Re: Dangerously high Load_Cycle_Count on my Thinkpad T61 Hitachi drive
On Mon, 3 Jun 2013 22:20:20 -0300
Henrique de Moraes Holschuh <hmh@debian.org> wrote:
> On Sun, 02 Jun 2013, Celejar wrote:
> > On Sun, 2 Jun 2013 21:54:09 -0300
> > Henrique de Moraes Holschuh <hmh@debian.org> wrote:
> > > http://www.thinkwiki.org/wiki/Problem_with_hard_drive_clicking
> > >
> > > > 2) What should I do to attempt to preserve whatever life it has left?
> > > > hdparm -B 254/255? Anything else?
> > >
> > > You need to issue that hdparm every time the box had any reason to change
> > > power policy to be sure it stuck. So make sure to reissue it when waking
> > > up, and when AC power is connected/disconnected.
> >
> > Thanks. Any ideas about why the load cycle count is so high and
> > increases so fast? I thought even -B 128 is a pretty standard value.
> > And would you recommend 254 or 255?
>
> It will unload the heads when left idle for a few seconds, presumably so as
> to power off the head actuator. Also, the HDD is much less prone to damage
> due to shock when the heads are unloaded...
Yes, I can understand why it would do that; I just thought that it
sounded rather aggressive, and I became chagrined about what that
portended for my HDD's life expectancy.
> > As to where to put the hdparm invocations, I suppose that writing an
> > "OnResume nn hdparm ..." directive in hibernate.conf will take care of
> > issuing it on wakeup, and setting all *_HD_POWERMGMT=254 in
> > laptop-mode.conf will take care of the AC power connect / disconnect
> > cases?
>
> Laptop mode was enough to deal with it on my T43, yes. Note that I
> configured it manually.
>
> As for the 600k head unload limit on the Hitachi 7K60 and 7K100 HDDs, that
> value comes directly from the data sheet.
Thanks.
Celejar
Reply to: