Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 1 Jun 1995 18:46:42 -0700 (PDT)
From:      "Rodney W. Grimes" <rgrimes@gndrsh.aac.dev.com>
To:        chuckr@Glue.umd.edu (Chuck Robey)
Cc:        jkh@freefall.cdrom.com, vince@penzance.econ.yale.edu, roberto@blaise.ibp.fr, FreeBSD-hackers@freefall.cdrom.com
Subject:   Re: Upgrading to 2.05A
Message-ID:  <199506020146.SAA06194@gndrsh.aac.dev.com>
In-Reply-To: <Pine.SUN.3.91.950601204700.7037B-100000@espresso.eng.umd.edu> from "Chuck Robey" at Jun 1, 95 08:48:19 pm

next in thread | previous in thread | raw e-mail | index | archive | help
> 
> On Thu, 1 Jun 1995, Jordan K. Hubbard wrote:
> 
> > > 	Really?  I was told that the 2.05A sources would allow me to 
> > > upgrade but I won't get any of the disk optimizations and other things 
> > > without doing a full install and I would need to do a full reinstall to 
> > > get properly optimized and coverted to the slice code since there will 
> > > probably be no backwards compatibility of this in 2.1.  Can anyone 
> > > confirm this?  Jordan?  Thanks.
> > 
> > This is all true.  Please folks, don't mislead people by telling them
> > that "upgrades" are a simple matter of holding your left foot in the
> > air while extracting the bindist over a system in single user mode
> > while "Muskrat Love" is playing on the radio.  Sure, it might _work_
> > but whether or not it works 100% is another matter, and WE get to pick
> > up the pieces afterwards in the questions list!
> > 
> > THERE IS NO UPGRADE PATH FROM 2.0R TO 2.0.5!  I'm sorry, I wish there
> > was, but there wasn't time to do the work.
> > 
> > 					Jordan
> 
> If this is true, I'm curious just what I'm gaining with the new 
> filesystem, besides a new way to refer to partitions.

Minfree change from 5% to 8% to correct for serious over allocation
problems (davidg knows more about this).  Newfs parameter changes to
effectivly eliminate all the rot delay optimizations that are actually
slow downs for almost every modern disk on the market.

A nice clean defrag'ed set of system binaries is another side effect
that can be a substantial gain if you have run a few 100 make worlds
on a system.  [If you don't belive the last one, do a clean install
on a disk then setting it running make worlds in a loop, come back
after about 5 times, reboot and look at the fsck fragmentation values]


-- 
Rod Grimes                                      rgrimes@gndrsh.aac.dev.com
Accurate Automation Company                   Custom computers for FreeBSD



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199506020146.SAA06194>