[Discuss] Partitioning scheme for SSD and HD

[Cy Schubert]

In message <5434A3EF.2010606 at darrenduncan.net>, Darren Duncan writes:
> On 2014-10-01 10:36 PM, Cy Schubert wrote:
> > SSD is great for reading but not as good for writes (it wears every time it
> > writes). Additionally your root filesystem is accessed every time you
> > traverse a path (though much of that is cached in the buffer cache).
> 
> Could this be a problem then with file systems that write access-time metadat
> a 
> every time a file is simply read?  I'm not sure how pervasive this problem is
>  
> though. -- Darren Duncan

That would contribute to it. The reason for the wear is that SSD technology 
has a sandwich of electrons that are eroded every time a write occurs. The 
microprocessor in the SSD will attempt to level wear however over a period 
of time -- usually years -- the SSD simply loses enough electrons to wear 
out. More expensive SSD technology has spare memory (sectors) which can be 
allocated to replace worn blocks. The more expensive the SSD the greater 
the number of spare blocks but eventually there is too much wear.

OTOH disks have bearings which also wear, causing the platters to wobble 
and eventually no longer write on the tracks. The consensus is that barring 
head crash of spinning disk, both will last approximately equivalent 
lengths of time.


-- 
Cheers,
Cy Schubert <Cy.Schubert at komquats.com>
FreeBSD UNIX:  <cy at FreeBSD.org>   Web:  http://www.FreeBSD.org

	The need of the many outweighs the greed of the few.