Re: Is my hard disk about to crash?
On Tue, 11 Jan 2005 09:15:29 -0500, H. S. <greatexcalibur@yahoo.com> wrote:
> Apparently, _Alvin Oga_, on 10/01/05 18:16,typed:
> > disk drive manufacturers spec their "max" operating temp,
> > and does NOT mean you can run at that temp 24x7x365
> >
> > the disks are warranted for 1,000,000 too .. but we're not worried
> > about that?? and we only care about their 1year warranty ( 6,000hrs )
> >
> > the life of the disks dies at a rate of 1/2 life cycle
> > for each 10C increase above 25C at "normal" operating condition
> > which could be "for use only for 2 hrs of the day" or something
> > silly ( everybody knows, laptop disks are not meant to be
> > used 24x7 )
Actually, the measures are usually in Mean Time Between Failures.
Take a look at http://www.hitachigst.com/hdd/technolo/drivetemp/drivetemp.htm
(hitachi)
and http://www.digit-life.com/articles/storagereliability/ (seagate).
Look at the temperature derating factor:
25C is normal operating conditions, 100% duty at 2400 POH/year.
At 34C a 35% decrease in the MTBF rate.
At 38C a 46% decrease in the MTBF rate.
> OMG, you have me worried:
> ~# hddtemp /dev/hda
> /dev/hda: WDC WD800JB-00FMA0: 41 C
I have the same model HD and it runs at 35 C, also 24/7.
I really suggest installing smartmontools.
You can read the smart error table from the HD, it contains the error
log of the disk. And the deamon runs scheduled self tests and warns in
case of problems. You can also see the error rate of your HD.
If you're really worried then buy a HD cooler. They are not that expensive.
The Power on Hours (POH) are a huge factor in the MTBF, normal
operating condirions are 2400 POH/year, 24/7 is 8760 POH/year (usually
a 50% decrease in MTBF).
All tests are done at 100% duty cycle, but my server does not have
that load. So that makes the MTBF go up again. Just read the second
article and do the math.
-Olaf
Reply to: