Path: utzoo!attcan!uunet!mcsun!sunic!draken!perand
From: perand@nada.kth.se (Per Andersson)
Newsgroups: news.sysadmin
Subject: Re: Machine room design
Message-ID: <2714@draken.nada.kth.se>
Date: 13 Jan 90 13:09:32 GMT
References: <13176@cit-vax.Caltech.Edu>
Reply-To: perand@nada.kth.se (Per Andersson)
Organization: Royal Institute of Technology, Stockholm, Sweden
Lines: 35

In article <13176@cit-vax.Caltech.Edu> ktl@wag240.caltech.edu (Kian-Tat Lim) writes:
>We may soon be able to design a new machine room.  Are there any
>generic (machine-independent) issues we should be sure to address?
>Examples might include raised floor, adequate air conditioning,
>earthquake resistance, cable trays, power outlet positioning, etc.  I
>think we have most bases covered; I'm just trying to make sure we're
>not missing anything.  Any suggestions, anecdotes, or warnings will be
>helpful.  I will summarize (not just batch together) any useful E-mail
>replies.

You don't mention what scale of money you have available, but:

Some kind of stabs on the power can be a quite cheap protection 
against diverse moonphasedependant trouble. This catches small 
peaks and disturbances om your incoming power.

A PDS (= Power Distribution System) can save you some trouble,
by having sequenced powerup after a powerfail. This way you
can give the disks power and time to aquire full speed before
giving power to the machine ( of course you might have small
machines with internal disks...)

If the situation motivates it, an UPS is the way to go 
(UPS =Uninterruptible i think). These usually have batteries for
15 or 30 minutes of lost power. This at least gives you time to
make a controlled shutdown. Quite expensive although.

Per


-- 
---
Per Andersson
Royal Institute of Technology, Stockholm, Sweden
perand@admin.kth.se, @nada.kth.se