I've currently got an java application which uses Quartz to schedule jobs. The jobs typically go to various regional databases and populate another consolidated database to get a consolidated view of the data. It also schedules routine tasks such as nightly batch activities and archiving.
The configuration data i.e job, trigger and database connection details are held in an XML config file and read by the java process.
However, today a problem occurred. A member of the team had accidental coped the production XML config file to UAT and changed some of the regional DB values to UAT values, but not the consolidated DB connection details. As a result we had erroneous UAT data in production and some details were duplicated (one from prod and one from UAT. As you can imagine this was a nightmare!
As a result I've proposed getting the DB connection details out of the xml config file and putting it in a unix vi properties file with property value pairs. This will split the job / triggers and connection details. The property file name would be propertyfile.properties.prod and propertyfile.properties.uat and the correct file read in depending on property set in the unix user .profile file.
This seems a bit complex. I was just wondering if anyone had similar problems and how the resolved it / if anyone has any advice to help create a fool-proof solution.
Your help ould be much appreciated