[GTALUG] Anyone familiar with SMART?

[D. Hugh Redelmeier]

On Wed, 23 Dec 2020 at 15:18, Giles Orr <gilesorr at gmail.com> wrote:
>
> I just bought two new, identical, external USB hard drives.  While
> I've used SMART on and off over the years, I think this is the first
> time I used it immediately on connecting a drive.  I'm using
> 'gsmartcontrol', a GUI that I'm liking.  In the "Statistics" tab I'm
> seeing, under "Rotating Media Statistics (rev 1)", "Number of
> Mechanical Start Failures" = "1".  This is straight out of the box,
> brand new.  Where this gets really weird is that both of the drives
> show exactly the same error.  These are WD easystore drives.  So ...
> is this something I should be worried about?

I'll respond since nobody else has.

I am no expert.  I tend to use S.M.A.R.T. when in distress.

- many USB chips in external drives do not implement SMART.  Yours
  seems to, so that's good

- I've found that SMART fields often don't mean what one would think
  that the mean.  (I look at them with smartctl -x)

- sometimes I capture the output of smartctl -x, do some things to the 
  disk, capture another smartctl -x, add then look at the diff of the two 
  logs.  Actually, "meld" is better than diff for this.

- smartctl has a table of drives but that doesn't always cover the
  drive I'm trying to study.

- I have a superstitious belief that some drives don't report their
  problems if they think that they can cover them up.  "Health" is
  such a nebulous concept.

Having said all that, I'd pretend that there is no problem.

I'd also try an experiment.  Can you actually bump that count by
unplugging the drive?  (This is assuming that you haven't actually got
anything of value on it.) (I don't know if your drives are USB-powered
or have a separate power supply.)

Are these "shingled" drives?  I don't know if they need more time for
a graceful power-down.