Date: Tue, 21 Aug 2001 20:15:12 -0700 From: Darryl Okahata <darrylo@soco.agilent.com> To: Hodge Podge <nicole@unixgirl.com> Cc: Michael VanLoon <MichaelV@EDIFECS.COM>, freebsd-hardware@FreeBSD.ORG Subject: Re: 3ware stuff not ready for heavy duty useage-followup Message-ID: <200108220315.UAA06753@mina.soco.agilent.com> In-Reply-To: Your message of "Tue, 21 Aug 2001 16:48:25 PDT." <XFMail.010821164825.nicole@unixgirl.com>
next in thread | previous in thread | raw e-mail | index | archive | help
[ Sorry about the long quoting, but the quoted parts really do need to
be emphasized. ]
Hodge Podge <nicole@unixgirl.com> wrote:
> On 21-Aug-01 Michael VanLoon wrote:
> >
> > The quality Enermax power supplies are a good place to look for some
> > guidance on this. Remember that cheap no-name power supplies are typically
> > going to handle this worse than well-made ones will. Their 300W model
> > (EG301P) has a max output rating of 15A at 12V. This means that if all
> > eight drives try to spin up at once, you are over-loading the power supply.
> > This can result in startup errors. But even as indicated above, if all
> > eight drives are doing full seek traffic all at once, even that might push
> > you above the power supply ratings (16A on drives > 15A on power supply).
> >
> > And remember this is only the drives -- this doesn't count other stuff
> > drawing 12V inside your case. Since the CPU and motherboard circuitry use
> > the 3.3V and 5V taps, this likely will not crash your machine, but will
> > cause hard drive errors.
> >
> > Even Enermax's biggest power supply, the 550W unit (EG651P) is just barely
> > capable of providing enough power for simultaneous spin-up of all eight
> > Maxtor drives (23.6A on drives just barely < 24A from PS). And that's
> > assuming you don't have more than 0.4A draw on 12V from everything else in
> > your system.
>
> I see.. So what you are saying is then that all those people who make and se
> ll
> systems with that many HD's more or less are all asking for trouble?
> That all the scsi systems I have built or seen running similiarly and withou
> t
> errors anywere near this are just lucky?
Maybe. The issue here is that you have not shown that there's an
high enough probability that the power supply is NOT the culprit. Now,
the problem could lie elsewhere, but you have *NOT* eliminated the power
supply. Michael's made a lot of good points in the above, and you have
not given enough evidence to "disprove" his points (or, at least, to
show that it's not likely).
Your points seem to be:
"Vendors sell it this way. It must be good."
Well, "previously-owned cars" are also sold. It doesn't mean
that they're any good. M$ Windows is also sold. That doesn't
mean that it's good for you, either.
In another post, you wrote:
> I would like to point to 1 example server that has 13 Drives in
> it, and only 1 300 Watt power supply.
> http://www.rackable.com/index-productsST.php3?checkit=noflash
Personally, I wouldn't touch this, if it has 13 high-performance
drives in it. As Michael's shown: 8 drives (without
motherboard!) can be marginal with a 300W power supply. I'd
hate to think what would happen with 13.
"It works for others."
Does it work with the exact same hardware configuration as yours
(without the 3ware card)? The issue is that the "others" may be
using SCSI drives (see the next point), they may be using
lower-power (and slower) drives, or they may not be banging on
all drives simultaneously.
"SCSI works."
Well, you really have not given any details. In the absence of
such, I'd have to say that:
1. SCSI systems with lots of drives are often in external,
stand-alone boxes without a motherboard. This means that
there is no power-hungry motherboard competing for power. A
"300W" power supply in a stand-alone external case will go
farther (but still may or may not be enough) than a "300W"
power supply in a PC.
2. IDE systems are often in large cases with a power-sucking
motherboard (because IDE cables are not meant to be strung
externally).
Michael's calculations show that, EVEN WITHOUT A MOTHERBOARD, a
300W power supply can be marginal if eight drives are banged
upon hard.
You have not shown why this cannot be a problem for you.
Now, the power supply may not be your problem, but you have not
given any concrete (and technical!) reasons why. Just "touchy-feely"
ones. ;-)
You need to refute Michael's technical arguments.
[ Going off on a slight tangent, it might be interesting to probe the
voltages with an oscilloscope, preferably a sampling one if available
(you've got to be careful about the usual electrocution, fire, and
component damage hazards, though ;-). You probably won't be able to
tell anything conclusive, however, unless the power's really dirty
(which would be a good indication that the power supply is the
culprit). Of course, you'd naturally have to do it under high-load
conditions. ]
[ Heh. I should talk. I've been meaning to put together a
3W-6800-based 8-drive box (I've had the parts for weeks, but haven't
had the time to put it together). I've got an Enermax 451 power
supply, and I've been *really* hoping that it'll work. We'll see how
it goes (probably in a few weeks, though). ;-( ]
--
Darryl Okahata
darrylo@soco.agilent.com
DISCLAIMER: this message is the author's personal opinion and does not
constitute the support, opinion, or policy of Agilent Technologies, or
of the little green men that have been following him all day.
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hardware" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200108220315.UAA06753>