|HOWL - Documentation|
Two sets of javadocs are available for the HOWL logger classes.
The HOWL logger exposes a limited set of public classes and methods supporting logging functionality. Programmers who are only interested in using the public interface of the HOWL logger should review the public API.
Most of the HOWL logger classes have package visibility. The javadocs for these classes provide implementation details and design notes that are of use to HOWL developers. Programmers interested in reviewing the documentation for all HOWL logger classes can review the complete project Javadocs.
JUnit test reports for the latest tagged version are available online.
CVS Difference Reports
A complete list of differences between the HOWL_1_0_1 TAG and the HOWL_1_0_2 TAG.
A complete list of differences between the HOWL_1_0_0 TAG and the HOWL_1_0_1 TAG.
A complete list of differences between the HOWL_0_1_11 TAG and the HOWL_1_0_0 TAG.
During restart, HOWL locates the active journal file then reads through the entire file looking for the logical end of the journal. The logical end is either the physical end, or the if the journal has been reused, the block with the largest block number. When a journal file is reused, the new block numbers are larger than the original block numbers. The logical end of the journal occurs where a block with a smaller block number than the previous block occurs. If HOWL is configured with very large journal files it may take a minute or two to process through the entire file. Given Murphy's law, the restart point will always be near the end of the file, so restart times will be noticeable if using large files.
If an application requires a large amount of journal space, rather than configure extremely large files, it is better to configure several smaller journal files.
|Copyright © 1999-2005, ObjectWeb Consortium | contact | webmaster | Last modified at 2007-04-02 08:38 AM|