Data backup and recovery in a data processing system

• US 5,381,545 A
• Filed: 05/28/1992
• Issued: 01/10/1995
• Est. Priority Date: 06/04/1991
• Status: Expired due to Term

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Baker, Stephen M., Lupton, Peter J.
• Primary Examiner(s): Canney, Vincent P.
• Application Number:
  US07/890,495
• Time in Patent Office: 957 Days
• Field of Search: 395/575, 395/600, 364/268.2, 364/282.1, 364/283.4, 364/920, 364/944.1, 364/945.3, 364/962.1
• US Class Current: 714/19
• CPC Class Codes: G06F 11/1451 by selection of backup contents
• CPC Class Codes: G06F 11/1466 to make the backup process non-disruptive
• CPC Class Codes: G06F 11/1471 involving logging of persistent data for recovery
• CPC Class Codes: Y10S 707/99953 Recoverability

Abstract

A data processing system for managing stored data is described. A log recording incremental updates to the data is provided for data recovery purposes in the event of a system failure. A backup copy of at least a portion of the stored data can be made without closing that portion for update. Any in-progress modifications at the time the backup is taken are allowed for by calculating a recovery point before the time of the backup, from which point recovery is to be performed using the recovery log. The recovery point depends at least in part on the start time of the oldest in-progress data file update.

Claims

1. Data processing apparatus comprising:

means for processing stored data as a series of logical units of work (LUWs), with updates to the stored data made during each LUW being committed at a commit point for that LUW;

means for maintaining a log of updates made to the stored data;

means for detecting the oldest uncommitted update made to the stored data;

means responsive to the detecting means for calculating a recovery point;

means for taking a backup copy of at least a portion of the stored data while that portion is available for update; and

means for recovering the database by reinstating the backup copy to provide a working copy and applying the updates contained in any log entries made after said recovery point.

2. Data processing apparatus as claimed in claim 1, in which the means for detecting the oldest uncommitted update comprises:

means for detecting the first log entry, corresponding to an update, made by each LUW;

means for storing the time of that first log entry;

means for detecting the earliest of at least a subset of the stored first log entry times; and

means for deleting the first log entry time corresponding to a completed LUW.

3. Data processing apparatus of claim 1, in which:

the stored data is organized as a plurality of data sets, each data set being addressable by means of one or more alias filenames.

4. Data processing apparatus as claimed in claim 3, in which:

each data set has an associated catalogue, for storing system housekeeping data; and

the recovery point corresponding to a data set is written to that data set'"'"'s catalogue; and

the backup copy of a data set includes a copy of the data set'"'"'s catalogue at the time the backup was taken.

5. Data processing apparatus as claimed in claim 3 or claim 4, further comprising:

means for associating with a data set a data item signifying whether that data set is eligible to be backed up while available for update.

6. Data processing apparatus of claim 3, further comprising:

means for writing a tie-up-record (TUR) to the log when a filename is associated with a data set, the TUR defining at least the association between the filename and the data set.

7. Data processing apparatus as claimed in claim 6, further comprising:

means for periodically writing to the log a set of TURs for at least some of the data sets.

8. Data processing apparatus as claimed in claim 7, in which:

the set of TURs comprises those TURs corresponding to filenames associated with open data sets eligible to be backed up while open.

9. Data processing apparatus as claimed in claim 7, in which:

the recovery point is defined as the start time of writing a previous set of TURs for which writing completed before the oldest uncommitted update was made.

10. Data processing apparatus of claim 7, in which:

a recovery point is calculated when a set of TURs has been written to the log.

11. Data processing apparatus of claim 1, in which:

the recovery point corresponding to a backup copy is calculated at the time that backup is taken.

12. Data processing apparatus of claim 1, in which the recovery point associated with the backup copy is equal to the start time of the oldest uncommitted updated at the time the backup copy was made.

13. A method of operation of a data processor, comprising the steps of:

processing stored data as a series of logical units of work (LUWs), with updates to the stored data made during each LUW being committed at a commit point for that LUW;

maintaining a log of updates made to the stored data;

detecting the oldest uncommitted update made to the stored data;

calculating a recovery point dependent upon the start time of the oldest uncommitted update;

taking a backup copy of at least a portion of the stored data while that portion is available for update; and

when necessary, recovering the database by reinstating the backup copy to provide a working copy and applying the updates contained in any log entries made after said recovery point.

14. A transaction management system having

means for processing stored data as a series of logical units of work (LUWs), with updates to the stored data made during each LUW being committed at a commit point for that LUW;

means for maintaining a log of updates made to the stored data;

means for detecting the oldest uncommitted update made to the stored data;

means responsive to the detecting means for calculating a recovery point;

means for taking a backup copy of at least a portion of the stored data while that portion is available for update; and

means for recovering the database by reinstating the backup to provide a working copy and applying the updates contained in any log entries made after said recovery point.

15. A transaction manager, for cooperating with a data processing apparatus, to provide means for

processing stored data as a series of logical units of work (LUWs), with updates to the stored data made during each LUW being committed at a commit point for that LUW;

maintaining a log of updates made to the stored data;

detecting the oldest uncommitted update made to the stored data;

calculating a recovery point dependent upon the start time of the oldest uncommitted update;

taking a backup copy of at least a portion of the stored data while that portion is available for update; and

when necessary, recovering the database by reinstating the backup copy and applying the updates contained in any log entries made after said recovery point.

16. Apparatus cooperable with data processing apparatus to provide means for

processing stored data as a series of logical units of work (LUWs), with updates to the stored data made during each LUW being committed at a commit point for that LUW;

maintaining a log of updates made to the stored data;

detecting the oldest uncommitted update made to the stored data;

calculating a recovery point dependent upon the start time of the oldest uncommitted update;

taking a backup copy of at least a portion of the stored data while that portion is available for update; and

when necessary, recovering tile database by reinstating the backup copy to provide a working copy and applying the updates contained in any log entries made after said recovery point.