fsck on FAT32 filesystem?

Robert Bonomi bonomi at mail.r-bonomi.com
Wed Jul 18 12:45:47 UTC 2012

> From owner-freebsd-questions at freebsd.org  Tue Jul 17 12:06:29 2012
> Date: Tue, 17 Jul 2012 19:02:19 +0200 (CEST)
> From: Wojciech Puchar <wojtek at wojtek.tensor.gdynia.pl>
> To: Robert Bonomi <bonomi at mail.r-bonomi.com>
> Cc: freebsd-questions at freebsd.org
> Subject: Re: fsck on FAT32 filesystem?
> >>
> >>
> >>> Surely SpinRite is "more clever" than that,
> >> i would bet otherwise. simple tools and free tools are always better
> >
> > You continue to demonstrate that you "don't know what you don't know".
> are you another sponsored by some "recovery tool" commercial producer?

What I am is an information systems professional with 45 years experience.
including 30 years with Unix, who does not suffer ignorant, ill-informed,
and arrogant, fools gladly.

You make pronouncements of your *opinions* as though they are God-given 
fact -- even on things which you _don't_ have actual knowledge.  You're 
entitled to have opinions, *BUT* the "Gospel According to Wojciech" is -not- 
'the answer' for everybody, in every situation. *IF* you ever learn that, 
realize that there _are_ other =legitimate= viewpoints on matters, and
qualify your statements with things like 'in my opinion', 'this might
help', 'have you considered trying' -- as opposed to dictating what the
reader must do, *especially* when you have missed critical facts in
the question you are responding to -- Then, and *ONLY*THEN*, are people 
likely to give your opinions about how to do things any serious consideration.

Case in point, your "I would bet otherwise" -- an implicit admission you
*don't* know how SpinRite actually works.  How much hard cash, US dollars,
do you have to 'put your money where your mouth is"?   Alternatively, you
can admit you were blowing bullshit -- that your words were merely
uninformed speculation, with no actual basis in fact.

As for my subject-expertise -- I have, personally, _written_ stand-alone code
that directly interfaces with hard-controller disk chips -- for purposes of 
evaluating the condition of damaged hard-disks.  I've had clients come to 
me for advice on data-recovery, having suffered catastrophic damage to their 
only copy of what was truly 'mission critical' data. (No, they weren't
existing clients -- if they had been, proper back-up procedures would have
been in place, and the disk crash would have been a 'non-event'.)

I have successfully recovered _every_byte_ of data from a damaged "State of 
The Art Compression" compressed disk volume, using custom device-driver code
that I wrote.

I've had clients that decided it WAS 'worth it' to pay one of the 'kilobuck
per megabyte of recovered data' (actual price) "Class 25 clean room" recovery
services -- where the damage to the drive was such that *ANY* attempt to
access anything on the drive would cause more damage.  Using "'simple, free
tools", like your 'dd' recommendation,  would (a) not have been successful,
and (b) *greatly* reduced what would be recoverable by the clean-room facility.

Your assertation that "free tools are always better" is pure, unadulterated
bullshit.  For 'simple' situations, they _may_ be adequate, or may not.

When there are various kinds of _serious_ problems, even -attempting- to 
use tools like 'dd' (or SpinRite, for that matter) can/will make things 
FAR worse.  Drive disassembly and platter cleaning _must_ be the first t
hing done in such situations.

_For_the_price_, SpinRite provides an amazing level of functionality. circa 
85-90% of what high-end professional tools costing 100x more can do.  It's 
not a FUS, but it is incredible 'bang for the buck', and does things that
*NO* Unix 'userland' application can do in reconstructing damaged data.
SpinRite _will_ recover data in a lot of situations where the 'dd' approach
is "less than effective".  Situations where SpinRite is ineffective, _and_
the "clean room" approach is _not_ required, are rare.  It's not perfect,
it won't fix "everything", but it is an incredibly inexpensive step up
(and a *LARGE* step up) from the 'dd' approach.  If the 'dd' type approach
you you recover 'what you need' that's great.  If _not_, SpinRite should
probably be the 'next step'.  If it _doesn't_ work, the cost/time for
trying it is 'inconsequential petty cash', elative to the cost of the _next_
approach.  And, if it -does- work, it paid for itself, a hundred times over,
by saving the cost of the really expensive approach.  "Cheap insurance'
even at several times the retail price.

More information about the freebsd-questions mailing list