GJS: KB0009

How do I Run 'Validate' on a Production Machine?

Applies to: RE/data

The following steps can be used to run the validate process on a production database which is inaccessible from the environment where RE/data has been installed.

Note: As with any procedure which involves several steps, some of which are not easily reversible, we recommend that backup copies of your database are made regularly and before each of the key steps in this procedure.

Method

Make a copy of your current raw dictionary using the Raw /Copy menu options in RE/data.
Perform a global save of the copy of the raw dictionary by saving the following globals:
- ^realias (if you have any alias definitions)
- ^redd(rawname) where 'rawname' is the name of the copy of the dictionary
- ^redd2(rawname)
Delete the copy of the raw dictionary using the Raw /Delete Whole menu option in RE/data.
Install a fresh copy of RE/data on your production machine. (Best performance will be obtained if you install it in the same UCI /namespace as the data you want to validate against, but make sure that the RE/data routine and global names do not conflict with any already in use in that area). The following routines and globals need to be installed, as a minimum:
- Routines: RE* and re*
- Globals: ^remnu, ^reopt, ^rehelp, ^redev, ^reio
Restore the global save of the raw dictionary into the RE/data area on the production system.
Perform the validation runs on the production system.
On completion of the validation runs, perform a global save of the following globals from the RE/data area on the production system:
- ^redd(rawname)
Restore the global save of the validated raw dictionary into the RE/data area on the original system.
Merge the validated raw dictionary back into the original raw dictionary using the Raw /Copy menu option in RE/data. This updates the original raw dictionary with the results of the validation process performed on the production machine.

Last reviewed: Jan 20, 2005