Subject: Re: A report on implementing runlevels in NetBSD
To: Guenther Grau <>
From: Ignatios Souvatzis <>
List: tech-userlevel
Date: 12/08/1999 10:15:44
On Tue, Dec 07, 1999 at 08:46:26PM +0100, Guenther Grau wrote:
> Ignatios Souvatzis wrote:
> > 
> > On Sat, Dec 04, 1999 at 08:33:37PM -0800, Greywolf wrote:
> > > Personally, I think Sun ought not have been bedded by AT&T in the first
> > > place.  Solaris 2.0 was a fscking disaster.
> Nice way to get around the violent language filter :-)

alt.sysadmin.recovery slang...

> > And 2.1, and 2.2, afaihh. My referrence in the other part of the building
> > started to say you could use 2.3 and 2.4 with dozens of patches. Apparently,
> > 2.5 was the first halfway decent, and in 2.6 already some of the performance
> > wins it had over 2.0-2.3 were lost again.
> Hmm, Solaris 7 feels a lot faster than 2.5.1 and 2.6 do, at least
> on the machines I work on. And booting is about the same speed, if
> not faster than NT on faster HW. It get's slower if you install
> lots of sw which needs daemons started at boot time, though,
> but this is equally true for NT, of course.

Just don't try to run Solaris 7/386 as a NFS server. I have witnessed that
it has very slow I/O. Don't know, how other systems would perform with the
same load (e.g., us or FreeBSD or BSDi or Linux; I don't count NT in there)...

I didn't dare to try 7 on our Sparcs because of the big load of new problems
reported... why care to upgrade (which is work) instead of using the machines
that have all necessary patches in by now for stable operation?

I'm very tempted to use 8, however, for IPv6 support.

 * Progress (n.): The process through which Usenet has evolved from
   smart people in front of dumb terminals to dumb people in front of
   smart terminals.  -- (obscurity)