When building a load module with exit routines and dynamic tables, you must decide whether you want to support dynamically loading or deleting these modules after initialization. This is especially useful in a test environment where new versions of a failing exit can be activated without a restart. Depending on the processing done in your exits, supporting dynamic loading and deleting might require no additional code or only a reorganization of your existing logic.

Making load modules dynamically loadable will increase the amount of testing you need to do. You need to not only test the function implemented by your modules, but also ensure that everything works after the module is dynamically loaded, refreshed or deleted. The advantage of dynamic load modules is that when you find a problem in your module, you can correct the problem and get a new version of your code running without major disruptions. If the problem is bad enough, you can delete the module, fix it, and load the fixed version.

If the code is tested and placed in a production environment, IBM suggests that you do not make loading, refreshing, or deleting load modules as a part of your normal operations. This is because it is not always possible to delete old modules from storage. JES2 only deletes a module from storage if it will surely not be used. Some of these old modules will take up space until JES2 or z/OS® is restarted (depending on where the module is loaded). Loading, refreshing, and deleting load modules in a production environment must be reserved for emergency situations (if it would save or delay an IPL) or for modules that IBM service has provided to collect additional diagnostic information. If a production environment needs to be altered on a regular basis, it would be better to alter the list of routines associated with an exit point rather than altering what modules are loaded.