Sigh, the /dev below really needs to be something like 
/mnt/dev.

cheers

bruce

Bruce ONeel <edoneel@sdf.lonestar.org> wrote:
> Date: Mon, 02 Feb 2004 17:22:54 +0100
> From: Bruce ONeel <edoneel@sdf.lonestar.org>
> Subject: Re: New 1.6.1 install notes; also pkg_add problem with 1.6.1
> To: Tim McNamara <timmcn@bitstream.net>
> Cc: port-mac68k@NetBSD.org
> 
> Tim McNamara <timmcn@bitstream.net> wrote:
> > Date: Sun, 1 Feb 2004 17:51:29 -0600
> > From: Tim McNamara <timmcn@bitstream.net>
> > Subject: New 1.6.1 install notes; also pkg_add problem with 1.6.1
> > To: port-mac68k@NetBSD.org
> > 
> > I installed 1.6.1 on a Quadra 700 yesterday- quite an ordeal due to the 
> > Installer bug (1.1h) with what are now "small" hard drives larger than 
> > 1 GB but smaller than 10GB.  This is the "Error on SCSIRead(), #5" 
> > problem, occurring in this case on a 4.3GB IBM drive.  Maximizing the 
> > memory to the Installer made no difference.  Fortunately the Installer 
> > would write the kernel into the root partition and enough of the 
> > base.tgz and etc.tgz files to boot into single user mode and finish if 
> > from there.
> > 
> > On boot I get "warning: no /dev/console" and then "mfs /dev" or 
> > something like that; boot seems to hang for about 2 or 3 minutes at 
> > this point and then proceeds.  Building devices with the Installer or 
> > from within /dev doesn't fix this.  It also hung once on boot at "adb0" 
> > but that hasn't been repeated.
> 
> Hi,
> 
> I think what you can do is the following.
> 
> Boot into the installer.
> 
> get a shell
> mount the hard disk 
> mv /dev /dev.old
> 
> reboot and let the hang happen while it builds a memory file
> system with /dev on it.
> 
> cd into /dev.old
> ./MAKEDEV all
> 
> reboot back into the installer
> remount the hard disk
> mv /dev.old /dev
> 
> and then when you reboot it should not make the memory file
> system with dev on it again.
> 
> cheers
> 
> bruce
> 
> > 
> > But it works and that's the main thing!  I used 1.4.2 on a IIci, 145B 
> > and this Quadra several years ago; from what I've seen so far, 1.6.1 
> > seems to be improved.  Seems faster, seems to handle DSL traffic over 
> > the sn0 better.  Very nice work and renders an obsolete machine useful 
> > again!
> > 
> > I am having a problem installing software to make it actually useful, 
> > though.  I want to install apache, Emacs, a browser and a few other 
> > applications.  Some of that might reasonably be a part of the standard 
> > distribution someday.  pkg_add fails as documented below; there was 
> > also a problem I was able to solve; the documentation says to do:
> > 
> > "% PKG_PATH=ftp://ftp.netbsd.org/pub/NetBSD/Packages/1.6.1/mack68k/All
> > % export PKG_PATH"
> > 
> > with the results that there is "no such command."
> > 
> > "setenv 
> > PKG_PATH=ftp://ftp.netbsd.org/pub/NetBSD/packages/1.6.1/mac68k/All" 
> > worked temporarily.
> > 
> > Also, I found that the URL in the documentation is not quite correct in 
> > that ".../Packages/..." should be ".../packages...".
> > 
> > Now, I can ftp to the directory and download packages manually, but I 
> > cannot do so with pkg_add.  For example:
> > 
> > # pkg_add -v emacs
> > 
> > parsing ftp://(correct URL with "=" appended)
> > path: (correct URL, with "=" appended)
> > increasing RLIMIT_NOFILE to max 956 open files
> > trying PKG_PATH (correct URL with "=" appended)
> > spawning FTP coprocess
> > ftp -detv (correct URL with "=/" appended)
> > ftp: connect to address 2001:4f8:4:7:2e0:81ff:fe21:6563: No route to 
> > host
> > pkg_add: expect1 failed, rc=0
> > pkg_add: ftp-start failed
> > Reusing FDs 4/5 for communication with FTP coprocess
> > 
> > ftp> nlist emacs.tbz /var/tmp/pkg.02731a
> > SIGPIPE!
> > pkg_add: short write: Broken pipe
> > pkg_add: nlist failed!
> > Reusing FDs 4/5 for communication with FTP coprocess
> > 
> > ftp> nlist emacs.tbz /var/tmp/pkg.02731b
> > SIGPIPE!
> > pkg_add: short write: Broken pipe
> > pkg_add: nlist failed!
> > Reusing FDs 4/5 for communication with FTP coprocess
> > 
> > ftp> nlist emacs.tbz /var/tmp/pkg.02731c
> > SIGPIPE!
> > pkg_add: short write: Broken pipe
> > pkg_add: nlist failed!
> > pkg_add: no pkg found for `emacs`, sorry
> > pkg_add: i package(s) failed
> > 
> > Any advice on what I have got miscued in my setup would be appreciated. 
> >   I don't know what address 2001:4f8:4:7:2e0:81ff:fe21:6563 is, it has 
> > nothing to do with any of my machines as far as I can tell.  Is the 
> > broken pipe an internal problem on my box?
> > 
> > Thanks for any pointers!