better disk reliability on a desktop machine

[Alex Zbyslaw]

lars wrote:

> -    /usr/ports/sysutils/smartmontools
>     can help you monitor your HDDs

But if your disk is a hardware RAID of any kind, and you cannot see 
through the controller to individual disks, then you'll only be told 
about one of the disks, I would presume.  That's where a CLI comes in, 
but I think they are scarce for the low-end controllers you see on 
desktop systems.  Or just rebooting daily and hoping the RAID BIOS will 
report a SMART error in time.

Even given that, smartmontools should be on everyone's list of "must 
have" ports.

> -    RAID 0 doubles the chances of HDD failure and thereby data loss
>
Agreed.  But my presumption is that the actual chances of hardware 
failure are pretty small.  Out of all the disks I've been responsible 
for in some way over the years (certainly hundreds), the actual number 
of failures I can remember is about a handful, and at least two of those 
came with some warning.

Actually, I suspect a RAID 0 more than doubles the chance something bad 
happening.  A single bad disk may be just good enough to be recoverable 
in some way, whereas the same errors in a RAID 0 could be curtains.  I 
certainly wouldn't do 0 without a frequent, automatic back-up strategy 
on anything which wasn't truly disposable.

>     e.g. Mini-ITX boards are cheap and fast enough for this purpose.

Or the PC you just replaced which now has an ebay value of 
not-enough-to-be-worth-it... :-)

The real expense is usually time.  Especially for home-based machines, 
backups become a chore, or you're up until 2am and just can't be 
bothered turning on the tape drive or whatever.  And a disk just drive 
knows when it hasn't been backed up recently ;-)

--Alex