Apache-Talk @lexa.ru 

Inet-Admins @info.east.ru 

Filmscanners @halftone.co.uk 

Security-alerts @yandex-team.ru 

nginx-ru @sysoev.ru 




      :: Filmscanners
Filmscanners mailing list archive (filmscanners@halftone.co.uk)

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[filmscanners] RE: HD failure [was RE: keeping the 16bitscans}

All excellent points.  I would suspect that Frank's response may have been
based on a poor choice of terms.  The MTBF is based on "observation" and
"observational data;" it just is not based on the uncontrolled conditions
and exigencies of everyday practival life as opposed to the controlled
contidions of experimental testing and simulations.

-----Original Message-----
From: filmscanners_owner@halftone.co.uk
[mailto:filmscanners_owner@halftone.co.uk]On Behalf Of Julian Robinson
Sent: Monday, March 31, 2003 7:33 PM
To: laurie@advancenet.net
Subject: [filmscanners] Re: HD failure [was RE: keeping the 16bitscans}

> > Well, the one thing you can say with absolutely certainty is
> > that the MTBF is not based on observational data.
>I can say with relative certainly that you are wrong.
>Frank Paris

Purist statistical theory aside, the MTBF (IF it was measured in real-life
conditions) does give you a very useful and practical estimate of how long
you can expect your drive to last.  And even in the purist domain, with the
given 300,000 hour example you can say accurately that the chance of a
drive failing in a given 12 month period is around 3% which is extremely
useful info.

This doesn't take into account the bathtub effect mentioned.

But the real trap is that the MTBF's are not measured in real life
conditions, which is I suggest why observation would not support the
300,000 figure.  Real life includes power surges and spikes, physical bumps
and electrical problems related to people "playing" with their systems -
installing, testing and uninstalling, as well as software-related and often
recoverable failures like corruption of your master boot record.

So when thinking about backup strategies, you need to consider the 3%
"native" chance of a drive failing in any year, increased, (possibly
greatly) by any physical, electrical or excess start-up stress.  All in all
a figure of 5 to 10% per year seems likely in average circumstances, and
that is pretty well I think what we see.

I leave mine on 24/7 and have had no probs at all with 6 disks over 6
years, until ... the other day when 2 died.  I suspect lightning related
surges (through the UPS/surge protector??Hmmm).  I wasn't backed up
properly (only one disk to another and a few semi-random CD's) so am
currently out to the forensic disc doctors.  Damn.


>In the intersts of keeping an interesting thread going here are some
>reliability figures for hard drives:
>Mean time Between Failure 300,000 hours (ie one failure every 34 years)
>Start/stops (at 40 deg C) 40,000
>On the basis of those figures I turned HD power management off on my PC. I
>was finding the disks were getting turned off maybe twenty times a day,
>which equates to one failure every 5 years or so.
>Best regards
>Mike Brown

Canberra, Australia

Unsubscribe by mail to listserver@halftone.co.uk, with 'unsubscribe
or 'unsubscribe filmscanners_digest' (as appropriate) in the message title
or body

Unsubscribe by mail to listserver@halftone.co.uk, with 'unsubscribe 
or 'unsubscribe filmscanners_digest' (as appropriate) in the message title or 


Copyright © Lexa Software, 1996-2009.