[ale] Which large capacity drives are you having the best luck with?

Wed Jan 5 11:23:47 EST 2011

Paul,

Thanks for the note.  I'm glad the info was helpful.  It does take a
while to compose an email like that.  I saved a bookmark to the article
you mentioned.  Note to self, do something today besides email.  Here
are a few tidbits that may not have been obvious or weren't stated
before.

Don't forget that you can run smartctl graphically with gsmartcontrol.
Both are available in the Ubuntu repository.  You can also access the
smart data and test from the Ubuntu Disk Utility.

Keep in mind I'm not promising results from Spinrite.  I'm just a
customer.  Also, some of my opinions about hard drives could be wrong!

You must run Spinrite from its own boot disc, which you create with the
installation program.  I would create at least 2 CD's.  Once the boot
disk is created, it can be run on disks with any OS on them, even TIVO
(by putting the drive in a pc).  You also have to make sure your BIOS is
set so you can boot from a CD.  On my Dell laptop, I press F12 at boot
time to get a menu to select the boot device.  On my desktop machine,
it's F11.  You could boot from a floppy, if you had to.  Each disc is
watermarked in the program with the owner's registration information.
Spinrite has to be the only thing running in the machine, and it brings
its own OS along.  Your machine will be unusable for anything else while
running tests.  Note that you CAN abort the test at any time and resume
where you left off with no harm.  So, if you want to run the machine in
the day time and restart the test every night, you can do that.

You may wish to back up your drives before running Spinrite.  It works
them very hard, and if they're on the verge of mechanical failure, it
could push them over the edge.  It tries to be as safe as possible
though.  Make as certain as possible that there are no power failures
while running disk tests.  This is very likely to cause corruption which
may not be correctable.  Ask me some other time about how NOT to test a
UPS.

I'm still looking for a comprehensive solution which will backup Linux
and Windows file systems without resorting to sector level copying.  I
like the idea of a bootable backup, where you just boot your backup
drive and work from it.  Then you replace and rebuild your main drive,
or use the backup as the main and rebuild the backup.

The Google hard drive study was way too technical for me to read
completely yesterday.  I mainly read summaries.  However, if I
understand it correctly, even though age for middle aged drives is not
as big a factor as you might think, they do tend to fail at or before
the 1 year time frame (infant mortality) and the 4 - 5 year time frame.
Your drives are definitely getting toward the old side.  You might
consider Jungledisk ( https://www.jungledisk.com/ ) for continuous
online backup.  They have versions for Windows and Linux.  Not sure
about Mac.  I have it running on my systems every 6 hours.

Note that, if the drive is not throwing errors, the result you want from
Spinrite is NOTHING.  If you ask a fireman if the house is on fire, you
want the answer to be NO.

However, even if there are not obvious results, the intensive read
rewrite process will refresh weak data and improve the reliability of
the drive.  If data is unreadable, it will go to EXTREME measures to
recover it.  Even normal operations are slow.  It just so happens that
I'm running Spinrite at the moment on my Son's laptop with a 250 GB
drive.  I'm using the most intensive level of analysis.  It's almost
finished after 20 hours with no errors.  I think my hard drives are a
bit faster than my Son's.

If it hits a snag, a patch of unreadable data, all bets are off.  It
will go into a statistical data recovery mode, and will try up to 2000
times to recover EACH piece of data.  This is repeated as many times as
necessary.  This process may take weeks to get through if the drive has
many errors.  Just let it run and ignore it, even if it looks like it's
frozen.  You can choose to abort, replace the drive, and restore a
backup, etc.  There is a high likelihood of recovering the data if you
let it run.  But, you cannot tell which "file" it's working on or tell
it to ignore the problem.  It's totally obsessive about recovering the
data.  It works at the sector level and doesn't care about files.  As
mentioned before, you can abort the test and restart later if desired.

Don't be alarmed if you see lots of seek errors on the SMART monitoring
screen.  Seagate drives, in particular, show huge numbers here.  I don't
think they are accurate.  Other drives may do that as well.  Apparently
it's fairly common.  It does NOT mean the drive is about to blow up.  I
would confirm any errors on the SMART monitoring screen with another
application, as that's not Spinrite's main purpose.  The video, linked
below, talks about how to interpret smart data.  One useful piece to
monitor is the error rate.

The usage of the Spinrite menus is non obvious at first.  Call me at
404-431-5472 and I'll be glad to walk you through the procedure and
follow along on my own machine, except for the waiting 4 days part.
Trying to describe it in email would be quite difficult.

Based on what I've learned here on the group.  After running Spinrite, I
would bring up Disk Utility under the system administration menu of
Ubuntu and run the smart long test.  Then I would look for any warning
messages.  If I get all green lights, I would feel good about relying on
the drive.  Although it's a pain, and time consuming, I try to run
Spinrite every 3-4 months.

http://www.grc.com/sr/spinrite.htm - You can find technical data about
Spinrite here.

http://www.grc.com/sr/themovie.htm - Here's a neat video interview with
the inventor.  The video is a little old, but still relevant.

https://encrypted.google.com/search?hl=en&as_q=spinrite&as_epq=&as_oq=&as_eq=&tbs=&num=100&lr=&as_filetype=pdf&ft=i&as_sitesearch=grc.com&as_qdr=all&as_rights=&as_occt=any&cr=&as_nlo=&as_nhi=&safe=images - Sorry for the length of that.  It will search Google for all references to Spinrite, in PDF files, on the Gibson research site, and show 100 entries at a time.  This can also be done manually from the advanced search page.  This is a good way to find specific technical data.  Any file name such as sn-127.pdf is a transcript of one of Steve's podcast.

If you have any questions, feel free to contact me through the list, or
personally by email or phone.  I don't always see ALE messages every
day.

Sincerely,

Ron

On Wed, 2011-01-05 at 06:21 -0500, Paul Cartwright wrote:
> Ron, wow, what an article!
> my problem is(was), I never figured out how to USE smartctl.. so after
> reading YOUR article I found this:
>  
> 
> 
>   Monitoring Hard Drive Health on Linux with smartmontools
> 
> 
> http://blog.shadypixel.com/monitoring-hard-drive-health-on-linux-with-smartmontools/
> 
> 
> I think I will give Spinrite a try.. since all of my HDs are 3 years old
> OR OLDER... and stay running 24/7/365. well, not the laptops, but my
> laptop drive has given me errors every time I install ubuntu.. It is a
> Dell XPS from 2006. BUT my new gig just dropped a nice new HP Probook
> 6440b on me, so my older laptop will now sit for a while:)
> 

-- 

(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 messages very quickly.)

Ron Frazier

770-205-9422 (O)   Leave a message.
linuxdude AT c3energy.com