This problem seems like it'd be well suited to a security-based approach. Do you have the z/OS Security Server (RACF)? Could you use RACF warning mode to get a handle on who's accessing (or not accessing) these fonts? I suppose you could even get slightly fancy and use some form of automation to shut off RACF authorization checking (in warning mode) for particular fonts once a warning is issued -- "Yep, that one's live," basically. You'd then take a look at logs to make your decisions about what's live and what isn't.
Of course you'd set this up so "everybody" would not be authorized to access the fonts you're screening for potential archiving, but RACF would allow the access to proceed with a warning. That's how warning mode works. If the user is authorized to access that particular resource, there's no warning, and in this case you do want a warning kicked out if there's any access. As far as I know RACF provides sufficient granularity to define access controlled resources for these purposes, so I think it'd work well. This technique is likely to offer comparatively low overhead, too, especially if you quickly inform RACF to stop warning, resource by resource, e.g. via automation. Or you can take a very slow, selective approach and tell RACF to control only one or a couple libraries, then see how that goes. It really depends on what you prefer and your particular situation. I'm using the word "resource" here because this same technique, if viable for you, should also work with more than just fonts. As someone else already suggested in this thread, fonts may not be the most "interesting" consumers of your DS6800 disk space. You might find all sorts of other cruft you can archive. If you (sadly) don't have the z/OS Security Server (RACF), then you can contact your friendly IBM representative to license it. Even for only a month, I suppose. Or your alternative security subsystem might be able to do something similar. This general class of problems (managing data growth) is full of automated and autonomic solution approaches nowadays. One of them is even called IBM Optim Data Growth Solution for z/OS, as it happens. However, as a general rule and in my view, you ought to have a "medium" or "large" data growth problem to solve before pulling out the more sophisticated, more powerful approaches. Many organizations do have such growth problems to solve, so I mention these other options for completeness. Did I come up with a good idea for you to explore? -------------------------------------------------------------------------------------------------------- Timothy Sipples VCT Architect Executive (Based in Singapore) E-Mail: [email protected] ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
