RELENG_5 installworld fails

[Kent Stewart]

On Saturday 26 February 2005 04:10 am, Velko Ivanov wrote:
> > So, you need to supply more info. Are you setting any special
> > parameters in /etc/make.conf? Did you follow UPDATING as far as the
> > sequence of buildworld, [build/install]kernel, boot to single user
> > mode and do the installworld? Before this build, when did you last
> > update your system?
>
> This points to /etc/make.conf:
> CPUTYPE=p3
> CFLAGS= -O -pipe -msse -mmmx
> COPTFLAGS= -O -pipe -msse -mmmx
>
> NO_BIND=true
> BOOTWAIT=0
>
> It is a dual PIII Xeon, 5.3-RELEASE freshly installed - installing
> cvsup and updating to RELENG_5 was the first thing I did.
> The procedure in UPDATING is followed strictly.

Some people don't follow it and strange things happen. FWIW, my upgrade 
worked as expected.

> I have done many installations on different machines and never had
> trouble with optimizations in make.conf (except when I specified the
> wrong CPU once :)).
> I just can't link problems with GCC optimization flags, to the fact
> that the path to uuencode is not set in a Makefile in some directory.
> I was just curious, thanks for the reply.

You never know but when something strange pops up. If you aren't using 
the defaults, killing the CPU and FLAGS are a place to start. It is too 
easy to add the "#" and then delete if nothing changes. For example, my 
5-stable is an athlon-xp and somewhere in time, I commented the CPU 
out. The default FLAGS is "-O -pipe" and so, I don't supply them 
either. 

I have been going to try and find what the boot parameter is. The delay 
at the start is not needed.

Kent

-- 
Kent Stewart
Richland, WA

http://users.owt.com/kstewart/index.html