[SunRescue] Q on "optimal" OS for Sun4c machines, now that So laris 8 won't run

GregoryLeblancGLeblanc at cu-portland.edu GregoryLeblancGLeblanc at cu-portland.edu
Thu Jul 13 12:06:48 CDT 2000


> -----Original Message-----
> From: Harry Regan [mailto:harry.regan at usa.net]
> Sent: Thursday, July 13, 2000 6:22 AM
> To: rescue at sunhelp.org
> Subject: Re: [SunRescue] Q on "optimal" OS for Sun4c 
> machines, now that
> Solaris 8 won't run
> 
> Chris' comment strikes a nerve with me.  I'm an old dude-- 
> I've been playing
> with computers since 1967-- "Why, back in my day, we had to 
> stoke the boilers
> before we did a compile!"  At the risk of sounding like a 
> geezer, I think the
> way people are educated in technology these days is 
> inadequate-- and a major
> culprit is our buddy Microsoft.
> 
> As a mainframe assembler programmer, you had to have a deep 
> understanding of
> the physical limits of the machine, what its peripherals 
> were, how to make
> them function, and most important, how to restore the damned thing to
> operating status when it blew up at 2 AM.

This is still required in a fairly large number of places.  The DBA groups
at CNF (a big shipping company) spend almost all of their time figuring out
how to use the hardware, and almost none desiging databases or anything like
that.  Hardware upgrades where he works are expensive now, especially if
they add another engine to an S/390 or some more processors to an UE 10000.
Purchasing the processor is cheap.  The extra cost on software licensing is
expensive.  So they tune things to death, and buy lots of fast "DASD".  This
sort of knowledge isn't dead, but it's less important in smaller shops.  I
could probably replace most of our P-II machines with SS20 or slower class
machines, and not lose any performance, because I know how to tune, and how
to NOT run NT (I know how to run it as well, but that's a different story).

> I get really stressed with new Microsoft Certified architects 
> and programmers
> who design applications that clearly have of understanding of 
> a production
> processing environment. "Dude! When it does that, just reboot 
> it!"  Yeah, and
> kill the other three apps running on the same server ("Oh! 
> Should I only run
> one app per server-- why, that would mean I'd have to run six 
> hundred servers
> to support the company...")

Unfortunately for me, I'm the "lower grade" admin, so I don't get a say in
this.  But, in NT shops, ours included, it IS 1 app per server.  We've got a
dedicated webserver, a dedicated mail server, a dedicated file server, a
dedicated print server, and so on.  I suspect that most of these tasks could
be better served by a single larger machine, but since we have to keep
rebooting things, we can't put everything on the same server.  
	Grego





More information about the rescue mailing list