[ale] Possibly bad hard drive

Ron Frazier (ALE) atllinuxenthinfo at techstarship.com
Thu Feb 21 16:31:20 EST 2013


By the way, dev null, why don't you post your name to your messages?

Anyway ... SIGH.  What I'm about to say is not meant to offend ... 
really.  However, almost everything you've said about Spinrite or Steve 
Gibson or GRC is wrong, so I'm going to address it.  I don't have any 
financial incentive to promote Spinrite.  I am a happy customer and hope 
others can use it to help recover data.  I would gladly buy the product 
again.  Comments inline.

On 2/21/2013 1:55 PM, dev null zero two wrote:
> Disclaimer: I used to do professional data recovery (clean room, head 
> and platter extractions and replacements, board rebuilds, RAID 
> rebuilds, etc) working for Scott Moulton.
>

I'm sure you were very knowledgeable and professional in your prior 
duties.  However, you don't have extensive knowledge of or experience 
with Spinrite.  I do.  I probably know more about it than anyone else on 
this forum, not from a coding perspective, but from a user perspective.  
I have used the product a number of times for personal data recovery.  I 
have had SUCCESS with using it several times for personal data 
recovery.  I have listened to every one of Steve Gibson's 391 podcasts, 
and when the topic of hard drive maintenance comes up, I listen 
carefully.  I have read much of the (admittedly somewhat dated) 
information he has about it online.

> Spinrite is no better than ddrescue (ddrescue, NOT dd_rescue, which is 
> much older and shittier). They both try and read sectors in reverse, 
> in PIO mode, coming from different sections of drive, retrying 
> multiple times, etc. But it's only as good as the ATA controller 
> itself and the firmware controlling it - which is to say, not that 
> great. that's why devices like the Deepspar Disk Imager exist: they 
> have custom controllers DESIGNED for failing drives (timeouts can be 
> changed, long reads can be used, read aheads can be turned off, you 
> can read sectors ignoring ECC, etc etc).

This is wrong.  Spinrite has far more advanced capabilities than 
ddrescue (based on my reading).  Spinrite will try to read a faulty 
sector up to 2000 times, and sometimes will spend days or weeks doing 
so.  This is far more than any other utility I'm aware of.  It is also 
far more than the OS will generally try.  He does try to read the sector 
from many different starting positions, as you alluded to, but Spinrite 
does this in a much more exhaustive manner, as far as I know.  Because 
the drive is a mechanical device, with wear and tear, minute variations 
in the head alignment when they come to a stop make small differences in 
the heads ability to read the sector.  Sometimes, this is enough to get 
a good read from the sector and fully recover all the data immediately.  
If so, it is relocated to a safe place.

IN ADDITION, Spinrite keeps in memory all the raw data that it has read 
on each of those 2000 attempts.  If a good read cannot be achieved 
during the 2000 attempts, Spinrite performs an exhaustive statistical 
analysis of those 2000 copies of the various sector reads and tries to 
extrapolate the most likely contents of the 512 bytes (or 4KB) of the 
sector.  It is thus able to sometimes recover the whole sector, and 
sometimes able to recover a PARTIAL sector.  However, that data is 
frequently enough to make the OS happy trying to read the file 
allocation table, or the user application happy trying to read the meta 
data of a document, etc.  If it works, it thus, it makes the drive 
usable, or at least more usable, where it wasn't before.

By the way, regarding specialized machines and controllers, this 
discussion is not about what a data recovery professional with a 
specialized lab can do with a drive.  It's about what a consumer with 
minimal to above average technical abilities can do with it in the field 
with little or no outside help.

>
> "One potential disadvantage of using dd-rescue first is that it could 
> force the HDD to spare out sectors that are not easily readable" 
> Sector remapping is done at the firmware level, Spinrite can't prevent 
> it, ddrescue can't prevent it, no software tool can prevent it (except 
> for maybe MHDD on certain Seagates if you have the firmware commands). 
> you need something lower level that can disable sector remaps at the 
> firmware level temporarily while you image the drive.
>

This is also wrong.  Spinrite has been preventing sector remapping since 
LONG ago, as a prelude to its data recovery work.  See the following link:

http://www.grc.com/srrecovery.htm

quote ->

The FIRST THING SpinRite does when it starts examining and working with 
a drive is to completely disable the drive's built-in automatic sector 
relocation. This way the drive can't whisk the sector away the first 
time it's not easily read, and SpinRite can study the sector to recover 
its data as much as necessary. (And as you might guess by now, no other 
utility is even aware of this problem.)

<-- end quote

ddrescue probably cannot do this, as you said, but Spinrite can and 
does.  One reason that this is possible is that, when Spinrite is 
running, it's the only thing running (on top of FreeDos).  It has 
exclusive control of the machine and the controller.  When ddrescue is 
running, it is limited in what it can do by what the OS will allow, even 
if running as super user.

For this reason alone, if I didn't think a drive had mechanical 
problems, only soft errors, I would probably be inclined to use Spinrite 
on it before trying something like ddrescue.  If I thought an exhaustive 
data analysis might make the drive worse, I might go the other way.

> The reason Spinrite sucks is that it recovers data back onto the drive 
> itself

Wrong.  First, the statement is inherently biased.  Second, Spinrite is 
targeted to average to above average consumers, but not uber geeks.  It 
is meant to help people which overwhelmingly (percentage wise) are 
Windows users, who know little or nothing about Linux, who probably 
wouldn't know (at least in the past) how to install a backup drive, and 
who don't really have access to any viable alternative to recover 
unreadable data.  Oh, and like most consumers, they probably don't have 
a backup.  In this context, restoring to the original drive, but back on 
to good sectors, is the only viable option for many of the target 
audience.  Having said that, it would be cool if a later version offered 
data saving to a different drive for the more advanced users.

> and is marketed as a magic bullet

Wrong.  Steve has never marketed the product this way in the 7 years 
I've been listening to the podcast.  He markets it as a potential way to 
recover otherwise unrecoverable and unusable hard drives which have soft 
failures (as opposed to the heads melting).  He never says it always 
works, but says that it does work in many cases.  And all that is 
exactly true.  The fact is, that this product has helped thousands of 
users, including myself, recover data and drives that they probably 
would have had to sacrifice otherwise.

> for data recovery, when in fact, it's incredibly disingenuous to sell 
> a tool that breaks the number one rule of all data recovery.

Wrong.  See discussion above about the logic of recovery in place.

> Spinrite can do nothing that ddrescue can't do, and in fact, ddrescue 
> is a bit more robust as far as error recovery and granularity goes, 
> plus it actually gives you a usable dd image.

Wrong again.  Spinrite does far more than ddrescue is capable of.  See 
discussion above.

> GRC is a blowhard idiot at data recovery and at networking

A) Wrong.  B) BIASED  C) EMOTIONAL  D) Not based on FACTS or LOGIC or 
EVIDENCE or EXPERIENCE or EXPERIMENT.

and therefore it is

E) IRRELEVANT

> (omg raw sockets in Win XP, the sky is falling!!).
>

Not relevant to this discussion.  HOWEVER, it is a FACT that, Steve was 
instrumental in persuading MS to drop raw sockets.  It is also a FACT 
that, should they have kept the feature in there, potentially millions 
of malicious users of Windows, with minimal training, minimal resources, 
and minimal difficulty, would have access to a much easier time 
perpetrating attacks on other users.

Perhaps you should do a bit more FACT CHECKING before trashing someone's 
product or their reputation in public.

Sincerely,

Ron


>
> On Thu, Feb 21, 2013 at 1:17 PM, Ron Frazier (ALE) 
> <atllinuxenthinfo at techstarship.com 
> <mailto:atllinuxenthinfo at techstarship.com>> wrote:
>
>     Wrong, sort of.  I am a Spinrite user and have had good results in
>     recovering data on a few occasions.  It is a valuable tool and has
>     data recovery capabilities that nothing else I'm aware of does.
>      HOWEVER, you have to use it the right way.  I'm away from my desk
>     right now and will write up a detailed description of how to use
>     Spinrite properly when I get back.
>
>     Spinrite can often recover sectors which are unreadable sectors
>     which other products cannot.  There are literally thousands of
>     accounts where Spinrite has recovered unusable drives enough where
>     they were either usable or they were at least where the data could
>     be recovered.
>
>     I will agree that you should first try to recover the readable
>     sectors and save them away.  Use dd-rescue, or GNU dd-rescue or
>     something else.
>
>     http://www.gnu.org/software/ddrescue/ddrescue.html
>
>     You may wish to google dd rescue.  I know I've read that there are
>     more advanced alternatives than the original dd-rescue.
>
>     Use one of these to read the readable sectors.  Save that data on
>     a different hard drive.
>
>     THEN, use Spinrite to potentially read the sectors that the rescue
>     program could not.  At this point, you have nothing to lose, since
>     you've recovered what you can and what remains is, by definition,
>     not recoverable by the rescue program.
>
>     One potential disadvantage of using dd-rescue first is that it
>     could force the HDD to spare out sectors that are not easily
>     readable.  Once that's done, even Spinrite won't be able to touch
>     them.  I don't know specifically how to prevent that behavior from
>     the drive.  But, I know that Spinrite prevents sector sparing
>     until it has finished its data recovery analysis.
>
>     I hope to write up a more detailed explanation of how to use
>     Spinrite later today.
>
>     Sincerely,
>
>     Ron
>
>
>
>
>     dev null zero two <dev.null.02 at gmail.com
>     <mailto:dev.null.02 at gmail.com>> wrote:
>
>     >you definitely shouldn't give spinrite a try if you value your data.
>     >recovering data BACK ONTO DAMAGED MEDIA is the number 1 no-no in data
>     >recovery. just buy a new drive and ddrescue your data to it.
>     >
>     >
>     >On Thu, Feb 21, 2013 at 11:43 AM, Scott Plante
>     ><splante at insightsys.com <mailto:splante at insightsys.com>>wrote:
>     >
>     >> You should definitely give Spinrite a try. Watch the video for a
>     >detailed
>     >> explanation of how it works.
>     >> http://www.grc.com/sr/spinrite.htm
>     >>
>     >> ------------------------------
>     >> *From: *"Jonathan Meek" <jonathan.l.meek at gmail.com
>     <mailto:jonathan.l.meek at gmail.com>>
>     >> *To: *"Atlanta Linux Enthusiasts" <ale at ale.org
>     <mailto:ale at ale.org>>
>     >> *Sent: *Thursday, February 21, 2013 8:31:33 AM
>     >> *Subject: *[ale]  Possibly bad hard drive
>     >>
>     >>
>     >> Hey guys,
>     >>
>     >> I believe my hard drive is about to give up the ghost hard.
>     Saturday
>     >the
>     >> system booted without issue then Monday, I booted the system and
>     >tried to
>     >> start Firefox. The taskbar freaked out and I had to do a hard
>     >shutdown.
>     >> After multiple restarts I was able to get the system back up but at
>     >all the
>     >> restarts, I got a error message that stated that it couldn't find a
>     >> particular directory and to press Ctrl+Alt+Del to restart the
>     system.
>     >>
>     >> Finally I got the system to give the prompt for entering in my
>     >harddrive
>     >> password (I have an encrypted hard disk that I setup when I did a
>     >fresh
>     >> install of lubuntu last time). It checked for errors and found
>     some,
>     >it
>     >> tried to repair them and hung at mounting /tmp.
>     >>
>     >> I restarted the system and this time it rebooted without issue
>     I got
>     >all
>     >> the way to the home screen and logged in. Launched Firefox without
>     >issue
>     >> and goofed around for a few minutes while I let my backup system
>     >backup for
>     >> the final time (in fear of never getting it back).
>     >>
>     >> Shutdown the system and restarted it with a Ubuntu 12.04 live CD in
>     >order
>     >> to do check the hard drive. Went into Disk Utility and the system
>     >> recognized I had a hard drive but when I tried any of the
>     benchmarks
>     >it
>     >> balked at me saying it couldn't read as well as the SMART
>     Status said
>     >"not
>     >> applicable". This might be from the encryption but I don't know.
>     >>
>     >> Exited out of the live CD, boot the system again, and it booted
>     >without
>     >> incident. Tried to do software update, it griped at me saying that
>     >there
>     >> was not enough room in /boot to do an update and to use sudo
>     apt-get
>     >clean.
>     >> Run sudo apt-get clean and tried the update again. Same error
>     >message.
>     >> Repeated this step 5 times before giving up.
>     >>
>     >> I am not sure what to do at this stage with it because I can't seem
>     >to
>     >> wipe the drive probably due to the encryption because I tried to
>     >install
>     >> Ubuntu 12.04 since I had a backup of all my data.
>     >>
>     >> All that backstory was to ask this one question: Is there anything
>     >else I
>     >> can do to give some level of assurance the actual status of the
>     hard
>     >drive?
>     >> I think it is busted but I am not sure.
>     >>
>     >> Regards,
>     >>
>     >> Jonathan
>     >>
>     >> _______________________________________________
>     >> Ale mailing list
>     >> Ale at ale.org <mailto:Ale at ale.org>
>     >> http://mail.ale.org/mailman/listinfo/ale
>     >> See JOBS, ANNOUNCE and SCHOOLS lists at
>     >> http://mail.ale.org/mailman/listinfo
>     >>
>     >>
>     >> _______________________________________________
>     >> Ale mailing list
>     >> Ale at ale.org <mailto:Ale at ale.org>
>     >> http://mail.ale.org/mailman/listinfo/ale
>     >> See JOBS, ANNOUNCE and SCHOOLS lists at
>     >> http://mail.ale.org/mailman/listinfo
>     >>
>     >>
>     >
>     >
>     >------------------------------------------------------------------------
>     >
>     >_______________________________________________
>     >Ale mailing list
>     >Ale at ale.org <mailto:Ale at ale.org>
>     >http://mail.ale.org/mailman/listinfo/ale
>     >See JOBS, ANNOUNCE and SCHOOLS lists at
>     >http://mail.ale.org/mailman/listinfo
>
>
>     --
>
>     Sent from my Android Acer A500 tablet with bluetooth keyboard and
>     K-9 Mail.
>     Please excuse my potential brevity if I'm typing on the touch screen.
>
>     (PS - If you email me and don't get a quick response, you might
>     want to
>     call on the phone.  I get about 300 emails per day from alternate
>     energy
>     mailing lists and such.  I don't always see new email messages
>     very quickly.)
>
>     Ron Frazier
>     770-205-9422 <tel:770-205-9422> (O)   Leave a message.
>     linuxdude AT techstarship.com <http://techstarship.com>
>
>
>     _______________________________________________
>     Ale mailing list
>     Ale at ale.org <mailto:Ale at ale.org>
>     http://mail.ale.org/mailman/listinfo/ale
>     See JOBS, ANNOUNCE and SCHOOLS lists at
>     http://mail.ale.org/mailman/listinfo
>
>
>
> _______________________________________________
> Ale mailing list
> Ale at ale.org
> http://mail.ale.org/mailman/listinfo/ale
> See JOBS, ANNOUNCE and SCHOOLS lists at
> http://mail.ale.org/mailman/listinfo
>    

-- 

(PS - If you email me and don't get a quick response, you might want to
call on the phone.  I get about 300 emails per day from alternate energy
mailing lists and such.  I don't always see new email messages very quickly.)

Ron Frazier
770-205-9422 (O)   Leave a message.
linuxdude AT techstarship.com

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.ale.org/pipermail/ale/attachments/20130221/293dedc4/attachment-0001.html>


More information about the Ale mailing list