| 1 |
On Wed, Jun 25, 2014 at 12:55 PM, Volker Armin Hemmann |
| 2 |
<volkerarmin@××××××××××.com> wrote: |
| 3 |
> |
| 4 |
> so without looking that drive up - you are using a desktop part for |
| 5 |
> non-stop setup? |
| 6 |
|
| 7 |
Honestly, I think it makes far more sense to build a fault-tolerant |
| 8 |
setup than to try to avoid faults by spending more on the parts. I've |
| 9 |
only run desktop hard drives on my 24x7 RAID. If they die I replace |
| 10 |
them under warranty - I've yet to have one die outside of warranty, |
| 11 |
and I'm usually upgrading for size by that timeframe anyway, and I can |
| 12 |
use the old drives for storage. |
| 13 |
|
| 14 |
By all means get better-grade components, but I wouldn't use that as |
| 15 |
an excuse for not having backups of some kind. ALL hard drives WILL |
| 16 |
fail, it is just a matter of when. ANY hard drive can fail the day |
| 17 |
after you buy it, a month after you buy it, and so on, though |
| 18 |
obviously the probability of a particular drive failing at any point |
| 19 |
in time may vary by what you pay for it. |
| 20 |
|
| 21 |
I'd buy a more expensive drive only if the TCO is actually lower. I'd |
| 22 |
engineer any system to accept the failure of at least one drive, and |
| 23 |
for any data I actually cared about I'd engineer the system to resist |
| 24 |
fire, the rm star, and so on. |
| 25 |
|
| 26 |
Rich |
Archives