SMART Reallocated Sector Count 269, Reallocated Event Count 20

I have an approximately 2-year Toshiba MK6476GSXN HD on my laptop (my only computer). Recently, I plugged in a faulty SD/USB converter, which caused all sorts of boot failures. Before I identified the failed converter, I thought that my HD was dying. I ran cygwin's smartmontools and compared it to results from 2 years ago. To interpret the results, I relied on google, and primarily the Wikipedia page http://en.wikipedia.org/wiki/S.M.A.R.T. . Higher normalized values are better, and it looks like 100 was selected as the healthy reference value for most of the attributes.

Selected columns from the smartctl output are shown at http://i58.tinypic.com/6psq3q.png, with the current output (as opposed to 2 years ago) being a hybrid of using smartctl switches -x and -a. I'm not entirely sure whether 3 of these attributes are abnormal, and these are marked with triple asterisks. The middle marked attribute (#191, G-Sense_Error_Rate) is probably not an indicator of HD health, as the errors are due to physical jarring. The remaining two attributes are #5 (Reallocated_Sector_Ct=269) and #196 (Reallocated_Event_Count=20). There is no clear consensus on whether this is bad (e.g., see http://superuser.com/questions/26842...ocate-sectors), so I intend to keep an eye on it and migrate off the HD in the long term.

I really dread the time commitment to do this, as I do occassional snapshots of user content (admin & non-admin accounts) by selectively using find, tar, and zipping to disc. I haven't gotten into the habit of using my $60 Norton backup utility (again from 2 years ago) because I want to see myself that I've selected all relevant content, and only relevant content. Where such content resides tends to shift over the years.

Can those with more experience with HD technology comment on this, based on the smartctl results? Am I missing something important that I haven't already noted?

wrote

I have an approximately 2-year Toshiba MK6476GSXN HD
on my laptop (my only computer). Recently, I plugged in a
faulty SD/USB converter, which caused all sorts of boot failures.
Before I identified the failed converter, I thought that my HD
was dying. I ran cygwin's smartmontools and compared it to
results from 2 years ago. To interpret the results, I relied on
google, and primarily the Wikipedia page
http://en.wikipedia.org/wiki/S.M.A.R.T. .
Higher normalized values are better, and it looks like 100 was
selected as the healthy reference value for most of the attributes.

It's the raw values that matter.

Selected columns from the smartctl output are shown at
http://i58.tinypic.com/6psq3q.png, with the current output
(as opposed to 2 years ago) being a hybrid of using smartctl
switches -x and -a.

I'm not entirely sure whether 3 of these attributes are
abnormal, and these are marked with triple asterisks.

Yes, the reallocated sectors is very bad and the reallocated event count is
bad.

The middle marked attribute (#191, G-Sense_Error_Rate) is probably
not an indicator of HD health, as the errors are due to physical jarring.

Yes, but may well be the cause of the reallocated sectors.

The remaining two attributes are #5 (Reallocated_Sector_Ct=269)

That is very bad.

and #196 (Reallocated_Event_Count=20).

And that is bad too.

There is no clear consensus on whether this is bad

Only because there is so much misinformation on the web.

(e.g., see
http://superuser.com/questions/26842...ocate-sectors),

A few are acceptable because some drives can
produce one with a mains failure, but not a laptop.

269 is an obscene number of reallocated
sectors and indicate a real problem.

And you certainly can not trust a drive with that many,
you should be backing up what matters all the time.

so I intend to keep an eye on it and migrate off the HD in the long term.

I really dread the time commitment to do this, as I do
occassional snapshots of user content (admin & non-admin
accounts) by selectively using find, tar, and zipping to disc.

That's dangerous because its easy to forget something and
only discover that you have forgotten something important
when the drive actually dies and you need to use the backup.

I haven't gotten into the habit of using my $60 Norton
backup utility (again from 2 years ago) because I want
to see myself that I've selected all relevant content, and
only relevant content. Where such content resides tends
to shift over the years.

And its easy to forget something.

Can those with more experience with HD technology
comment on this, based on the smartctl results?

The drive is dying.

Am I missing something important that I haven't already noted?

Yes, that the drive is dying.

On Saturday, August 30, 2014 6:23:15 PM UTC-4, Rod Speed wrote:
andymhancockATgmailDOTcom wrote
I ran cygwin's smartmontools and compared it to results from 2
years ago. To interpret the results, I relied on google, and
primarily the Wikipedia page
http://en.wikipedia.org/wiki/S.M.A.R.T. . Higher normalized values
are better, and it looks like 100 was selected as the healthy
reference value for most of the attributes.

It's the raw values that matter.

The threshold seems to be for the normalized values. Are the thresholds of any value in terms of information?

The remaining two attributes are #5 (Reallocated_Sector_Ct=269)

That is very bad.

and #196 (Reallocated_Event_Count=20).

And that is bad too.

...snip...

269 is an obscene number of reallocated sectors and indicate a real
problem.

And you certainly can not trust a drive with that many,
you should be backing up what matters all the time.

...snip...

Yes, that the drive is dying.

For a 600MB HD, are thre reasonable valuse be for Reallocated_Sector_Ct and Reallocated_Event_Count?

wrote
Rod Speed wrote
andymhancockATgmailDOTcom wrote

I ran cygwin's smartmontools and compared it to results from 2
years ago. To interpret the results, I relied on google, and
primarily the Wikipedia page
http://en.wikipedia.org/wiki/S.M.A.R.T. . Higher normalized values
are better, and it looks like 100 was selected as the healthy
reference value for most of the attributes.

It's the raw values that matter.

The threshold seems to be for the normalized values.

Yes.

Are the thresholds of any value in terms of information?

Only with at few odd attributes, most obviously with
the ones that Seagate does differently where the raw
values don't have any useful information from them.

The remaining two attributes are #5 (Reallocated_Sector_Ct=269)

That is very bad.

and #196 (Reallocated_Event_Count=20).

And that is bad too.

...snip...

269 is an obscene number of reallocated
sectors and indicate a real problem.

And you certainly can not trust a drive with that many,
you should be backing up what matters all the time.

...snip...

Yes, that the drive is dying.

For a 600MB HD, are thre reasonable valuse be for
Reallocated_Sector_Ct and Reallocated_Event_Count?

Low single digit reallocated sectors is fine, but that's it.

The reallocated event count isnt significant, it's
the number of reallocated sectors that matters.

The other one which can be significant is the
pending sector count but isnt with your drive.

On Sunday, August 31, 2014 3:07:34 AM UTC-4, Rod Speed wrote:
andymhancockATgmailDOTcom wrote
Rod Speed wrote
andymhancockATgmailDOTcom wrote
I ran cygwin's smartmontools and compared it to results from 2
years ago. To interpret the results, I relied on google, and
primarily the Wikipedia page
http://en.wikipedia.org/wiki/S.M.A.R.T. . Higher normalized
values are better, and it looks like 100 was selected as the
healthy reference value for most of the attributes.

It's the raw values that matter.

The threshold seems to be for the normalized values.

Yes.

Are the thresholds of any value in terms of information?

Only with at few odd attributes, most obviously with the ones that
Seagate does differently where the raw values don't have any useful
information from them.

269 is an obscene number of reallocated sectors and indicate a
real problem. And you certainly can not trust a drive with that
many, you should be backing up what matters all the time.

...snip...

For a 600MB HD, are thre reasonable values for
Reallocated_Sector_Ct and Reallocated_Event_Count?

Low single digit reallocated sectors is fine, but that's it. The
reallocated event count isnt significant, it's the number of
reallocated sectors that matters.

The other one which can be significant is the pending sector count
but isnt with your drive.

Thanks, Rod.

On Sunday, August 31, 2014 11:00:50 AM UTC-4, andymh...ATgmailDOTcom wrote:
On Sunday, August 31, 2014 3:07:34 AM UTC-4, Rod Speed wrote:
Low single digit reallocated sectors is fine, but that's it. The
reallocated event count isnt significant, it's the number of
reallocated sectors that matters.

The other one which can be significant is the pending sector count
but isnt with your drive.

Thanks, Rod.

I should add, this is the drive that replaced the HD 2 years ago,
which exhibited problems after a few months. I'm utterly disappointed
by the fragility of HDs. Whatever happened to the concept of brand
reputation?

wrote
andymh...ATgmailDOTcom wrote
Rod Speed wrote

Low single digit reallocated sectors is fine, but that's
it. The reallocated event count isnt significant, it's
the number of reallocated sectors that matters.

The other one which can be significant is the
pending sector count but isnt with your drive.

Thanks, Rod.

I should add, this is the drive that replaced the HD 2
years ago, which exhibited problems after a few months.

Thought we had had previous history but didn't think to check.

I'm utterly disappointed by the fragility of HDs.

I haven't ever lost one myself in the last decade or so.
The kids did manage to kill one by kicking it in a very
small Atom powered desktop brick thing that they
had on the floor where it was kickable.

The G shock value is a worry with yours, did you drop it ?

Whatever happened to the concept of brand reputation?

You don't see those Toshibas dying like flys.
Laptops do have a very hard life.

On Sunday, August 31, 2014 3:15:26 PM UTC-4, Rod Speed wrote:
andymhancockATgmailDOTcom wrote:
I'm utterly disappointed by the fragility of HDs.

I haven't ever lost one myself in the last decade or so. The kids
did manage to kill one by kicking it in a very small Atom powered
desktop brick thing that they had on the floor where it was
kickable.

The G shock value is a worry with yours, did you drop it ?

Whatever happened to the concept of brand reputation?

You don't see those Toshibas dying like flys. Laptops do have a
very hard life.

This is good to know, Rod. I won't be paranoid about laptops now.

As for whether I dropped it, I think this is in fact the one I
dropped. But I think the SMART reports I posted back then (which is
where I pulled the old data from for this post) did not indicate a
problem. Someone, might have been you, explained that it might be
because the drive was off -- I can't be sure, I just know that someone
more knowledgeable than I proposed that explanation.

Since then, the laptop has been given the princess treatment. I make
sure that the Toshiba shock notification app is active so that I get
immediate feedback of when I'm not being gentle enough. It hardly
ever leaves the home (a 1-bedroom apartment) and hardly ever moves
around in the apartment. So I'm not sure what could have happened
since the to give it the G shock value. However, I'll take your
anecdote of robustness as a good "data point".

Anyway, I finished the backup after reading all the past caveats on
using Norton Ghost 15. It boots and I'm using it now. The drive
health seems much better (http://i57.tinypic.com/pwvoz.png).
Raw values of key attributes from the S.M.A.R.T. wikipedia link:

ID#5 = Reallocated_Sector_Ct = 0

ID#10 = Spin_Retry_Count = 0

ID#184 = End-to-End_Error = 0

ID#188 = Command_Timeout = 0

ID#196 = Reallocation Event Count -- DOESN'T EXIST !!

ID#197 = Current_Pending_Sector = 0

ID#198 = Offline_Uncorrectable = 0

ID#201 =
Soft Read Error Rate or TA Counter Detected -- DOESN'T EXIST !!

ID#230 = Drive Life Protection Status -- DOESN'T EXIST !!


Really odd that 196 doesn't exist in the new drive, but it does on the
old drive.

Also, it's odd that there are no threshold values. Perhaps thresholds
aren't useful, since none of the problematic values in the old drive
were close to their thresholds.

Thanks for all your help!

wrote:
I have an approximately 2-year Toshiba MK6476GSXN HD on my laptop (my only
computer). Recently, I plugged in a faulty SD/USB converter, which caused
all sorts of boot failures. Before I identified the failed converter, I
thought that my HD was dying. I ran cygwin's smartmontools and compared
it to results from 2 years ago. To interpret the results, I relied on
google, and primarily the Wikipedia page
http://en.wikipedia.org/wiki/S.M.A.R.T. . Higher normalized values are
better, and it looks like 100 was selected as the healthy reference value
for most of the attributes.

Selected columns from the smartctl output are shown at
http://i58.tinypic.com/6psq3q.png, with the current output (as opposed to
2 years ago) being a hybrid of using smartctl switches -x and -a. I'm not
entirely sure whether 3 of these attributes are abnormal, and these are
marked with triple asterisks. The middle marked attribute (#191,
G-Sense_Error_Rate) is probably not an indicator of HD health, as the
errors are due to physical jarring. The remaining two attributes are #5
(Reallocated_Sector_Ct=269) and #196 (Reallocated_Event_Count=20). There
is no clear consensus on whether this is bad (e.g., see
http://superuser.com/questions/26842...ocate-sectors),
so I intend to keep an eye on it and migrate off the HD in the long term.

I really dread the time commitment to do this, as I do occassional
snapshots of user content (admin & non-admin accounts) by selectively
using find, tar, and zipping to disc. I haven't gotten into the habit of
using my $60 Norton backup utility (again from 2 years ago) because I want
to see myself that I've selected all relevant content, and only relevant
content. Where such content resides tends to shift over the years.

Can those with more experience with HD technology comment on this, based
on the smartctl results? Am I missing something important that I haven't
already noted?

The key question is whether these reallocations showed up due to
an external problem or due to a fault with the disk. Now you have
that converter as possible source of the problems. Keep monitoring
the rellocatiosn and if they grow, even if only slowly, replace
the disk. But they could stay constant as well.

I once had a disk in a server-room that after 1 year suddenly
had about 300 reallocations. As it turned out, they were likely
due to somebody using a power-tool direcly next to the disk while
it was writing data. It continued to run fine 24/7 for another
two years until it was replaced for other reasons.

From the information you have, this can go either way.

Side note: You should monitor smart-attibutes continuously
and run long SMAR selfchecks about every 14-30 days. On
Windows, the "Hard Disk Sentinel" does a good job of that
and it will record the history of attributes.

Arno

On Sunday, August 31, 2014 10:49:03 PM UTC-4, Arno wrote:
andymhancockATgmailDOTcom wrote:
I have an approximately 2-year Toshiba MK6476GSXN HD on my laptop
(my only computer). Recently, I plugged in a faulty SD/USB
converter, which caused all sorts of boot failures. Before I
identified the failed converter, I thought that my HD was dying. I
ran cygwin's smartmontools and compared it to results from 2 years
ago. To interpret the results, I relied on google, and primarily
the Wikipedia page http://en.wikipedia.org/wiki/S.M.A.R.T. .
Higher normalized values are better, and it looks like 100 was
selected as the healthy reference value for most of the attributes.

Selected columns from the smartctl output are shown at
http://i58.tinypic.com/6psq3q.png, with the current output (as
opposed to 2 years ago) being a hybrid of using smartctl switches
-x and -a. I'm not entirely sure whether 3 of these attributes are
abnormal, and these are marked with triple asterisks. The middle
marked attribute (#191, G-Sense_Error_Rate) is probably not an
indicator of HD health, as the errors are due to physical jarring.
The remaining two attributes are #5 (Reallocated_Sector_Ct=269) and
#196 (Reallocated_Event_Count=20). There is no clear consensus on
whether this is bad (e.g., see
http://superuser.com/questions/26842...ocate-sectors),
so I intend to keep an eye on it and migrate off the HD in the long
term.

...snip...

Can those with more experience with HD technology comment on this,
based on the smartctl results? Am I missing something important
that I haven't already noted?

The key question is whether these reallocations showed up due to an
external problem or due to a fault with the disk. Now you have that
converter as possible source of the problems. Keep monitoring the
rellocatiosn and if they grow, even if only slowly, replace the
disk. But they could stay constant as well.

I once had a disk in a server-room that after 1 year suddenly had
about 300 reallocations. As it turned out, they were likely due to
somebody using a power-tool direcly next to the disk while it was
writing data. It continued to run fine 24/7 for another two years
until it was replaced for other reasons.

From the information you have, this can go either way.

Side note: You should monitor smart-attibutes continuously and run
long SMAR selfchecks about every 14-30 days. On Windows, the "Hard
Disk Sentinel" does a good job of that and it will record the
history of attributes.

Understood that the bad SMART readings may be due to external factors. The thing is, I've already migrated off of it to a 500GB Momentus XT hybrid SSD/spinning drive. So the questionable Toshiba HD is no longer in the computer and will not see much action. I did this out of fear instead of staying the course and continuing to use the Toshiba HD with constant monitoring of the SMART parameters. It is in an enclsoure, but in the long term, I'll be trying to keep the Toshiba HD tucked away (if not recycled) simply because I have a very tiny space (which is also the reason why I have this laptop as the sole computer in the apartment).

You mentioned in another thread that one could do conclusive testing. What kind of testing would be appropriately conclusive? Right now, I am thinking of the following (Toshiba does not provide it's own testing tools):

* Windows 7 Disk Check, http://support.microsoft.com/kb/2641432/en-us
* Hitachi Drive Fitness tool
* Seagate's SeaTools
* smartctl long self-test
(I'm only now discovering all the smartctl options!)

There is a hockey sock full of tools at http://pcsupport.about.com/od/toolso.../tophddiag.htm, but I'm thinking that one could spend a great deal of time trying out different tools unless one focuses on identifying the few well-trusted tools. Also, I'm not a big fan of going extreme and downloading a large number of things from a sight that I'm unfamiliar with, just from a cautionary perspective.

You mentioned HD Sentinel -- I think I can set up cygwin's smartctl server to do this and email me, so the history will be stored in gmail, but I have yet to fiddle with it (a long term project,considering other competing priorities and the availability of manual testing). Since the questionable HD will not be connected to the computer in the long term, however, I'm wondering if such a plan is appropriate. I can always use it as a precautionary practice on the new Momentus XT; however, it's partially SDD and regular scans/defrags are not recommended, so I'm wondering if regular integrity tests are also not recommended.