[Discuss] SSD

[blu at nedharvey.com (Edward Ned Harvey)]

> From: discuss-bounces+blu=nedharvey.com at blu.org [mailto:discuss-
> bounces+blu=nedharvey.com at blu.org] On Behalf Of Stephen Adler
> 
> The bit I'm worried about is disk crashes with
> the SSDs not working due to what ever with the memory system of the
> drives. (I take it they run parity memory?) 

All HDD's and SSD's use FEC or something equivalent.  This is more powerful
and robust than just parity, but same idea.  Even so, it's a relatively weak
checksum, so that's why things like ZFS and BTRFS implement OS-level
checksumming of filesystems...

USB flash drives don't all have FEC or any error correction at all.  I have
actually experienced enough bit errors from USB fobs, that I simply don't
use them for anything anymore.  If I need usb storage, I use a hard drive or
nothing.


> Does it make sense to have a
> raid setup for redundancy? Or just trust the drives will work fine? 

If you would use redundant HDD's in the same situation, then use redudnant
SSD's.  You choose the redundancy level of your storage purely by the value
of your data...  I am sure you will never encounter the situation where $100
makes the redundancy worth while, and $200 makes it not worth while.


> the OS. My home directory would still be on the old trusty magnetic
> "analogue" disk drives which I run a raid system on.

heheh.  "old trusty" I would disagree with, and "analog" ...  HDD's and
SSD's are equally analog...  If anything, the SSD's might be more analog,
because they have multiple bits of information per cell...  so the cell
stores an analog value, which is then fed into an A2D to produce 3-bits of
information...

Anyway, Magnetic is the word.   ;-)