SSDs are scary

Digimer lists-5ZoueyuiTZiw5LPnMra/2Q at public.gmane.org
Sat Jan 4 18:54:42 UTC 2014


On 04/01/14 12:17 PM, D. Hugh Redelmeier wrote:
> Via slashdot and other postings, I got to this.  Boy, most SSDs seem
> to have unconscionable flaws:
>
> <lkcl.net/reports/ssd_analysis.html>
>
> This will certainly affect my use of SSDs.

It's a young technology, there is a reason it's called "the bleeding 
edge". :)

I've bought and used a couple dozen SSDs in the last 2~3 years and I've 
had three or four failures so far. Not bad for a young technology.

To the specific power-loos induced corruption; That speaks to use-case. 
In laptops (and tablets/phones), the device's internal battery helps to 
protect against power events.

It's similar to BBWC/FBWC in high-end RAID arrays; It's a trade-off you 
make. Now, running write-back caching without a BBU/FBU, that is insane. 
Equivalently, running SSDs in a desktop without a UPS is along the same 
lines of risky behaviour. Thankfully, half-decent UPSes aren't that 
expensive.

All this said, that document specifically states that the Intel S3500 
and 320 have internal capacitance sufficient to flush a full cache out 
to NAND in a power loss event. So if this issue concerns you, it sounds 
like you have a valid option if you want to use SSDs still 
(http://www.newegg.ca/Product/Product.aspx?Item=N82E16820167163 - well 
under $2/GB).

-- 
Digimer
Papers and Projects: https://alteeve.ca/w/
What if the cure for cancer is trapped in the mind of a person without 
access to education?
--
The Toronto Linux Users Group.      Meetings: http://gtalug.org/
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://gtalug.org/wiki/Mailing_lists





More information about the Legacy mailing list