Test farm news - I think my cat is forcing me to acquire Legos
Eric S. Raymond
esr at thyrsus.com
Sat Jun 4 03:43:57 UTC 2016
Hal Murray <hmurray at megapathdsl.net>:
> Newer Pis have the power going in the side rather than end. That's going to
> make a similar rack difficult to assemble.
I noticed that when I was watching the University of Southampton video
on their Lego cluster rack - those were all Pi 2s and it nakes a
difference. I can design around that, I think.
> > But I'd hate like hell to have results from a multi-day profiling run
> > ruined. Preventive measures seem required.
> Get used to it. There are all sorts of reasons why multi-day runs don't go
> the way you plan.
Indeed, but that doesn't mean I shouldn't minimize the hazards. :-)
> > The only way I can see to head off further mishaps is to put the farm
> > machines into a weighted rack that takes up enough of the Windowsill to deny
> > Zola a landing pad. With, like, spiky bits sticking up so he won't want to
> > jump on top of *it*.
> Double sided sticky tape might be enough to keep things from getting knocked
> off the window sill.
True, but I'll occasionally need to remove them myself to swap SDs,
attach or detach cabled GPSes, etc. You give me the idea to use
doublesided tape to secure the *rack* on the sill. Probably a better
plan than putting weights on the rack - less hazardous to my desk,
> Wood is a pretty good construction material. You could build a shelf that
> sits an inch or two above the sil.
Indeed. Unfortunately, carpentry was never among my skills and I have
neither the tools nor the materials for that.
One reason Legos are a good plan is that I can build shaped bays to hold
the SBCs, with egress slots for the cabling, and still leave them easy to
lift out when I need to service them. That kind of detail work would be
significantly trickier in wood. And a Lego design will be easier to change
if I have a better idea.
That University of Southampton video was a boon. If I hadn't had "you
can use Legos" lurking in my memory, the design constraints for this
rack would have been pretty hard to satisfy. Assembled metalwork
(assuming I could find suitable parts, which isn't a given) would be
heavy enough to damage the equipment on my desk if it fell off the
sill, I don't have the skill to carpenter something, and I only have
5.5" of sill depth to work with (4.5" if I want to clear the Venetian
blinds, which I do).
Some people will think "that's silly" becuse Lego == kids' toys in
their minds - they don't consider that it makes a perfectly sound
modular construction material for uses like this. I seriously think
it's optimal here. I'll probably write it up as a no-joke maker build.
<a href="http://www.catb.org/~esr/">Eric S. Raymond</a>
More information about the devel