Nice looking 'disk array'

[Lennart Sorensen]

On Mon, Jan 29, 2007 at 02:21:57PM -0500, Byron Sonne wrote:
> > My experience has been that software raid performed a lot better (maybe
> > the hardware raid cards weren't very good)
> 
> Probably the case; I'm not talking about typical College St. gear.
> Consumer stuff is often a poor representation of the techonologies that
> underlie the product.

My experience is with IBM ServeRaid 4M cards.  Performance was a lot
worse than linux software raid.  I was very surprised by it.

> Negatory on all counts, good buddy! I'll bet $20 that Compaq/HP hardware
> raid cards are in more boxes... think of all the Proliant and Prosignia
> servers out there. Now think of the IBM and Dell product equivalents.
> I'm definately certain that with all of those combined there's more of
> them, and they've had more testing for longer, than any linux software raid.
> 
> Software raid is a toy :)

Given companies like 3ware actually claim most of their users run
software raid rather than hardware raid, I doubt it.  A lot of people
run software raid under linux.  I can't imagine any hardware raid card
has the same number of users as software raid under linux.

Software raid is not a toy.  It is a very useful tool.  Many servers
don't need the cpu power they have and can outperform dedicated hardware
processors in many cases as a result.  And for some people avoiding
vendor lockin-in is a major advantage in addition to the potential
performance gains and the cost savings (Well some people use raid cards
as JBOD controllers, so no cost savings there).

--
Len Sorensen
--
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