Release 10.1A: OpenEdge Getting Started:
Database Essentials

Modifying applications

Changes to applications require careful planning to reduce interruptions to users. Although there might be a process to test application changes at your site, database administrators should consider it there responsibility to verify expected application changes. The most effective way to do this testing is to have a test copy of your database that is an exact image of what you have in production and a thorough test plan that involves user participation.

Making schema changes

Schema changes can take hours to apply if they are not done properly. If the developers tested the application of schema changes against a small database, they might not have noticed a potential problem. A small database could apply an inefficient schema update in a short period of time and would not raise any red flags. If you have a full-size test environment, you can apply the schema change and know approximately how long it will take to complete. It is important to understand how long this process takes, since the users of the application will be locked out of parts of your system during the schema update.

Making application code changes

The amount of time it takes to apply application code changes can be greatly reduced by an advance compilation of your code against a CRC-compatible copy of your production database. To maintain CRC compatibility, start by creating a basic database, which is one that contains no data—only schema definitions. Use the basic database to seed a production database and a development database. The basic database is also saved, so you will have three copies of your database. If you already have a production database in place, the basic database is obtained by dumping the schema from that database.

As development occurs on a test copy of the database, the production and basic databases remain unmodified. When you are ready to promote your schema changes from development to production, first make an incremental data definition dump from the OpenEdge Data Dictionary by comparing the development schema with the basic database. The incremental data definitions can be applied to the basic database, and you can compile your application against that database. Second, the incremental data definitions can be applied at a convenient time on the production database (after appropriate testing). While the incremental data definitions are being applied, you can move the r-code you created against the basic database into place avoiding additional downtime to compile the application code.