The system I inherited has hundreds of files that aren’t being used. I’ve keep a spreadsheet of the adabas commands issued, and these files have not had a command issued against them for 8 months. I want to delete these files, but of course I need to be able to restore them just in case.
My plan:
Run an adarep and save it.
Do an adauld.
Make a copy of the tape.
Restore the file to a test database to verify it can be done. Use the adarep for the NAME, UISIZE, NISIZE, DSSIZE, MAXISN etc parameters.
Especially when you need to restore the file in a couple of years with a different Adabas version
The safest option would be to go one step further and DECOMPRESS as well.
Remember to save the ADAREP on the tape too. Also the JCL you used to load them again. Include ADACMP cards if you decompress. Wait. There’s more. At least a text listing of the DDMs, if not a NATUNLD of the DDMs and/or unload of Predict DDEs. Text listing is good especially when you need to restore the files in a couple of years with a different Natural and/or Predict version .
I know its a little late for anymore tidbits, but one thing to be wary of on inherited systems is some of those files might be only used at year-end. And I know in some of the places I’ve worked many of the files were required by law to be kept for a certain number of years.
Your plan sounds fine. You can shorten the decompress process by using the INFILE= parameter of the decompress and take the file straight from the database to a decompressed dataset on tape (thus skipping the unload step).
I would also highly recommend unloading the predict objects to tape and keeping them all together. The adarep will only give you the short names. If you are really going to clean things up, you might as well archive the metadata and get it out of your active data dictionary.
Also, make a dual copy of these so you can keep on on site and one in your off site storage for any possible disaster recovery scenarios.
