Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!wuarchive!rex!doerschu From: doerschu@rex.cs.tulane.edu (David Doerschuk) Newsgroups: comp.sys.ibm.pc Subject: Re: puzzler Message-ID: <1659@rex.cs.tulane.edu> Date: 12 Dec 89 03:08:44 GMT References: <9214@hoptoad.uucp> <28301@amdcad.AMD.COM> <1989Dec11.221205.11140@sj.ate.slb.com> Reply-To: doerschu@rex.UUCP (David Doerschuk) Distribution: usa Organization: Computer Science Dept., Tulane Univ., New Orleans, LA Lines: 64 Summary: Expires: Sender: Followup-To: Keywords: >>|"No room for system on destination drive". All of my old system files >>|are still there on the hard disk, and it won't let me delete them prior >>|to trying to replace them: "Error: io.sys (or msdos.sys) is read only". > >The problem is not with removing the old files from the drive, but the fact >that these system files must occupy the first contiguous block on the disk. >Once user data is written to the disk, those blocks after the boot files will >become occupied, therefore leaving no 'contiguous' space big enough for larger >boot programs. > >Russ Poffenberger DOMAIN: poffen@sj.ate.slb.com Sorry, lost the original poster attribute. There is a fiddle for this if you own a copy of the Norton Utilites. In your root directory, create 5 or 6 directories with names beginning with: "AAAA". Then do Norton's Directory Sort (DS) on the root, and move those directories (empty) to the "top" of the root directory list. Now enter the main Norton Utility program (NU) and in the Choose Menu select cluster 2 and 3 and then select the "Information on item" choice. Norton tells you what file is occupying those clusters. As long as its one of the "AAAA" directories you created, you're fine. Go on checking clusters two at a time (4 and 5, 6 and 7, etc.) until you're up to about cluster 18 or so. If you find any clusters that contain files or directories other than your special "AAAA" ones, you must 1) exit Norton 2) make a temporary directory to hold the contents of the offending directory 3) transfer the contents of the offending directory to the new temporary directory 4) Make a note of the name of the offending directory (its now empty and will be deleted later). When you have all the clusters from 2 to 18 or so filled with either "AAAA" directories or empty "offending" directories, you're set. Here's the last step: Delete all the "AAAA" directories and empty "offending" directories. Now you've got a "hole" at the beginning of your disk drive that the sys command can fill with IBMBIOS.COM and IBMSYS.COM (not sure if those names are correct, but I mean the two hidden system files that DOS needs at the beginning of your drive.) Wasn't that easy? :-) You can get into additional trouble with the above method if one of the clusters you want empty has a file from a non-root directory in it: copy the file to a new temporary location but don't delete it until you're ready to delete all the "AAAA" and "offending" directories, otherwise one of your next "copy to a temporary location" moves will copy it into that cluster you just freed up (since DOS tries to use the smallest-numbered empty cluster first). Fun, eh? I've actually done this successfully, although it was probably not a particularly effective use of time! If you don't have the Norton Utilities (or similar), or don't feel like being a Disk Hero (tm), or feel that your time is worth more than $.23/hour, Plan B is: Back up the entire hard disk, then delete everything on it. Do NOT reformat it. Boot from a floppy, and "sys" the DOS files onto the hard disk. With an empty hard disk, you are guaranteed that the system files will land in the correct place. Re-load (I hesitate to use the word "restore") the hard disk from your backup floppies, and you're ready to go! BTW, I believe it is not actually essential that both IBMBIOS.COM and IBMSYS.COM (again, names?, not sure) occupy the first clusters on the disk in their entirety. I know that at least part of one of them MUST be in the first user clusters (it is cluster #2, isn't it?) but isn't it possible for the files to be non-contiguous after a certain number of clusters? Good Luck! Dave doerschu@rex.cs.tulane.edu