frankb@usource.UUCP (Frank Bicknell) (05/26/89)
In article <241@vector.Dallas.TX.US> chip@vector.Dallas.TX.US (Chip Rosenthal) writes: > In article <752@mccc.UUCP> pjh@mccc.UUCP (Pete Holsberg) writes: > > [...cte] we were both wondering how to make backups of the > > distribution floppy disks? Can you use DISKCOPY or are they > > *NIX-formatted disks? [etc...] > > There is no such thing as a "unix-formatted" disk. > [much interesting information about formatting deleted] Your statements are true, but there are little details about the usage of that formatting which differ from unix to unix. It depends on what unix you're talking about. Some unixes, such as AT&T (3b2,3b1) do something weird with the first track or two. While there was no difference in the formatting of those tracks, they decided to skip them and reserve them for booting even when you are placing a tar image on the disk! AT&T are not the only folks to do this: I've run across a BSD-based Unix which does the same sort of thing. BTW... it's common to find 640K (96ds8) on other unix floppies. We are fortunate that the kind folks at SCO decided to not do such monkey business and chose the straightforward approach when using the disk for a non-bootable floppy. They were even forward-thinking enough to allow almost complete flexibility with the floppy drive (8,9,15 spt, 48,96,135 tpi, even ds,ss! Anyone still use single-sided floppy disks? :) ). So, between using the flexibility of the floppy device driver and the ability to skip tracks/sectors/whatever with the 'seek/skip' options of 'dd', I've found I've been able to read other unix's floppies very nicely with SCO. Unfortunately, the reverse is not true: I don't know how to read a 48ds9 (360K) floppy on one of those other machines, for example, and it's mighty inconvenient (though I presume possible using the raw device and 'dd') to create or read a tar disk which begins at track 0. And now a question: why is it that floppy operations are so incredibly slow under unix. SCO and AT&T are both bad about this. Do something on the floppy drive and you may as else go to lunch while it's chugging along. A 3b2 (older models, at least) takes about two minutes to read or write one 80-track floppy!!! That's about 1.5 seconds a track. It's academic, of course... I haven't used the floppy as the swap device in some time now... -- Frank Bicknell UniSource; 1405 Main St, Ste 709; Sarasota, FL 34236 killer!usource!frankb || frankb@usource.UUCP
mhlevy@sbee.sunysb.edu (Mark Levy) (05/31/89)
> > And now a question: why is it that floppy operations are so > incredibly slow under unix. SCO and AT&T are both bad about > this. Do something on the floppy drive and you may as else go > to lunch while it's chugging along. A 3b2 (older models, at > least) takes about two minutes to read or write one 80-track > floppy!!! That's about 1.5 seconds a track. It's academic, of > course... I haven't used the floppy as the swap device in some > time now... If you read the manual, you will find that in the later versions of SCO XENIX, you can specify the interleave. A 1:1 interleave speeds things up tremendously, if you controller can handle it. Sorry, but I'm not sure when that feature was implemented, and I don't have my manual handy right now to give you the switch setting. Mark -- ~~~~~~~~~ Disclaimer??? We don't need no stinkin' disclaimer!!! ~~~~~~ Mark Levy { mhlevy@sbccvm.BITNET } % Debbie Gibson is pregnant, with { mhlevy@sbee.sunysb.edu } % my two headed love child.... { mhlevy@ccvm.sunysb.edu } % Mojo Nixon