Archives


- Beowulf
- Beowulf Announce
- Scyld-users
- Beowulf on Debian

Back-UPS garbage? (was: Quick survey -- UPSs on slave nodes?)

Many of your questions may have already been answered in earlier discussions or in the FAQ. The search results page will indicate current discussions as well as past list serves, articles, and papers.

Search

Robert G. Brown rgb at phy.duke.edu
Mon Feb 10 16:30:53 PST 2003


On Mon, 10 Feb 2003, Drew Raines wrote:

> <alvin at Maggie.Linux-Consulting.com> writes:
> 
> > $40 el cheapo APC for each node...
> 
> I assume you're referring to the ``Back-UPS'' line.  The only thing
> I've found those to be able to support on battery power is my clock
> radio, which I could more easily protect with a 9-volt.
> 
> Has any one of you actually used these to sustain a PC for any
> measurable length of time?  I'm not sure at what market this junk
> is aimed, perhaps suckers.

In the Durham NC residential grid we have a power outage that lasts
approximately one second approximately once a month.  One second is just
about exactly what it takes to crash a computer (or a digital clock, or
a digital coffee maker, or...grrr).  Maddening, as somebody is ALWAYS
doing something critical right when the power bobbles -- just for a
second.  Crash.  Boom.  There goes the homework.

I've recently started putting my home computers (which are all members
of my home cluster:-) on $50 cheap UPS because they are enough to
sustain a system across these frequent, short, outages.  I don't need
even minutes -- just a few seconds worth of protection.  Minute or
longer outages are MUCH rarer, and tend to be associated with things
like ice storms, hurricanes, nuclear disasters.  Well, not THAT much
rarer...

I also was motivated by a home system that blew power supply,
motherboard, hard disk and at least three cards.  Maybe it was a surge,
maybe not, but a cheap UPS is also a good surge protector, and only $20
more than a good surge protector alone.

So far, the two I've gotten have worked fine through a few of these
short outages, and I'll probably get one more (as I still have one
computer without one).  I don't know if I'd try scaling up to a whole
"real" cluster this way, though -- there are more cost effective ways to
do cluster nodes;-)

   rgb

-- 
Robert G. Brown	                       http://www.phy.duke.edu/~rgb/
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567  Fax: 919-660-2525     email:rgb at phy.duke.edu






More information about the Beowulf mailing list