Version 3.0

Removed MongoDB option. Removed database reload from filenames capability. Changed SQL database from embedded Derby to external Postgress. Added support for Tomcat database connection pool.

Version 2.60

Upgrade dependencies and move to Java 21.

Version 2.50

Remove ability to delete data.

Version 2.40

Major change to update all dependencies to most recent versions. New software for handling file uploads.

Version 2.35

Update dependent libraries to fix issue with log4j vulnerability.

Version 2.34

Change from DIGEST to BASIC authentication.

Version 2.33

Fix a bug in the cache trimming thread that caused folders to be left open, eventually causing the Tomcat process to run out of file handles.

Version 2.32

Update UNIX helper scripts to fix problems with syntax of the 'tr' command. Automate the deployment of the application (internal change). Inclusion of ability to use a Mongo backend database (in addition to existing SQL capability).

Version 2.09

Fix a bug that prevents CDF files being uploaded correctly on UNIX servers.

Version 2.08

Added the ability to handle data in the new IMCDF format (INTERMAGNET CDF - see Discussion Document 22). Updated the GINCache client program and library (which can be downloaded from the help pages).

Version 2.07

Fixed bug that prevented cache status being displayed properly when the cache is empty.

Version 2.06

Various bug fixes, including creating cache directories properly and formatting a delete request correctly in plain text view.

Version 2.0

Internal re-write of the software to address problems with database that stores metadata.

Version 1.51

Added a configuration option that allows the thread that deletes files from the cache to be switched off - this is the default option. This is to remove the problems associated with the deletion thread while the database issues (slow response) are investigated.

Version 1.5

Changes made to the internal cache file structure. Files are now stored in batches of 1000 in subdirectories of the main cache directory. The subdirectory name is the file's ID number divided by 1000. Subdirectories are automatically created and destroyed. This was implemented to try to help with problems associated with having tens of thousands of files in a single directory.

Version 1.4

Fixed bugs to do with trimming the cache to the configured maximum size. Added an option to allow trimming the cache to be run manually (it also runs automatically once an hour). Changed the default cache size from 20Gb to 5Gb.

Version 1.3

Support for quasi-definitive data added.

Version 1.2

Check for type of incoming files removed, because IMF and IAGA files put the observatory name as the file type, which some browsers then convert to an application type (e.g. .mpt ==> applcation/vnd.ms-project).

Version 1.1

Metadata moved to a JDBC compliant database - previously held as an in-memory array, but the size grew too large. System tested using an embedded Derby database. See InstallConfigInstructions.txt for more details.

Option added to allow deletion of data. This is achieved by passing a command through to the GIN software that runs the GIN's deletion programme. This option can only be used by an administrative user.

The process of deleting cache files had been moved to a background thread. The cache size is checked once an hour and the oldest files deleted until the cache size is within the user specified limits. A second background thread allows cache metadata to be reloaded from the cache. The reload process is started by an 'admin' user from the Administration page. The Administration page shows details of the last deletion process as well as progress of the metadata reload.

Rules preventing data being uploaded have been relaxed. Specifically you may now upload IAGA-2002 variation data and duplicate data files.

Version 1.0

Initial release