Method and apparatus for logging journal data using a log write ahead data set

• US 4,507,751 A
• Filed: 06/21/1982
• Issued: 03/26/1985
• Est. Priority Date: 06/21/1982
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Gawlick, Dieter, Gray, James N., Obermarck, Ronald L., Iimura, Wallace M.
• Primary Examiner(s): Zache, Raulfe B.
• Application Number:
  US06/390,485
• Time in Patent Office: 1,009 Days
• Field of Search: 364/200 MS File, 364/900 MS File, 364/300
• US Class Current: 1/1
• CPC Class Codes: G06F 11/1471 involving logging of persistent data for recovery
• CPC Class Codes: G06F 16/00 Information retrieval; Database structures therefor; File system structures therefor
• CPC Class Codes: G06F 9/44 Arrangements for executing specific programs
• CPC Class Codes: G06F 9/463 Program control block organisation
• CPC Class Codes: Y10S 707/99953 Recoverability

Abstract

The invention provides a computing apparatus and method for maintaining a journal log. The computing apparatus includes volatile storage for storing a log buffer and a non-volatile storage for storing a journal log. Non-volatile storage means are provided for storing in a write-ahead data set a plurality of short data blocks. The log buffer contents are written to the write-ahead data set responsive to a process epoch occurring before the log buffer is filled. The log buffer contents are written to the journal log upon the log buffer being filled. The redoing or undoing of data base changes is made with reference to the write ahead data set only in the case of a system failure resulting in loss of log buffer data not yet written to the journal log; otherwise data base changes are redone or undone with reference to the log buffer or journal log.

Claims

1. In a system in which data base references are first recorded in a buffer and upon said buffer being filled, transferring its contents to a first log of references and bounding synchronization points of the first (begin transaction) or second types (commit or rollback transaction) stored together with the data base in a nonvolatile memory, a method for data base recovery in the event of system failure, comprising the steps of:

(a) formatting a second log in the memory;

(b) writing data base references into the buffer;

(c) monitoring buffer capacity and the occurrence of synchronization points;

(d) transferring any accumulated buffer contents to the second log when a synchronization point occurs prior to the buffer being filled; and

(e) in the event the data base references in the buffer are lost during system failure, redoing or undoing any changes to the data base utilizing the second log, otherwise redoing or undoing any changes utilizing the first log.

2. A method according to claim 1, comprising the further step of:

(f) transferring the buffer contents to either the first or second log prior to the writing of any changes to the data base.

3. A method according to claim 1, wherein the step of redoing any data base changes includes reincorporating all changes to the data base starting from a known prior information state (checkpoint) using the first and second logs as a trace, said redoing step being invoked whenever a synchronization point of the second type concludes a sequence of references in the interval between the occurrence of the known prior information state and the system failure.

4. A method according to claim 1, wherein the step of undoing any data base changes includes returning the affected data base portion to a state just prior to the advent of a synchronization point of the first type starting from the state at the time of the system failure using the first and second logs as a trace to the synchronization point, said undoing step being invoked whenever a synchronization point of the first type initiates a sequence of references and there not being any termination of said sequence prior to the occurrence of the system failure.

5. In a system in which transactions denominating a sequence of data base references bounded by synchronization points of a first (begin transaction) and a second (commit or rollback transaction) type are either completed or aborted, the data base references being first recorded in a buffer and upon said buffer being filled transferring said references to a first log stored together with the data base in a nonvolatile memory, a method for ensuring the redoing and undoing of any changes to the data base in the event a system failure occurs, comprising the steps of:

(a) formatting a second log in the memory;

(b) writing data base references into the buffer;

(c) monitoring the capacity of the buffer and the occurrence of synchronization points;

(d) transferring any accumulated buffer references to the second log strobed only by the occurrence of synchronization points; and

(e) given the unavailability of the buffer contents, redoing those transactions having synchronization points of the second type occurring in the interval between a known prior information state of the data base and the system failure and undoing those transactions having synchronization points of the first type but not of the second type occurring in said interval.