Method and a system for processing a log record

• US 5,638,508 A
• Filed: 03/09/1994
• Issued: 06/10/1997
• Est. Priority Date: 07/17/1987
• Status: Expired due to Fees

• Current Assignee: Hitachi, Ltd., Hitachi Microcomputer Engineering Limited
• Original Assignee: Hitachi, Ltd., Hitachi Microcomputer Engineering Limited
• Inventors: Sumiyoshi, Takashi, Kanai, Sadasaburoh, Kitajima, Hiroyuki, Tsuboi, Toshiaki
• Primary Examiner(s): Chan, Eddie P.
• Assistant Examiner(s): Ellis, Kevin L.
• Application Number:
  US08/208,044
• Time in Patent Office: 1,189 Days
• Field of Search: 395/425, 395/600, 395/575, 395/182.18, 395/182.13, 395/440, 395/446, 395/456, 395/460, 395/462, 395/486
• US Class Current: 714/20
• CPC Class Codes: G06F 11/1471 involving logging of persistent data for recovery

Abstract

A data processing system for processing transactions, wherein a log record to be used for recovery of the system is written into a log file for system recovery in synchronism with the end of a transaction, and log records other than resident information are is written for a plurality of transactions into a log file for archives.

Claims

1. A data processing method for storing a plurality of log records of transaction processing in a computer system, comprising the steps of:

classifying said log records into at least a first type of log record and a second type of log record based on uses for said log records, said first type of log record being used for recovery of said computer system, and said second type of log record being used for archival purposes;

storing said first type of log record into a first log buffer upon generation of said first type of log record in a transaction;

storing said second type of log record into a second log buffer upon generation of said second type of log record in said transaction;

storing said first type of log record in said first log buffer to a first storage medium, at least when said transaction ends; and

storing said second type of log record in said second log buffer to a second storage medium when said second log buffer is full, wherein size of said second log buffer is a predetermined times as large as an average amount of data for said second type of log record produced in one transaction.

2. A data processing method according to claim 1, wherein said first type of log record includes at least one of log records for updating a data base or a table in a main storage/virtual storage, a log record concerning an input/output message and a log record indicating completion of a transaction.

3. A data processing method according to claim 1, further comprising the steps of:

storing resident data from a main storage/virtual storage at a predetermined check point into a third storage medium of said computer system;

wherein a capacity of said first storage medium of said first type of log record is large enough for storing log records produced within an interval of the checkpoints and said first storage medium for storing said first type of log record is repeatedly used in a wrap-around manner; and

recovering said computer system by using said first type of log record from said first storage medium and said resident data from said third storage medium.

6. A data processing method according to claim 1, wherein said first storage medium for storing said first type of log record is a nonvolatile semiconductor storage.

7. A data processing method according to claim 1, wherein said second type of log record further includes at least one of a log record for database recovery, a log record for recording business processing contents and a log record to be used in user programs for business processings.

8. A data processing method according to claim 1, wherein accessing speed of said first storage medium is higher than accessing speed of said second storage medium.

4. A data processing system for storing a plurality of log records of transaction processing in a computer system, comprising:

means for classifying said log records into at least a first type of log record and a second type of log record based on uses for said log records, said first type of log record being used for recovery of said computer system, and said second type of log record being used for archival purposes;

means for storing said first type of log record into a first log buffer upon generation of said first type of log record in a transaction;

means for storing said second type of log record into a second log buffer upon generation of said second type of log record in said transaction;

means for storing said first type of log record in said first log buffer to a first storage medium, at least when said transaction ends; and

means for storing said second type of log record in said second log buffer to a second storage medium when said second log buffer is full, wherein size of said second log buffer is a predetermined times as large as an average amount of data for said second type of log record produced in one transaction.

5. A data processing system according to claim 4, further comprising:

third means for storing resident data from a main storage/virtual storage medium;

wherein a capacity of said first storage medium of said first type of log record is large enough for storing log records produced within an interval of the checkpoints and said first storage medium is repeatedly used in a wrap-around manner; and

fourth means for recovering said computer system by using said first type of log record stored in said first storage medium and said resident data stored in said third storage medium.

9. A data processing system according to claim 4, wherein accessing speed of said first storage medium is higher than accessing speed of said second storage medium.