Subject: Re: e450 as a modern server
To: None <port-sparc64@netbsd.org>
From: Sean Caron <caron.sean@gmail.com>
List: port-sparc64
Date: 10/30/2006 16:21:49
hi miles -- i definitely feel what you're saying. the risc platforms
definitely aren't for everyone. there are a lot of things that they
won't really do, or won't really do well enough that you'd want to use
them. i personally love my risc gear because although it might be a
little slower than some junk x86 gear that is easily obtainable from
any garbage bin these days and might use a little more power, i feel
it is much more trustworthy. i can toss that sparc in the corner and
it'll run for years and years and years (the e450 i discuss earlier is
the one exception to this rule i have *ever* encountered. i've run or
am currently running tons of other various sparc systems as servers
and never had a failure -- and these were systems that were probably
running hard for 10 years *before* i bought them). to me, thats worth
it. even if i have to run all my servers in single-threaded mode, and
run into lots of hardware support issues and all the other
inconveniences. but a lot of people would think i'm straight up out of
my mind to waste my time with this old stuff...

i tried to run netbsd/sparc64 as a desktop platform on my ultra here
in my office at work and it did just not work out, as much as i
would've liked. i spent almost two weeks trying to get mozilla
running. and yeah some of this is because i hate using ports
collections, but things can still be a lot of trouble. its a shame,
because i really would have rather ran netbsd than gentoo linux on it
(i try to avoid solaris if i can help it)

commercial apps.. yeah, that's hopeless.

its sad, but you're entirely right. x86 won. thanks to the magic of
the free market *gag*, as usual, the cheapest and crappiest standard
wins. nobody cares about these old risc platforms anymore. even the
manufacturers hardly care. they're just selling legacy stuff to tide
people over until they drop for the opteron gear. the sad fact is that
there just aren't a lot of people who have the knowledge of sparc
assembly and machine internals who have time to burn on this project.
i wish i knew how to do real systems programming because i would love
it if netbsd/sparc64 could even catch up to what other platforms are
doing in terms of being able to run a multiprocessor kernel -- or even
a threaded application for the love of god. but since i can't, i'm
just grateful for the software that is provided by the few people who
put a moment or two of their time into this occasionally.

i don't know what i'm gonna do when my current line of equipment goes
south. i don't think netbsd is ever gonna support the ultrasparc III.
but the thought of running my whole network on x86 stuff is so lame.
pity nobody cares about real computers anymore. this industry just
gets less and less interesting every year.

-S