HD SMART question

[tmetro-blu-5a1Jt6qxUNc at public.gmane.org (Tom Metro)]

David Rosenstrauch wrote:
> I...decide to check out the drive's status in smartctl before I start
> dumping data onto it. ...it looks like the drive is showing a load of
> smart errors in the "raw read error rate" and the "seek error rate" -
> and that they're increasing!
...
> ATTRIBUTE_NAME      VALUE WORST THRESH WHEN_FAILED RAW_VALUE
> Raw_Read_Error_Rate 100   100   006        -       66806
> Raw_Read_Error_Rate 100   100   006        -       74278

Yeah, but you're looking at the raw value. As I recall, the raw value is 
generally meaningless, as the mechanism by which it gets converted into 
the "cooked" value is proprietary to the manufacturer. The WD drive may 
have a similar raw internal counter that they're just choosing not to 
expose.

Notice that the cooked values are consistent between runs, and appear to 
be at their factory initialized values.

Read through the smartd FAQ (http://smartmontools.sourceforge.net/). I 
believe it addresses this point on raw vs. cooked values.


> Self-tests seem to succeed without error

Because nothing has failed yet. Notice the WHEN_FAILED column just shows 
"-". Until the cooked value for that setting gets below 6, it isn't 
considered a failed state. (Some smart attributes decrement.)


> can I trust this new drive with my backup data, or is this
> new one a lemon too...

See the "drive burn-in" thread from last Fall:
http://www.nabble.com/hard-drive-burn-in-td19576394s24859.html

I'd recommend using some of the techniques recommended there (some of 
which you're already doing) to exercise the drive for a week or so 
before you trust it with data. And of course, make sure you've set up 
smartd to send you notifications if anything changes.

  -Tom

-- 
Tom Metro
Venture Logic, Newton, MA, USA
"Enterprise solutions through open source."
Professional Profile: http://tmetro.venturelogic.com/