Possibly OT: 1U 'rack'

[The Natural Philosopher]

messagenews: [e-mail]...

will do too..

conduction is not the ideal way of getting heat out of a box, but it does work.

Even through a plank of wood.

Rack design and rack kit design is 'interesting' heatwise.

MOST boxes suck from front or back and blow backwards.

Heat extraction by hot air extraction is the method..until you have seriously high power when forced blowing on the hot bits is needed. MOST racks are happy with floor ingress of air, and some form of overall either chimney or extractor fans in the top.

Generally the kit inside is blown well enough to stand ambient at

30C-40C: The job of the rack fans and the machine room aircon is to make sure they get that air temp.

Or else put aircon in the racks themselves..

[The Natural Philosopher]

Dont be daft. One linux server would have replaced the lot..whenever I found MASSES of servers it was always because 'we have this app, and it needs windows server to run on, and its really crap so it needs a BIG windows server to run on, and the company that installed the software insists its on its own server'

sigh. These days its one blade virtual machine chassis and as many crappy windows servers as you want.

[Tim Watts]

Nope - Centos. This is High Energy Physics - they do NOT use windows for anything heavy...

[Tim Watts]

HEP is an entirely different game. You really do want 100+ multicored servers running flat out if you want to be a decent Tier 2 player with CERN's data. When I say "decent" - that is because you are still merely part of the GRID and there are a large number of other sites of similar or bigger size participating.

That's not including storage (a petabyte counts as "decent" currently) and the extra servers to manage that. Total power for said site, around 80kW.

You'll also be needing a several gigabit link to the Internet (5gigs is a reasonable number in the case above - and that's typical average consumption of HEP data on a good day, not the actual link speed, so a Tier 2 university will ideally be wanting a 10 gig link to JANET.

To put this in perspective, CERN's output is reckoned to be in the order of

200MB/sec continuous average (save for shutdowns) but given that gets fed to Rutherford Labs who also hold a lot of historic data for the UK, you have to have more bandwidth for when some experiment wants to pull down archived datasets for cross analysis (often).

The numbers are just loony in that field.

[Jeremy Nicoll - news posts]

This is not altogether true; some rack mount stuff is meant to be mounted with a 1u space above it, perhaps fronted by a ventilation grill.

[Jeremy Nicoll - news posts]

That's right BUT it doesn't mean that you have to populate every single u with equipment. The racks just give versatility.

[dennis]

Well if you ever chose to challenge me about something I am wrong about I will admit it. However when its always you that is wrong there isn't much point in me admitting I was correct. Now why don't you do as you say and admit that you blew it and got it wrong again?

[dennis]

Do you have an example? I have always found that they would then come as a 2U module to ensure they were fitted correctly.

[Tim Watts]

In the old days, we found it wise not to over crowd certain equipment and to deliberately leave a 1U gap for every n items. But I would worry less with modern equipment as long as it is well engineered. You wouldn't have a problem with Dell, not with the 8 odd Delta fans those bastards put in.

But some of the cheaper stuff I've had in the past was not as well provisioned and did benefit from a bit of spacing.

[ARWadsworth]

Is it panto season already?

Everyone "Oh no you won't [admit when you are wrong]"

[Jeremy Nicoll - news posts]

Not off the top of my head; I'd have to check manuals for the products I own (which aren't necessarily such anyway).

I tended to leave such a space in flightcased rackmount equipment and when they were being moved around kept an A4 binder in the space with photocopied manuals and use notes for the items concerned; the binder would be out of the space when actually using the equipment.

[Dave Plowman (News)]

A radio mic receiver with fixed aerials and run from a wall wart is likely only 'rack mountable' to make it appear professional...

[Grimly Curmudgeon]

Jesus, Aitch Kerist. What kind of shit is this?

[Grimly Curmudgeon]

You utterly ignorant little man.

[dennis]

More likely so it will fit into a flight cased rack. There is a rack version of my midi sound generator but I bought the nice package as it was cheaper and I don't need a rack.

Anyway to the OP, if its a metal case check to see if its got a Kensington lock slot in it and use a Kensington lock to secure it.

[charles]

not quite true - we have 12 in our theatre. Putting them on shelves would be silly. Putting them in a rack means that you can see the indicator lights from the sound desk - as well as allowing for neat cable runs - both in and out.

[Dave Plowman (News)]

And the aerials are mounted on the receivers? You're very lucky if this is satisfactory.

[Dave Plowman (News)]

Yes. Rack mounting of pro units is of course common in studios. But the aerials will be elsewhere. And pro units have internal power supplies.

Many don't realise the difference decent aerials make to radio mics. With the supplied ones they have a range of perhaps 100 yards. With a decent array you'll get several times that.

[Man at B&Q]

:

Conduction is not venting.

MBQ

[Theo Markettos]

That's why they cost 5 euro. It's remarkably strong paper honeycomb, and does what it says on the tin (be a coffee table and (presumably) take someone sitting on it). What more did you want?

Theo