[comp.sys.amiga] help my hard drive & diDiskSalv problems

itch@cbnews.att.com (richard.m.brack) (08/31/90)

My hard drive probem:
   Last weekend my hard drive freaked out.  I have a 40meg SupraDrive
   in two 20meg partitions.  My amiga decided that dh1: was no longer a dos
   drive, so it refused read it.  Incedentaly this seems to have
   happened right after I saved a song with MED 2.0 (coincidence?).

DiskSalv 1.42 problems:
   So I used DiskSalv 1.42 to recover my files from dh1: to floppies.
   It took me about 5 trys to get disksalv to work.  It would start
   counting blocks, and then just hang.  It finally worked with
   this invocation:  'disksalv dh1: df0: ffs format nodos' .  I'm sure
   it wasn't the first time I tried those arguments either.  At this
   point disksalv wouldn't even format the floppies.  Was this because of
   the 'nodos' argument?  My format command is sys:system/format which 
   is where I thought it was supposed to be.  Anyway...I did finally get
   all my stuff from dh1: onto (20) floppies.

What now???
   I'm not sure what I need to do now.  Please don't tell me I have
   to reformat the entire drive.  I don't think I can format just
   a partition, can I?  My drive manual says something about 'zeroing'
   my drive or partition.  This sounds like what I should do but the
   manual really doesn't say that much about what 'zeroing' is.  

My questions:
   1) What might have caused my dh1: to no longer be a dos drive?
   2) What was the deal with DiskSalv?
   3) Now what do I do to my hard drive?

ANY comments would be greatly appreciated!!  Thank You Thank You Thank You.


RichBrack  
-- 
{ the itchman cometh   /-/         _          i don't want to be your angel }
{ itch@cbnews.att.com /-/        _|_|_          i want to be your witch!    }
{ att!cbnews!itch \-\/-/         ( * )tch                 -yello            }
{                  \/\/           /^\                                       }

daveh@cbmvax.commodore.com (Dave Haynie) (09/07/90)

In article <1990Aug30.191714.10892@cbnews.att.com> itch@cbnews.att.com (richard.m.brack) writes:

>DiskSalv 1.42 problems:
>   So I used DiskSalv 1.42 to recover my files from dh1: to floppies.
>   It took me about 5 trys to get disksalv to work.  

OK, so CLI-based programs aren't as user-friendly as Intuition-based programs.
This gets fixed in DiskSalv 2.0, which is ever-so-slowly on the way.

>   It would start counting blocks, and then just hang.  

I've never seen that happen.  If you find a problem I can reproduce, you have
found a problem I will fix.

>   It finally worked with this invocation:  'disksalv dh1: df0: ffs format 
>   nodos' .  I'm sure it wasn't the first time I tried those arguments 
>   either.  At this point disksalv wouldn't even format the floppies.  Was 
>   this because of the 'nodos' argument?  My format command is 
>   sys:system/format which is where I thought it was supposed to be.  

It sounds to me like most of your problems stem from your system disk (the
one SYS:, L:, C:, S:, LIBS:, FONTS:, and DEVS: points at) being the one 
you're trying to recover.  DiskSalv V1.42 counts on various operating
system things, like the Format program and the Disk-Validator being present
for formatting and similar functions to work correctly.  There's really no
good way around this, doing what V1.42 does.  The V2.00 release solves the 
same kinds of problems in a totally different way, and thus won't need 
anything from SYS: once its up and running.

Currently, I know of two significant bugs in DiskSalv V1.42:

	- The Out Of Memory trapping function actually needs memory to run
	  correctly.  If you trap to this routine in very low memory 
	  situations (where, at this point, the salvage operation is aborted
	  anyway), the trap winds up trapping itself, and you get an unending
	  loop of "ERROR: Out of Memory" or some-such echoing at your
	  console.

	- DiskSalv gets confused in a rather complicated situation in which
	  it locates still-recoverable files on a disk which belong to a 
	  directory node that has been overwritten by a file node.  This 
	  case could result in a crash or possibly a hangup.

There's also one problem we recently encountered at Commodore.  Someone here
accidently formatted the start of a hard disk that wasn't supposed to get
formatted.  DiskSalv V1.42 seemed to get real confused about that disk; it
would chug for awhile, start restoring, crash, and upon reboot, all of the
files it found were of 0 length.  The problem here was actually that the Format
command had overwritten block 0 of that disk, so it didn't appear to be an FFS
disk anymore.  Salvaging an FFS disk in SFS mode, or visa-versa, is a bad idea
and may even crash the program.  The FFS/NOFFS switch can force DiskSalv into
the right mode if it can't properly detect the mode from your damaged disk.
V2.00 may grow a more robust filesystem typing routine if I get around to it.

>What now???
>   I'm not sure what I need to do now.  Please don't tell me I have
>   to reformat the entire drive.  I don't think I can format just
>   a partition, can I?  My drive manual says something about 'zeroing'
>   my drive or partition.  This sounds like what I should do but the
>   manual really doesn't say that much about what 'zeroing' is.  

You can certainly reformat just the partition that's damaged.  For example,
"Format Drive DH1: ..." will only format the DH1: partition, not the whole
disk (unless the DH1: partition occupies the whole disk, of course).

>My questions:
>   1) What might have caused my dh1: to no longer be a dos drive?

A few things.  Some program bugs can damage a hard disk, though it's rare.
Using the mg1a program in text wrap mode managed to clobber a hard disk in
my system once.  Usually a crash, powerdown, or other reset condition while 
the disk is being written to is the most likely cause of a disk error.  Such
a crash can damage a file or directory header; if the directory header is
your root directory, the filesystem may have trouble reading you disk at all
upon reboot.

>   2) What was the deal with DiskSalv?

Explained above.  You seem to be having more trouble than most folks.  The
GUI-based DiskSalv should make things a bit more foolproof.

>   3) Now what do I do to my hard drive?

If you have examined your output floppies, and they all look good (eg, files
are there that you expect, they seem the right length and all), you can
reformat.  You may wish to try a fix-in-place tool like FixDisk, or even
DiskDoctor, since you should have nothing to lose if you files are safe on
floppy.  Or you can wait a few months for DS 2.00, and see if it can fix
you up any better...

>RichBrack  


-- 
Dave Haynie Commodore-Amiga (Amiga 3000) "The Crew That Never Rests"
   {uunet|pyramid|rutgers}!cbmvax!daveh      PLINK: hazy     BIX: hazy
      Get that coffee outta my face, put a Margarita in its place!

billsey@agora.uucp (Bill Seymour) (09/09/90)

In article <14249@cbmvax.commodore.com> daveh@cbmvax (Dave Haynie) writes:
>In article <1990Aug30.191714.10892@cbnews.att.com> itch@cbnews.att.com (richard.m.brack) writes:
>
>>DiskSalv 1.42 problems:
>>   So I used DiskSalv 1.42 to recover my files from dh1: to floppies.
>>   It took me about 5 trys to get disksalv to work.  
>
>OK, so CLI-based programs aren't as user-friendly as Intuition-based programs.
>This gets fixed in DiskSalv 2.0, which is ever-so-slowly on the way.
>
>>   It would start counting blocks, and then just hang.  
>
>I've never seen that happen.  If you find a problem I can reproduce, you have
>found a problem I will fix.

	This isn't a problem with Disksalv, it's a problem with his HD
combination... There's an occasional situation in the Supra software
where the device driver doesn't seem to correctly handle a grown error
on the hard disk. What you see is the device driver, and scsi bus, hung
while it tries to access the sector that has the error. you can get it
either by getting a new error on a disk, or by not completely mapping
the old errors out.
	The only real short term solution is to back the partition up
using something that doesn't do a full block search first, then low-level
format the drive while mapping all bad blocks out. The long term solution
is for Supra to get a new programmer in that can update their software
to handle these types of errors. Having Willie do their new software
updates while also doing their new hardware is perhaps just a bit too
much for one guy to handle.

>>What now???
>>   I'm not sure what I need to do now.  Please don't tell me I have
>>   to reformat the entire drive.  I don't think I can format just
>>   a partition, can I?  My drive manual says something about 'zeroing'
>>   my drive or partition.  This sounds like what I should do but the
>>   manual really doesn't say that much about what 'zeroing' is.  
>
>You can certainly reformat just the partition that's damaged.  For example,
>"Format Drive DH1: ..." will only format the DH1: partition, not the whole
>disk (unless the DH1: partition occupies the whole disk, of course).

	Doing a 'zero' of the offending partition would certainly
work if his problem were only and AmigaDOS bad block. 'Zero' from
within SupraFormat is the same as 'Format drive foo: name FooBar Quick'
from CLI. The problem is that as soon as he tried writing again
to that bad block, he'd start hanging the system just as he was
during the block mapping that DiskSalv does.
	One other thing I just thought of that could be his problem...
If he's got termination problems on his SCSI bus, he could be getting
lots of soft errors which could show as that same type of grown error.
In that case, zeroing and restoring would work, as long as he fixed
the cabling/termination problem first.

>
>>RichBrack  
>
>-- 
>Dave Haynie Commodore-Amiga (Amiga 3000) "The Crew That Never Rests"
>   {uunet|pyramid|rutgers}!cbmvax!daveh      PLINK: hazy     BIX: hazy
>      Get that coffee outta my face, put a Margarita in its place!


-- 
     -Bill Seymour             ...tektronix!reed!percival!agora!billsey
=============================================================================
Bejed, Inc.       NES, Inc.        Northwest Amiga Group    At Home Sometimes
(503) 281-8153    (503) 246-9311   (503) 656-7393 BBS       (503) 640-0842

itch@cbnews.att.com (richard.m.brack) (09/10/90)

In article <1990Sep8.192254.5960@agora.uucp>, billsey@agora.uucp (Bill Seymour) writes:
> In article <14249@cbmvax.commodore.com> daveh@cbmvax (Dave Haynie) writes:
> >In article <1990Aug30.191714.10892@cbnews.att.com> itch@cbnews.att.com (richard.m.brack) writes:
> >
> >>DiskSalv 1.42 problems:
> >>   So I used DiskSalv 1.42 to recover my files from dh1: to floppies.
> >>   It took me about 5 trys to get disksalv to work.  
> >>   It would start counting blocks, and then just hang.  
> 
> 	This isn't a problem with Disksalv, it's a problem with his HD
> combination...

Thanks for your help, guys.  In the mean time I have gotten my drive up
and running.  As I stated I did finally get DiskSalv to work.  I then
zeroed my dh1: partition and copied the disksalv'ed stuff back on.
It has been up for about a week with no problems.

> 	Doing a 'zero' of the offending partition would certainly
> work if his problem were only and AmigaDOS bad block. 'Zero' from
> within SupraFormat is the same as 'Format drive foo: name FooBar Quick'
> from CLI. The problem is that as soon as he tried writing again
> to that bad block, he'd start hanging the system just as he was
> during the block mapping that DiskSalv does.
> 	One other thing I just thought of that could be his problem...
> If he's got termination problems on his SCSI bus, he could be getting
> lots of soft errors which could show as that same type of grown error.
> In that case, zeroing and restoring would work, as long as he fixed
> the cabling/termination problem first.

Are you trying to tell me that my is just sitting around waiting to
crash again?  This is really the only trouble I have had with my drive,
and as I said above, it has been up for a week or so with no problems.
Do you suggest I do more than what I have done?  


Thanks again for listing to my problems.....

RichBrack

-- 
{ the itchman cometh   /-/         _          i don't want to be your angel }
{ itch@cbnews.att.com /-/        _|_|_          i want to be your witch!    }
{ att!cbnews!itch \-\/-/         ( * )tch                 -yello            }
{                  \/\/           /^\                                       }

billsey@agora.uucp (Bill Seymour) (09/11/90)

In article <1990Sep10.164805.23863@cbnews.att.com> itch@cbnews.att.com (richard.m.brack) writes:
:In article <1990Sep8.192254.5960@agora.uucp>, billsey@agora.uucp (Bill Seymour) writes:
:> 	Doing a 'zero' of the offending partition would certainly
:> work if his problem were only and AmigaDOS bad block. 'Zero' from
:> within SupraFormat is the same as 'Format drive foo: name FooBar Quick'
:> from CLI. The problem is that as soon as he tried writing again
:> to that bad block, he'd start hanging the system just as he was
:> during the block mapping that DiskSalv does.
:> 	One other thing I just thought of that could be his problem...
:> If he's got termination problems on his SCSI bus, he could be getting
:> lots of soft errors which could show as that same type of grown error.
:> In that case, zeroing and restoring would work, as long as he fixed
:> the cabling/termination problem first.
:
:Are you trying to tell me that my is just sitting around waiting to
:crash again?  This is really the only trouble I have had with my drive,
:and as I said above, it has been up for a week or so with no problems.
:Do you suggest I do more than what I have done?  

	You're certainly running the risk of having the same problem
if you only did a 'Zero' of that partition. I've found that the best
(or at least easiest) way to check the partition after Zeroing is to
go ahead and do a Format from AmigaDOS. This will read and write each
block on the partition and pretty much assure you that you don't have
a grown error waiting to bite you. If format hangs or aborts somewhere
on the partition, it's an indication that you should go ahead and
backup the whole drive then do a low level format with bad block mapping
turned on.
	If your problem is termination/cabling related, the errors will,
in most cases, be recoverable. Just click on 'Retry' when you see an
error. The risk you run there is during a write, some corrupted data
during a read usually doesn't hurt worse than a crash, corrupted data
during a write will leave you with something on the drive that you don't
want. :-{

:Thanks again for listing to my problems.....
:
:RichBrack
:
:-- 
:{ the itchman cometh   /-/         _          i don't want to be your angel }
:{ itch@cbnews.att.com /-/        _|_|_          i want to be your witch!    }
:{ att!cbnews!itch \-\/-/         ( * )tch                 -yello            }
:{                  \/\/           /^\                                       }


-- 
     -Bill Seymour             ...tektronix!reed!percival!agora!billsey
=============================================================================
Bejed, Inc.       NES, Inc.        Northwest Amiga Group    At Home Sometimes
(503) 281-8153    (503) 246-9311   (503) 656-7393 BBS       (503) 640-0842