Spinrite

[Rod Speed]

But why would it be more or less clear for other sectors?

Basically spinrite tells you to bugger off if
the bad is in the fat. It doesnt if it isnt.

If it doesn't do that in the FAT -- and we just agreed
that it could in theory -- then isn't it logical to conclude
that it doesn't force relocation on any sectors?

It claims to be doing more than just marking
bad clusters as bad at the OS level.

Its unlikely even Gibson is lying as flagrantly as that.

 

[Joep]

Hey,

Yes, impressive, isn't it? All that from a 94 kB program, 'pure genius'.
Try and beat that, van der Steen.

Well, he DOES do it all in assembly, that's cool and does result in compact
code normally. But I also *think* he uses UPX and I also read that somewhere
.... so do I (use UPX) ... that makes the current DiskPatch build 68 kB only.
;-)

Yes, and I told you what it does in the part that you snipped.
Cut the obstinacy and read it again.

Okay, well even then that's only *slightly* less dumb - doesn't even come
close to magic. If you read their pages you'd (or at least I) expect voodoo
and magic to take place on my machine. They make it sound like it's God's
gift to mankind (see http://grc.com/sroverview.htm) ...

Anyway - to avoid a too negative atmosphere here, I am actually glad you
are trying to explain this, just so you know!

regards

 

[Folkert Rienstra]

What did you pay for this program?

It doesn't do datarecovery on the logical level (file system, etc), only on
the physical level although it is a bit of a hybrid.
It can however recover sector data and copy that to another free cluster
if sector reassigment is unavailable. This is why it works on a partition basis.
It relocates the cluster data and adjusts the FAT accordingly.

The FAT can be repaired if different sectors of each FAT copy are bad
sectors.
How?


If it is the "bad sectors at write problem", where all sectors written
in one second become bad, typically the same sectors of each FAT copy
will be damaged.

But then 'typically' the drive is dead.

 

[Folkert Rienstra]

Hey,


Well, he DOES do it all in assembly, that's cool and does result in compact
code normally. But I also *think* he uses UPX and I also read that somewhere

It was probably me. Like I said, the problem is more with the author.

... so do I (use UPX) ... that makes the current DiskPatch build 68 kB only. ;-)

Shrinker recognizes SpinRite5 as Shrinker.
PKlite recognizes it as 'PKlite, version undetermined'.
It's not UPX, as that has 'UPX' in the code.
Not that it is important. What is, is that S.G. lies about it.

Okay, well even then that's only *slightly* less dumb - doesn't even come
close to magic. If you read their pages you'd (or at least I) expect voo-
doo and magic to take place on my machine. They make it sound like it's
God's gift to mankind (see http://grc.com/sroverview.htm) ...

Yes, but then you know that Steve Gibson is a God, don't you?
One wouldn't expect less!
(Did I tell you "the problem is more with the author"?).

Anyway - to avoid a too negative atmosphere here, I am actually glad you
are trying to explain this, just so you know!

regards

Oeps. Van de arme tak? ;-)

 

[Svend Olaf Mikkelsen]

How?

By copying the good FAT sectors to the bad?

 

[Folkert Rienstra]

By copying the good FAT sectors to the bad?

The discussion was about those sectors being bad (and staying bad)
so I expected some hocus pocus with a single new FAT in another
location or something.

But you are probably correct:

Spinrite doesn't touch them because it can't read them and that is
why they stay, not that they are necessarily physically bad or would
not reassign when written to. It just stops because it thinks it can't
do anything about it. That is the common problem with unrecoverable
data, you have to leave it as is or discard it and take your loss.

If it was smarter it would realize that it doesn't need to read them
as it can copy the fat-copy alternative to the bad sectors and then
retry them for check. If it fails again, then it has exhausted it's options.

Actually, we have no way of knowing whether it isn't already doing that.

Btw, like to add that since SpinRite works on partitions it
knows which sectors contain userdata and which ones don't.
It doesn't need to be able to read the ones that don't to be
able to discard them if necessary by writing to them directly.

 

[Rod Speed]

I finally tried emailing Steve Gibson, and he said it can and
often does work. His tech was just wrong to say it doesn't
work on the FAT. So his tech just caused a big waste of time.

Quite bluntly I wouldnt believe anything Gibson says after
the obvious con job in the spruik for spinrite on that site.

Gibson said I could have my money back if it didn't work, so I tried it.

It never complained about the location of the bad sectors. It
just did whatever it did, and now the visible surface is 100% good.

Thats not surprising if what it does got the drive to remap bad sectors.

Disk Doctor finds no bad sectors on it,
although it still had a lot of logical errors to fix.

Presumably because some of the bads were in the FAT.

 

[Rod Speed]

I finally tried emailing Steve Gibson, and he said it can and
often does work. His tech was just wrong to say it doesn't
work on the FAT. So his tech just caused a big waste of time.

Quite bluntly I wouldnt believe anything Gibson says after
the obvious con job in the spruik for spinrite on that site.

Gibson said I could have my money back if it didn't work, so I tried it.

It never complained about the location of the bad sectors. It
just did whatever it did, and now the visible surface is 100% good.

Thats not surprising if what it does got the drive to remap bad sectors.

Disk Doctor finds no bad sectors on it,
although it still had a lot of logical errors to fix.

Presumably because some of the bads were in the FAT.

I'm still VERY dubious about the value of spinrite tho because
if it did just force the drive to remap at least one bad in the
FAT, spinrite should have been able to work out which sector
was bad, get the data from the other fat in that sector, force
the drive to remap the bad and then put the data that it had
obtained from the other copy of the FAT into the replacement
sector and so there shouldnt have been any logical errors.

 

[Rod Speed]

Do you think it's technically wrong, or that
it uses scare tactics to sell the program?

I know it vastly overhypes what they claim it can do.

In spades with what Gibson claims is unique to spinrite.

He's a pathological liar.

Yeah but I think it is surprising that the drive didn't remap
them before, because it kept failing to read the sectors.

Sure, most drives only remap when writes fail, not reads.

If Spinrite was the only way to force the remap,
for whatever reason, then it's worth something.

The hard drive manufacturer's ute can obviously do that.

The logical errors might have been made worse by Scandisk,
since it ran automatically at the first boot after the disk went bad.

Sure. The whole basics of scandisk is that you do have things
properly backed up and it really just fixes the basic problems
that are the result of not shutting the system down properly.

It was never intended to handle a dying drive.

I wonder why it and NDD do their logical testing before the physical,
since that only screws things up when there are bad sectors.

Sure, but it was never intended to handle a dying drive.

 

[Tom Del Rosso]

The hard drive manufacturer's ute can obviously do that.

Non-destructively?

 

[Rod Speed]

Non-destructively?

I consider that non destructively to be a furfy. You're sposed
to have proper backups and they are needed if the drive just
stops working completely. I think its stupid to try to kludge
around the lack of proper backups by doing what SpinRite
claims to do, attempt to get as much of the data out of a
bad sector as possible before forcing the drive to remap
that bad sector and then putting back what data it has
allegedly recovered from the sector once its obviously bad.

I dont believe it actually does that anyway when it clearly
doesnt do whats MUCH simpler, just get the data from
the second copy of the FAT with a bad sector in the FAT.
It clearly didnt do that with your bad because if it had,
you wouldnt have seen the logical errors.

 

[Tom Del Rosso]

Non-destructively?

I consider that non destructively to be a furfy. You're sposed
to have proper backups [.....]

Yes, but this was a client machine where a server is present, so it gets
all the attention and the backups. Unfortunately programs like Word
default to the silly 'My Documents' folder, and while it's obvious that
the save folder should be changed in a network environment, most
installers never think of changing it for every program. I didn't
install this machine but I don't fault the guy who did for that. Often
when you set up a computer for someone you don't even know what programs
they will use, and they often start using other programs later on (like
in this case), without telling their tech that they've been saving files
locally.

I dont believe it actually does that anyway when it clearly
doesnt do whats MUCH simpler, just get the data from
the second copy of the FAT with a bad sector in the FAT.
It clearly didnt do that with your bad because if it had,
you wouldnt have seen the logical errors.

I think that would have made it worse, since the other copy of the FAT
was also corrupted by that time. When Scandisk ran automatically it
must have made a mess before it realized there were bad sectors and
halted. Otherwise, if one copy of the FAT was still valid, it wouldn't
have taken 5 passes of both NDD and Scandisk (which seems better at
fixing LFN errors) to sort it out. If Spinrite had made a simple copy
from one FAT to the other it would have been even harder to sort out.
That's why I think it's better that programs specialize in treating
either logical or physical problems.

 

[Rod Speed]

Yes, but this was a client machine where a server is present, so it
gets all the attention and the backups. Unfortunately programs like
Word default to the silly 'My Documents' folder, and while it's obvious
that the save folder should be changed in a network environment, most
installers never think of changing it for every program.

All completely irrelevant to what spinrite can or cannot do.

The only thing that makes any sense is to configure a system like
that properly and if the client is so clueless that they cant do that
for themselves, they should be encouraged to get someone to
configure the system for the properly in the first place.

I didn't install this machine but I don't fault the guy who did for that.

You should, thats a completely stupid way to have it configured.

Often when you set up a computer for someone you
don't even know what programs they will use, and they
often start using other programs later on (like in this case),
without telling their tech that they've been saving files locally.

Anyone setting a system like that up for them should be
making it clear that someone who knows what they are
doing should be configuring that basic stuff for them.

I think that would have made it worse, since the other
copy of the FAT was also corrupted by that time.

Dont believe that. If the drive developed a bad sector in
one of the copys of the FAT, scandisk should have noticed
the mismatch in the FAT copys and complained about that.

When Scandisk ran automatically it must have made a
mess before it realized there were bad sectors and halted.

Dont believe that, see above.

Otherwise, if one copy of the FAT was still valid, it
wouldn't have taken 5 passes of both NDD and Scandisk
(which seems better at fixing LFN errors) to sort it out.

Or you managed to produce that mess by running them
both. NDD particularly is notorious for turning what is
a recoverable situation into an unrecoverable mess.

If Spinrite had made a simple copy from one FAT to
the other it would have been even harder to sort out.

Bullshit. If there was a bad sector in one of the copys
of the FAT, it should have noticed the mismatch
between the FATs and complained about that.

What may well have happened is that there isnt actually
a bad sector at the platter level at all, just a fault in the
drive that produces what appear to be random bad
sectors, random in the sense that the location of the
purported bad sector varys from run to run, usually
due to a crack in the flexible connection to the heads.

No program like scandisk, ndd or spinright can
do a damned thing about that sort of drive error.

The only viable protection against that
sort of drive fault is proper backups.

That's why I think it's better that programs specialize
in treating either logical or physical problems.

Not even practical when the absolute vast bulk of problems
are logical ones due to the system not be shut down properly
for whatever reason. Since thats by far the most common
real world situation, whats used to handle those has to be
aware of the possibility of physical problems and cant just
throw up its hands and give up when any problem is seen.
The most thats viable is to do the basics like stop when
the app gets recoverable read errors from the drive and
suggest that the drive has a physical problem.

Whats MUCH better is real backups.

 

[Svend Olaf Mikkelsen]

If Spinrite had made a simple copy
from one FAT to the other it would have been even harder to sort out.
That's why I think it's better that programs specialize in treating
either logical or physical problems.

It of course should only copy FAT sectors that were good in one FAT
copy, and bad in the other.

Also no remapping was probably involved, since bad sectors do not
occur by the causes presumed when the bad sector reassignment
procedure was designed.

 

[Folkert Rienstra]

Non-destructively?

I consider that non destructively to be a furfy. You're sposed
to have proper backups [.....]

Yes, but this was a client machine where a server is present, so it gets
all the attention and the backups. Unfortunately programs like Word
default to the silly 'My Documents' folder, and while it's obvious that
the save folder should be changed in a network environment, most
installers never think of changing it for every program. I didn't install
this machine but I don't fault the guy who did for that. Often when you
set up a computer for someone you don't even know what programs they
will use, and they often start using other programs later on (like in this
case), without telling their tech that they've been saving files locally.

I dont believe it actually does that anyway
when it clearly doesnt do whats MUCH simpler,

There is no such indication that it didn't nor that it did.

There is no such indication at all.

I think that would have made it worse, since the other
copy of the FAT was also corrupted by that time.

Which could just as easily explain what you saw if that's what it did:
copy from the other FAT. If SpinRite had made-up false data or just
cleared the sector, the FATs would have been different.

When Scandisk ran automatically it
must have made a mess before it realized there were bad sectors and
halted. Otherwise, if one copy of the FAT was still valid, it wouldn't
have taken 5 passes of both NDD and Scandisk (which seems better at
fixing LFN errors) to sort it out.

Scandisk will tell when FATs are not the same. Scandisk can't tell which
one is false unless the signature is missing or similar obvious indicators.

If Spinrite had made a simple copy from one FAT to
the other it would have been even harder to sort out.

No, just different.

 

[Folkert Rienstra]

I finally tried emailing Steve Gibson, and he said it can and often does
work. His tech was just wrong to say it doesn't work on the FAT. So
his tech just caused a big waste of time.

Gibson said I could have my money back if it didn't work, so I tried it.

It never complained about the location of the bad sectors. It just did
whatever it did, and now the visible surface is 100% good. Disk Doctor
finds no bad sectors on it, although it still had a lot of logical errors to fix.

So much for "Basically spinrite tells you to bugger off if the bad is in the
fat. It doesnt if it isnt".

 

[Fred]

just the pathetic excuse for a juvenile troll any 3 year old could leave for dead.

 

[Folkert Rienstra]

process, i.e. the remapping of a sector

doing that.

i.e. read successfully

That para is an unintelligible obscenity. Try again.

Makes perfect sense, Rodney, to anyone that_can_ and _wants_ to read.

And its terminally pig ignorant to imply that all dri-
ves behave identically on those specifics anyway.

Terminal to claim the opposite, Rodney, since there are no other ways.
Except for remapping the sector with loss of data and claim all is well
and nothing happened. A drive that is happily corrupting your data.
Wake-up, Rodney.

Mindlessly silly, as always with you.

If you say so, Rodney. Obviously you don't have a clue
about how harddrives work. Never had, never will have.

Obviously not true when a ute is deliberately
attempting to encourage the drive to remap a bad.

Nothing obvious about that, Rodney. More so, when you made it
abundantly clear that you don't even believe that it does that.

What _is_ obvious is that you made a terminal stupid remark about
how harddrives do Read after Write and are now frantically try-
ing to escape out of that wet paper bag that you found yourself in.
So terminally stupid that you didn't dare use your own name to reply.

Its terminally pig ignorant to imply that all
drives behave identically on those specific.

Easy to claim Rodney. Care to back it up?

 

[Folkert Rienstra]

just the pathetic excuse for a juvenile troll any 3 year old could leave for dead.

Couldn't find anything to fault, eh Rodney. Bummer.
Try harder, Freddles, but watch those brainfarts.

 

[Fred]

process, i.e. the remapping of a sector

Didnt help, try again.

i.e. read successfully

Didnt help, try again.

Makes perfect sense, Rodney, to
anyone that_can_ and _wants_ to read.

Even you should be able to bullshit your way out
of your predicament better than that pathetic effort.

Presumably it was intended all along to be
just another of you peurile cryptic trolls that
was never meant to make any sense at all.

Terminal to claim the opposite, Rodney,
since there are no other ways.

Obvious lie.

Except for remapping the sector with loss of data and
claim all is well and nothing happened. A drive that is
happily corrupting your data. Wake-up, Rodney.

Even you should be able to bullshit your way out
of your predicament better than that pathetic effort.

If you say so, Rodney. Obviously you don't have a clue
about how harddrives work. Never had, never will have.

Even you should be able to bullshit your way out
of your predicament better than that pathetic effort.

Nothing obvious about that, Rodney. More so, when you made
it abundantly clear that you don't even believe that it does that.

Even you should be able to bullshit your way out
of your predicament better than that pathetic effort.

I never ever said anything even remotely resembling
anything like that, pathological lying child.

What _is_ obvious is that you made a terminal stupid
remark about how harddrives do Read after Write

Another lie that anyone can check for themselves.

and are now frantically trying to escape out of
that wet paper bag that you found yourself in.

Even you should be able to bullshit your way out
of your predicament better than that pathetic effort.

Easy to claim Rodney. Care to back it up?

YOU made that terminally stupid claim, ****nert.
YOU get to do the backing up, ****nert.
THATS how it works, ****nert.