Date: Fri, 21 Nov 1997 12:24:31 -0500 From: Charles Henrich <henrich@crh.cl.msu.edu> To: "Jordan K. Hubbard" <jkh@time.cdrom.com> Cc: freebsd-hackers@freebsd.org Subject: Re: Serious performance issue with 2.2.5-RELEASE Message-ID: <19971121122431.30389@crh.cl.msu.edu> In-Reply-To: <3057.880101505@time.cdrom.com>; from Jordan K. Hubbard on Fri, Nov 21, 1997 at 12:38:25AM -0800 References: <199711210611.BAA01289@crh.cl.msu.edu> <3057.880101505@time.cdrom.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On the subject of Re: Serious performance issue with 2.2.5-RELEASE, Jordan K.
Hubbard stated:
> Well, now hang on a second.. If you're going to be an ISP, at least one who
> takes himself and his services halfway seriously, then surely cost of an
> extra PC (or, as I noted, even just an extra *disk*) is going to pale in
> comparison to the cost of inflicting instability on your user base.
Well, that depends really. The cost of an extra PC in this case is an ALR
Dual Revolution MP-Pro system, which if I just mirror the system is going to
cost around $10k. I can rebuild the entire system in an hour when neccesary.
It does depend on the function of the system, of course. If its an access
authorization machine you have a much more critical system than say a news
server.
> stuck in my craw as one of those "doctor, doctor, it hurts when I do *this*"
> sort of bug reports which indicated a far more serious degree of dysfunction
> elsewhere.
I understand that. Hot-Upgrading (by which I mean doing a complete OS install
on top of my filesystem after booting from the install floppy) can cause
strange problems that cease to exist if you clean upgrade... I havent run into
a problem doing this with FreeBSD yet however.. Actually, if you want to be
really frightened, I've upgraded entire operating systems from 2.1 to 2.1.7
including kernel while the system was running live (using sup), a quick reboot
and we're now on 2.1.7 and it worked wonderfully well. Scary stuff maynard :)
> And if that's the case then you're one of the lucky ones. Does that make it
> excusable as a general operating principle? No. :-)
Well, as long as when you do it you accept the fact that (and have the ability
to)
1) It may break completely
2) You will have to rebuild from scratch in a very short time
> Well, thanks. Still, don't you think we should be trying to set a proper
> example for all those ISPs wandering around lost in the woods out there? ;)
Maybe, maybe we should call this Extreme-Computing :) For me it always comes
down to the critical nature of the system. If my entire business ceases
because this system is down, I'll take an entirely different approach, than if
a user cant read news for a few hours..
-Crh
Charles Henrich Michigan State University henrich@msu.edu
http://pilot.msu.edu/~henrich
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19971121122431.30389>