references the record field.

6. The method of claim 5, wherein the record data element references more than one record field.

7. The method of claim 4, further comprising:

maintaining a record field in storage upon a condition that the record field is referenced by at least one of (a) at least one of the plurality of database archives and (b) the current database index.

8. The method of claim 7, further comprising:

allotting, one of for each record field and for each group of record fields forming a record, a corresponding bit to indicate whether the record field is referenced by any of the plurality of database archives;

one of;

for each record field, during generation of the record field, setting the record field'"'"'s corresponding bit to a first value; and

for each group, during generation of the group, setting the group'"'"'s corresponding bit to the first value;

for each of the plurality of database archives, during generation of the database archive, one of;

setting, for each record field referenced by the database archive, the record field'"'"'s corresponding bit to a second value if not previously set to the second value; and

setting, for each group referenced by the database archive, the group'"'"'s corresponding bit to the second value if not previously set to the second value; and

one of;

removing a particular record field from storage upon a condition that the corresponding bit of the particular record field is set to the first value and the particular record field is not referenced by the current database index; and

removing a particular group from storage upon a condition that the corresponding bit of the particular group is set to the first value and the particular group is not referenced by the current database index.

9. The method of claim 1, further comprising:

modifying a current database index that, for representation of a current state of the database, references one or more of a plurality of record fields including the at least one record field, by at least one of deleting from the current database index, inserting into the current database index, and updating in the current database index a representation of a record;

wherein;

deleting the representation of the record includes one of (a) removing from the current database index a record entry corresponding to the record and (b) annotating the record entry corresponding to the record to indicate its deletion;

inserting the representation of the record includes inserting into the current database index the record entry corresponding to the record; and

updating the representation of the record includes at least one of (a) removing from the current database index the record entry corresponding to the record and inserting a new record entry in place of the removed record entry and (b) annotating the record entry corresponding to the record to indicate its deletion and inserting the new record entry immediately following the record entry corresponding to the record.

10. The method of claim 9, wherein inserting the record includes annotating the record entry corresponding to the record to indicate its time of entry, the method further comprising:

copying the current database index to obtain the index of the first database archive; and

subsequent to the copying of the current database index;

removing from the current database index each record entry that is annotated to indicate its deletion; and

clearing all time of entry annotations.

11. The method of claim 10, further comprising:

retrieving one of a second database archive and the current database index; and

based on record entry annotations of the one of the second database archive and the current database index, generating the index of the first database archive.

12. The method of claim 11, wherein:

the first database archive is associated with a time t;

a particular record entry of the one of the second database archive and the current database index is included in the first database archive upon a condition that the one of the second database archive and the current database index indicates that a record to which the particular record entry corresponds was part of a state of the database at time t.

13. The method of claim 1, further comprising:

copying a current database index to obtain the index of the first database archive.

14. The method of claim 1, further comprising:

storing a log of record operations; and

for generation of the first database archive;

retrieving a copy of a second database archive generated prior to generation of the first database archive; and

modifying the copy of the second database archive in accordance with the log;

wherein, subsequent to the modification of the copy, the copy is stored as the first database archive.

15. The method of claim 14, wherein:

the modification includes;

inserting a new record entry into the copy if the log indicates one of an insert operation and an update operation; and

removing a record entry from the copy if the log indicates one of a deletion operation and the update operation.

16. The method of claim 15, wherein:

a state of the database at a particular time is reconstructible based on the first database archive;

each entry of the log includes a time of a corresponding logged record operation;

modification of the copy based on a particular record operation indicated by the log is made upon a condition that the particular record operation is associated by the log with a time that is one of at and earlier than the particular time.

17. The method of claim 16, wherein:

the database includes a plurality of tables;

the log includes a plurality of logs, each log corresponding to a different one of the plurality of tables;

the second database archive includes a plurality of index portions, each portion corresponding to a different one of the logs; and

modifying the copy includes separately modifying each index portion of the copy according to its corresponding log.

18. The method of claim 14, further comprising:

for generation of a third database archive associated with a time t that is prior to a time with which the first database archive is associated;

retrieving a copy of the first database archive;

for each record entry of the copy, removing from the record entry from the copy if and upon a condition that the first database archive indicates that a record to which the record entry corresponds was not part of a state of the database at the time t; and

subsequently storing the copy as the third database archive.

19. The method of claim 18, wherein a record entry of the first database archive is included in the third database archive if and upon a condition that:

(A) it is one of;

(a) associated with a time-of-deletion timestamp that one of (i) indicates a time subsequent to time t and (ii) is NULL; and

(b) not associated with a time-of-deletion timestamp; and

(B) it is one of;

(a) associated with a time-of-entry timestamp that one of (i) indicates a time one of at and earlier than the time t and (ii) is NULL; and

(b) not associated with a time-of-entry timestamp.

20. The method of claim 19, wherein the first database archive is used as a basis for generating the third database archive upon a condition that of all previously generated database archives associated with times subsequent to time t, the time with which the first database archive is associated is closest to time t.

21. A database archiving method, comprising:

generating a first database archive representing a state of a database at a first time; and

subsequent to the generation of, and based on data of, the first database archive, generating a second database archive representing a state of the database at a second time that is earlier than the first time.

22. A database archiving method, comprising:

storing a plurality of record fields; and

storing a plurality of database indices, wherein;

each of the indices corresponds to a different time;

each of the indices includes at least one corresponding record entry, each record entry corresponding to a record included in a state of a database at the time to which the index corresponds;

each record entry references at least one of the record fields; and

each record field is referenceable by more than one of the indices.

23. A database archiving method, comprising:

storing a first record index referencing record data elements as a first database archive representing a state of a database at a first time;

storing a log of record operations including at least one record operation corresponding to a period beginning one of at and prior to the first time;

selecting a second time that is after the first time and within the time period;

generating a second record index based on the first record index and the log; and

storing the second record index as a second database archive representing a state of the database at the second time.

24. A computer-readable medium having stored thereon instructions adapted to be executed by a processor, the instructions which, when executed, cause the processor to perform a database archiving method, the database archiving method comprising:

storing in a first database archive an index that includes at least one record entry, wherein each of the at least one record entry references at least one record field of a database; and

storing each of the at least one record field independently of the database archive.

25. A database archive, comprising:

an index that;

includes a plurality of record entries that each references at least one corresponding record field stored independently of the database archive; and

when loaded by a processor, causes the processor to reconstruct a state of a database that existed at a time to which the database archive corresponds by retrieving all record fields referenced by the index.