[ale] disk drive diagnostics nirvana - NOT - I have questions

[Ron Frazier (ALE)]

On 10/23/2012 9:17 AM, mike at trausch.us wrote:
> On 10/22/2012 06:12 PM, Ron Frazier (ALE) wrote:
>    
>>    Since the HDD
>> controller can usually only discover read / write problems when you
>> actually access the sector, I've developed a practice over the years to
>> read and write every sector on the hard drive a few times per year.
>>
>> If the sector
>> doesn't read, Spinrite uses advanced statistical algorithms to try up to
>> 2000 times to recover the data.
>>      
> Not really.  It sends the same command over and over again, because it
> has no direct control of anything (it is, just like BIOS, just like the
> Linux kernel, only allowed to talk to the IDE/SATA controller, not to
> the drive directly).  It is therefore constrained by the controller!
>
> Anyway, that's all I'm going to say since SpinRite isn't what this
> thread is about.

Hi Mike T,

I also don't want to totally derail this thread.  However, according to 
the documentation on Steve's website (as I recall), what you stated is 
not correct.  According to him, SR flies the heads to the target sector 
that's giving trouble many many times from different starting positions 
and at different velocities (I think) which slightly change the 
alignment when the heads stop.  SR reads the sector data, whether or not 
the system is reporting that it has an error.  It saves all this data in 
memory.  If it gets a good read, it relocates the data to a safe place.  
If it cannot read the sector correctly after 2000 tries, which means no 
other program likely would either, it does exhaustive statistical 
analysis on the 2000 samples it has accumulated and extrapolates the 
most likely contents of the sector and writes those to the disk.  Often, 
this will recover all the sector's data.  Sometimes it recovers only 
some of it correctly.  Sometimes, it recovers none, in which case, the 
sector is marked as unrecoverable.  The point is, the data was 
completely unreadable, and therefore totally lost by any other 
conventional means.  Sometimes, even a partial recovery of data in the 
sector to its original state is good enough to allow a non booting drive 
to boot, or an unreadable file or folder to be read, which is probably 
better than not having that item read at all.  I have personally had 
this experience with a failing drive, where the drive was inaccessible, 
and SR recovered it enough to allow me to boot it and get the data off, 
then replace it.  Running SR periodically (or badblocks in non 
destructive read write mode, I presume) will, hopefully, allow the 
controller to find weak sectors and map them out before they become a 
problem.  As far as I know, badblocks does not have the extensive data 
recovery function in case the sector refuses to read by conventional 
tactics.

By the way, the thread is about data reliability and HDD maintenance, so 
this is still related.

Sincerely,

Ron


-- 

(To whom it may concern.  My email address has changed.  Replying to former
messages prior to 03/31/12 with my personal address will go to the wrong
address.  Please send all personal correspondence to the new address.)

(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