On Tue, 2003-01-07 at 10:52, Tom Brinkman wrote:
> On Monday January 6 2003 08:00 pm, Andrew Miller wrote:
> > I built a server recently for web server, email server, file server
> > and other uses  and am running Mandrake 9.0 on it.  I used a pair
> > of 100 GB IBM Deskstars (120GXP's I believe).  IBM is having
> > problems with these drives failing and now recommends only using
> > them 330 hours a month, something less than 50% of the time (see
> > http://www.sheller.com/ibmpress.htm).  My question is whether my
> > use of these drives in a server application is something I should
> > avoid? Also, is there a setting for power savings that would sleep
> > the drive when it is not getting any hits (most of the time). 
> > Looks l could add a decent SCSI drive that is meant for always-on
> > use. Your feedback is welcome.
> 
>      There's two schools on this, I side with those that believe 
> spinning the drives down is a bad idea.  Sort'a like a lightbulb, 
> they last longer if you're not always turnin 'em off and on ;)
> -- 
>     Tom Brinkman                  Corpus Christi, Texas

I'm in the school that turns stuff off.  I'm very uncomfortable with the
idea of comparing a hard drive with a short circuit.  I'm more amiable
to thinking of a hard drive as having "miles", like a car; which is a
much closer anology.  It's closer because of bearing surfaces; the
number of rotations that the bearings have is finite,  just as the
number of rotations that your car crankshaft has is finite.  You don't
want to leave your car running all the time just to keep it operational.

Most of the time a drive fails not from it's controller electronics, but
rather from sealed bearing failure.  I know a fellow in the data
recovery business, and he is constantly getting old drives spinning long
enough to extract the data.  He does not spend a comparatively long time
replacing drive electronics.

Yes the temperature differentials have an impact, but only in those
cases that don't have good enough ventilation.  Also the impact of
temperature differentials is directly proportional to the scope or
variation bandwidth of those temperature differentials.  In a light bulb
scenario you are talking about probably the most extreme variance
possible in an appliance, since it goes from room temperature to several
hundred degrees in a fraction of a second within power-on.  In a direct
current solid state device, such as a hard drive, this is far from the
case, and even less so with adequate airflow.

I've been involved with hardware since the late 80's, and I still have
hard drives operational from that time.  They are not powered on 24/7.

Just my wooden nickel...

--LX
 

-- 
���������������������������������������������������
Kernel  2.4.18-6mdk     Mandrake Linux  8.2
Enlightenment 0.16.5-11mdk    Evolution  1.0.2-5mdk
Registered Linux User #268899 http://counter.li.org/
���������������������������������������������������


Want to buy your Pack or Services from MandrakeSoft? 
Go to http://www.mandrakestore.com

Reply via email to