I think that's why I hadn't given this much thought prior to this -- I'm used to the idea of redeploying the same DASD and just formatting it once to erase previous guests data. So I agree there's a difference between doing this - where the customer has only logical access to the data - and where physical access is changing. When storage units leave a data center - 'then' I think multiple wipes make sense -- but outside of that, I'm not seeing it's applicability on a mainframe supplying a virtual Linux environment (maybe cleaning up after a DR exercise/realdeal if sensitive data was used).
Very interesting discussion - I had not understood (or thought very hard about) the physics and why multiple passes with different bit patterns would be done -- but now I get it. Thanks for educating a 'software' guy on the mechanics! Scott On Thu, Oct 8, 2009 at 8:29 AM, Thomas Kern <tlk_sysp...@yahoo.com> wrote: > In this discussion there should be a differentiation between reusing a DASD > allocation while still under your physical control (giving the mdisk to > another user), and relinquishing physical control (return to vendor during > an upgrade or selling it to 3rd-party or turning it over to the federal > excess list). > > I use one pass of format by DIRMAINT to clean an allocation before allowing > reuse withing my own installation. But it the DASD has to leave the > building, I use ICKDSF or FDR/Erase several times before it gets powered > off. And if the data was really sensitive, the platters get removed (fun > job > on real 3380s), shredded into confetti and then melted. > > /Tom Kern > /301-903-2211 > > >