landay@cory.Berkeley.EDU (James A. Landay) (08/16/88)
This evening I was trying to speed up my hard drive by unfragmenting it.
I used Simon Poole's REGORG (v. .04) First I did partition C. No Problem
(I had FOLDER999.PRG in case of 40 folder problems) I then rebooted.
Did partition D. After 10 mins. or so, strange things started to happen.
The amount done indicator reached the end of the space for it and kept
going. Went past the end of the dialog box. Kept going. To the edge of
the screen. CRASSSSSSSSSSSSH!!!!!!!!! BOMBSSSSSSSSS!!!
Rebooted! Partition D was fried!!! Luckily I had backed up with Turtle
the night before (great program George!) I used the BMS utility to
zero out the partition and proceeded to restore partition D (how are you
supposed to restore from Turtle backed up disks? I had to use some program
run from Gulam that someone posted a while back called RESTORE.)
I don't know where I screwed up, but I ended up with an directory with a
blank name that when entered showed everything in the root. (and the same if
I entered the blank one in that directory) Seemed like a circular pointer
problem. I used a disk editor and some docs someone posted on the floppy
FATS a long time ago and wiped out that entry. It seems to work. Is that
the proper way to go? (I couldn't get any programs to fix it, DLII or Disk
Doctor.)
Is there a new (safe) version of REORG?
Thanks,
D
D
D
D
D
D
James A. Landay
ARPA: landay@cory.berkeley.edu
..!ucbvax!cory!landaypoole@forty2.UUCP (Simon Poole) (08/19/88)
In article <5037@pasteur.Berkeley.EDU> landay@cory.Berkeley.EDU (James A. Landay) writes: >This evening I was trying to speed up my hard drive by unfragmenting it. >I used Simon Poole's REGORG (v. .04) First I did partition C. No Problem >(I had FOLDER999.PRG in case of 40 folder problems) I then rebooted. >Did partition D. After 10 mins. or so, strange things started to happen. >The amount done indicator reached the end of the space for it and kept >going. Went past the end of the dialog box. Kept going. To the edge of >the screen. CRASSSSSSSSSSSSH!!!!!!!!! BOMBSSSSSSSSS!!! > >Rebooted! Partition D was fried!!! Luckily I had backed up with Turtle >the night before (great program George!) I used the BMS utility to ...... >Is there a new (safe) version of REORG? First a comment on the specific problem: (skip this part if you are not interested in this, more about the status of DL II after it) ----------------------------------------------------------------- Here's the relevant part from the (never finished) DL II Manual: - Appendix A: ReOrg, a disk reorganizer - -IMPORTANT: REORG IS NOT A STANDALONE PROGRAM! YOU NEED AT LEAST A -IMAGE BACKUP PROGRAM AND A DISKCHECK PROGRAM BEFORE YOU CAN USE -IT! - -IMPORTANT: ALWAYS, ALWAYS MAKE AN IMAGE BACKUP OF YOUR DISK -BEFORE RUNNING REORG ON IT! - -ReOrg improves disk performance on harddisk and floppys through -better placement of subdirectories and files on disk and through -defragmentation of the files themselves, it will also remove -deleted and zero length entries. - -As the actual reorganizing is NOT an incremental process, during -the time from the start of the reorganization to the end of it, -your disk is in a unusable condition (at least from the viewpoint -of the operating system). This means, that a power failure, -computer malfunction, disk error or program failure during the -reorganization will leave you with a corrupted disk! - -To run ReOrg: backup your disk, run the diskcheck program on it -(DL II will work ok) and ONLY run ReOrg after all crosslinked -and lost clusters have been disposed of! ReOrg has NO protection -against incorrect disks and will probably run into a infinite -loop if you do not take these precautions! - -After you have run ReOrg and reorganized your disk, the computer -will reboot, this is normal and serves the purpose of -reinitializing the TOS internal directory tree. [A file by file backup will do, it's just not as quick] There is per definition no 'safe' version of this kind of program. ------------------------------------------------------------------ Second Part: Now and then people ask me about updates to two programs (DLII and ReOrg) I wrote in 86/87 and had in Beta test for a few months. The versions were never released officialy in any way to the public (I did allow a bit wider distribution in one case), and if people would actually read copyright notices, it probably wouldn't have gone unnoticed that there is no mention of free copying etc...... I more or less decided a few months ago never to distribute these programs in an enhanced and bug-fixed form (something my personal versions are). At the time I had two options: - produce a very much enhanced version as a commercial product: I decided not to do this, as it would have ment another 6 Months or so of extra developement and programming (6 months being optimistic), with rather unsure financial rewards (not to mention that working on a project also means an initial investment (you still have to live on something) of $$$$$$ which I don't have). - distribute it under the same terms as UniTerm: which would have ment: - finishing the manual (~1 months work) (this naturally also translates into $$$$$) - support ~1h/day for the next two years (extrapolation from time I spend on UniTerm plus questions about the beta test version of DL II (you would all be suprised how many buggy ramdisks are out there.....)) - near zero income As neither of there possibilities are really viable, I just decided to to NOTHING. -- ---------------------------------------------------------------------------- UUCP: ...mcvax!cernvax!forty2!poole Simon Poole BITNET: K538915@CZHRZU1A ----------------------------------------------------------------------------
landay@cory.Berkeley.EDU (James A. Landay) (08/21/88)
In article <441@forty2.UUCP> poole@forty2.UUCP (Simon Poole) writes: >Now and then people ask me about updates to two programs (DLII and >ReOrg) I wrote in 86/87 and had in Beta test for a few months. The >versions were never released officialy in any way to the public (I >did allow a bit wider distribution in one case), and if people >would actually read copyright notices, it probably wouldn't have >gone unnoticed that there is no mention of free copying etc...... > >I more or less decided a few months ago never to distribute these >programs in an enhanced and bug-fixed form (something my personal >versions are). Maybe these programs should be taken off the file servers (where I got it) that store st binaries. Thank You for the help James A. Landay ARPA: landay@cory.berkeley.edu ..!ucbvax!cory!landay
Thomas_E_Zerucha@cup.portal.com (08/22/88)
Simon Poole writes> "There is no safe version of this kind of program" This is ultimately true, but then again, you could have a random bit error tell your hard disk controller to do a format. There is a commercial alternative - in fact I wrote it, so I stand to profit from this, so the information may contain some of these biases. The program is "Hard Disk Sentry" and is available from Beckemeyer Development Tools. It is as safe as any such utility can be. First it performs a thourough check of the disk structure, and if you select optimize (It is still wise to back things up), it will do it by maintaining a correct backup image of each file, so the worst that will happen is that you will have a bunch of clusters which don't belong to any file - i.e. it first copies the images of the files to an unused portion of the disk, then it links these clusters, then it finally unlinks the original area. You can interrupt such a process anytime without losing data, unless lightning strikes your computer. And any potential errors are easily recovered from using the repair feature. I tried more than a few times hitting reset right in the middle of the operation and haven't lost a byte of data (again, the worst thing were the leftover backup images which the repair would collect into one folder which I simply deleted). *Don't try this at home*. In short, I did most of the extra work to create a "safe" repair/optimize utility. But it is commercial and you should contact Beckemeyer if you are interested in a copy. (Disclaimer: I have tested it extensively, but it can have problems and create trashed disks if your hardware is marginal and writing a sector from any program may not work, you cannot expect my program to recover from such instances. Also, I use it with all kinds of TSRs and Accessories including Caches, although I specifically warn against it since I can't predict what any given program will do in combination with mine. I am not claiming that it will *never* trash any disk under any circumstances, just that I have taken every reasonable precaution to prevent it in the algorithms I chose. You still should backup).
ripley@netmbx.UUCP (Hans-Ch. Eckert) (08/25/88)
[ The REORG / DL_II Dilemma ] What about the Norton's Speed-Disk algorithm ? I've seen it on one of those with three big and blue letters on it :-) and it works VERY fine and SAFE. Nobody there, capable of managing this on the ST ? (What about asking P.Norton to port all his utilities to the ST ? As the OS's are nearly the same, and I think the programs are not all written in 100% machine-code, it should be possible...) BTW, what I DO need is a disk-editor (-doctor), which is capable of handling with read-errors and non-standard formatted disks (e.g. fat formats) Any suggestions ? -- == Greetings from : == Hans-Ch.Eckert, Regensburger Str. 2 == == *** ripley@netmbx.UUCP *** == 1000 Berlin 30 [Tel: 030/246292] == ============================================================================== == Superkalifragilistikexpialigetisch ! (Mary P.) ==
poole@forty2.UUCP (Simon Poole) (08/30/88)
In article <5157@pasteur.Berkeley.EDU> landay@cory.Berkeley.EDU.UUCP (James A. Landay) writes: [refering to DLII and ReOrg] > >Maybe these programs should be taken off the file servers (where I got >it) that store st binaries. > I don't really care, it's just a bit much when people expect me to support these programs.... -- ---------------------------------------------------------------------------- UUCP: ...mcvax!cernvax!forty2!poole Simon Poole BITNET: K538915@CZHRZU1A ----------------------------------------------------------------------------
poole@forty2.UUCP (Simon Poole) (08/30/88)
In article <8364@cup.portal.com> Thomas_E_Zerucha@cup.portal.com writes: >Simon Poole writes> >"There is no safe version of this kind of program" [describes the way his program works, claims it is safe, but still want's you to make backups :-)] Just that nobody gets a wrong impression, ReOrg works quite well, and is as safe as possible, IF you follow the instructions (then it is as safe as your backup program). While it would have been trivial to add all the fixing and checking code to ReOrg, I've got something against kitchen-sink programs. ReOrg is supposed to be a TOOL to be used together with other programs: "Power tools are not toys!" BTW, while your program sounds quite nice, it would fail miserably on two of the four partitions on my HD (guess why...). -- ---------------------------------------------------------------------------- UUCP: ...mcvax!cernvax!forty2!poole Simon Poole BITNET: K538915@CZHRZU1A ----------------------------------------------------------------------------
Thomas_E_Zerucha@cup.portal.com (09/01/88)
Simon Poole writes: >while your program sounds quite nice, it would fail miserably on two of the >four partitions on my hard drive (guess why?) If they are some kind of partitions for a Mac emulator or Idris or OS/9, m program wouldn't recognize them at all. There is a HD optimizer for the Mac, (Symantic Utilities) but I haven't tried it. There are many for the PC, so one that runs under PC-Ditto should work too. I don't know about anything for Idris or OS/9 or any other type. If you check out most of the optimizers for the PC (those that I know about), no matter how "safe" they are they recommend backing up first and rebooting afterword, just because no matter how "safe" my program might be, I can't be sure that someone else has figured out a way to compromise the safety, for example a non-writethrough cache would wreak havoc if my program aborted aborted before it had a chance to update. Also if directories happened to be cached and it didn't recognize a media change or something and rewrote the old information over the reorg'ed FAT's, it would also trash the disk. As for my utility, I *HAVE* hit reset right in the middle of the optimization without data loss, and the worst that can *theoretically* happen is extra :"orphaned" files which I have also gotten when hitting reset. It is designed to be this robust and foolproof, but fools can be even more clever. We also have something in this country called tort liability which I have never had the chance to be part of, and never want to, but I have to be careful with my claims simply because of this. Even if it is 99.98% reliable, I can't take a chance on not disclaiming all this stuff.
lbl@druhi.ATT.COM (Barry Locklear) (09/03/88)
In article <8627@cup.portal.com>, Thomas_E_Zerucha@cup.portal.com writes: > Simon Poole writes: > >while your program sounds quite nice, it would fail miserably on two of the > >four partitions on my hard drive (guess why?) > > If they are some kind of partitions for a Mac emulator or Idris or OS/9, m > program wouldn't recognize them at all. I think what Simon meant was that two of his partitions are either full or contain files larger than the free space on the partition. If your program works like I think it does, then this would cause it to fail. Otherwise, it sounds great. Barry Locklear
Thomas_E_Zerucha@cup.portal.com (09/06/88)
> "If it works like I think it does..."
It doesn't. It can handle anything but nearly completely full disks (I set
a limit of 8K free space minimum when it will inform you that it can't
optimize). I had no trouble with a 2.5 MEGABYTE file with only 250K left
on the partition. It also rearranges the directories and sets up files
for PC Ditto bootability (without using the MS-DOS format command).
Hard Disk Sentry was designed to handle nearly everything, and on the things
it couldn't, to not lose any data. Large files aren't a problem. Even
fairly full partitions aren't, as long as there is some free space (and
moving one file is usually enough to create adequate space). It isn't
blindingly fast, especially with little freespace to play with (since it
is copying every byte on the hard disk twice, and adjusting directories
and FAT's as it goes - a *lot* of thrashing), but that is because I chose
safety and functionality over speed.