Method of mirroring log datasets using both log file data and live log data including gaps between the two data logs

• US 6,178,427 B1
• Filed: 05/07/1998
• Issued: 01/23/2001
• Est. Priority Date: 05/07/1998
• Status: Expired due to Term

• Current Assignee: CA, Inc. (d/b/a CA Technologies) (Broadcom, Inc.)
• Original Assignee: PLATINUM Technology IP, Inc. (Broadcom, Inc.)
• Inventors: Parker, Christopher F.
• Primary Examiner(s): Black, Thomas G.
• Assistant Examiner(s): DO, THUY T
• Application Number:
  US09/074,080
• Time in Patent Office: 992 Days
• Field of Search: 707/202, 707/204, 707/203
• US Class Current: 1/1
• CPC Class Codes: G06F 11/2066 Optimisation of the communication load
• CPC Class Codes: G06F 11/2071 using a plurality of controllers
• CPC Class Codes: Y10S 707/99953 Recoverability

Abstract

A method of mirroring log data in real time. The method can be used to mirror all log data from a source database or only that log data pertaining to a subset of objects in the source database. The log data is written to one or more datasets which are created and stored in a target database system which is typically remotely located. The log datasets in the target system can be used for disaster recovery of the source system. The method monitors log data from the source system and automatically updates target log datasets accordingly.

Claims

1. A method of mirroring log data for a database wherein the log data includes live log data and log file data contained in a log file, comprising the steps of:

receiving log file data between a starting log point and an ending log point;

writing the log file data into a dataset;

receiving live log data;

determining if there is a gap between the live log data and the log file data;

performing a gap recovery if there is a gap, the gap recovery including receiving gap log file data and writing the gap log file data into the dataset; and

writing the live log data into the dataset.

2. The method of claim 1, wherein the steps of writing log data into the dataset include determining whether the dataset is full and if the dataset is full, writing the log data into a new dataset.

3. The method of claim 2, wherein determining whether the dataset is full includes detecting an I/O error.

4. The method of claim 2, wherein determining whether the dataset is full includes counting a number of log records written into the dataset.

5. A method of mirroring log data for a database, comprising:

applying log file data between a starting log point and an ending log point into a dataset;

receiving live log data continuously;

determining whether log data gap exists between the live log data and the log file data;

buffering data in the live log data if the log data gap exists;

applying gap log file data into the dataset if the log data gap exists; and

applying the buffered data into the dataset.

6. A method of mirroring log data for a database as claimed in claim 5, further including:

applying the live log data.