[comp.sys.amiga] What do you do about a trashed hard disk?

chris@genly.UUCP (Chris Hind Genly) (08/21/90)

My disk was just clobbered again.  I ran diskdoctor and was notified
of unreadable files, blocks used twice, and duplicate keys.  Yet in
the end the disk could not be validated.  If diskdoctor can't
fix the disk sufficiently so the disk can be validated, then whats
the point of diskdoctor?

I don't think the disk is as in bad a shape as diskdoctor says.  In
the past I have made a backup of the hard disk after it went bad,
reformatted and reloaded and everything was fine.

Is there anything I can do besides reformat?

Is there something better than diskdoctor?  I've heard that disksalv
is better.  But it doesn't work on hard disks.

Is there anyway to minimize the risk of trashing the disk when running
a program known to be unstable.  Perhaps:
  - Reduce the number of buffers
  - No other processes running and using the disk. 
  - Run out of the ram disk, so nothing will be flushed to the hard disk?

(I'm getting tired of reloading my disk.  The problem is caused
by a program I'm developing which occasionally runs away.  When I
first heard of the Amiga and its unprotected memory, I was horrified.
After a while I got used to the idea.  Its FAR better the MSDOS on
an IBM PC.  Its days like these that I lean towards the horrified
side again.)


       *                        *                        *
                                                            \|/
               *           _______                         --O--
                      ____/ KC1VP \____        *            /|\
 *      _____________/  (203) 389-8680 \_______________
 ______/  95 Fountain Terr., New Haven, CT, USA, 06515 \_______
/ Chris Hind Genly    chris@genly.uucp   uunet!hsi!genly!chris \
----------------------------------------------------------------

a218@mindlink.UUCP (Charlie Gibbs) (08/22/90)

In article <chris.5660@genly.UUCP> chris@genly.UUCP (Chris Hind Genly)
writes:

>Is there something better than diskdoctor?  I've heard that disksalv
>is better.  But it doesn't work on hard disks.

     Although DiskSalv can't fix up a disk in place like DiskDoctor
does (or claims to do :-) it can pull stuff from hard disks; it just
can't, of course, recover an entire hard disk to a floppy or RAM disk.
The FILE keyword lets you specify which files to recover.

     Another option is FixDisk, by Werner Gunther.  This one has
saved my bacon several times.  It can repair hard disk partitions
up to 48 megabytes in place, much like DiskDoctor, but it doesn't
make things worse, unlike DiskDoctor.  Recommended.

     Between these two programs, I've been able to recover from just
about anything.  In really bad cases, either program might hang, but
if I would reboot and try again things would usually straighten out.

Charlie_Gibbs@mindlink.UUCP
[If the line eater were to move, where wou

chris@genly.UUCP (Chris Hind Genly) (08/22/90)

>In article <chris.5660@genly.UUCP> chris@genly.UUCP (Chris Hind Genly) writes:
>
>  Is there something better than diskdoctor?  I've heard that disksalv
>  is better.  But it doesn't work on hard disks.
>
>In article <3617@teksce.SCE.TEK.COM> dales@teksce.SCE.TEK.COM (Dale Snell) writes:
>
>     Say what?  DiskSalv most certainly *does* work on hard drives;

Sorry, I wasn't very clear.  I meant disksalv doesn't actually fix the
file system on the hard disk so that it can be validated.  I keep good
backups, so there isn't really a need to make a backup using disksalv.

What I want, since I suspect there is actually very little damage, is
something that will repair the file system on the hard disk.  I was looking
for something like the Unix fsck, or the MSDOS chkdsk.  Diskdoctor doesn't
leave the disk in a state where it can be validated.

       *                        *                        *
                                                            \|/
               *           _______                         --O--
                      ____/ KC1VP \____        *            /|\
 *      _____________/  (203) 389-8680 \_______________
 ______/  95 Fountain Terr., New Haven, CT, USA, 06515 \_______
/ Chris Hind Genly    chris@genly.uucp   uunet!hsi!genly!chris \
----------------------------------------------------------------

dales@teksce.SCE.TEK.COM (Dale Snell) (08/22/90)

[woof!]

In article <chris.5660@genly.UUCP> chris@genly.UUCP (Chris Hind Genly) writes:
|
|Is there something better than diskdoctor?  I've heard that disksalv
|is better.  But it doesn't work on hard disks.
|

     Say what?  DiskSalv most certainly *does* work on hard drives; I've
used it myself several times.  You'll need a stack of floppies, and a
lot of time, but it works quite well.

|Is there anyway to minimize the risk of trashing the disk when running
|a program known to be unstable.  Perhaps:

     The only thing I could suggest would be to not write to the disk. 
You could use Lock, I think.  Of course, if the program bypasses the
OS...

                                    Seizure later,
				    --dds

Dale D. Snell      dales@teksce.SCE.TEK.COM      74756.666@compuserve.COM
"Welcome to Oregon, home of the California Raisins."

jerry@truevision.com (Jerry Thompson) (08/23/90)

Disksalv DOES work on hard drives!  It has saved my butt a couple of times.
Stop with this DiskDoctor stuff already and get DiskSalv.
-- 
Jerry Thompson                 |     // checks  ___________   | "I'm into S&M,
"What I want to know is, have  | \\ //   and    |    |    |   |  Sarcasm and
 you ever seen Claude Rains?"  |  \X/ balances /_\   |   /_\  |  Mass Sarcasm."

andy@cbmvax.commodore.com (Andy Finkel) (08/23/90)

In article <chris.5660@genly.UUCP> chris@genly.UUCP (Chris Hind Genly) writes:
>My disk was just clobbered again.  I ran diskdoctor and was notified
>of unreadable files, blocks used twice, and duplicate keys.  Yet in
>the end the disk could not be validated.  If diskdoctor can't
>fix the disk sufficiently so the disk can be validated, then whats
>the point of diskdoctor?

Diskdoctor is designed to make the disk as readable as possible,
so you can recover your data.  It's not really designed to bring the
disk back to 100% condition.  It will, quite often, do this.
However, it always says you should now copy your files off to a new
disk and reformat the old one after being run to recover the disk.
 

                 andy

-- 
andy finkel		{uunet|rutgers|amiga}!cbmvax!andy
Commodore-Amiga, Inc.

"Of course it's the murder weapon.  Who would frame someone with a fake?"

Any expressed opinions are mine; but feel free to share.
I disclaim all responsibilities, all shapes, all sizes, all colors.

mcmahan@netcom.UUCP (Dave Mc Mahan) (08/24/90)

 In a previous article, chris@genly.UUCP (Chris Hind Genly) writes:
>
>Is there anything I can do besides reformat?

You can try all the other remmedies people have posted, but I find that I
eventually have to just bite the bullet and reformat after doing a backup.
So far, I have only had to do this once.  I was able read the disk, but not
write to it.  I had a recent backup from MRBackup (Thank goodness!) and I
just copied the new files to floppy by hand.  MRBackup won't run if it can't
write to the disk, I guess.  That was version 3.x or maybe 2.4, so it may have
been fixed by now.

>
>Is there anyway to minimize the risk of trashing the disk when running
>a program known to be unstable.  Perhaps:
>  - Reduce the number of buffers
>  - No other processes running and using the disk. 
>  - Run out of the ram disk, so nothing will be flushed to the hard disk?

You can try to run with just a floppy based system.  You do this by not mounting
your harddrive and just booting off of a floppy.  Kind of a pain, but think of
the alternatives.....  

I'm not sure if there is an 'unmount' command that allows you to protect your
disk or not.  It sure would be convienient!  If there is, you can do development
on the hard disk, copy the new test program over to floppies or RAM disk,
unmount the hard drive, and do your testing with no hard drive enabled. 
Trashing the disk is still possible, but MUCH harder, since your program would
have to be VERY devious to get at the drive and cause damage.  Maybe you could
set up just a spare partition on the hard drive of 1 or 2 megabytes and do
your testing there after unmounting your main drive.  That way, if you trash
the partition you will be able to easily re-format it and still have your
main partition left intact.  Of course, this assumes that you can easily
unmount your main partition without having to re-boot.

 
>/ Chris Hind Genly    chris@genly.uucp   uunet!hsi!genly!chris \

   -dave

cosell@bbn.com (Bernie Cosell) (08/24/90)

chris@genly.UUCP (Chris Hind Genly) writes:

}What I want, since I suspect there is actually very little damage, is
}something that will repair the file system on the hard disk.  I was looking
}for something like the Unix fsck, or the MSDOS chkdsk.  Diskdoctor doesn't
}leave the disk in a state where it can be validated.

I happened to notice in an Amiga World maybe two months ago that
Central Coast Software [the folks that do Quarterback] claim to have an
fsck-like disk repairer.  Anyone know anything about it?

  /Bernie\

dales@teksce.SCE.TEK.COM (Dale Snell) (08/25/90)

In article <chris.5928@genly.UUCP> chris@genly.UUCP (Chris Hind Genly) 
writes:
|>In article <3617@teksce.SCE.TEK.COM> dales@teksce.SCE.TEK.COM 
|>(Dale Snell) writes:
|>
|>     Say what?  DiskSalv most certainly *does* work on hard drives;
|
|Sorry, I wasn't very clear.  I meant disksalv doesn't actually fix the
|file system on the hard disk so that it can be validated.
| ...
|I was looking
|for something like the Unix fsck, or the MSDOS chkdsk.  

     Ah, I see.  I misunderstood.  You're right, DiskSalv won't fix
file-systems in place (though Dave Haynie has dropped a hint or two). 
An fsck-like facility is something that is really needed for the Amiga,
especially with hard drive capacities getting larger all the time.  I
sure could have used one, several times.  Dave?  How 'bout it?

                                    Seizure later,
				    --dds

Dale D. Snell      dales@teksce.SCE.TEK.COM      74756.666@compuserve.COM
"Welcome to Oregon, home of the California Raisins!"

mrr@mrsoft.Newport.RI.US (Mark Rinfret) (08/26/90)

>In article <12518@netcom.UUCP> mcmahan@netcom.UUCP (Dave Mc Mahan) writes:
>
> In a previous article, chris@genly.UUCP (Chris Hind Genly) writes:
>>
>>Is there anything I can do besides reformat?
>
>You can try all the other remmedies people have posted, but I find that I
>eventually have to just bite the bullet and reformat after doing a backup.
>So far, I have only had to do this once.  I was able read the disk, but not
>write to it.  I had a recent backup from MRBackup (Thank goodness!) and I
>just copied the new files to floppy by hand.  MRBackup won't run if it can't
                                               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>write to the disk, I guess.  That was version 3.x or maybe 2.4, so it may have
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>been fixed by now.
 ^^^^^^^^^^^^^^^^^^

Yes, this was addressed a few releases ago.  MRBackup now has a CLI command
line  parameter  or a WorkBench TOOLTYPES entry which supports an alternate
"working directory".  This is where MRBackup writes backup volume "context"
information.

By  the  way...   Version 4.0b has an apparent problem with compression and
files which expand when compression is applied (as happens with certain IFF
files,  etc.).   The problem is due to a failure to propogate an error code
to  the  appropriate  level.  I would have released a fix by now, but a bug
introduced in Aztec C 5.0b prevented a quick fix (short parameters).  I had
to  rework  my  entire  support  library and all MRBackup code modules as a
result.  Sorry for the delay.

My  recommendation  is  that  you  forego  the use of compression until the
release  of  version 5.0, which will be released in about a month (though I
may release a beta before then).

--
#################################################################
# Mark R. Rinfret, MRSoft               Home: 401-846-7639      #
# mrr@mrsoft, galaxia!mrsoft!mrr        Work: 401-849-9930 x301 #
#################################################################