On 08-Dec-99 David Wolfskill wrote:
>>From: "Sameer R. Manek" <[EMAIL PROTECTED]>
>>Date: Tue, 7 Dec 1999 21:05:38 -0800
> Unless I'm confused rather more than is usual -- which is by no means
> especially unlikely -- that also assumes that the "spare box" hardware is
> sufficiently close to that of the "production systems" that the resulting
> system can be tested with sufficient confidence (for some value of
> "sufficient") that the results will apply to the production system -- or,
> better yet, that the "spare box" can actually *become* the (new)
> "production system". (Then the just-retired system becomes the "spare
> box"... after sufficient burn-in....)
Indeed the latter scenario is ideal (ideal is that the spare box and
the production box are identical).
> Of course, this also presumes that the sysadmin(s) can tell when one set
> of hardware is "sufficiently close" to some other (sometimes arbitrary)
> set of hardware, which is certainly not the case with my dealings with
> PC hardware. :-( Others may well have different perspectives, certainly.
Two identical boxes is IMHO the only way to be sure.
> And that count of "3 upgrades/year" is per machine; I doubt that I'm all
> that unusual in having some dozen or so servers and about 30 desktops to
> cope with, as well as a few other tasks to occupy my time. (In
> practice, I'm upgrading *far* less often than 3/year for any machine
> that I rely on. Not, of course, that I'm happy with the upgrade
> schedule I've been accomplishing; I'm working in "triage" mode.)
This kind of conservative approach is most appropriate to critical
servers. For workstations I just tend
-------------------------------------------------------
Tell a computer to WIN and ...
... You lose
-------------------------------------------------------
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-stable" in the body of the message
