Text 7000, 68 rader
Skriven 2006-08-11 01:29:44 av Maurice Kinal (1:140/13.1)
   Kommentar till text 6996 av Paul Rogers (1:105/360.0)
Ärende: You gotta smile
=======================
Hey Paul!

Aug 10 16:43 06, Paul Rogers wrote to Maurice Kinal:

 MK>> It should fail but I didn't bother reading it back since

 PR> There's that word again.  You often get that way?

Yes.  For the smaller drives then dd write/read cycle(s) is the way I'd go. 
With regular ide HDs, such as in the case with the drives I tested, then I
preferred the mke2fs's -cc switch.  It mostly depends on size and somewhat on
type to which method I'd choose to go with.

 PR> Regardless of what did or didn't affect Wayne's World,
 PR> the issue here is whether dd can be trusted to wipe an
 PR> entire drive, does a destructive write to all sectors.

If the drive has no other problems then it works great but if the drive is
large AND one requires verification then I'd go with at least one 'c' in the
mke2fs method.  If it passes that test AND you wish a different type then 83 it
could be re-fdisked later and formatted to whatever type required.  Bottomline
is that I'd feel confident that the drive is indeed clean and all those blocks
are fine, readable and writable.  However that isn't even an issue here and
thus I end up with a properly formatted drive just using mke2fs with at least
-c which gives me two write/read cycles with two different patterns.

 PR> I use a slight variant of a well known acronym to remember
 PR> my litany of primary issues.
 PR>         I manage computing to be:
 PR>         - Functional
 PR>         - Usable
 PR>         - Compatible
 PR>         - Adaptable
 PR>         - Reliable.
 PR> and it sounds like his salvage fails three of the five.

For me it would be that I lack confidence in any of them and thus nothing else
really matters beyond that.  If I had no other choice I see a couple that
potentially I could squeeze some life out of until they could be replaced by
something that would renew my confidence level to a higher state but thank
goodness that isn't the case ... today.  Out of the five you list above
'Compatible' needs only apply to Linux.  As far as any other OS compatibilty
that only applies to what gets sent from here.  So far that has never been an
issue.  Incoming is a whole other issue but over the years I have learned to
compensate for other OS's lack of concern in the realm of compatibilty where it
TRULY matters. 

 PR> It's less about testing than it is about sterilization. IMNSHO

Right.  All the error free blocks got a very good scrubbing and dd did it's
part.  However the badblocks remained badblocks and in the case of all the
drives in the stack they were chock full of those.  My neighbour just liberated
them as he is under the impression that he can do better.  I am confident he
cannot.  :-)

I'll let you know what he discovers if indeed he goes that far with them all
and learns anything beyond any of them being trustworthy for anything.

As for my needs I am happy with straight dd as a 'magnet' with a dd based
verification for all sorts of flashdisks.  The largest ones I use and have are
256M (two compactflash and one usb) but really only need 16M which gives me
elbow room for what I use and want these things for.  This is great.

Life is good,
Maurice

--- Msged/LNX 6.2.0
 * Origin: The Pointy Stick Society XXII - A two bit outfit (1:140/13.1)